AE3-302

Flight Dynamics
Lecture Notes J.A. Mulder / W.H.J.J. van Staveren / J.C. van der Vaart / E. de Weerdt

March 7, 2007

Flight Dynamics
Lecture Notes J.A. Mulder / W.H.J.J. van Staveren / J.C. van der Vaart / E. de Weerdt March 7, 2007

Faculty of Aerospace Engineering

Delft University of Technology

Delft University of Technology

Copyright c Control and Simulation division, Delft University of Technology All rights reserved.

Table of Contents

Nomenclature 1 Introduction 1-1 Introduction to ight dynamics and control . . . . . . . . . . . . . . . . . . . . 1-1-1 1-1-2 Flight control surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . Flight control systems . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xxiii 1 2 2 6 16 18

1-2 Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-3 Book outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Equations of Motion

19
21 22 22 24 25 26 27 30 30 32 33 33 38 42 43

2 Reference Frames 2-1 Overview of reference frames . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1-1 Inertial reference frame, FI . . . . . . . . . . . . . . . . . . . . . . . . . 2-1-2 2-1-3 2-1-4 2-1-5 2-1-6 2-1-7 Earth-centered, Earth-xed reference frame (FC ) . . . . . . . . . . . . . Normal Earth-xed reference frame, FE . . . . . . . . . . . . . . . . . . Vehicle carried normal Earth reference frame, FO . . . . . . . . . . . . . Body-xed reference frame, Fb . . . . . . . . . . . . . . . . . . . . . . . Aerodynamic (air-path) reference frame, Fa . . . . . . . . . . . . . . . . Kinematic (ight-path) reference frame, Fk . . . . . . . . . . . . . . . .

2-1-8 Vehicle reference frame, Fr . . . . . . . . . . . . . . . . . . . . . . . . . 2-2 Transformations between reference frames . . . . . . . . . . . . . . . . . . . . . 2-2-1 Euler angles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-2-2 2-2-3 2-2-4 Transformation matrices . . . . . . . . . . . . . . . . . . . . . . . . . . Transformation from FI to FE . . . . . . . . . . . . . . . . . . . . . . . Transformation from FI to FO . . . . . . . . . . . . . . . . . . . . . . .

Flight Dynamics

iv 2-2-5 2-2-6 2-2-7 2-2-8 2-2-9

Table of Contents Transformation from FE to FO . . . . . . . . . . . . . . . . . . . . . . . Transformation from FO to Fb . . . . . . . . . . . . . . . . . . . . . . . Transformation from FO to Fa . . . . . . . . . . . . . . . . . . . . . . . Transformation from FO to Fk . . . . . . . . . . . . . . . . . . . . . . . Transformation form Fb to Fa . . . . . . . . . . . . . . . . . . . . . . . 45 46 49 50 52 53 55 56 63 66 66 67 68 68 69 70 70 72 77 83 83 91 95 99 99 102 102 103 104 105 106 111 112 112 113

2-2-10 Transformation form Fb to Fk . . . . . . . . . . . . . . . . . . . . . . . 2-2-11 Transformation from Fk to Fa . . . . . . . . . . . . . . . . . . . . . . . 2-3 Rotating reference frames . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3-1 2-3-2 2-3-3 2-3-4 2-3-5 2-3-6 2-3-7 2-4-1 2-4-2 2-4-3 Derivation of the angular velocity vector . . . . . . . . . . . . . . . . . . Derivation of E . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . EI Derivation of O . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . OI Derivation of O OE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Derivation of b . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . bO Derivation of a and k aO kO . . . . . . . . . . . . . . . . . . . . . . . . Derivation of a and k . . . . . . . . . . . . . . . . . . . . . . . . . ab kb Euler angle singularity . . . . . . . . . . . . . . . . . . . . . . . . . . . . Quaternion transformation matrix . . . . . . . . . . . . . . . . . . . . . Quaternion propagation through time . . . . . . . . . . . . . . . . . . .

2-4 Quaternions, a brief introduction . . . . . . . . . . . . . . . . . . . . . . . . . .

3 Derivation of the Equations of Motion 3-1 Newtons laws of physics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-2 Derivation of body translational acceleration . . . . . . . . . . . . . . . . . . . . 3-3 Derivation of body angular momentum derivative . . . . . . . . . . . . . . . . . 3-4 External forces and moments . . . . 3-4-1 External forces . . . . . . . . 3-4-2 External moments . . . . . . 3-5 Composition of equations of motion 3-5-1 3-5-2 Moment equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Force equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3-5-3 Kinematic relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-6 Rotating masses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7 Simplications of the equations of motion . . . . . . . . . . . . . . . . . . . . . 3-7-1 3-7-2 3-7-3 Non-rotating Earth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Flat Earth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Flat and Non-rotating Earth . . . . . . . . . . . . . . . . . . . . . . . .

Table of Contents 4 Linearized Equations of Motion

v 117

4-1 Linearization about arbitrary ight condition in arbitrary body-xed reference frame 119 4-1-1 4-1-2 4-1-3 4-2-1 4-2-2 Linearization of states . . . . . . . . . . . . . . . . . . . . . . . . . . . . Linearization of forces and moments . . . . . . . . . . . . . . . . . . . . Linearization of kinematic relations . . . . . . . . . . . . . . . . . . . . . Linearized set of equations . . . . . . . . . . . . . . . . . . . . . . . . . Moments and products of inertia . . . . . . . . . . . . . . . . . . . . . . 121 122 128 130 131 132 134 135 138 148 151

4-2 Linearization about steady, straight, symmetric ight condition . . . . . . . . . .

4-3 Equations of motion in non-dimensional form . . . . . . . . . . . . . . . . . . . 4-3-1 4-3-2 Symmetric Motions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Asymmetric Motions . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4-4 Symmetric equations of motion in state-space Form . . . . . . . . . . . . . . . . 4-5 Asymmetric equations of motion in state-space form . . . . . . . . . . . . . . .

II

Static Stability Analysis

153
155 156 156 169 175 184 195 206 219 220 226 233 238 241 255 258 262

5 Analysis of Steady Symmetric Flight 5-1 Aerodynamic center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1-1 5-1-2 5-1-3 5-1-4 5-1-5 5-1-6 5-1-7 Aerodynamic forces and moments acting on a wing . . . . . . . . . . . . A compact description for the aerodynamic moment characteristics . . .

The role of the aerodynamic center and Cmac in static stability . . . . . . The position of the aerodynamic center and Cmac of a wing . . . . . . . Inuence of wing shape on the aerodynamic center and Cmac . . . . . . . Characteristics of wing-fuselage-nacelle congurations . . . . . . . . . . . Aerodynamic eects of nacelles . . . . . . . . . . . . . . . . . . . . . . .

5-2 Equilibrium in steady, straight, symmetric Flight . . . . . . . . . . . . . . . . . . 5-2-1 5-2-2 5-2-3 5-2-4 5-2-5 5-2-6 5-2-7 Conditions for equilibrium . . . . . . . . . . . . . . . . . . . . . . . . . . Normal force on the horizontal tailplane . . . . . . . . . . . . . . . . . . Hinge moment of the elevator . . . . . . . . . . . . . . . . . . . . . . . Flow direction and dynamic pressure at the horizontal tailplane . . . . . . Eect of airspeed and center of gravity on tail load . . . . . . . . . . . . Elevator deection required for moment equilibrium . . . . . . . . . . . . Stick forces in steady ight . . . . . . . . . . . . . . . . . . . . . . . . .

Flight Dynamics

vi 6 Longitudinal Stability and Control Derivatives

Table of Contents 267 268 270 270 271 272 276 276 278 279 280 291 295 295 295 295 296 299 299 303 303 307 312 324 328 329 329 331 336 339 340 343 345 347 348 348

6-1 Aerodynamic forces in the nominal ight condition . . . . . . . . . . . . . . . . 6-2 Derivatives with respect to airspeed . . . . . . . . . . . . . . . . . . . . . . . . 6-2-1 6-2-2 6-2-3 6-3-1 6-3-2 6-3-3 Stability derivative CXu . . . . . . . . . . . . . . . . . . . . . . . . . . . Stability derivative CZu . . . . . . . . . . . . . . . . . . . . . . . . . . . Stability derivative Cmu . . . . . . . . . . . . . . . . . . . . . . . . . . . Stability derivative CX . . . . . . . . . . . . . . . . . . . . . . . . . . . Stability derivative CZ . . . . . . . . . . . . . . . . . . . . . . . . . . . Stability derivative Cm . . . . . . . . . . . . . . . . . . . . . . . . . . .

6-3 Derivatives with respect to angle of attack . . . . . . . . . . . . . . . . . . . . .

6-4 Derivatives with respect to pitching velocity . . . . . . . . . . . . . . . . . . . . 6-5 Derivatives with respect to the acceleration along the top axis . . . . . . . . . . 6-6 Derivatives with respect to the elevator angle . . . . . . . . . . . . . . . . . . . 6-6-1 6-6-2 6-6-3 Control derivative CXe . . . . . . . . . . . . . . . . . . . . . . . . . . . Control derivative CZe . . . . . . . . . . . . . . . . . . . . . . . . . . . Control derivative Cme . . . . . . . . . . . . . . . . . . . . . . . . . . .

6-7 Symmetric inertial parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Lateral Stability and Control Derivatives 7-1 Aerodynamic force and moments due to side slipping, rolling and yawing . . . . . 7-2 Stability derivatives with respect to the sideslip angle . . . . . . . . . . . . . . . 7-2-1 7-2-2 7-2-3 7-2-4 7-2-5 7-3-1 7-3-2 7-3-3 7-4-1 7-4-2 7-4-3 Stability derivative CY . . . . . . . . . . . . . . . . . . . . . . . . . . . Stability derivative C . . . . . . . . . . . . . . . . . . . . . . . . . . . Stability derivative Cn . . . . . . . . . . . . . . . . . . . . . . . . . . . Stability derivative Cn . . . . . . . . . . . . . . . . . . . . . . . . . . . Stability derivative Cm 2 . . . . . . . . . . . . . . . . . . . . . . . . . . Stability derivative CYp . . . . . . . . . . . . . . . . . . . . . . . . . . . Stability derivative Cp . . . . . . . . . . . . . . . . . . . . . . . . . . . Stability derivative Cnp . . . . . . . . . . . . . . . . . . . . . . . . . . . Stability derivative CYr . . . . . . . . . . . . . . . . . . . . . . . . . . . Stability derivative Cr . . . . . . . . . . . . . . . . . . . . . . . . . . . Stability derivative Cnr . . . . . . . . . . . . . . . . . . . . . . . . . . .

7-3 Stability derivatives with respect to roll rate . . . . . . . . . . . . . . . . . . . .

7-4 Stability derivatives with respect to yaw rate . . . . . . . . . . . . . . . . . . . .

7-5 The forces and moments due to aileron, rudder and spoiler deections . . . . . . 7-6 Aileron control derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-6-1 Control derivative CYa . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table of Contents 7-6-2 7-6-3 Control derivative Ca . . . . . . . . . . . . . . . . . . . . . . . . . . . Control derivative Cna . . . . . . . . . . . . . . . . . . . . . . . . . . .

vii 350 351 352 355 355 355 357 359 359 360 365 368 377 379 381 385 388 390 392 403 408 410 411 412 417 420 426 431 437 442 445 445 445 448 448 450 450

7-7 Spoiler control derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-8 Rudder control derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-8-1 Control derivative CYr . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-8-2 7-8-3 Control derivative Cr . . . . . . . . . . . . . . . . . . . . . . . . . . . Control derivative Cnr . . . . . . . . . . . . . . . . . . . . . . . . . . .

8 Longitudinal Stability and Control in Steady Flight 8-1 Stick Fixed Static Longitudinal Stability . . . . . . . . . . . . . . . . . . . . . . 8-1-1 8-1-2 8-1-3 8-1-4 8-1-5 8-1-6 8-2-1 8-2-2 8-2-3 8-2-4 8-2-5 8-2-6 8-2-7 8-3-1 8-3-2 8-3-3 8-3-4 8-3-5 8-3-6 Stick xed static longitudinal stability in gliding ight . . . . . . . . . . . Neutral point, stick xed . . . . . . . . . . . . . . . . . . . . . . . . . . Elevator trim curve and elevator trim stability . . . . . . . . . . . . . . . Inuence of various parameters on elevator trim curve . . . . . . . . . . . Static longitudinal stability of tailless aircraft . . . . . . . . . . . . . . . Determination Cme from measurements in ight . . . . . . . . . . . . . Stick free static longitudinal stability in gliding ight . . . . . . . . . . . Neutral point, stick free . . . . . . . . . . . . . . . . . . . . . . . . . . . Elevator stick force curves and elevator stick force Stability . . . . . . . . Eect of center of gravity and mass unbalance on stick force stability . . Inuence of design variables on control forces . . . . . . . . . . . . . . . Control forces a pilot can exert . . . . . . . . . . . . . . . . . . . . . . . Airworthiness requirements for steady straight symmetric ight . . . . . . Characteristics of longitudinal control in turning ight . . . . . . . . . . . The stick displacement per g . . . . . . . . . . . . . . . . . . . . . . . . The manoeuvre point, stick xed . . . . . . . . . . . . . . . . . . . . . . The stick force per g . . . . . . . . . . . . . . . . . . . . . . . . . . . . The manoeuvre point, stick free . . . . . . . . . . . . . . . . . . . . . . Non-aerodynamic means to inuence the stick force Per g . . . . . . . .

8-2 Stick Free Static Longitudinal Stability . . . . . . . . . . . . . . . . . . . . . . .

8-3 Longitudinal Control in Pull-Up Manoeuvres and Steady Turns . . . . . . . . . .

9 Lateral Stability and Control in Steady Flight 9-1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-2 Equations of Equilibrium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-3 Steady Horizontal Turns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-3-1 9-3-2 9-3-3 Turns Using the Ailerons Only, r = 0 . . . . . . . . . . . . . . . . . . . Turns Using the Rudder Only, a = 0 . . . . . . . . . . . . . . . . . . . . Coordinated Turns, = 0 . . . . . . . . . . . . . . . . . . . . . . . . . .

Flight Dynamics

viii

Table of Contents 9-3-4 Flat Turns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-4 Steady, Straight, Sideslipping Flight . . . . . . . . . . . . . . . . . . . . . . . . 9-5 Steady Straight Flight with One or More Engines Inoperative . . . . . . . . . . . 9-6 Steady Rolling Flight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-7 Control Forces and Hinge Moments for Lateral Control . . . . . . . . . . . . . . 9-7-1 9-7-2 Roll Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rudder Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453 453 458 463 470 470 472

III

Dynamic Stability Analysis

473
475 477 489 490 494 503 503 507 513 517 519 535

10 Analysis of Symmetric Equations of Motion 10-1 Solution of the equations of motion . . . . . . . . . . . . . . . . . . . . . . . . 10-2 Stability criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-3 The complete solution of the equations of motion . . . . . . . . . . . . . . . . . 10-4 Approximate solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Analysis of Asymmetric Equations of Motion 11-1 Solution of the equations of motion . . . . . . . . . . . . . . . . . . . . . . . . 11-2 General character of the asymmetric motions . . . . . . . . . . . . . . . . . . . 11-3 Routh-Hurwitz stability criteria for the asymmetric motions . . . . . . . . . . . . 11-4 Spiral and Dutch roll mode, the lateral stability diagram . . . . . . . . . . . . . 11-5 Approximate solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A List of Symbols

B Aircraft Parameters 547 B-1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 547 B-2 Linearized Equations of Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . 547 B-3 LTI-System Representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 548

C MATLAB Files 559 C-1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 559 C-2 Root Finding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 559 C-3 Eigenvalue Computation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C-4 Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 560 560

List of Figures

1 2 3

Denition of the mean aerodynamic cord and related parameters . . . . . . . . . Parameters dening the geometry of the wing . . . . . . . . . . . . . . . . . . . Examples of denitions of the vertical tailplane area (see also NACA TN 775) .

xxviii xxix xxx 3 5 6 6 7 7 9 10 11 12 14 15 21 23 23 24 25 26 27 28

1-1 The Wright Flyer I [136] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2 Boeing 767 3D-view [113] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-3 Concorde 3D-view [113] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-4 X-29 in ight [35] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-5 737-400 Flight and ground spoiler deployment on landing [32] . . . . . . . . . . 1-6 Boeing 737 slats [32] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-7 Mechanical ight control system example [132] . . . . . . . . . . . . . . . . . . 1-8 KC-135/707 Aileron Balance Panel Installation [132] . . . . . . . . . . . . . . . 1-9 Boeing 707 Flight Control Surfaces [132] . . . . . . . . . . . . . . . . . . . . . . 1-10 Boeing 707 Stabilizer-Actuated Tab Mechanism [132] . . . . . . . . . . . . . . . 1-11 Boeing 727 Aileron Control and Trim Systems [132] . . . . . . . . . . . . . . . . 1-12 Boeing 747 Elevator Control System [132] . . . . . . . . . . . . . . . . . . . . . 2-1 Aircraft velocity vector decomposition . . . . . . . . . . . . . . . . . . . . . . . 2-2 Denition of ecliptic plane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3 Denition of earth-centered inertial reference frame . . . . . . . . . . . . . . . . 2-4 Polar motion - precession track . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-5 Denition of earth-centered, earth-xed reference frame . . . . . . . . . . . . . . 2-6 Earth geoid and latitude denitions . . . . . . . . . . . . . . . . . . . . . . . . . 2-7 Normal earth-xed reference frame for spherical earth . . . . . . . . . . . . . . . 2-8 Vehicle carried normal earth reference frame for spherical earth . . . . . . . . . .

Flight Dynamics

x 2-9 Body-xed Reference Frame

List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 29 31 32 33 36 37 37 38 40 40 41 44 48 50 53 57 59 61 64 71 73 73 74 85 86 101 107 120 123 125 127 133 141

2-10 Stability Reference Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-11 Aerodynamic Reference Frame in relation to Body-xed Reference Frame . . . . 2-12 Left-handed (!) Aircraft Reference Frame Fr . . . . . . . . . . . . . . . . . . . 2-13 Euler angles dene the attitude of Fb irrespectively of translation . . . . . . . . . 2-14 Euler angles tutorial: initial and end body orientation . . . . . . . . . . . . . . . 2-15 Euler angles tutorial: rotation sequence z y x . . . . . . . . . . . . . . 2-16 Euler angles tutorial: rotation sequence x y z . . . . . . . . . . . . . . 2-17 Simultaneous vector decomposition in two reference frames . . . . . . . . . . . 2-18 Vector decomposition for rotation about X-axis . . . . . . . . . . . . . . . . . . 2-19 Vector decomposition for rotation about Y -axis . . . . . . . . . . . . . . . . . . 2-20 Vector decomposition for rotation about Z-axis . . . . . . . . . . . . . . . . . . 2-21 Transformation from inertial reference frame to normal earth-xed reference frame 2-22 Transformation from vehicle carried normal earth reference frame FO to the bodyxed reference frame Fb . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-23 Transformation from vehicle carried normal earth reference frame FO to the aerodynamic reference frame Fa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-24 Transformation from the body-xed reference frame Fb to the aerodynamic reference frame Fa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-25 General situation for vector dierentiation . . . . . . . . . . . . . . . . . . . . . 2-26 Change of a vector due to rotation vector . . . . . . . . . . . . . . . . . . . . 2-27 Example - derivation of pilot velocity with respect to FO . . . . . . . . . . . . . 2-28 Angular velocity vector E EI explanation . . . . . . . . . . . . . . . . . . . . . . 2-29 Euler angle singularity: Aircraft in vertical climb . . . . . . . . . . . . . . . . . . 2-30 Direction cosines equality for Euler axis rotations . . . . . . . . . . . . . . . . . 2-31 Rotating a reference frame about the Euler axis . . . . . . . . . . . . . . . . . . 2-32 Rotating an arbitrary vector about the Euler axis . . . . . . . . . . . . . . . . . 3-1 Vector denition for point mass P in a body. . . . . . . . . . . . . . . . . . . . 3-2 Moment about center of mass due to force on point mass . . . . . . . . . . . . . 3-3 Earth gravity vector denition . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4 Engine rotation axis denition . . . . . . . . . . . . . . . . . . . . . . . . . . .

4-1 Linearization about a single point and a trajectory . . . . . . . . . . . . . . . . . 4-2 Example of a time-function: u(.) . . . . . . . . . . . . . . . . . . . . . . . . . . 4-3 Airow tutorial: example of time-dependent aerodynamic characteristics . . . . . 4-4 The impossibility of a force in the XB -direction arising from a variation in velocity dv along the YB -axis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-5 The angles between the principal inertial axes and the aircraft reference axes . . 4-6 Relation between moment of inertia Iy and the mass . . . . . . . . . . . . . . .

List of Figures 5-1 The aerodynamic forces and the moment acting on a wing in symmetric ight .

xi 156 158 159 159 160 161 162

5-2 CN , CL and CT , CD as functions of for the case of a Fokker F-27 wing (from reference [149]) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-3 CN as a function of CT for a Fokker F-27 wing (from reference [149]) . . . . . 5-4 Rectangular wing conguration with zero-twist, = 1 , as used for linear potential 3 ow simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-5 Simulated CN , CL and CT , CD as functions of for the wing conguration depicted in gure 5-4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-6 Simulated CN as a function of CT and CL as a function of CD for the wing conguration depicted in gure 5-4 . . . . . . . . . . . . . . . . . . . . . . . . 5-7 The variation of the moment with changes in the reference point . . . . . . . .

5-8 Location of point (x, z) with respect to the position of the mean aerodynamic chord 162 5-9 Moment curves for various positions of the reference point for the Fokker F-27 wing (from reference [149]) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-10 Simulated moment curves for various positions of the reference point for the wing conguration of gure 5-4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-11 Denition of the angle 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 165 166

5-12 Denition of the center of pressure relative on the mean aerodynamic chord (mac) 167 5-13 The magnitude of C R and the position of the line of action of C R as a function of the angle of attack for Fokker F-27 wing (from reference [149]) . . . . . . . . 5-14 The position of the center of pressure as a function of the angle of attack for Fokker F-27 wing (from reference [149]) . . . . . . . . . . . . . . . . . . . . . . 5-15 The aerodynamic forces and moment about an arbitrary point at two adjacent values of the angle of attack . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-16 The line of action of the dierence between aerodynamic force vectors at adjacent values of the angle of attack . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-17 The position of the aerodynamic center . . . . . . . . . . . . . . . . . . . . . . 5-19 The moment about an arbitrary point (x, z) if the position of the ac and the Cmac are known . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-20 Equilibrium and stability of a wing, suspension point ahead of the ac, Cmac negative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-21 Equilibrium and stability of a wing, suspension point ahead of the ac, Cmac positive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-22 Equilibrium and stability of a wing, suspension point in the aerodynamic center, Cmac equals 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-23 Equilibrium and stability of a wing, suspension point in the aerodynamic center, Cmac negative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 168 169 170 172

5-18 The aerodynamic forces and moment, using the ac as the moment reference point 174 176 178 179 180 181

5-24 Equilibrium and stability of a wing, suspension point behind the ac, Cmac negative 182 5-25 Equilibrium and stability of a wing, suspension point behind the ac, Cmac positive 183 5-26 The basic, additional and total lift distribution of a wing with negative twist . . 185

Flight Dynamics

xii

List of Figures 5-27 Numerical simulation of the basic lift distribution (cb c at = CL =0 ), additional lift distribution (ca c) and total lift distribution (c c) for a wing without wingtwist (left, = 0o ) and a wing including wing-twist (right, = 3o ); NACA 0012 airfoil, b = 16 m, S = 32 m2 , c = 2 m, = 0o , = 0o , = 1, r = 0o , = 3o , = 5o . Results from a source/doublet singularity panelmethod, linearized potential ow. . . . . . . . . . . . . . . . . . . . . . . . . . 5-28 Numerical simulation of the basic lift distribution (cb c at = CL =0 ), additional lift distribution (ca c) and total lift distribution (c c) for a wing without wingtwist (left, = 0o ) and a wing including wing-twist (right, = 3o ); NACA 0012 airfoil, b = 16 m, S = 32 m2 , c = 2 m, = 0o , = 37o , = 1, r = 0o , = 3o , = 5o . Results from a source/doublet singularity panelmethod, linearized potential ow. . . . . . . . . . . . . . . . . . . . . . . . . . 5-29 Numerical simulation of the basic lift distribution (cb c at = CL =0 ), additional lift distribution (ca c) and total lift distribution (c c) for a wing without wingtwist (left, = 0o ) and a wing including wing-twist (right, = 3o ); NACA 0012 airfoil, b = 16 m, S = 32 m2 , c = 2 m, = 0o , = 0o , = 1 , 3 r = 0o , = 3o , = 5o . Results from a source/doublet singularity panelmethod, linearized potential ow. . . . . . . . . . . . . . . . . . . . . . . . . . 5-30 Numerical simulation of the basic lift distribution (cb c at = CL =0 ), additional lift distribution (ca c) and total lift distribution (c c) for a wing without wingtwist (left, = 0o ) and a wing including wing-twist (right, = 3o ); NACA 0012 airfoil, b = 16 m, S = 32 m2 , c = 2 m, = 0o , = 37o , = 1 , 3 r = 0o , = 3o , = 5o . Results from a source/doublet singularity panelmethod, linearized potential ow. . . . . . . . . . . . . . . . . . . . . . . . . . 5-31 Numerical simulation of the basic lift distribution (cb c at = CL =0 ), additional lift distribution (ca c) and total lift distribution (c c) for a wing without wingtwist (left, = 0o ) and a wing including wing-twist (right, = 3o ); NACA 0012 airfoil, b = 16 m, S = 32 m2 , c = 2 m, = 5o , = 0o , = 1 , 3 o , = 3o , = 5o . Results from a source/doublet singularity panelr = 0 method, linearized potential ow. . . . . . . . . . . . . . . . . . . . . . . . . . 5-32 Numerical simulation of the basic lift distribution (cb c at = CL =0 ), additional lift distribution (ca c) and total lift distribution (c c) for a wing without wingtwist (left, = 0o ) and a wing including wing-twist (right, = 3o ); NACA 0012 airfoil, b = 16 m, S = 32 m2 , c = 2 m, = 5o , = 37o , = 1 , 3 r = 0o , = 3o , = 5o . Results from a source/doublet singularity panelmethod, linearized potential ow. . . . . . . . . . . . . . . . . . . . . . . . . . 5-33 The calculation of the position of the ac and Cmac for a wing with arbitrary geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-34 The inuence of camber on the moment curves (from references [119, 118]) . .

187

188

189

190

191

192 193 196 198 199 201 202

5-35 The variation of Cmac with wing sweep and angle of twist (from reference [49])197 5-36 The positions of the local acs of swept wings and delta wings . . . . . . . . . . 5-37 The inuence of wing sweep on the moment curves of wings of aspect ratio A = 4, t = 0.6, c = 0.06, M = 0.40 (from reference [169]) . . . . . . . . . . . . . . . 5-38 The inuence of taper ratio on the moment curves of swept wings, A = 4, = t c = 0.06, M = 0.40 (from reference [87]) . . . . . . . . . . . . . . . . . . . . . 45o ,

5-39 The position of the ac of delta wings and concept wings as a function of aspect ratio, = 45o (from reference [159]) . . . . . . . . . . . . . . . . . . . . . . .

List of Figures 5-40 The inuence of aspect ratio on the moment curve of swept wings, = 45o , t = 0.6, c = 0.06, M = 0.40 (from reference [97]) . . . . . . . . . . . . . . . . 5-41 The moment curve of a straight wing and a swept wing, A = 5, = 1, Re = 1106 (DUT measurements) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-42 Boundaries for the combinations of aspect ratio and sweep angle for which destabilizing changes in the moment curves may be expected (from references [58, 75]) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-43 Pressure distribution over a fuselage in inviscid ow . . . . . . . . . . . . . . . 5-44 Numerical simulation of particle lines (top) and pressure distribution over a fuselage in inviscid ow, = 10o , linearized potential ow . . . . . . . . . . . . . . . . . 5-45 The variation of the angle of attack and the normal force along the fuselage axis in a wing induced ow eld . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-46 Wing-fuselage conguration as used in the ow eld simulations depicted in gures 5-47 and 5-48 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-47 Numerical simulation of the velocity eld of a wing-fuselage conguration, = 10o 5-48 Numerical simulation of the velocity eld of a wing-fuselage conguration, = 10o , detail of gure 5-47 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
f 5-49 The inuence of aspect ratio on the variation of d along the fuselage axis ( = 0) (from reference [143]) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . f 5-50 The inuence of sweep angle of wings with A = on d along the fuselage () (from reference [143]) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xiii

203 204

205 208 209 211 212 .212 213 214 214

5-51 Fuselage induced variations of the local angle of attack along the wing span . . 215 5-52 Numerical simulation of the fuselage induced upwash along the wing span, = 5o 215 5-53 The c c-distribution along the span of a wing with and without a fuselage (from reference [54]) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-54 The change in wing moment due to loss in lift over the wing center part . . . . 5-55 Shift of ac due to fuselage eects, as a function of wing sweep angle(from reference [141]) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-56 Flow simulations over a nacelle, = computed with a source/doublet singularity panel-method, linearized potential ow . . . . . . . . . . . . . . . . . . . 5-57 Generic Large Transport Aircraft (GLTA), no nacelles . . . . . . . . . . . . . . . 5-58 Particle lines in an XB OZB -plane (top) and non-dimensional sectional pressure distribution Cp (bottom) for the left wing of a generic transport aircraft model, no nacelles, = 10o , computed with a source/doublet singularity panel-method, linearized potential ow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-59 Particle lines in an XB OZB -plane (top) and non-dimensional sectional pressure distribution Cp (bottom) for the left wing of a twin-engined transport aircraft , = 10o , computed with a source/doublet singularity panel-method, linearized potential ow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-60 Spanwise lift distribution c c at an angle-of-attack = 10o twin-engined transport aircraft conguration , source/doublet singularity panel-method, linearized potential ow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-61 Particle lines in an XB OZB -plane and non-dimensional sectional pressure distribution coecients Cp for the left wing of a four-engined transport aircraft , = 10o , source/doublet singularity panel-method, linearized potential ow . . . . . . . . 15o , 216 217 218 219 220

221

222

223

224

Flight Dynamics

xiv

List of Figures 5-62 Four-engined transport aircraft with spanwise lift distribution c c at an angle-ofattack = 10o , source/doublet singularity panel-method, linearized potential ow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-63 The equilibrium in steady, symmetric ight . . . . . . . . . . . . . . . . . . . . 5-64 The forces and moments in steady, straight, symmetric ight . . . . . . . . . . 5-65 Geometry parameters of the horizontal tailplane, elevator and trim tab, and their positive deections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-66 Simplied version of the equilibrium of forces and moments . . . . . . . . . . . 5-67 Pressure distributions over the tailplane chord due to h , e and te . . . . . . . 5-68 The normal force coecient CNh as a function of h , e and te for the tailplane of the Fokker F-27 (Wind tunnel measurements from reference [17]) . . . . . . 5-69 The quotient reference [21])
CNh CNh

225 227 229 231 234 236 237

as a function of

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ce ch

for various trailing edge angles (from 238 239 242 243 244 244

5-70 The hinge moment coecient Che as a function of h , e and te for a tailplane of the Fokker F-27 (wind tunnel measurements from reference [17]) . . . . . . . 5-71 Overbalance with respect to angle of attack and elevator angle. . . . . . . . . . 5-72 The angle of attack h of the horizontal tailplane, h = + ih . . . . . . . 5-73 The contribution of the lifting and free vortices to the induced vertical velocities in front of and behind a wing . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-74 The position of the vortex plane behind the wing . . . . . . . . . . . . . . . . . 5-75 The vertical displacement of the wake and the tailplane relative to the ow eld
xh

due to an increase in angle of attack ; zh =

xh xw

dx = (xh xw ),

zwake =
xw

(x) dx, zwake < zh as (x) is smaller than . With 245 246 247

increasing angle of attack the horizontal stabilizer moves downwards through the wake . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-76 Numerical simulation of the deformation and wake roll-up of the vortex sheet behind a wing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
d 5-77 Computed values of d behind a wing of elliptical lift distribution, as a function of aspect ratio and distance behind the wing . . . . . . . . . . . . . . . . . . . . .

5-78 Numerical simulation of for wings with varying taper ratio (CL = 1.0, = 0o , A = 6), computed with a source/doublet singularity method, also see reference [143] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
c 5-79 Numerical simulation of c2b for wings with varying sweep angle (CL = 1.0, = 1, A = 6), computed with a source/doublet singularity method, also see reference [143] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

c c 2b

248

249

5-80 Numerical simulation of downwash angles in the plane of symmetry behind a wing ( = 13o , NACA 23014 airfoil, A = 6, = 0.5 and CL = 1.1615, , computed with a source/doublet singularity method, see also reference [146]) . . . . . . . 5-81 Cessna Ce550 Citation II conguration with and without nacelles . . . . . . . . .
d 5-82 The eect of nacelles on d (at X = 5.0 m, Z = 1.5 m) aft of the wing of a Cessna Ce550 Citation II, computed with a source/doublet singularity panelmethod, linearized potential ow . . . . . . . . . . . . . . . . . . . . . . . . . .

250 252

253

List of Figures 5-83 Calculation of ground eect . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xv 254 256 257 258

5-84 The ground eect on the downwash angle and the location of the wake relative to the horizontal tailplane of a Siebel 204 D-1 aircraft . . . . . . . . . . . . . . 5-85 Simplied picture of the equilibrium of moments . . . . . . . . . . . . . . . . . 5-86 The variation of the tail toad with airspeed at dierent c.g. positions and values of Cmac . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-87 The tail load as a function of airspeed at two c.g. positions for De Havilland Mosquito II F (from reference [26]), W = 6800 kg, S = 41.8 m2 , lh = 8.0 m and c = 2.81 m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-88 The positive direction of control deections, control forces, control surface deections and hinge moments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-89 The relation between the control force and the hinge moment . . . . . . . . . . 5-90 Contributions to the hinge moment . . . . . . . . . . . . . . . . . . . . . . . .

259 263 263 264 268

6-1 The attitude of the stability reference frame relative to the Xr - and Zr -axis, after a disturbance from the equilibrium ight condition . . . . . . . . . . . . . . . . 6-2 The denition of aircraft parts wing, horizontal stabilizer, pylon, nacelles, vertical n and fuselage for the Cessna Ce550 Citation II model (starting from the left top gure, in clockwise direction) . . . . . . . . . . . . . . . . . . . . . 6-3 Calculated contributions of various aircraft parts to the force curve CX versus for the Cessna Ce550 Citation II . . . . . . . . . . . . . . . . . . . . . . . . . 6-4 Calculated contributions of various aircraft parts to the force curve CZ versus for the Cessna Ce550 Citation II . . . . . . . . . . . . . . . . . . . . . . . . . 6-5 Calculated contributions of various aircraft parts to the moment curve Cm versus for the Cessna Ce550 Citation II . . . . . . . . . . . . . . . . . . . . . . . . 6-6 The pure q-motion, with =
xxc.g. R

277 278 280 281 282 283 284 285 286 286 287 287 289 291 292 300 301 302

xxc.g. q c c V

. . . . . . . . . . . . . . .

6-7 Harmonic q-motion, = constant, = . . . . . . . . . . . . . . . . . . . . . 6-8 Harmonic -motion, q = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-9 Combined harmonic q- and -motion, q = , = constant 6-10 Aircraft in a curved ow eld . . . . . . . . . . . . 6-11 Equivalent curved aircraft in a straight ow eld with Fr . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-12 Equivalent curved aircraft in a straight ow eld . . . . . . . . . . . . . . . . . . . . . . . . . . . . aircraft frame of reference, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
q c V q c V

6-13 Equivalent curved aircraft in a straight ow eld, 3-dimensional view

6-14 Calculated contributions of various aircraft parts to the force curve CX versus for the Cessna Ce550 Citation II . . . . . . . . . . . . . . . . . . . . . . . . . 6-15 Calculated contributions of various aircraft parts to the force curve CZ versus for the Cessna Ce550 Citation II . . . . . . . . . . . . . . . . . . . . . . . . . 6-16 Calculated contributions of various aircraft parts to the moment curve Cm versus q c V for the Cessna Ce550 Citation II . . . . . . . . . . . . . . . . . . . . . . . . 7-1 The six degrees of freedom of a rigid aircraft 7-3 Asymmetric force and moments . . . . . . . . . . . . . . . . . . .

7-2 Denition of the angle of attack and sideslip angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Flight Dynamics

xvi

List of Figures 7-4 CY as a function of measured on a model of the Fokker F-27 in gliding ight (from references [148, 150, 24]) . . . . . . . . . . . . . . . . . . . . . . . . . . 7-5 Relation between the sideslip angle , the sidewash angle and the angle of attack v at the vertical tailplane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-6 Aerodynamic force coecient CY as a function of for the Cessna Ce550 Citation II, = 0o . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-7 The origin of the dierence in angles of attack at the left and right wing for a wing with dihedral in sideslipping ight . . . . . . . . . . . . . . . . . . . . . . . . . 7-8 The origin of the dierence in the velocities over the left and right wing for a swept wing in sideslipping ight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-9 The variation of C with CL for swept wings . . . . . . . . . . . . . . . . . . . 7-10 High-wing-fuselage conguration with the calculated velocity eld in sideslipping ight ( = 10o , = 0o ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-11 Low-wing-fuselage conguration with the calculated velocity eld in sideslipping ight ( = 10o , = 0o ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-12 Calculated particle traces around the bottom half of the Ce550 Citation II, sideslipping ight = 10o , = 0o related to the magnitude of the airows velocity along traces) . . . . . . . . . . . . . . . . . . . . . . . . . . . fuselage of the Cessna (note: the speedbar is the calculated particle . . . . . . . . . . . . .

304 306 307 309 310 310 313 314

315 316 316 317 319 320 321

7-13 Origin of a rolling moment caused by wing-fuselage interactions in sideslipping ight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-14 Aerodynamic moment coecient C as a function of for the Cessna Ce550 Citation II, = 0o . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-15 The position of the point of action of (CYv )v relative to the X- and Z-axis in the stability reference frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-16 The contribution of the slipstream to C . . . . . . . . . . . . . . . . . . . . . 7-17 Cn as a function of measured on a model of the Fokker F-27 in gliding ight (from references [148, 150, 24]) . . . . . . . . . . . . . . . . . . . . . . . . . . 7-18 The sidewash induced by the fuselage at the vertical tailplane . . . . . . . . . . 7-19 Calculated particle traces around the fuselage of the Cessna Ce550 Citation II, sideslipping ight = 10o , = 0o (note: the speedbar is related to the magnitude of the airows velocity along the calculated particle traces) . . . . . . . . . . . 7-20 Calculated particle traces around the fuselage of a large 4-engined transport aircraft, sideslipping ight = 10o , = 0o (note: the speedbar is related to the magnitude of the airows velocity along the calculated particle traces) . . . . . 7-21 The eect of wing-fuselage interactions on the sidewash at the vertical tailplane of a low-wing aircraft in sideslipping ight . . . . . . . . . . . . . . . . . . . . . 7-22 The eect of wing-fuselage interactions on the sidewash at the tailplanes and the derivative Cn (from reference [142]) . . . . . . . . . . . . . . . . . . . . . . . 7-23 The contribution of the sideforce on the propeller to Cn . . . . . . . . . . . . 7-24 Aerodynamic moment coecient Cn as a function of for the Cessna Ce550 Citation II, = 0o . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-25 The pitching moment as a function of sideslip angle (from reference [126]) . . . 7-26 The inuence of the vertical position of the horizontal tailplane on the variation of Cm in sideslipping ight for a fuselage with tailplanes (from reference [126]) . .

322

323 325 326 327 327 328 330

List of Figures 7-27 The variation of the local geometric angle of attack at the wing and the tailplane of a rolling aircraft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-28 Aerodynamic force coecient CY as a function of for the Cessna Ce550 Citation II, = 0o . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-29 The variation of the local geometric and eective angle of attack along the span of a rolling wing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-30 The roll-damping as a function of aspect ratio and sweep for various wing taper ratios (from reference [34]) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-31 The ow at the tailplanes of a rolling aircraft . . . . . . . . . . . . . . . . . . .
pb 7-32 Aerodynamic moment coecient C as a function of 2V for the Cessna Ce550 o Citation II, = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-33 The origin of a negative yawing moment about the Z-axis of the stability reference frame for a wing having a positive rate of roll (attached ow) . . . . . . . . . . pb 2V

xvii

330 331 332 333 335 336 337 338 339 341 342 342 343 344 345 346 347 349 350 351 353 354 356 358 362 363 364

7-34 The side force and yawing moment on a rolling, swept back wing

. . . . . . . .

pb 7-35 Aerodynamic moment coecient Cn as a function of 2V for the Cessna Ce550 o Citation II, = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-36 The pure r- motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-37 The change in geometric ow direction at an arbitrary point of the aircraft due to an r-motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-38 The variation in airspeed in spanwise direction due to an r-motion . . . . . . . 7-39 The side forces on the vertical tailplane and the propeller due to an r-motion. . rb 7-40 Aerodynamic force coecient CY as a function of 2V for the Cessna Ce550 Citao . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . tion II, = 0

7-41 Aerodynamic moment coecient C as a function of for the Citation II, = 0o . . . . . . . . . . . . . . . . . . . . . . . . 7-42 The eects of the size of the vertical tailplane and the taillength reference [88]) . . . . . . . . . . . . . . . . . . . . . . . . . . .

rb 2V

Cessna Ce550 . . . . . . . . on Cnr (from . . . . . . . .

rb 7-43 Aerodynamic moment coecient Cn as a function of 2V for the Cessna Ce550 o . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Citation II, = 0 7-44 The positive direction of control deections, control forces, control surface deections and hinge moments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-45 The eect of an aileron deection on the lift distribution in spanwise direction . 7-46 The Frise aileron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-47 Spoiler types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-48 The local wing twist of a wing cross-section due to elastic deformation caused by an aileron deection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-49 The side force and yawing moment due to a rudder deection . . . . . . . . . . 7-50 Asymmetric stability and control derivatives . . . . . . . . . . . . . . . . . . . .

8-1 The contribution of various parts of the aircraft to the moment curve Cm . 8-2 Measured contributions of various aircraft parts to the moment curve of the Fokker F- 27, reference point at 0.346 c (from reference [147]) . . . . . . . . . . . . . 8-3 Calculated contributions of various aircraft parts to the moment curve of the Cessna Ce550 Citation II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Flight Dynamics

xviii

List of Figures

8-4 Inuence of the tailplane incidence on the moment curve of the Fokker F-27, reference point at 0.346 c (from reference [147]) . . . . . . . . . . . . . . . . . 8-5 Inuence of the position of the reference point (center of gravity) on the moment curves of the Fokker F-27, (from reference [147]) . . . . . . . . . . . . . . . . . 8-6 Change in the moment dCm due to a change in the angle of attack 8-7 Moment curves and corresponding trim curves . . . . . . . . . . . . . . . . . . . . . .

365 367 368 370 372 373 374 375 378 380 382 384 386 387 390 392 394 397 398 399 404 406 407 407 408 411 415 416 419

8-8 Initial control surface deections . . . . . . . . . . . . . . . . . . . . . . . . . . 8-9 Initial and ultimate control displacement and control surface deection for the transition to a lower airspeed . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-10 Aircraft responses to a control deection . . . . . . . . . . . . . . . . . . . . . 8-11 Inuence of cg position on the trim curve . . . . . . . . . . . . . . . . . . . . . 8-12 Inuence of the tailplane angle of incidence on the trim curve (aircraft has control position stability) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-13 A positive Cma.c is obtained by combining sweepback and negative wing twist . 8-14 The determination of elevator eciency from ight tests . . . . . . . . . . . . . 8-15 Trim curves for two cg positions of the Fokker F-27 (from reference [51]) . . . . 8-16 Some stability characteristics derived from the measured trim curves shown in gure 8-15 (Fokker F-27) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-17 The equilibrium of the free elevator . . . . . . . . . . . . . . . . . . . . . . . . 8-18 The change in the moment dCmf ree due to a change in angle of attack . . . 8-19 The behaviour of the free elevator after a change of the angle of attack . . . . . 8-20 Schematic form of the elevator control force curve, Fe as a function of (a) dynamic pressure 1 V 2 and (b) Fe as a function of equivalent airspeed Ve . . . . . . . . 2 8-21 Measured trim curves and elevator control force curves for the De Havilland D.H.98 Mosquito M II F in gliding ight (from reference [27]) . . . . . . . . . . . . . 8-22 Schematic representation of the concept of positive feel,
dFe dse

. . . . . . . . . . . . . . . . .

8-23 Uncertainty in trim speed due to friction in the control mechanism

8-24 Measured elevator control force curves for the North American Harvard II B in gliding ight (from reference [15]) . . . . . . . . . . . . . . . . . . . . . . . . . 8-25 Trim curves, hinge moment coecients and required trim tab angles as functions of CL for the Siebel 204-D-1 aircraft (from reference [16]) . . . . . . . . . . . . 8-26 The inuence of a bobweight in the control mechanism . . . . . . . . . . . . . 8-27 The inuence of a spring in the control mechanism . . . . . . . . . . . . . . . . 8-28 The inuence of a spring or bobweight on the elevator control force curve . . . . 8-29 Maximum control forces as a function of the duration (from reference [140]) . . 8-30 Response curves due to an elevator step deection, Lockheed 1049 C Super Constellation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-31 Response curves due to an elevator step deection, Auster J-5B Autocar (from reference [23]) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-32 Description of the character of the control feel for various ratios of the control force per g to the control displacement per g (from reference [170]). . . . . . .

List of Figures 8-33 The inuence of the stick displacement per g and the stick force per g on the pilots opinion of the control characteristics at constant airspeed (from reference [86]). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-34 The forces and angular velocity q in the idealized pull-up manoeuvre . . . . . . 8-35 The forces and angular velocity in a steady horizontal turn . . . . . . . . . . . . 8-36 The incremental elevator deection e as a function of the incremental load factor n in pull-up manoeuvres, Auster J-5B Autocar (from reference [23]). . 8-37 The incremental elevator angle e as a function of the incremental load factor in turns, Auster J-5B Autocar (from reference [23]) . . . . . . . . . . . . . . . . 8-38 Calculated positions of the manoeuvre point, stick xed, and the stick displacement per g of the Fokker F-27 Friendship . . . . . . . . . . . . . . . . . . . . . . . 8-39 The variation of the angle of attack in an arbitrary point of the aircraft caused by a pure q-motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-40 The incremental control force Fe as a function of the incremental load factor n in pull-up manoeuvres, Auster J-5B Autocar (from reference [23]) . . . . . 8-41 The incremental control force Fe as a function of the incremental load factor n in turns, Auster J-5B Autocar (from reference [23]) . . . . . . . . . . . . . 8-42 Calculated positions of the manoeuvre point, stick free, and the stick force per g of the Fokker F-27 Friendship . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-43 Inuence of altitude and magnitude and sign of Ch on the position of the manoeuvre point, stick free, and the stick force per g . . . . . . . . . . . . . . . . 8-44 Increasing the range of permissible cg positions by decreasing |Ch | . . . . . . . 8-45 The incremental hinge moment due to a bobweight in the control mechanism in a pull-up manoeuvre at a load factor n . . . . . . . . . . . . . . . . . . . . . . .

xix

419 423 424 427 428 430 432 438 438 440 441 443 444

9-1 The forces along the YB -axis of an aircraft in steady, horizontal, asymmetric ight 446 9-2 Steady turns using ailerons only, North American Harvard II B, gliding ight, CL = 0.31, xc.g. = 0.304 c, V = 78 m/sec (from reference [14]) . . . . . . . . . 9-3 Steady turns using rudder only, North American Harvard II B, gliding ight, CL = 0.31, xc.g. = 0.304 c, V = 78 m/sec (from reference [14]) . . . . . . . . . 9-4 Steady coordinated turns, North American Harvard II B, gliding ight, CL = 0.31, xc.g. = 0.304 c, V = 78 m/sec (from reference [14]) . . . . . . . . . . . . . . . 9-5 Steady at turns, North American Harvard II B, gliding ight, CL = 0.31, xc.g. = 0.304 c, V = 78 m/sec (from reference [14]) . . . . . . . . . . . . . . . . . . . 9-6 Steady, straight, sideslipping ight, North American Harvard II B, gliding ight, CL = 0.31, xc.g. = 0.304 c, V = 78 m/sec (from reference [14]) . . . . . . . . . 9-7 The magnitude of Cne due to an engine failure, as a function of the location and the sense of rotation of the propeller and the vertical position of the wing, measured on a model of a twin-engined propeller-driven aircraft (from reference [105]). . . 9-8 The magnitude of Cne due to an engine failure, as a function of the location and the sense of rotation of the propeller and the vertical position of the wing, measured on a model of a twin-engined propeller-driven aircraft (from reference [105]). . . 9-9 Steady, straight, single-engined ight at = 0 9-10 Steady, straight, single-engined ight at = 0 9-11 Minimum control speed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449 451 454 456 457

459

461 462 463 464

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Flight Dynamics

xx

List of Figures 9-12 Steady, straight, sideslipping ight with the right propeller feathered, Fokker F-27, h = 1850 m, CL = 1.92, xc.g. = 0.28 c, V = 46.3 m/sec (from reference [51]) . 9-14 Variation of the wing twist angle along the wing span due to aileron deection for two locations of the ailerons . . . . . . . . . . . . . . . . . . . . . . . . . . 9-15 The average value of along the aileron span in steady, rolling ight . . . . . 9-16 Roll-rate p as a function of airspeed V for various values of a and Fa . . . . .

465

9-13 The limitation in the maximum attainable roll-rate due to elastic wing deformation 467 468 469 470 483 486 493 495 496 497

10-1 Aperiodic motions corresponding to a real eigenvalue . . . . . . . . . . . . . . 10-2 Periodic motion corresponding to a pair of complex, conjugate eigenvalues 1,2 = j . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-3 The location of the eigenvalues for the symmetric motions . . . . . . . . . . . . 10-4 Response curves for a step elevator deection (e = 0.005 [Rad]) for the Cessna Ce500 Citation, phugoid response . . . . . . . . . . . . . . . . . . . . . . . . 10-4 (Continued) Response curves for a step elevator deection (e = 0.005 [Rad]) for the Cessna Ce500 Citation, phugoid response . . . . . . . . . . . . . . . . 10-5 Response curves for a step elevator deection (e = 0.005 [Rad]) for the Cessna Ce500 Citation, magnication of gure 10-4, short period response . . . . . . .

10-5 (Continued) Response curves for a step elevator deection (e = 0.005 [Rad]) for the Cessna Ce500 Citation, magnication of gure 10-4, short period response 498 11-1 The location of the eigenvalues = b V for the asymmetric motions . . . . . . b 11-2 Response curves for a pulse-shaped rudder deection (r = +0.025 [Rad] during 1 second) for the Cessna Ce500 Citation . . . . . . . . . . . . . . . . . . . . . 11-2 (Continued) Response curves for a pulse-shaped rudder deection (r = +0.025 [Rad] during 1 second) for the Cessna Ce500 Citation . . . . . . . . . . . . . . 11-2 (Continued) Response curves for a pulse-shaped rudder deection (r = +0.025 [Rad] during 1 second) for the Cessna Ce500 Citation . . . . . . . . . . . . . . 11-3 Response curves for a pulse-shaped aileron deection (a = +0.025 [Rad] during 1 second) for the Cessna Ce500 Citation . . . . . . . . . . . . . . . . . . . . . 11-3 (Continued) Response curves for a pulse-shaped aileron deection (a = +0.025 [Rad] during 1 second) for the Cessna Ce500 Citation . . . . . . . . . . . . . . 11-3 (Continued) Response curves for a pulse-shaped aileron deection (a = +0.025 [Rad] during 1 second) for the Cessna Ce500 Citation . . . . . . . . . . . . . . 11-4 The spiral motion of an aircraft . . . . . . . . . . . . . . . . . . . . . . . . . . 11-5 Characteristics of the Dutch Roll motion . . . . . . . . . . . . . . . . . . . . . 11-6 Roll- and yaw-rate characteristics of the Dutch Roll motion . . . . . . . . . . . 11-7 Lateral stability diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507 508 509 510 511 512 513 514 515 516 518

List of Tables

4-1 Non-dimensional parameters in the equations of motion, symmetric motions

. .

142 143 144 144 145 145 146 147 147 150 152 235 240

4-2 Non-dimensional parameters in the equations of motion, asymmetric motions . . 4-2 (Continued) Non-dimensional parameters in the equations of motion, asymmetric motions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-3 Stability derivatives, symmetric motions . . . . . . . . . . . . . . . . . . . . . . 4-4 Control derivatives, symmetric motions . . . . . . . . . . . . . . . . . . . . . . . 4-5 Moments and products of inertia, symmetric motions . . . . . . . . . . . . . . . 4-6 Stability derivatives, asymmetric motions . . . . . . . . . . . . . . . . . . . . . 4-7 Control derivatives, asymmetric motions . . . . . . . . . . . . . . . . . . . . . . 4-8 Moments and products of inertia, asymmetric motions . . . . . . . . . . . . . . 4-9 Symbols appearing in the general state-space representation of equation (4-46) . 4-10 Symbols appearing in the general state-space representation of equation (4-50) . 5-1 Shorthand notation for the normal force derivatives on a tailplane . . . . . . . . 5-2 Abbreviated notation for the hinge moment derivatives . . . . . . . . . . . . . . 6-1 Simplied formulae for the calculation of stability and control derivatives and coecients in the initial, steady ight condition for the symmetric motions (without the eects of propellers and jets, aeroelasticity and compressibility of the air). . 6-1 (Continued) Simplied formulae for the calculation of stability and control derivatives and coecients in the initial, steady ight condition for the symmetric motions (without the eects of propellers and jets, aeroelasticity and compressibility of the air). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2 Inuence of the center of gravity position on some stability derivatives . . . . .

297

298 298 413 418

8-1 Maximum permissable values of the required control forces, according to the U.S. and British Civil Airworthiness Regulations . . . . . . . . . . . . . . . . . . . . 8-2 Permissable values of the stick force per g in kg . . . . . . . . . . . . . . . . .

Flight Dynamics

xxii

List of Tables 10-1 Symmetric stability and control derivatives for the Cessna Ce500 Citation . . . 11-1 Asymmetric stability and control derivatives for the Cessna Ce500 Citation . . . B-1 Symmetric and asymmetric stability and control derivatives for the Cessna Ce500 Citation, Cruise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-2 Symmetric and asymmetric stability and control derivatives for the Fokker F-27 Friendship, Cruise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-4 Symmetric stability and control derivatives for the Learjet I, Approach . . . . . . B-5 Symmetric stability and control derivatives for the Beechcraft M99, Cruise . . . B-6 Symmetric stability and control derivatives for the Boeing 747-100, Approach . . B-7 Symmetric stability and control derivatives for the Boeing 747-100, Holding, aps up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-8 Symmetric stability and control derivatives for the Boeing 747-100, Approach, aps 33o . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-9 Symmetric stability and control derivatives for the Boeing 747-100, Landing . . B-10 Asymmetric stability and control derivatives for the Lockheed L1049C stellation, Cruise . . . . . . . . . . . . . . . . . . . . . . . . . . . B-11 Asymmetric stability and control derivatives for the Lockheed L1049C stellation, Approach . . . . . . . . . . . . . . . . . . . . . . . . . Super Con. . . . . . . Super Con. . . . . . . 492 506 550 551 552 552 553 553 554 554 555 555 556 556 557

B-3 Symmetric stability and control derivatives for the Cessna Ce-172 Skyhawk, Cruise 551

B-12 Asymmetric stability and control derivatives for the BAC-Aerospatiale Concorde, Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-13 Asymmetric stability and control derivatives for the North-American X-15 experimental aircraft, Cruise (unspecied) . . . . . . . . . . . . . . . . . . . . . . . . B-14 Asymmetric stability and control derivatives for the De Havilland Canada DHC-2 Beaver, Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Nomenclature

Before reading the main chapters of this book, we advise the reader to start with this chapter. We will introduce several key notations and characters which will be used throughout the book. Knowing these notations and characters will facilitate the reading of the book and will improve the understanding of the text. We have tried to keep the reasoning behind the notations the same throughout the book. Only one exception exists and is with respect to the use of subscripts (see page xxv).

Vectors
All vectors in this book are in three dimensional space, i.e. R3 , unless dened otherwise. All vectors are in boldface. The superscript denotes the reference frame in which the vector is expressed while the subscript indicates the particular parameter in question. For example:
b Vk

is the kinematic (subscript k) velocity (vector V) expressed in the body-xed reference frame (superscript b). The legend for the subscripts en superscripts are given in the section on page xxv. Sometimes a second subscript is added. In that case the subscript denotes the point of origin of the vector or to which point or body the vector properties belong. For example:
b Vk,G

is the kinematic velocity of point G expressed in the body-xed reference frame. Four types of vectors have xed component notations. Position vectors have components x, y, z while velocity vectors have components u, v, w. Acceleration components are denoted by ax , ay , az . Finally, the rotational velocity vector has components p, q, r. Other vectors do not have a xed components notation. The same rules regarding the superscript
Flight Dynamics

xxiv and the subscript notation apply for the components of the vectors.

Nomenclature

A vector can be dierentiated with respect to time which is always coupled to a reference frame (without a reference frame time dierentiation only has a pure theoretical meaning). The time derivative of vector with respect to reference frame Fb is written as: dX dt

Note that the derivative of a vector is again a vector. Thus the previous expression can be appended by a superscript denoting the reference frame in which the time derivative is expressed. In cases where the meaning is obvious, the superscript and/or the subscripts are sometimes dropped, cleaning the text of any unnecessary notations which would only cloud the text. If any notation is simplied, it is explicitly mentioned so in the text.

Reference frames
A crucial element in this book is the denition of reference frames. All reference frames are right-handed orthogonal unless stated otherwise and are dened by their origin and orientation of the axes. A reference frame is denoted by F and its axes by X, Y, Z. The subscript appended to these letters indicates the reference frame we are talking about. For example: Fb (GXb Yb Zb ) is the body-xed reference frame which has the origin in the vehicle center of mass G and has the axes Xb , Yb , Zb . A list of all reference frames used in this book is given next.

FI FE FO Fb Fs Fp Fz Fa Fk Fr

Inertial reference frame Normal Earth-xed reference frame Vehicle carried normal Earth reference frame Body-xed reference frame Stability reference frame Principle axis reference frame Zero-lift body axis reference frame Aerodynamic reference frame Kinematic reference frame Vehicle reference frame

(section (section (section (section (section (section (section (section (section (section

2-1-1) 2-1-3) 2-1-4) 2-1-5) 2-1-5) 2-1-5) 2-1-5) 2-1-6) 2-1-7) 2-1-8)

After chapter 2 the notations Fx , Fy , Fz are used to denote the total aerodynamic forces along the X, Y, Z-axis respectively. Although the context in which these parameters are used is dierent, care should be taken with the interpretation of the parameter F . Transformation matrix The orientation between reference frames is dened by a maximum of three Euler angles. The

Nomenclature

xxv

sequence of rotation combined with the set of angles enables us to transform any coordinate from one reference frame to another. A transformation matrix T is used to quickly transform a complete vector. The subscripts indicate the reference frames involved in the transformation. For example: Xb = Tba Xa where Xb is the vector X expressed in reference frame Fb , Tba is the matrix for the transformation from frame Fa to Fb , and Xa is the vector X expressed in reference frame Fa . Angular velocity vectors When the orientation between reference frames is time-variant, one can dene a rotation vector describing this change through time. The variable dening a rotation vector is . The subscripts indicate which reference frames are involved while the superscript denotes the reference frame in which the vector is expressed. For example: b ba is the rotation vector describing the angular velocity of reference frame Fb with respect to reference frame Fa (subscripts) expressed in frame Fb (superscript). One should be careful not to mix-up the interpretation of the subscripts for the angular velocity vectors and for the transformation matrices.

Superscripts and subscripts

At this point a distinction must be made between chapters 1 to 3 and the remainder of the book. Chapters 1 to 3 For the rst three chapters the main subject is the derivation of the equations of motion for the most general case of a spherical, rotating Earth. In those chapters the superscripts appended to a vector indicate the reference frame in which the vector is expressed. The possible superscripts are:

I E O b s p z a k r

Inertial reference frame Normal Earth-xed reference frame Vehicle carried normal Earth reference frame Body-xed reference frame Stability reference frame Principle axis reference frame Zero-lift body axis reference frame Aerodynamic reference frame Kinematic reference frame Vehicle reference frame
Flight Dynamics

xxvi

Nomenclature

The subscripts in chapters 1 to 3 denote either reference frames (in case of transformation matrices and angular velocity vectors) or indicate a particular form of the vector. The subscript for the reference frames are the same as for the superscripts given above. The other subscripts are: a k aerodynamic kinematic

Remaining chapters In the remainder of the book superscripts are not used often. One exception which occurs in chapter 4, is the superscript indicating the derivative with respect to parameter i, e.g. f = df . di The function of the subscript changes considerably. A subscript in chapter 4 and subsequent chapters indicates the partial derivative of the function with respect to the parameter indicated by the subscript. For example: Xu = X u

Geometric of aircraft parameters

The geometric aircraft parameters are used to determine the aerodynamic force and moments and to make the equations of motion dimensionless. Figure 2 illustrates the various parameters dening the geometry of the wing. The reference axes used in this gure are those of the Vehicle reference frame (see section 2-1-8). Wing area, S The area of the wing projection on the Xr OYr -plane. Often the wing is partially covered by the fuselage and the engine nacelles. The wing area is then calculated using straight line extensions of the wing leading and trailing edges through the fuselage and the nacelles. The wing area can be expressed as,
b +2

S=
b 2

c dy

where y is the coordinate in the Yr -direction. Wingspan, b The distance in Yr -direction between the wing tips. Mean aerodynamic chord, c (mac) is dened as,

Nomenclature

xxvii

1 c= S

b +2

c2 dy
b 2

One can dene the location and orientation of the mean aerodynamic cord within the OXr Zr plane (see gure 1). Four coordinates, xo , xe , z0 , ze ,determine the location and orientation of the mac. There denition is similar that of the mean aerodynamic cord: xo = zo = =
1 S +b/ 2

xo (y) c (y) dy xe =
b/ 2 +b/ 2

1 S

+b/ 2

xe (y) c (y) dy
b/ 2 +b/ 2

1 S

zo (y) c (y) dy
b/ 2 +b/ 2

ze =

1 S

ze (y) c (y) dy
b/ 2

1 S

(y) c (y) dy
b/ 2

Mean or geometric chord, cm Is very often used in the literature and is dened as: cm = S b

Taper ratio, A measure of the variation in chord length along the span. It is expressed by, = Aspect ratio, A Dened as, A= ct cr

b2 S

Wing sweep, The angle between the Yr -axis and the projection of the 1/4-chord line on the Xr OYr plane (gure 2). In some cases such as delta wings, only the angle between the Yr -axis and the projection of the wings leading edge on the Xr OYr -plane is given. Dihedral, The dihedral of a wing is the angle between the Yr -axis and the projection of the 1/4chord line on the Yr OZr -plane. Washout or wing-twist, Expresses the variation in direction of the local wing chord relative to the direction d of the chord line at the wing root. Neither the gradient of the washout ( d y ) nor the magnitude of wing sweep or dihedral need to be constant along the span. If necessary, these parameters are given as functions of the coordinate in span direction.
Flight Dynamics

xxviii

Nomenclature

Zr c zo ze

xo

xe

Xr

c(y) y(< 0) xo (y) xe (y) Xr xo Yr

Figure 1: Denition of the mean aerodynamic cord and related parameters

xe

Nomenclature

xxix

Yr

1 4 -chord

line

cr

c dy ct

y Xr b Zr

Yr O

Figure 2: Parameters dening the geometry of the wing

Flight Dynamics

xxx

Nomenclature

Figure 3: Examples of denitions of the vertical tailplane area (see also NACA TN 775)

Wing airfoil The shape of the cross section of the wing parallel to the plane of symmetry. The above geometric parameters apply not only to wings, but to tailplanes as well. Denitions of the geometric parameters for the elevator, or control surfaces in general, are given in chapter 7. It proves to be dicult to dene the geometry of vertical tailplanes in a way applicable to all aircraft. In particular the distinction between fuselage and vertical tailplane is often hard to make. In many instances the division between vertical tailplane and dorsal n is also more or less arbitrary. Usually the surface of the dorsal n is not considered to be part of the vertical tailplane. Figure 3 shows examples of denitions of vertical tailplanes, see reference [11]. Similar and other denitions are given in references [8] and [101].

Aircraft Congurations
The ight dynamic characteristics are dependent on the aircraft conguration. A full description of the aircraft conguration gives the aircrafts weight or mass, center of gravity position and internal as well as external loading, undercarriage position, control surface deections, ap angle, airbrake and spoiler deections. A description of the engine operating condition, such as throttle position, engine speed etcetera, is also required.

Nomenclature

xxxi

Some denitions as used in U.S. military requirements are briey described below. CR (Cruising Flight) Engine thrust or power for level ight at cruising speed, aps in the position for cruising ight, undercarriage retracted. L (Landing) Throttle closed, undercarriage down, aps in the position for landing. PA (Powered Approach) Undercarriage down, aps and airbrakes in the normal position for the powered approach, engine thrust or power for level ight at 1.15 VSL or the normal airspeed in the powered approach, if the latter is lower. NOTE: VSL is the stall speed (in the aircraft conguration for landing).

Flight Conditions
By specifying a specic ight condition, we specify (a part of) the state of the aircraft, i.e. we indicate which aircraft states are varying. This information can be used to deduce the equations of motion for that ight condition. The following denitions of ight conditions will be used throughout this book. Steady ight An aircraft is in steady ight if the components of the aerodynamic velocity vector VG (u, v, w) and the components of the body rotation vector (p, q, r) in the body-xed reference frame Fb are constant. In this ight condition the aerodynamics force vector components in Fb and the aircrafts pitch and roll attitude (, ) are constant too. Straight ight An aircraft is in a straight ight condition, if the ightpath vector is straight. Symmetric ight An aircraft is in symmetric ight, if the velocity vector of any point of the aircraft is parallel to the plane of symmetry (roll angle = 0). Slipping ight An aircraft is in slipping ight, if the velocity vector of the aircrafts center of gravity is not parallel to the plane of symmetry of the aircraft.

Flight Dynamics

xxxii

Nomenclature

Chapter 1 Introduction

This book discusses the theory of stability and control of aircraft at subsonic airspeeds. This book handles the theory of ight dynamics which describes the aircraft/spacecraft velocities (translational and rotational), position, and orientation in four dimensional space. Velocities, position, and orientation through time can be computed from the accelerations in time. The analysis of the ight dynamics is based on Newtons Second law: F = ma. If we know the forces acting on the aircraft/spacecraft and the mass distributions we can determine the accelerations. By integrating the accelerations with respect to time and by knowing the initial velocities, the velocities at each moment in time can be calculated. Likewise, by integrating the velocities with respect to time and knowing the initial position and orientation, the position and orientation at each time instant can be determined. The dynamics, i.e. the behavior through time, of the aircraft or spacecraft can be inuenced by using controls, e.g. aerodynamic surfaces or thrusters. There are two types of ight dynamics analysis. The rst concerns the stability. How does the aircraft react to movements of the controls or other types of disturbances? Does it have inherent stability? This type of analysis concerns the characteristics of the aircraft without the pilot aspect in it. The second type is about the aircraft/pilot connection. It is the analysis of the dynamics of the aircraft or spacecraft with respect to the ability of the pilot to control the craft. This is the eld of handling qualities. There are specic regulations that state the required handling qualities for each type of aircraft. These regulations are usually dened in terms of responses of the aircraft to control inputs, like elevator, rudder or throttle inputs. For aviation these regulations have matured trough the years and are described in detail in regulation documents, such as the Joint Aviation Regulations (JAR) set up by the Joint Aviation Authorities (Europe) and the Federal Aviation Regulations set up by the Federal Aviation Administration (USA). The handling qualities of an aircraft are related to the control surfaces and control mechanism. There is a rich history on the development of control surfaces and control mechanism. The following section gives an overview. After that a list of terms and denitions used in ight dynamics is given. The setup of the book is to keep everything as general as possible and to simplify things through the use of assumptions
Flight Dynamics

Introduction

where necessary. The assumptions which are used throughout the book are listed in section 1-2. Finally, the book outline will be given in section 1-3.

1-1

Introduction to ight dynamics and control

The Wright Flyer was the rst powered piloted aircraft which had full attitude control (see gure 1-1). It had a double rudder to control yaw, a double elevator to control pitch and used warping of the wings to control roll. The control technique of using mechanical links connected to aerodynamic control surfaces was the founding of modern controlled ight. Modern aircraft still use the idea of pitch, roll, and yaw control through means of deectable control surfaces. (Warping of wings requires exible wings therefore limiting its eld of application. Soon rigid wings with control surfaces where used.) However, over the years, many additional (control) surfaces have been developed. The location, shape, and purpose of each surface varies. Also the mechanisms to move the ight control surfaces, i.e. the ight control systems, have changed over the years. To fully comprehend the inuences these development have on the eld of ight dynamics and control, a short introduction is given in this section. First the control surfaces which have been developed throughout the years are addressed (section 1-1-1). Thereafter, in section 1-1-2, the mechanisms of translating pilot commands to the control surface deections are discussed.

1-1-1

Flight control surfaces

Flight control surfaces have been developed throughout the years and many variants exist. A classication has been made for control surfaces. Two types are dened: Primary ight control surfaces Flight-critical control surfaces. If control of these surfaces is lost, control over the aircraft is (partially) lost and the aircraft is likely to crash. Secondary ight control surfaces Non-ight-critical control surfaces. If control of these surfaces is lost, complete control over aircraft is still possible. The most obvious examples of primary ight control surfaces are the elevator, aileron, and rudder. These basic control surfaces are needed to control the aircraft about all three axis of the aircraft. Without any redundancy of these ight control surfaces, malfunction of one of these control surfaces would make the aircraft hard to control. Stable ight may still be possible if the pilot has thorough knowledge of the ight dynamics of the aircraft. Examples of secondary ight control surfaces are: speed brakes, lift dumpers, slats, aps, and trim surfaces. Control over the orientation of the aircraft is still possible when these surfaces malfunction. As said, many control surfaces have been developed. The dierence between surfaces are, apart from their dimensions, the location on the aircraft and their purpose. In the following, a list is presented in which dierent types of control surfaces are explained briey.

1-1 Introduction to ight dynamics and control

(a) First ight (Kittyhawk England, 10:35 AM, 17 Dec 1903)

Elevator

Rudder
(b) Prole view

Figure 1-1: The Wright Flyer I [136]

Ailerons Purpose: provide roll control. Location: on the trailing edge of the main wing. Extra information: ailerons are placed either near the tip of the wing (Out-board ailerons) or near the root of the wing (In-board ailerons). Out-board ailerons are only active at low speeds. In-board ailerons are active at all speeds. Elevators Purpose: provide pitch control. Location: trailing edge of the horizontal stabilizer. Rudders Purpose: provide yaw control. Location: trailing edge of the vertical stabilizer. Canards Purpose: provide pitch control. Location: in front of the main wing. Canards can consist of a xed surface with trailing edge control surface or the whole canard surface can be rotated and controlled. Elevons
Flight Dynamics

Introduction Purpose: provide pitch and roll control. Location: trailing edge of the main wing. Extra information: to provide pitch control, the elevons are extended symmetrically. For roll control the elevons are extended asymmetrically. Elevons are typically used on tailless delta-wing aircraft.

Flaperons Purpose: provide roll control and additional lift. Location: trailing edge of main wing. Extra information: the aperons is a combination of the ailerons and trailing edge aps. The ap function is obtained by symmetrical extension of the aperons. Roll control is obtained through asymmetric extension. Flaps Purpose: increase lift. Location: either on the leading-edge of the wing or the trailing edge of the wing. Extra information: the leading-edge and trailing-edge aps are used to deform the shape of the wing cross-section (extending the chamber line). The increase in wing surface causes an increase in lift. Therefore slower take-o and landing speeds are possible. Slots Purpose: increase lift (at higher angles of attack). Location: trailing edge of the wing. Extra information: a slot is an alternative to the trailing-edge ap. Slots outperform trailing-edge aps by letting air ow through the wing such that, at high angles of attack, the collapse of airow on the upper surface of an airfoil is reduced, thus maintaining maximum lift at high angles of attack. Slats Purpose: increase lift (at higher angles of attack). Location: leading edge of the wing. Extra information: the slat is actually a leading-edge slot. It delays wing stall at higher angles of attack. Spoilers Purpose: spoil lift, increase drag. Location: on main wing surface. Extra information: Spoiler have two main eects. By symmetric extension on both wings in ight, the spoilers act as speed brakes. During landing, symmetric extension will lead to a smaller roll-out distance. Asymmetric extension in ight will provide additional roll control without inducing wing twist to cause roll-reversal. There are two types of spoiler known: ground spoilers and ight spoilers. Ground spoilers are used during landing and extend further than ight spoilers (which are used in ight). Speed brakes Purpose: add drag to decelerate aircraft. Location: on wing surface. Extra information: speed brakes are aerodynamic surfaces which pop out of the wing at a near perpendicular angle to the airow. This increases drag to the maximal possible extent. Stabilators, Stabilons, and Tailerons (names are synonymous) Purpose: provide pitch and roll control. Location: These surfaces are the two halves (left and right) of the horizontal stabilizer. Extra information: pitch control is achieved by rotating the horizontal stabilizer symmetrically. Roll control is achieved through asymmetric rotation. Stabilators are used to augment the surfaces which provide pitch (elevators, elevons, canards) and roll (ailerons, aperons, spoilers) control.

1-1 Introduction to ight dynamics and control

Trim surfaces Purpose: remove required control force exerted by the pilot. Location: trailing-edge of the vertical stabilizer (or part of the rudder) and/or trailing-edge of the horizontal stabilizer (or part of the elevator). Extra information: trim surfaces are used to balance the aerodynamic moment on control surfaces (during steady ight) such that the loads on the controls are eliminated . The pilot can then y hands-free. Two main possibilities exist for the horizontal trim surface. Either the entire horizontal stabilizer acts as a trim surface (so-called variable- incidence horizontal stabilizer) or a smaller surface is used (trailing-edge surface of part of the elevator, creating the so-called variable camber horizontal stabilizer).

Each aircraft can have a dierent conguration regarding its control surfaces. In gures 1-2 to 1-6, several aircraft are depicted. Each gure shows the location of several specic control surfaces.

Rudder

Aileron Outboard trailingedge ap Flight spoiler Slats

Elevator

Stabilizer Inboard trailingedge ap Ground spoiler

Leadingedge ap

Figure 1-2: Boeing 767 3D-view [113]

Flight Dynamics

Introduction

Elevons

Figure 1-3: Concorde 3D-view [113]

Carnard

Flaperon

Elevator

Figure 1-4: X-29 in ight [35]

1-1-2

Flight control systems

How do you move the control surfaces of an aircraft and what does a pilot feel when moving a control stick or wheel? Why have the ight control systems changed so much through time? These are just a few questions which one can ask when looking at the history of ight control systems. In the development of control systems their have been two main drives. The rst is the desire to build larger aircraft. Larger aircraft means larger aerodynamic surfaces which in turn means larger aerodynamic forces and thus larger control forces. The very rst control systems where purely mechanical consisting of wires, springs, and wheels, which relied on the pilot strength to move them. The control forces a pilot can generate (especially for larger periods of time) is limited, so when the aircraft size increased, something had to be done to the controls.

1-1 Introduction to ight dynamics and control

Aileron - Ground spoiler Outboard t railing-edge ap

- Flight spoiler

Inboard t railing-edge ap - Ground spoiler

Figure 1-5: 737-400 Flight and ground spoiler deployment on landing [32]

Figure 1-6: Boeing 737 slats [32]

Flight Dynamics

Introduction

The second drive in the development of control systems is the invention of the autopilot. With the development of new technical systems, like computers and electric actuators, it became possible to control the aircraft electrically. Looking at the past, several types of ight control systems can distinguished: Mechanical human powered system (reversible) Mechanical hydraulic powered system (irreversible/reversible) Fly-by-wire system (irreversible) These dierent types will be briey discussed in the next three sections. Mechanical human powered ight control system The rst powered aircraft, the Wright Flyer, used a simple form of a mechanical human powered control systems. The pilot lay head-forward on the lower wing. With his left hand the pilot controlled the elevator pitch by moving a lever attached to the elevator. By using a cable system attached to a saddle, in which the pilot would lay, the pilot could control the warping of the wings by moving his hips sideways. This cable system was also connected to the rudder, thus creating two control surface movements with one movement of the pilot. This enabled the pilot to balance the aircraft and provided directional steering. Although advanced in its time, the control system of the Wright Flyer was very simple indeed. An example of a more complicated system is the mechanical ight control system given in gure 1-7. One characteristic of a mechanical control system is that a direct linkage exists between the pilot and the control surfaces. Another characteristic is that it is reversible, i.e. the amount of control force needed to move an aerodynamic surface is directly felt by the pilot. Mechanical ight control systems are found in small aircraft where control loads are not excessive. For ying larger aircraft, ight control systems had to be changed. Two mechanical solutions were developed. The rst one is an attempt to extract the maximum mechanical advantage through levers and pulleys. The maximum reduction in forces in such a system is limited by the strength of the cables, levers, and pulleys. The second solution is the introduction of so-called control tabs (also known as servo tabs). These are small surfaces at the end of the control surface which reduce the required control force exerted by the pilot by using aerodynamic properties of the aircraft, see gure 1-8. In this case the pilot controls are directly linked to the control tabs. The control tab generates an aerodynamic force which in turn moves the aerodynamic surface. The Boeing 707 used the concept of control tabs in its ight control system (see gure 1-9 and 1-10). The ideas of extracting maximum mechanical advantage and using control tabs worked well for middle sized aircraft. However, when the aircraft became even larger, neither of the two solutions did suce. This lead to the use of hydraulic power.

1-1 Introduction to ight dynamics and control

(a) Elevator control - ghter aircraft

(b) Aileron control - commercial aircraft

Figure 1-7: Mechanical ight control system example [132]

Flight Dynamics

10

Introduction

Figure 1-8: KC-135/707 Aileron Balance Panel Installation [132]

1-1 Introduction to ight dynamics and control

11

Figure 1-9: Boeing 707 Flight Control Surfaces [132]

Flight Dynamics

12

Introduction

Figure 1-10: Boeing 707 Stabilizer-Actuated Tab Mechanism [132]

1-1 Introduction to ight dynamics and control

13

Mechanical hydraulic powered ight control systems The development of large aircraft lead to the need for hydraulic powered ight control systems. The pilot could no longer generate the required control forces by using purely mechanical links. Hydraulic-power became the source of the control forces. The mechanical hydraulic powered ight control system consists of two parts: 1. A mechanical part (the same as for fully mechanical FCS) 2. A hydraulic part The mechanical part transports the pilot commands to the hydraulic system. This is done by opening the appropriate set of servo valves of the hydraulic system. The hydraulic system then generates the forces for the actuators which move the aerodynamic surfaces. The Boeing 727 and 737 used such a ight control system (see gure 1-11). They both have a manual, mechanical linkage backup, in case of any failure of the primary ight control system. The backup system is called manual reversion since the control is done by the pilot and the force is directly felt by the pilot (thus noted as a reversible system). The mechanical hydraulic powered FCS is an irreversible system because the forces are not felt by the pilot. The feedback of forces tells the pilot a great deal about the current airspeed. In order to still keep the feeling of controlling the aircraft, articial feel devices were installed that generate forces on the mechanical part of the FCS. The amount of force exerted on the mechanical system is made dependent on airspeed. The Boeing 747 was the rst aircraft in the Boeing series to have a fully powered actuation system. The manual reversion system was omitted from the FCS because the required control forces for any ight condition would have been too large to be generate by the pilot. The schematic layout of part of the ight control system of the Boeing 747 can be seen in gure 1-12. The benets of hydraulic-powered control surfaces with respect to human-powered control surfaces are two-fold. Firstly, the drag on the control surfaces is reduced be eliminating the need for control tabs. Also the control surface eectiveness is increased by eliminating control tabs. Secondly, the control surface utter characteristics have improved since the hydraulic system introduces high mechanical stiness. The main drawback of hydraulic-powered control surfaces is that it is an irreversible system.

Flight Dynamics

14

Introduction

Figure 1-11: Boeing 727 Aileron Control and Trim Systems [132]

1-1 Introduction to ight dynamics and control

15

Figure 1-12: Boeing 747 Elevator Control System [132]

Flight Dynamics

16

Introduction

Fly-by-wire ight control system With the invention of the (analog) computer it became possible to control an aircraft electronically. Initially the layout of the ight control system did not change very much. The mechanical part of the hydraulic powered control system was replaced by an electronic circuit and the cockpit controls now operate signal transducers which generate the appropriate commands. The only change to the hydraulic part of the system was the replacement of mechanical servo valves with electronic servo valves (the valves were operated by an electronic controller). This initial setup of the Fly-by-wire (FBW) FCS is known as the analog FBW FCS. The force feedback to the pilot is controlled by the electronic controller which drives an electric feel device. The main advantage of the (initial) analog y-by-wire FCS was that there was no longer a need for a complex, fragile, and heavy mechanical linkage between the pilot and the hydraulic system. The advantages of analog FBW FCS further increased when analog computers were developed. By replacing the electronic controller with analog computers the designer of the FCS has more exibility with regard to setting the ight control characteristics. The pilot commands no longer needed to be interpreted as control surface deection but more as intentions, like ying straight (column center) or commanding a constant pitch rate. The analog computer would then translate the pilot commands to actual control surface deections. This concept opened the way for aircraft which where inherently unstable but could still be controlled as stable aircraft, the so-called relaxed stability concept. The development of the digital computer expanded the application of the FBW FCS considerably. The analog computers had limited capabilities, but the possibilities with the digital computers were vast. Now, the ying and handling characteristics of an aircraft could be set precisely. More and more exotic (inherently unstable) aircraft could be own with the help of a digital FBW FCS. The next step in de development of ight control systems was to replace the heavy and high maintenance hydraulic mechanism. They were replaced with electrical power circuits. These circuits power electrical or self-contained electro-hydraulic actuators that are controlled by the digital ight control computers. This type of ight control system is also known as Power-bywire FCS. Other developments are Fly-by-optics FCS which use light pulses instead of electric pulses to transport pilot commands. Even more futuristic developments are the designs of intelligent FCS which can recongure themselves when the aircraft sustains damage. The FCS can choose the best set of control surfaces to perform the desired/commanded motion.

1-2

Assumptions

In this book the derivation for the equations of motion will be kept as general as possible. The equations of motion are derived in chapter 3. Two assumptions are made during that derivation. These are: Assumption 1: Spherical Earth In reality the Earth is an ellipsoid. Assuming that the Earth is spherical will simplify

1-2 Assumptions

17

the denition of latitude (see section 2-1-3). This is turn simplies the transformation between reference frames which simplies the equations of motion. Assumption 2: Vehicle is a rigid body of constant mass It is assumed that the vehicle under consideration does not deform under normal ight conditions. When this is true and the mass of the vehicle is constant then the matrix of inertia will be constant. This assumption implies no fuel consumption and no vehicle elastic modes. After the general equations of motion are derived, other assumptions are used to simplify the equations. This reduced set of equations will be used to clarify certain aspects of ight dynamics in chapter 5 till 11. Whenever an assumption is used it will be clearly stated in the text. The following assumptions will be used in this book (after Chapter 3): Assumption 3: Flat Earth When the duration of the motion under consideration is short, the inuence of the Earths curvature is negligible. In such a case we can assume that the Earth is at. This will make the vehicle carried normal Earth reference frame coincide with the normal Earth xed reference frame. Assumption 4: Non-rotating Earth By neglecting the angular velocity of the Earth one neglects the inuence of two types of acceleration, i.e. coriolis acceleration and centripetal acceleration. If real-life measured accelerations are used as input for simulations, one should remove the coriolis and centripetal components. Large errors are introduced if the time-span of the motion is large (in the order of hours). Assumption 5: The vehicle has a plane of symmetry If this is true then the orientation of the body-xed reference frame can aligned with the principle axis of the vehicle in the symmetry plane. This will cause Ixy and Iyz to be zero. Assumption 6: Aircraft has a conventional conguration An aircraft with conventional conguration has one main wing, a horizontal and vertical stabilizor, ailerons, elevators, and one rudder. Such an aircraft is best suited for the explanation of ight dynamics principles. The reader can deduce the inuence of other control systems (like the use of aperons, elevons, etc.) by themselves. The basic principle of pitch, roll, and yaw control, remains. The only dierence is that (stronger) coupling eects may be introduced. Assumption 7: Zero wind velocity No wind means that the undisturbed air is a rest relative to the Earths surface. This means that the kinematic velocity is equal to the aerodynamic velocity. Assumption 8: Resultant thrust lies in the symmetry plane The means that the aerodynamic force caused by the engine(s), i.e. thrust, only inuences the symmetric aerodynamic forces X, Z and the symmetric aerodynamic moment M.
Flight Dynamics

18

Introduction

1-3

Book outline

In part I, the derivation of the equations of motion for aircraft is given. Before the derivation is performed the reference frames which will be used are dened (see section 2-1). The building blocks of the equations of motion are the orientation and angular velocity of one reference frames to another. These block will be given in the remainder of chapter 2. The general equations of motion are derived in chapter 3. Simplication of the equations is performed thereafter to dene the set of equations which will be used in the remainder of the book. Linearization of this set is performed in chapter 4. In part II, static stability and control is considered. First several aspects of aircraft aerodynamics, like the aerodynamic center and the process of deriving an state of equilibrium, are discussed (chapter 5). The aircraft characteristics for symmetric ight are explained and analyzed (chapter 6 and 8). Stability and Control derivatives will be derived analytically. An analysis of the aircraft characteristics is performed for symmetric, steady ight, followed by an analysis on the characteristic modes for symmetric ight from the view of static stability. The same analysis and derivations as for symmetric ight will be done for asymmetric ight (chapter 7 and 9). In Part III we will perform simulations with the symmetric (chapter 10) and asymmetric (chapter 11) linearized equations of motion.

Part I

Equations of Motion

Flight Dynamics

Chapter 2 Reference Frames

This chapter denes the reference frames used in ight dynamics. First of all one needs at least one frame of reference to describe any motion or position. The need for multiple reference frames arises from two considerations. The rst is that the denition of a vector makes more sense for one particular reference frame than for another. The aircrafts velocity vector expressed with respect to the Earth surface has a direct physical meaning while another decomposition will have no direct meaning (see gure 2-1, where VN is the velocity component due North, VE is the velocity component due East, and VZ is the velocity component perpendicular the to Earth surface). The other consideration is that the use of additional reference frames will make the derivation of the equations of motion easier. Certain motions are more easily described in a particular reference frame before a translation is made to the reference frame for which the equations of motion are derived.

North

North

VN

V Versus
VX VY Earth surface VZ

VE Earth surface

VZ

Figure 2-1: Aircraft velocity vector decomposition

Flight Dynamics

22

Reference Frames

Several reference frames will be used in subsequent chapters to derive the equations of motion. The orientation and state of the a vehicle (aircraft or spacecraft) can be dened in these reference frames. An important aspect when using multiple reference frames is the transformation of vector coordinates from one frame to another. A set of rotation angles is dened for each transformation. The rotation matrices given in this chapter are all dened using so-called Euler angles. The Euler angles are useful in many (normal-ight) cases. However, when certain situations occur such as in acrobatic ight, the Euler angle transformations will fail. A solution to this problem is the use of Quaternions. In the last part of this chapter the subject of vector dierentiation is discussed. In particular the cases where the set of reference frames used during the derivation of the equations of motion are rotating relative to one another. The reference frames will be described in section 2-1. In section 2-2, the transformations between the reference frames will be given using Euler angles. Transformation matrices will be established for every transformation between reference frames. The eect of changes in orientation between reference frames over time is an important aspect of deriving the equations of motion. This eect will be described in section 2-3. Finally, a short introduction to Quaternions is given in section 2-4, p. 70.

2-1

Overview of reference frames

In this section a total of nine reference frames will be dened. The reason for dening multiple reference frames is to make the derivation of the equations of motion easier and more meaningful. In principle only one reference frame (an inertial reference frame) would suce, but this would not make the understanding of ight dynamics easier.

2-1-1

Inertial reference frame, FI

The inertial reference frame (AXI YI ZI ), also known as the Earth-Centered Inertial (ECI) reference frame, is a right-handed orthogonal axis-system. The origin is at the center of mass of the earth. The ZI -axis is directed to the north along the spin-axis of the earth which goes through the poles of the earth. The XI -axis passes through the equator at the location where the ecliptic and the equator cross: the vernal equinox (see gure 2-2, p. 23). 1 The YI passes through the equatorial plane perpendicular to the XI end ZI to complete the coordinate axes system. The spin-axis of earth does not remain in the same direction through time. Due to gravitational inuences of the moon and other celestial bodies, the spin-axis wobbles about a ctive mean spin-axis. This motion is known as polar motion (see gure 2-4, p. 24). The denition of the inertial reference frame still holds although the choice of spin-axis orientation can dier per country. For purpose of clarity an International Celestial Reference System (or
The ecliptic is dened as the great circle formed by the intersection of the plane of the earths orbit with the celestial sphere (it can be seen as the apparent annual path of the sun in the heavens).
1

2-1 Overview of reference frames

23

Conventional Celestial Reference System) has been dened. The coordinate system is dened by the mean equator and equinox of J2000.0. (The standard epoch J2000.0 is 12hr on January 1, 2000.) The mean equator and equinox are the ctitious equator and equinox, derived after removing the eects of nutation. For more information about polar motion the reader is directed to Ref. [93].

Figure 2-2: Denition of ecliptic plane

ZI
North

Equatorial plane

YI

Noval equinox

XI
Ecliptic plane

South

Figure 2-3: Denition of earth-centered inertial reference frame

Flight Dynamics

24

Reference Frames

Figure 2-4: Polar motion - precession track

2-1-2

Earth-centered, Earth-xed reference frame (FC )

The Earth-centered, Earth-xed reference frame (ECEF) (AXC YC ZC ) is a right-handed coordinate axis-system which is very similar to the Earth-centered inertial reference frame. The dierence between the two reference frames are the locations where the X-axis and Y -axis pass through the equatorial plane. The XC -axis crosses the Greenwich meridian (see gure 2-5, p. 25) while the XI -axis passes through the vernal equinox (see gure 2-3, p. 23). The ZC -axis is directed along the spin-axis of the earth, equal to the ZI -axis. The YC passes through the equatorial plane perpendicular to the XC end ZC to complete the coordinate axes system. Since the XC -axis is xed to the Greenwich meridian, FC is xed to the Earth and will thus rotate about its ZC -axis with the angular velocity of the Earth. This in contrast with the inertial reference frame which has no angular velocity. Similar to the denition of the International Celestial Reference System, there is the denition of the International Terrestrial Reference System (or Conventional Terrestrial Reference System). It takes the spin-axis as the true celestial pole during the period between 1900-1906 (this period has been chosen for historical reasons). The Greenwich meridian remains the same. This reference frame is widely used for navigational purposes (like GPS and GLONASS).

2-1 Overview of reference frames ZC

North Greenwich meridian

25

YC A t
Noval equinox

Equatorial plane

XC
Ecliptic plane

South

Figure 2-5: Denition of earth-centered, earth-xed reference frame

2-1-3

Normal Earth-xed reference frame, FE

The normal Earth-xed frame of reference (OXE YE ZE ) is a right-handed orthogonal axissystem with the origin at an arbitrary location in the Inertial reference frame. The reference frame is xed to the earth, meaning that it rotates with the same velocity as the earth itself (with respect to the Inertial reference frame). The XE YE plane is tangent to the earth surface. The average earth surface, the earth geoid2 , is taken as a reference since the earth surface changes constantly (see gure 2-6, p. 26). The XE -axis is directed to the north and the YE -axis is directed 90 degrees to the right of the XE axis. When the earth is considered as a sphere, the ZE -axis point to the center of the earth (see gure 2-7). The earth is actually an ellipsoid such that the gravitational attraction vector gr is no longer directed to the center of the earth. Due to the rotational acceleration generated by the earths spin velocity , the apparent gravitational pull is always tangent to the earth geoid. A specic latitude can be dened for each vector in gure 2-6 on page 26: gc Geocentric latitude
Geoid: the particular equipotential surface that coincides with mean sea level and that may be imagined to extend through the continents. This surface is everywhere perpendicular to the force of gravity.
Flight Dynamics
2

26

Reference Frames Dened as angle between AO and the equatorial plane. It denotes the direction of the center of the earth.

gr Gravitation latitude Indicates the direction of the gravitational attraction g r . gd Geodesic latitude Indicates the direction of apparent gravity (which is a combination of the gravitational attraction and the centripetal acceleration due to earth rotational speed ). The direction of the apparent gravity is always perpendicular to the earth geoid. However, in the remainder of this book it is assumed that the earth is a perfect sphere such that latitudes all become equal to each other and thus only one latitude angle exists. The location of the origin of the normal Earth-Fixed reference frame is arbitrary but often chosen as the aircrafts center of gravity at the beginning of the considered motion. Another common choice is the location of the aircraft position projected along the ZE -axis on the earth geoid (see gure 2-7). The advantage of the latter choice of origin is that the altitude of the origin is always zero. In the remainder of this book the origin of the Normal Earth-Fixed reference frame will always be placed at the earth geoid.
Polar axis (North) Earth surface

p o ( R)

gr A gc gr gd

g a

Equatorial plane Earth geoid

Figure 2-6: Earth geoid and latitude denitions

2-1-4

Vehicle carried normal Earth reference frame, FO

The vehicle carried normal Earth reference frame is similar to the normal earth-xed reference frame. It has the same origin O as the normal earth-xed reference frame, but the ZO -axis points in the direction of the local gravity vector seen by the aircraft center of gravity G. Also, the XO -axis points to the direction of north as seen from the aircraft center of gravity, thus from point G instead of O. As a result the XO - and XE -axis are mostly not equal.

2-1 Overview of reference frames ZI

North

27

XE
Equatorial plane

O ZE
Noval equinox

YE Lt0

Rt

YI

t0 XI
South

Figure 2-7: Normal earth-xed reference frame for spherical earth

The same holds for the ZO - and ZE -axis and thus also for the YO and YE -axis. This can clearly be seen in gure 2-8 (p. 28). The FO and Fo reference frames are parallel to each other. This frame is used to demonstrate which terms will disappear in the equations of motion when a at earth is assumed.

2-1-5

Body-xed reference frame, Fb

The body-xed reference frame (GXb Yb Zb ) is a right-handed orthogonal axis-system with the origin at the aircrafts reference point. In general the center of mass of the aircraft is chosen as the aircrafts reference point. If the gravity eld is constant than this point coincides with the center of gravity of the aircraft. The reference frame remains xed to the aircraft even in perturbed motion. The choice of the direction of the axes is rather arbitrary, the most commonly used orientation of the axes-system is depicted in gure 2-9 (p. 29). The Xb -axis is in the symmetry plane of the aircraft 3 and points forward. The actual direction is still arbitrary which inuences the denition of the angle of attack (which will be explained in future sections). The Zb -axis also lies in the symmetry plane and points downward. Finally the Yb -axis is directed to the right, perpendicular to the symmetry plane. Because the X-axis direction can be chosen arbitrarily, an innite number of possible body-xed reference frames exist. The most commonly used body-xed reference frames denitions are:
3

This is only possible under the assumption that such a plane exists.
Flight Dynamics

28 ZI

Reference Frames

North

G h Xo o Zo LtG Yo XO ZE XE o LtO

Equatorial plane

YE YO YI

Noval equinox

Rt

A tO tG

ZO

LgG

XI
South

Figure 2-8: Vehicle carried normal earth reference frame for spherical earth

Principle axes, Fp When a symmetry plane exist, the direction of the reference frame axis can be chosen such that most cross-products of inertia are zero. For a symmetric aircraft, two axes, i.e. Xp and Zp , will lie in the symmetry plane. The direction of the axes is chosen such that they lie parallel to the axes dened by the geometry of the aircraft such that the dierence between moments of inertia are largest. This usually means that the Xp axis is pointed towards the nose of the aircraft. Zero-lift body axes, FZ Taking a look at the relative wind in the symmetry plane, there is a wind direction for which the total lift is zero. The XZ -axis lies in this direction. The remaining two axis can be chosen arbitrarily. Stability reference frames, FS For this reference frame to be dened, a reference ight condition must be chosen which is in most cases a condition of steady ight. For this condition the relative wind can be projected onto the plane of symmetry. The XS -axis is chosen in this direction. The ZS -axis is also in the plane of symmetry but perpendicular to the XS -axis pointing downward, The YS -axis completes the reference frame. Once the orientation of the reference frame relative to the aircraft is chosen, it will remain xed thereafter. A

2-1 Overview of reference frames

29

graphical description is given in gure 2-10 (p. 29). The stability reference frame will be used in subsequent sections of this book. The origin is chosen in the center of mass of the aircraft.

Aircraft center of gravity: G

Yb

Xb

Zb

Figure 2-9: Body-xed Reference Frame

Plane of symmetry

YS Va ZS

XS

Figure 2-10: Stability Reference Frame

Flight Dynamics

30

Reference Frames

2-1-6

Aerodynamic (air-path) reference frame, Fa

The aerodynamic (air-path) reference frame is coupled to the aerodynamic velocity V a (also known as air-path velocity of air velocity). The aerodynamic velocity is dened as the velocity of the center of mass G relative to the undisturbed air. The origin of the reference frame is in general the same as that of the body-xed reference frame Fb . The Xa -axis is in the direction of the aerodynamic velocity V a . The Za -axis is in the symmetry plane of the aircraft. The Ya -axis is perpendicular to the GXb Zb plane to complete the (right-orthogonal) axis system (see gure 2-11, p. 31). The angles a and a are the aerodynamic angle of attack and the aerodynamic sideslip angle respectively. They denote the orientation of the aerodynamic reference frame with respect to the body-xed reference frame. The aerodynamic velocity expressed in the aerodynamic reference frame is denoted as: a Va ua a a Va = va = 0 (2-1) a wa 0

Common practice is to express the aerodynamic velocity vector in the body-xed reference frame: b u ua b b (2-2) Va = va = v b wa w

The parameters (u, v, w) are commonly used to denote the components of any velocity vector along the axis of any reference frame. If the parameters are stated without any superscript then they are equal to the components of the aerodynamic velocity (in this book).

2-1-7

Kinematic (ight-path) reference frame, Fk

The kinematic (ight-path) reference frame is coupled to the kinematic velocity V k (also known as ight-path velocity or ground velocity, i.e. the velocity relative to the ground4 ). The kinematic velocity is dened as the derivative of the position of the center of mass OG relative to the normal earth-xed reference frame FE : Vk,G = Vk = dOG dt

The origin of the reference frame is in general the same as that of the body-xed reference frame Fb . The Xk -axis is in the direction of the kinematic velocity V k . The Zk -axis is in the symmetry plane of the aircraft. The Yk -axis is perpendicular to the GXk Zk plane to complete the (right-orthogonal) axis system. The angles k and k are the kinematic angle of attack and the kinematic sideslip angle respectively. They are used to describe the orientation of the kinematic reference frame with respect to the body-xed reference frame. The graphical representation of the kinematic reference frame is the same as for the aerodynamic reference frame (gure 2-11, p. 31). If there is no wind, there is no dierence between these two
The ground velocity is always tangent to the ground track, i.e. the projection of the aircraft trajectory on the ground
4

2-1 Overview of reference frames

31

a a Ya a Yb Va a Xa Aircraft symmetry plane a Za a Xb

Wing plane

Zb

Figure 2-11: Aerodynamic Reference Frame in relation to Body-xed Reference Frame

reference frames. Further clarication on this matter will be given in section 2-2-11 (p. 55). The kinematic velocity expressed in the kinematic reference frame is denoted as: uk Vk k k k Vk = vk = 0 k wk 0

(2-3)

Common practice is to express the kinematic velocity vector in the vehicle carried normal earth reference frame: O uk VN O O Vk = vk = VE (2-4) O wk VZ Where VN is the velocity due local North, i.e. the North seen from point G, VE is the velocity due local East, and VZ is the velocity in local vertical direction.
Flight Dynamics

32

Reference Frames

2-1-8

Vehicle reference frame, Fr

The vehicle frame of reference (Or Xr Yr Zr ) is a left-handed orthogonal axis-system with the origin at an arbitrary, yet xed and invariable, position. The reference frame is xed to the vehicle. The usual denitions, in case of an aircraft, for the axes directions are:

Xr -axis Parallel to the plane of symmetry, in a direction xed and invariable relative to the aircraft. The positive Xr -axis points to the rear of the aircraft. Yr -axis Perpendicular to the plane of symmetry. The positive Yr -axis points to the left. Zr -axis Perpendicular to the Or Xr Yr -plane. The positive Zr -axis points upward in normal, upright, ight. This frame of reference is used to prescribe locations of important parts of the aircraft, e.g. elevator hinge lines, location of aerodynamic centers, dimensions and positions of control surfaces, etcetera. An example of the aircraft frame of reference is given in gure 2-12 (p. 32).

Zr

Xr Or

Yr

Figure 2-12: Left-handed (!) Aircraft Reference Frame Fr

2-2 Transformations between reference frames ZI Xb O r Zb A XI

(a)

33 ZI

Yb

Yb Xb

YI

=
XI Zb

A, O

YI

(b)

Figure 2-13: Euler angles dene the attitude of Fb irrespectively of translation

2-2

Transformations between reference frames

The reference frames dened in the section 2-1 will be used to dene position parameters (longitude, latitude, etc.) and vector parameters (airspeed, forces, moments, etc.). They are needed to express the motion of the aircraft (or spacecraft) by dierential equations. To make equations of motion more readable, dierent motions will be expressed in dierent reference frames. To couple the motions in each reference frame to get the overall aircraft motion through time in one particular reference frame, a transformation between reference frames is necessary. The transformation between any two reference frames at any point in time can be described by three angles, known as Euler angles. Mathematically, the rotation using the three angles results in a particular transformation matrix. The Euler angles and the relevant transformation matrices will be explained in section 2-2-1 and 2-2-2 (p. 38) respectively. Thereafter the most important transformations between reference frames and the angles (including the rotation sequence) which are needed to accomplish the transformation will be discussed (section 2-2-3 till 2-2-11, p. 42 till p. 56).

2-2-1

Euler angles

The orientation of a body in Euclidean space relative to an inertial reference frame can be described by three successive rotations through three body-referenced Euler angles5 . Therefore we need a body-xed reference frame is necessary6 . Note that only the orientation of a reference frame is described by Euler angles, not the translation of one reference frame relative to another. This means that, in terms of Euler angles, position (a) in gure 2-13 (p. 33) is equivalent to position (b). In principle the transition from the orientation of an inertial reference frame FI to the
5 Named after Leonhard Euler (April 15, 1707 till September 18, 1783), a Swiss mathematician and physicist. He is considered to be on of the greatest mathematician of all times. 6 The rotation method given here is valid for any set of reference frames. To make the explanation easier to understand a body-xed (non-inertial) and inertial reference frame are taken here.

Flight Dynamics

34

Reference Frames

orientation of an body-xed reference frame Fb can be performed in 12 dierent ways. When starting from the orientation of the inertial reference frame, three consecutive rotations are performed. Each rotation is performed about one of the axes of the body-xed reference frame. A maximum of three rotations are needed to arrive from the attitude of the inertial reference frame at the orientation of the body-xed reference frame. The only condition required to guarantee that transition to any attitude of the body-xed reference frame is that no two consecutive rotations are about the same axis. Bearing this in mind one can set up six symmetric sets of Euler angles:

x y x , x z x ,

y x y , y z y ,

z x z z y z

and six sets of asymmetric Euler angles:

x y z , x z y ,

y x z , y z x ,

z x y z y x

Where x , y , z denote rotations around the X-axis, Y -axis, and Z-axis. Any of these sets of Euler rotations allows transition to any attitude of Fb . One should bear in mind, however, that dierent sets of Euler angle rotation sequences will result in dierent magnitudes of the corresponding Euler angles. An example is given in the mini-tutorial on page 35. Transformations using Euler angles are very often used because one can easily visualize the orientation changes per Euler angle. However, a major drawback exists, the so-called Euler angle singularity or gimbal lock. The exact meaning of this drawback will not be explained until section 2-4, but it suces to say that the singularity occurs at an pitch angle of 90 and 90 . During most (normal) aircraft manoeuvres these pitch angles will not be reached such that the singularity is of no concern.

2-2 Transformations between reference frames Mini-tutorial: Rotation sequence using Euler angles This tutorial shows why the specication of the rotation sequence is important. It will show that when a dierent rotation sequence is applied using a xed set of angles will lead to a dierent orientation. Suppose one wants to rotate a body about the origin of an inertial reference frame such that a transition is made from the initial orientation (gure 2-14a, p. 36) to the nal orientation (gure 2-14b). This transition will be done using Euler angles of 30 , 30 , and 60 for x , y , z , respectively. The correct rotation sequence belonging to this set of Euler angles is the a-symmetric sequence: z y x , the so-called 3-2-1 rotation. This sequence is the most common sequence in the aerospace industry. The rotations are taken about the axis of the body-referenced axis system. After the rst and the second rotation an intermediate reference frame is dened. The complete rotation sequence is stated as: FI FI FI Fb The rotations are visualized in gure 2-15a till c (p. 37). Now, what will happen if the Euler angles stay the same but the rotation sequence is altered? To show the consequences a second rotation sequence is applied: FI FI FI Fb The rotations are visualized in gure 2-16a till c (p. 37). As one can see, the end orientation of the body clearly diers from the previous case. The same end orientation can be reached using the second rotation scheme, but it would require a dierent set of Euler angles!
x y z z y x

35

In subsequent sections the rotation sequences and the names of the Euler angles for each transition between specic reference frames are given. Only the transition between reference frames, mentioned in the previous section, will be discussed. Using the Euler angles and the specic rotation sequences, one can set up so-called transformation matrices. These matrices can be used to transform vector coordinates in one reference frame to vector coordinates in another reference frame. The method of deriving such a transformation matrix will be discussed in the next section. In subsequent sections, the transformation matrices for each transition between reference frames will be specied.

Flight Dynamics

36

Reference Frames

ZI

XI
(a) Initial orientation of body

YI

ZI Zb

Yb Xb

XI
(b) Final orientation of body

YI

Figure 2-14: Euler angles tutorial: initial and end body orientation

2-2 Transformations between reference frames

37

ZI , ZI

ZI

ZI

YI YI x

+
XI z XI YI z XI , XI YI

(a) z rotation in FI (60 )

ZI y ZI , ZI

(a) x rotation in FI (30 )

ZI

ZI

ZI XI YI , YI YI , YI

XI

+
y YI

y XI YI

+
XI XI , XI

(b) y rotation in FI (30 )

ZI ZI

(b) y rotation in FI (30 )

ZI Zb , ZI Yb z x XI

Zb

+
Xb YI z

Xb , XI

YI XI YI XI YI Yb

(c) x rotation in FI (30 )

(c) z rotation in FI (60 )

Figure 2-15: Euler angles tutorial: rotation sequence Figure 2-16: Euler angles tutorial: rotation sequence x y z z y x

Flight Dynamics

38

Reference Frames

2-2-2

Transformation matrices

The transition between the orientation of one reference frame to another can be described in terms of Euler angles. These angles can be used to set up a so-called transformation matrix T. Any vector which is expressed in one reference frame can be expressed in a second reference frame (see gure 2-17, p. 38). Z1 Z2 v X2 Y2
O

X1
(a) Vector denition

Y1

Z1
1 vz

Z2 v

Z1

Z2 v
2 vy

X2
O

X2 Y2
1 vy 2 vx

2 vz

Y2 Y1

X1

1 vx

Y1

X1
(c) Vector decomposition F2

(b) Vector decomposition F1

Figure 2-17: Simultaneous vector decomposition in two reference frames

The transition matrix contains the information necessary to calculate the coordinates of vector v in the second reference frame F2 when knowing the coordinates of vector v in the rst reference frame F1 : 1 2 vx vx 2 1 vy = T21 vy 2 1 vz vz

(2-5)

The transformation matrix T2,1 must thus contain information regarding the orientation of reference frame F1 with respect to reference frame F2 . It does not contain any information about a possible translation between reference frame origins. When using Euler angles (see previous section) the transformation matrix can be dened as a multiplication of three (sub) transformation matrices. Each (sub-)transformation matrix describes one rotation in the

2-2 Transformations between reference frames

39

entire rotation sequence used in combination with the Euler angles. Thus the complete transformation matrix can be dened as: 1 2 vx vx 2 1 vy = T21 T1 1 T1 ,1 vy 2 1 vz vz

(2-6)

where T1 1 is the transformation matrix for the rst rotation in the overall sequence expressing the vector v1 in terms of the axes of the intermediate reference frame F1 , i.e. v1 . T1 1 is the matrix for the second rotation expressing the vector v1 , which is expressed in F1 , into to the second intermediate reference frame F1 (v1 ). Finally the transition from F1 to F2 is described by the transition matrix T21 (v1 v2 ). Equation (2-6) is obtained by substitution: v1 v1 v2

The information on a single rotation depends on the particular axis around which the rotation is performed. If the axis of rotation is the X-axis of the reference frame in which the vector is currently decomposed (F1 ), then the coordinates of vector v in F2 are given by (see gure 2-18, p. 40)):
2 1 1 1 +0.vy +0.vz vx = 1.vx 2 1 1 1 vy = 0.vx + cos x .vy + sin x .vz 2 = 0.v 1 1 + cos .v 1 vz x z x sin x .vy

= T1 1 v1 v2 = T21 T1 1 T1 ,1 v1 = T1 1 v1 1 = T21 v

(2-7)

(2-8)

The transformation matrix is then dened as: 1 0 0 = 0 cos x sin x 0 sin x cos x

T21

(2-9)

If the axis of rotation is the Y -axis of the reference frame in which the vector in currently decomposed (F1 ), then the coordinates of vector v in F2 are given by (see gure 2-19, p. 40)):
2 1 1 1 vx = cos y .vx +0.vy sin y .vz 2 = 1 +1.v 1 1 vy 0.vx +0.vz y 2 = sin .v 1 +0.v 1 + cos .v 1 vz y x y z y

(2-10)

The transformation matrix is then dened as: cos y 0 sin y = 0 1 0 sin y 0 cos y

T21

(2-11)

Flight Dynamics

40

Reference Frames

A-A

a = Z1 sin x

x Z2

Z1 Z2 Y2 x O Y1 x
Z1 cos x

Z1

b = Y1 cos x

Y2
b a

x O
Y1 sin x

Y1

X1 , X2

A-A
(a) 3D view of rotation (b) 2D view of rotation

Figure 2-18: Vector decomposition for rotation about X-axis

y Z2

Z1

B-B

Z1 Z2
b a y

a = X1 sin y b = Z1 cos y

O X1 y X2
(a) 3D view of rotation

Z1 sin y

X1 Y1 , Y2
B-B

O
X1 cos y

X2
(b) 2D view of rotation

Figure 2-19: Vector decomposition for rotation about Y -axis

If the axis of rotation is the Z-axis of the reference frame in which the vector in currently decomposed (F1 ), then the coordinates of vector v in F2 are given by (see gure 2-20, p. 41):

2 1 1 1 vx = cos z .vx + sin z .vy +0.vZ 2 1 1 1 vy = sin z .vx + cos z .vy +0.vz 2 = 1 1 1 vz 0.vx +0.vy +1.vz

(2-12)

2-2 Transformations between reference frames The transformation matrix is then dened as: cos z sin z 0 = sin z cos z 0 0 0 1
C-C
C-C +

41

T21

(2-13)

Y1 Y2
Y1 cos z

a = Y1 sin z b = X1 cos z

Za , Zb

X2 z
b a

z Y2 O X1 z X2
(b) 2D view of rotation

z Y1

O
X1 sin z

X1

(a) 3D view of rotation

Figure 2-20: Vector decomposition for rotation about Z-axis

If the rotation sequence is z y x the transformation for the entire rotation is described by: v2 = T21 v1 = T21 T1 1 T1 1 v1 1 0 0 cos y 0 sin y = 0 cos x sin x 0 1 0 0 sin x cos x sin y 0 cos y cos z sin z 0 sin z cos z 0 v1 0 0 1 cos y cos z cos y sin z sin y sin x sin y cos z sin x sin y sin z sin x cos y cos sin + cos x cos z = x z cos x sin y sin z cos x sin y cos z cos x cos y + sin x sin z sin x cos z

(2-14)

v1

Transformation matrix T21 depend on the selected set of Euler rotations. This follows easily from the algebraic multiplication of the transformation matrices for a dierent sequence, for
Flight Dynamics

42

Reference Frames

instance x y z , resulting in a dierent matrix then given by equation (2-14). Transformation matrices have special characteristics. These characteristics stem from two criteria. First of all, one wishes to apply a orthogonal transformation, i.e. a rotation from one orthogonal reference frame to another, thus the transformation matrix must also be orthogonal. Secondly, after the rotation of a reference frame, a vector must still have its original length. This implies that the determinant of the transformation matrix must be equal to 1. This leads to the following criteria for transformation matrices (see Ref. [166]): 1. The eigenvalues of the transformation matrix must satisfy one of the following (orthogonality property): All eigenvalues are 1.

One eigenvalue is 1 and the other two are -1.

One eigenvalue is 1 and the other two are complex conjugates of the form ei and ei . 2. The determinant of the transformation matrix must be: det (T) 1 When these criteria are met for a transformation matrix T two important properties are guaranteed: 1. T1 = Tt = Tba ab ab 2. Tt Tab = T1 Tab = I ab ab Proof: v2 = T21 v1 v1 = T12
(taking the inverse)

(orthogonality) v2

v1 = T1 v2 21 v 1 = Tt v 2 21

T1 = Tt 21 21

(2-15)

From property one follows property two. These properties will be used during the derivation of the equations of motion and kinematic relations. In the following sections the transformation matrices will be derived which are frequently used in practice. As explained above, the particular form of each matrix strictly depends on the specic order of the rotation sequence. Each matrix belongs to one particular transformation between two reference frames and is thus only valid for the rotation sequence as stated per transformation.

2-2-3

Transformation from FI to FE

The transformation from the inertial reference frame FI to the normal earth-xed reference frame is performed using two consecutive rotations (see gure 2-8 p. 28 and 2-21 p. 44). For

2-2 Transformations between reference frames the rotation sequence one intermediate reference frame will be dened, FI : FI FI FE The two rotations in sequence are: 1. Rotation tO about the ZI -axis. 2. Rotation LtO
2

43

about the YE -axis (= YI -axis).

The rst rotation brings to YI -axis to the YE -axis. The rotation is equal to: tO = t tO (2-16)

Where t is the rotational speed of the earth and tO is known as the stellar time of point O. The second rotation consists of two parts. First the reference frame is rotated by 90 = [rad] about the YE -axis to bring the XE -axis in northern orientation, i.e. in ZI -direction. 2 Thereafter a rotation of LtO is performed to point the XE -axis to the geographical north (seen from point O on the earth surface). Angles sign and limits tO is positive if point O is west of the XI ZI -plane. LtO is positive if point O is on the northern hemisphere. The ranges of both angles are: < tO < < LtO < 2 2 Transformation matrix XE = TEI XI TEI = TEI TI I sin LtO 0 cos LtO cos tO sin tO 0 sin tO cos tO 0 0 1 0 = cos LtO 0 sin LtO 0 0 1 sin LtO cos tO sin LtO cos tO cos LtO sin tO cos tO 0 = cos LtO cos tO cos LtO sin tO sin LtO (2-17)

(2-18)

2-2-4

Transformation from FI to FO

The transformation from the inertial reference frame FI to the vehicle carried normal earth reference frame FO is similar to the transformation from FI to the normal earth-xed reference frame FE (described in the previous section, see also gure 2-8, p. 28). Two consecutive rotations are required:
Flight Dynamics

44

Reference Frames

ZI ZI North Equatorial plane

ze

XE

xe ye

A Noval equinox XI tO South

LtO

YE YI ZE XI A o tO LtO YI

(a) Location and orientation of FI and FE

ZI

(b) Origin of FE shifted to A

ZI XE

ZI
2

YI A o tO XI XI YI ZE XI A o

YE

LtO

YI

(c) 1st rotation

(d) 2nd rotation

Figure 2-21: Transformation from inertial reference frame to normal earth-xed reference frame

2-2 Transformations between reference frames 1. Rotation tG about the ZI -axis. 2. Rotation LtG
2

45

about the YO -axis.

The rst rotation brings the YI -axis to the YO -axis. The rotation is equal to: tG = t tG (2-19)

Where t is the rotational speed of the earth and tG is known as the stellar time of point G. The second rotation consists of two parts. First the reference frame is a rotation of 90 = [rad] about the YG -axis to bring the XG -axis in northern orientation, i.e. in ZI 2 direction. Thereafter a rotation of LtG is performed to point the XG -axis to the geographical north (seen from point G, see gure 2-8, p. 28).

Angles sign and limits tG is positive if point G is west of the XI ZI -plane. LtG is positive if point G is on the northern hemisphere. The ranges of both angles are: < tG < < LtG < 2 2 Transformation matrix XO = TOI XI (2-20)

TOI

= TOI TI I sin LtG 0 cos LtG cos tG sin tG 0 sin tG cos tG 0 = 0 1 0 cos LtG 0 sin LtG 0 0 1 sin LtG cos tG sin LtG cos tG cos LtG = sin tG cos tG 0 cos LtG cos tG cos LtG sin tG sin LtG

(2-21)

2-2-5

Transformation from FE to FO

The transformation matrix for the case of the normal earth-xed reference frame FE to the vehicle carried normal earth reference frame FO can be easily deduced using the results from the previous two sections: XE = TEO XO = TEI TIO XO = TEI Tt XO OI

(2-22)

Flight Dynamics

46

Reference Frames

Performing the multiplication using equation (2-18) and (2-21) and by using the summation rules: sin(A B) = sin A cos B cos A sin B cos(A B) = cos A cos B + sin A sin B with LgG = tG tO , one will come to the following transformation matrix: = T EO sin LtO sin LtG cos LgG + cos LtO cos LtG sin LtG sin LgG cos LtO sin LtG cos LgG sin LtO cos LtG (2-23)

sin LtO sin LgG cos LgG cos LtO sin LgG

sin LtO cos LtG cos LgG cos LtO sin LtG cos LtG sin LgG sin LtO cos LtG cos LgG + sin LtO sin LtG

(2-24)

2-2-6

Transformation from FO to Fb

The transformation from the vehicle carried normal earth reference frame to the body-xed reference frame is done in three consecutive rotations. For the rotation sequence two intermediate reference systems will be dened, i.e. FO and FO : FO FO FO Fb The three rotations in sequence are (see gure 2-22, p. 48): Rotation yaw angle about the ZO -axis Rotation pitch angle about the YO -axis Rotation roll angle about the Xb -axis (= XO -axis) In the aerospace industry the most commonly used rotation sequence is that of . Other sequences are also possible (see section 2-2-1, p. 33). The yaw, pitch, and roll angle are the rotations about the Z, Y , and X-axis respectively. This set of angles is commonly confused with another set of angles, namely: the azimuth, climb, and bank angle. However the denitions of these angles are: azimuth angle: angle between Xb -axis and its projection on a predened (local) vertical plane. climb angle: angle between Xb -axis and its projection on the (local) horizontal plane. bank angle: angle between Yb -axis and its projection on the (local) horizontal plane.

2-2 Transformations between reference frames

47

For the current rotation sequence these two sets of angles are related in the following way: azimuth angle = yaw angle climb angle = pitch angle bank angle = roll angle x cos (pitch angle) Two denitions for the azimuth angle can be stated depending on the orientation of the XO -axis: XO is the direction of the geographic North: v true heading XO is the direction of the magnetic North: m magnetic heading The dierence between the two headings is the magnetic declination dm, which is positive when the magnetic North is east of the geographic North. Angles sign and limits The positive rotation direction of the angles are in accordance with the denition of right-hand reference systems. In gure 2-22 (p. 48) a negative yaw angle () is taken to enhance the graphical representation. The ranges of the angles according to convention are: < < < < 2 2 < < Transformation matrix Xb = TbO XO TbO = TbO TO O TO O 1 0 0 = 0 cos sin 0 sin cos cos sin 0 sin cos 0 0 0 1 cos cos sin sin cos = cos sin cos sin cos + sin sin (2-25)

cos 0 sin 0 1 0 sin 0 cos cos sin sin sin sin + cos cos cos sin sin sin cos sin

(2-26)

sin cos cos cos

Flight Dynamics

48

Reference Frames

GYb Zb -plane

Xb

Vertical plane

YO

XO Zb Yb ZO Horizontal plane

(a) Aircraft orientation

Horizontal plane XO

YO

YO

XO

ZO ,ZO

(b) 1st rotation:

GYb Zb -plane XO Vertical plane YO YO XO Horizontal plane ZO ZO YO XO Yb ZO Zb Xb Horizontal plane Vertical plane

(c) 2nd rotation:

(d) 3rd rotation:

Figure 2-22: Transformation from vehicle carried normal earth reference frame FO to the body-xed reference frame Fb

2-2 Transformations between reference frames

49

2-2-7

Transformation from FO to Fa

The transformation from the vehicle carried normal earth reference frame to the aerodynamic reference system is done in three consecutive rotations. For the rotation sequence two intermediate reference systems will be dened, FO and FO : FO FO FO Fa The three rotations in sequence are (see gure 2-22, p. 48): Rotation a aerodynamic yaw angle about the ZO -axis Rotation a aerodynamic pitch angle about the YO -axis Rotation a aerodynamic roll angle about the Xa -axis (=XO -axis) The rst two rotations, a and a , are used to bring the Xa -axis parallel with the aerodynamic velocity vector. The third rotation, a , is used to place the Za -axis into the symmetry plane of the aircraft. The rotation sequence is similar to that of the previous transformation FO Fb . The complete transformation can be seen in gure 2-23 , p. 50 (similar to gure 2-22, p. 48). For the current rotation sequence the relation between yaw, pitch, roll and azimuth, climb, bank angle is: azimuth angle = yaw angle climb angle = pitch angle bank angle = roll angle x cos (pitch angle) The aerodynamic azimuth angle, climb angle, and bank angle, are sometimes called the airpath azimuth or air-path track angle, air-path climb or air-path inclination angle, and air-path bank angle respectively.

Angles sign and limits The positive rotation direction of the angles are in accordance with the denition of right-hand reference systems. In gure 2-23 (p. 50) a negative azimuth angle (a ) is taken to enhance the graphical representation. The ranges of the angles according to convention are: < a < < a < 2 2 < a < Transformation matrix Xa = TaO XO (2-27)
Flight Dynamics

50

Reference Frames

GYa Za -plane symmetry plane

Xa Va YO a a Ya ZO a Za a XO a a

Vertical plane

Horizontal plane

Figure 2-23: Transformation from vehicle carried normal earth reference frame FO to the aerodynamic reference frame Fa

TaO = TaO TO O TO O 1 0 0 cos a 0 sin a = 0 cos a sin a 0 1 0 0 sin a cos a sin a 0 cos a cos a sin a 0 sin a cos a 0 0 0 1 cos a cos a cos a sin a sin a sin a sin a cos a sin a sin a sin a sin a cos a cos sin + cos a cos a = a a cos a sin a sin a cos a sin a cos a cos a cos a + sin a sin a sin a cos a

(2-28)

2-2-8

Transformation from FO to Fk

The transformation from the vehicle carried normal earth reference frame FO to the kinematic reference frame Fk is similar to the transformation from FO to the aerodynamic reference frame Fa in the previous section. The rotation sequence is the same, only the magnitudes and names of the angles are dierent. For the rotation sequence two intermediate reference systems will be dened, FO and FO :

2-2 Transformations between reference frames

51

FO FO FO Fk The three rotations in sequence are: Rotation k kinematic yaw angle about the ZO -axis Rotation k kinematic pitch angle about the YO -axis Rotation k kinematic roll angle about the Xk -axis (=XO -axis) The rst two rotations, k and k , are used to bring the Xk -axis parallel with the kinematic velocity vector. The third rotation, k , brings the Zk -axis into the symmetrical plane of the aircraft. For the current rotation sequence the relation between yaw, pitch, roll and azimuth, climb, bank angle is: azimuth angle = yaw angle climb angle = pitch angle bank angle = roll angle x cos (pitch angle) The kinematic azimuth angle, climb angle, and bank angle, are sometimes called the ightpath azimuth or ight-path track angle, ight-path climb or ight-path inclination angle, and ight-path bank angle respectively. Angles sign and limits The positive rotation direction of the angles are in accordance with the denition of right-hand reference systems. The ranges of the angles according to convention are: < k < < k < 2 2 < k < Transformation matrix Xk = TkO XO Analogue to equation (2-28) (p. 51): TkO = TkO TO O TO O cos k cos k sin k sin k cos k = cos k sin k cos k sin k cos k + sin k sin k (2-29)

cos k sin k sin k sin k sin k + cos k cos k cos k sin k sin k sin k cos k

sin k cos k cos k cos k

sin k

(2-30)

Flight Dynamics

52

Reference Frames

2-2-9

Transformation form Fb to Fa

The transformation from the body-xed reference frame Fb to the aerodynamic reference frame Fa consists of two consecutive rotations. For the rotation sequence one intermediate reference frame will be dened, Fb : Fb Fb Fa The two rotations in sequence are (see gure 2-24, p. 53): Rotation a aerodynamic angle of attack about the Yb -axis. Rotation a aerodynamic sideslip angle about the Za -axis (= Zb axis). The rst rotation brings the Xb -axis to the Xb -axis and the Zb -axis to the Zb of the intermediate reference frame Fb . Since the second and last rotation is about the Zb -axis, the Zb -axis is equal to the Za -axis. The second rotation bring the Xb -axis to the Xa -axis, thus into the direction of the aerodynamic velocity.

Angles sign and limits The positive rotation direction of the angles are in accordance with the denition of right-hand reference systems. The angle of attack is positive if the body axis Xb lies above the Xa , thus if the nose of the aircraft is pointed above the aerodynamic velocity vector. The sideslip angle is positive when the body axis Xb lies left from the aerodynamic velocity vector (seen from the origin of the reference frame). The ranges of the angles according to convention are: < a < < a < 2 2 Transformation matrix Xa = Tab Xb (2-31)

Tab = Tab Tb b cos a sin a 0 = sin a cos a 0 0 0 1 cos a cos a sin a sin a cos a cos a = sin a 0

cos a 0 sin a 0 1 0 sin a 0 cos a cos a sin a sin a sin a cos a

(2-32)

2-2 Transformations between reference frames

53

Symmetry plane

Wing plane Ya a Yb , Yb Va Xa a Xb

a Xb

Za , Zb

Zb

Figure 2-24: Transformation from the body-xed reference frame Fb to the aerodynamic reference frame Fa

Useful relations The aerodynamic velocity vector can be related to the aerodynamic angle of attack and aerodynamic angle of sideslip. The aerodynamic velocity vector can be expressed in the body-xed reference frame as stated in equation (2-2). Using this relation the magnitude of the aerodynamic velocity vector can be dened as: Va =
b b ub2 + va2 + wa2 a

The aerodynamic angle of attack is dened as: a = tan1

b wa ub a

Finally the aerodynamic angle of sideslip is dened as: a = sin1

b va Va

2-2-10

Transformation form Fb to Fk

The transformation from the body-xed reference frame Fb to the kinematic reference frame Fk consists of two consecutive rotations, similar to those for the transformation of Fb to the aerodynamic reference frame Fa (see previous section):
Flight Dynamics

54 Rotation k kinematic angle of attack about the Yb -axis.

Reference Frames

Rotation k kinematic sideslip angle about the Zk -axis (= Zb axis). The rst rotation brings the Xb -axis to the Xb -axis and the Zb -axis to the Zb of the intermediate reference frame Fb . Since the second and last rotation is about the Zb -axis, the Zb -axis is equal to the Zk -axis. The second rotation bring the Xb -axis to the Xk -axis, thus into the direction of the kinematic velocity. Angles sign and limits The positive rotation direction of the angles are in accordance with the denition of right-hand reference systems. The kinematic angle of attack is positive if the body axis Xb lies above the Xk , thus if the nose of the aircraft is pointed above the kinematic velocity vector. The sideslip angle is positive when the body axis Xb lies left from the kinematic velocity vector (seen from the origin of the body-xed reference frame). The ranges of the angles according to convention are: < k < < k < 2 2 Transformation matrix Xk = Tkb Xb (2-33)

Tkb = Tkb Tb b cos k sin k sin k cos k = 0 0 cos k cos k sin k cos k = sin k

0 cos k 0 sin k 0 0 1 0 1 sin k 0 cos k sin k cos k sin k cos k sin k sin k 0 cos k

(2-34)

Using this relation the magnitude of the kinematic velocity vector can be dened as: Vk =
b b ub + vk + wk k
2 2 2

Useful relations The kinematic velocity vector can be related to the kinematic angle of attack and kinematic angle of sideslip. The kinematic velocity vector can be expressed in the body-xed reference frame as: b uk b Vk = vk b wk

2-2 Transformations between reference frames The aerodynamic angle of attack is dened as: k = tan1
b wk ub k

55

Finally the aerodynamic angle of sideslip is dened as: k = sin1

b vk Vk

2-2-11

Transformation from Fk to Fa

The transformation from the kinematic reference frame Fk to the aerodynamic reference frame Fa can be done several ways. The rst one is to use the Tab and Tkb to derive the transformation matrix Tak : Tak = Tab Tbk = Tab Tt kb Tak = Tab Tbk = Tab Tt kb cos a cos a sin a cos a sin a = sin a cos a cos a sin a sin a sin a 0 cos a cos k cos k sin k cos k sin k sin k cos k 0 cos k sin k sin k sin k cos k

The second method is to go straight from the rst to the second reference frame using three consecutive rotations. For the rotation sequence two intermediate reference frames will be dened, Fk and Fk : Fk Fk Fk Fa The three rotations in sequence are: Rotation w wind angle of attack about the Yk -axis. Rotation w wind sideslip angle about the Zk -axis. Rotation w wind roll angle about the Xk -axis (= Xa -axis) The need for a third rotation can be explained by saying that the rst rotation about the Yk -axis will move the Zk -axis to the Zk -axis which lies outside the vehicle symmetry-plane. The second rotation is about the Zk -axis such that the Zk -axis is also outside the vehicle symmetry-plane. Since the Za -axis is in the vehicle symmetry-plane, a third rotation is necessary to move the Zk -axis to the Za -axis.
Flight Dynamics

56

Reference Frames

Wind relations The dierence between the aerodynamic velocity vector and the kinematic velocity vector is due to the wind velocity vector. The relation between the three velocity vectors can be written as: Vk = Va + Vw The wind velocity vector is dened as the velocity of an undisturbed particle W in the center of gravity G expressed in the vehicle carried normal earth reference frame. Thus: uw dOW = vw Vw = VW,G = dt O ww The magnitude of the wind velocity vector can be dened as:
O Vw = O O uO2 + vw 2 + ww 2 w

The aerodynamic angle of attack is dened as: w = tan1

b ww ub w

Finally the aerodynamic angle of sideslip is dened as: w = sin1

b vw Vw

2-3

Rotating reference frames

At any given moment in time we can express any vector in any reference frame. Vector expressions can be related through the relation r2 = T21 r1 (2-35)

But what will happen if we look at this relation through time? What additional relations will we have? In this section we will incorporate the aspect of time by looking at the time derivatives of vectors. We will see that the properties of the reference frames, i.e. location, orientation, and velocities, with respect to each other must all be used to correctly specify vector derivative relations. We will start this discussion by giving a general situation specied in gure 2-25. Imagine that we have the inertial reference frame FI (O, I, J, K) and an body xed reference frame Fb (G, i, j, k) attached to an aircraft. As we can see, we have the following relation RP = R + r (2-36)

We can express this relation in either reference frame. Expressing it in the inertial reference frame we obtain RP = RI .I + RJ .J + RK .K + rI .I + rJ .J + rK .K = RI .I + RJ .J + RK .K + TIb (ri .i + rj .j + rk .k) = RI .I + RJ .J + RK .K + ri . + rj . + rk .k i j (2-37)

2-3 Rotating reference frames

57

j K RP O J I P i R k r G

Figure 2-25: General situation for vector dierentiation

where k are the unit vectors of reference frame Fb expressed in the inertial reference frame i, j, = TIb i = TIb i = TIb j = TIb j k = TIb k = TIb 1 0 0 0 1 0 0 0 1
T T T

(2-38)

The derivative of the vector RP , using the chain rule, is given by

dRP dt

d dt

i j RI .I + RJ .J + RK .K + ri . + rj . + rk .k
d dt

= RI .I + RJ .J + RK .K +

ri . + rj . + rk .k i j
rj . dj dt

(2-39) + rk . dk dt

d i = RI .I + RJ .J + RK .K + ri . + rj . + rk .k + ri . dt + i j

since dI = dJ = dK = 0. This derivative is the derivative with respect to the inertial reference dt dt dt frame since the vector RP is expressed in FI during the derivation (and is expressed in FI after the derivation). This fact is incorporated in the notation as follows I I I I dRP di dj dk = RI .I + RJ .J + RK .K + ri . + rj . + rk .k + ri . i j + rj . + rk . dt I dt dt dt
I I I

(2-40) The subscript next to the sign | denotes the reference frame in which the derivative is taken. The previous relation is a very important one and we will analyze it piece by piece. The rst three terms on the right hand side denote the translation of the origin of the body xed reference frame. We can also write these terms as dR RI .I + RJ .J + RK .K = dt
I I = VG I

(2-41)

The second group of terms, denote the relative time derivative of r with respect to Fb . It is the change in vector r as we would see it if we stand at the origin of Fb . We will write these
Flight Dynamics

58 terms as dr ri . + rj . + rk .k = i j dt

Reference Frames

=
b

r t

(2-42)

where t denotes the derivative with respect to the body xed reference frame. It is called the relative derivative. The last group of terms contain information regarding the rotation of Fb as we will see later on.

Moving, non-rotating reference frame As we can already guess, the relation for the time derivative when the body xed reference frame is non-rotating is given by
I dRP dt I

= RI .I + RJ .J + RK .K + ri . + rj . + rk .k i j =
dR I dt I

dr I dt b

(2-43)

since the orientations (and the lengths) of the unit vectors i, j, k do not change. Non-moving, rotating reference frame Now let us look at the situation where the body xed reference frame is non-moving but rotating ( dR = 0, angular velocity vector = ). It that case we have dt I I I dRP I d i d j dk = ri . + rj . + rk .k + ri . i j + rj . + rk . (2-44) dt I dt dt dt
I I I

We need the expressions of the unit vector time derivatives. To derive these expressions we will take a closer at the change of a vector b (of constant length) caused by a single rotation about the vector (see gure 2-26). In the case that the rotation angle goes to zero we have the following relation ( is replaced by the innitely small angle d)
0

lim {|db| = d |QA| = d |b| sin }

(2-45)

We also have the following relation by denition d = || d = || dt dt Combining relations we obtain lim {|db| = || |b| sin dt} db = || |b| sin dt (2-47) (2-46)

The latter equation is a scalar equation dening the relations between vector magnitudes and angles. The direction of db is perpendicular to the plane spanned by the vectors and dt QA, thus we have sin db db = cos (2-48) dt dt 0

2-3 Rotating reference frames

59

k Q b b j G B db A

i
(a) 3D view

Q B A i
(b) Top view

db

Figure 2-26: Change of a vector due to rotation vector

Flight Dynamics

60 Substituting the obtained relation for

db dt

Reference Frames gives

sin || |b| sin sin db = || |b| sin cos = || |b| sin cos dt 0 0 This relation is given exactly by the cross product of and b 0 sin cos || |b| sin sin b = || 0 |b| sin sin = || |b| sin cos 1 0 0

(2-49)

(2-50)

Thus we have the following important relation for the time derivative of a vector (of constant length) due to a rotation vector db =b (2-51) dt Using this result in equation 2-44 yields
I dRP dt I

= ri . + rj . + rk .k + ri . i j = = =
dr dt dr dt dr dt I b I b I b

I d i dt I

+ rj .

+ ri .I + rj .I + i j bI bI I + bI ri . + rj . + rk .k i j + I rI bI

I I d j + rk . dk dt I dt I I rk .bI k

(2-52)

The latter result is an important relation which is often used during the derivation of the equations of motion. Some remarks are in order: The time derivative dr b is the time derivative of the vector r relative to the moving, dt rotating reference frame! It is the derivative of r as seen by an observer standing in the origin of Fb . The rotation vector bI denotes the relative angular velocity of reference frame Fb with respect to FI . This angular velocity vector can be expressed in any reference frame as it is a vector. As usual, the superscript denotes in which reference frame it is expressed. The rotation vector has some useful properties: For any set of reference frames we have 2 = 2 12 21 and 2 = 2 + 2 20 21 10 (2-54) (2-53)

For mathematical proof of these two properties see [30], p. 228 and the mini tutorial 2-3 on page 63.

2-3 Rotating reference frames q G r P i

61

R RP O I

Figure 2-27: Example - derivation of pilot velocity with respect to FO

Moving, rotating reference frame If we combine the two relations obtained in the previous two sections, we get dRP dt =
I

dR dt

+
I

dr dt

+ bI r

(2-55)

This relation is valid for any two reference frames. Take one reference frame as the xed (inertial) reference frame and let the other frame move with respect to the rst. The most important aspect of this relation is that all terms should be expressed in the same reference frame otherwise the relation is not valid.

Example 2.1
Consider an aircraft with pitch angle (10 deg) and a pitch up rate q (5 deg/s) as depicted in gure 2-27. The aircraft is ying with a speed of 200 m/s in the direction of I at an altitude T of 11 km at a distance of 5 km from point O, i.e. RO = 5.103 , 0, 11.103 . The location of T the pilot (point P ) given with respect to the body xed reference frame is rb = [10, 0, 5] . The vehicle carried normal earth reference frame FO is given by the quadruple (O, I, J, K) and is xed. The moving, rotating reference frame Fb is xed to the aircraft in the center of gravity (G, i, j, k). We will derive the velocity of the pilot with respect to FO . The location of the pilot expressed in FO is given by RO P = RO + rO = RI .I + RJ .J + RK .K + rI .I + rJ .J + rK .K = RI .I + RJ .J + RK .K + TIb (ri .i + rj .j + rk .k) = RI .I + RJ .J + RK .K + ri . + rj . + rk .k i j

(2-56)

Flight Dynamics

62

Reference Frames
This relation is given in inertial coordinates. The unit vectors of FI are o course given by I J K = = = 1 0 0 0 0 1 0 0 1
T T T

(2-57)

Since we have a single rotation, the transformation matrix is simply (see section 2-2-6, = = 0) cos 0 sin 0 1 0 TOb = Tt = (2-58) bO sin 0 cos We can now express the unit vectors dening Fb in terms of Fo = TOb i = TOb i = TOb j = TOb j k = TOb k = TOb 1 0 0 1 0 0 0 0 1
T T T T

= = =

cos 0 sin T 0 1 0 T sin 0 cos

(2-59)

Using previous results, the location of the pilot expressed in FO is RO P = RO + rO = RO + TIb rb T 5.103 cos 10 0 + 0 0 1 = 11.103 sin 10 0 5.103 + 10 cos 10 5 sin 10 0 = 11.103 10 sin 10 5 cos 10 5008.98 0 11006.66 0 = q 0

sin 10 10 0 0 cos 10 5

(2-60)

The angular velocity of Fb with respect to FO expressed in FO is given by: O bO

(2-61)

We now have all the information required to determine the velocity of point P with respect to FO and expressed in FO :
O VP O dR = dtP O dR O + dr + O = dt r bO dt b 200 0 0 10 cos 10 5 sin 10 5 0 = 0 + 0 + 180 0 0 0 10 sin 10 5 cos 10 5 200 + 180 (10 sin 10 5 cos 10 ) 0 = 5 180 (10 cos 10 5 sin 10 ) O O O

(2-62)

2-3 Rotating reference frames Mini-tutorial: angular velocity vector interpretation The goal of this mini-tutorial is to demonstrate the principles of the angular velocity vector. Consider the situation of the rotating earth. Two reference frames are considered, i.e. the inertial reference frame FI and the normal earth-xed reference frame FE at longitude tO and latitude LtO = 0 , i.e. point O is on the equator (see gure 2-28). The goal is to determine the angular velocity vector E , i.e. to determine EI the angular velocity of reference frame FE with respect to reference frame FI expressed in frame FE . Since point O is xed on earth, the only rotation between the reference frames is the rotation of the earth. The rotation can be described in the inertial reference frame by: 0 I = 0 EI t The angular velocity vector can be transformed to another reference frame through multiplication with a transformation matrix (since it is simply a vector). The desired angular velocity vector thus becomes: E EI = TEI I EI sin LtO cos tO sin LtO cos tO cos LtO 0 0 = sin tO cos tO 0 cos LtO cos O cos LtO sin tO sin LtO t t cos LtO = t 0 sin LtO

63

For the simpler case considered here where LtO = 0 the angular velocity vector becomes: E = [1 0 0]t , which states that the angular velocity vector is in the EI direction of the XE -axis. Looking at gure 2-28 (p. 64) one can see that this is also the case: the FE frame is rotation counterclockwise about the XE -axis with respect to the inertial reference frame.

2-3-1

Derivation of the angular velocity vector

Two methods exist to derive the angular velocity vector. The rst method is to directly perform the mathematics when considering: a = Tab ab dTba dTba = Tt ba dt dt

This method works ne but can lead to complex expressions which must be simplied by hand. The second method is more simple and comprehensable. Imagine a set of reference frames consisting of two reference frames Fa and Fb . The transformation in orientation from Fa to Fb is done according to the rotation sequence z y x . Consider the angular
Flight Dynamics

64

Reference Frames

t A-A North ZI

Equatorial plane p3 Noval equinox p4 XI South A ZE t0

XE p2 YI O, p1 YE

A-A p3

Time: tp1 p2

A-A p3

Time: tp4 p2

ZE p4 YE p1 XI

YI p4

ZE

YI

YE XI

p1

A-A p3

Time: tp2 YE p2 ZE YI p4

A-A p3 YE

Time: tp3 p2

ZE

YI p4

p1 XI

p1 XI

Figure 2-28: Angular velocity vector E explanation EI

2-3 Rotating reference frames

65

velocity vector b , i.e. the angular velocity of Fb with respect to Fa expressed in Fb . During ba the rotation sequence two intermediate reference systems are encountered:

Fa Fa Fa Fb According to equation (2-54), the angular velocity vector in reference frame Fb can be described by the summation of the angular velocities per rotation step considered in reference frame Fb : b = b + b a + b a ba ba a a The angular velocities between each of the reference frames is described by: b ba b a a b a a = Tba a ba = Tba a a a = Tba a a a

= Tba x 0 0 = Tba 0 y 0 = Tba 0 0 z

t t t

(2-63)

These relations all use the following simple principles: 1. Consider the angular velocity of the next reference frame with respect to the current reference frame. 2. Express the angular velocity vector in the current reference frame. 3. Transform the obtained angular velocity vector to Fb using the proper transformation matrix. If the transformation matrices are considered to be composed of sub-transformation matrices belonging to each single rotation: Tba Tba Tba equation (2-63) can be rewritten as: b ba = T x = T x T y = T x T y T z

Equation (2-64) is a very useful relation for the determination of any angular velocity vector. An interesting point is that equation (2-64) can be simplied to: 0 0 x b = Tx Ty 0 + Tx y + 0 (2-65) ba z 0 0

0 0 x = Tx Ty Tz 0 + Tx Ty y + Tx 0 z 0 0

(2-64)

This is because the multiplication of Ti i is always equal to i for any i since the rotation axis for i is the same as the angular velocity axis, i.e. i -axis. This also means that the relation a a = a a is valid for any single rotation transition from reference frame Fa to a a frame Fa . In the following sections, relation (2-65) will be used to derive several angular velocity vectors concerning particular combinations of reference frames.
Flight Dynamics

66

Reference Frames

2-3-2

Derivation of E EI

E is the angular velocity of the normal earth-xed reference frame FE with respect to the EI inertial reference frame FI , expressed in FE . The transformation from FI to FE is performed using two consecutive rotations (see section 2-2-3, p. 42): 1. Rotation tO = t tO about the ZI -axis. 2. Rotation LtO
2

about the YE -axis.

The derivative of these angles are:

d dt d dt

(tO ) LtO
2

= I I I = I EI

= tO = t
y

= LtO 0

The derivative of the latitude is zero since the reference frame FE is earth-xed. Using equation (2-65) one can derive the following expression for the angular velocity: 0 E = TEI 0 EI tO 0 sin LtO 0 cos LtO 0 = 0 1 0 tO cos LtO 0 sin LtO cos LtO = tO 0 sin LtO cos LtO = t 0 sin LtO

2-3-3

Derivation of O OI

O is the angular velocity of the vehicle carried normal earth reference frame FO with OI respect to the inertial reference frame FI , expressed in FO . The transformation from FI to FO is performed using two consecutive rotations (see section 2-2-4, p. 43): 1. Rotation tG = t tG about the ZI -axis. 2. Rotation LtG
2

about the YE -axis.

The derivative of these angles are:

d dt d dt

(tG ) LtG
2

= I I I = I OI

= tG
y

= LtG

2-3 Rotating reference frames Using equation (2-65) one can derive the following 0 0 O 0 + LtG OI = TOI tG 0 sin LtG 0 cos LtG 0 1 0 = cos LtG 0 sin LtG tG cos LtG = LtG tG sin LtG expression for the angular velocity: 0 0 0 + LtG tG 0

67

2-3-4

Derivation of O OE

O is the angular velocity of the vehicle carried normal earth reference frame FO with OE respect to the normal earth-xed reference frame FE , expressed in FE . The determination of O can be performed using the following relation: OE
O O O O O OE = OI + IE = OI EI = O TOI I OI EI

Using the results from the previous two section one can form the following relation: tG cos LtG sin LtG 0 cos LtG 0 0 O 0 1 0 LtG OE = t cos LtG 0 sin LtG tG sin LtG tG cos LtG t cos LtG = 0 LtG t sin LtG tG sin LtG (tG + t ) cos LtG = LtG (tG + t ) sin LtG

The relations between FO and FE can be described by stating the latitude LtG and longitude LgG of point G (aircraft center of mass) with respect an observer in point O (the origin of FE ). These relations are: LgG = tG tO LtG = Lt + LtO From this follows: tG = LgG + tO = LgG t LtG = Lt LgG cos LtG = LtG G sin LtG Lg
Flight Dynamics

Thus: O OE

68

Reference Frames

2-3-5

Derivation of b bO

b is the angular velocity of the body reference frame Fb with respect to the vehicle carried bO normal earth reference frame FO , expressed in FB . The transformation from FO to Fb is performed using three consecutive rotations (see section 2-2-6, p. 46): 1. Rotation azimuth angle about the ZO -axis
2. Rotation inclination angle about the YO -axis

3. Rotation bank angle about the Xb -axis The derivative of these angles are:
d dt d dt d dt

() = O O O () = () =

z O O O
O bO

= = =

Using equation (2-65) one can derive the following expression for the angular velocity:
O O O b bO = TbO O O + TbO O O + bO cos 0 sin = sin sin cos sin cos cos sin sin cos cos 0 1 0 0 0 cos sin + 0 sin cos 0 sin = cos + sin cos sin + cos cos

0 0 + 0 0

2-3-6

Derivation of a and k aO kO

O is the angular velocity of the aerodynamic reference frame FX with respect to the vehicle aO carried normal earth reference frame FO , expressed in Fa . The transformation from FO to Fa is performed using three consecutive rotations (see section 2-2-7, p. 49): 1. Rotation a aerodynamic azimuth angle about the ZO -axis 2. Rotation a aerodynamic climb angle about the YO -axis 3. Rotation a aerodynamic bank angle about the Xa -axis

2-3 Rotating reference frames The derivative of these angles are:

d dt d dt d dt

69

(a ) = O O O (a )

a
y

= O O O

a a

(a ) = O aO

Using equation (2-65) one can derive the following expression for the angular velocity:
O O O a aO = TaO O O + TaO O O + aO 0 sin a cos a 0 = sin a sin a cos a sin a cos a 0 cos a sin a sin a cos a cos a a 0 a 1 0 0 0 cos a sin a a + 0 0 sin a cos a 0 0 a a sin a = a cos a + a sin a cos a a sin a + a cos a cos a

The derivation of the angular velocity vector k is analogue to the derivation of a . The aO kO nal expression is given by: k kO k k sin k = k cos k + k sin k cos k k sin k + k cos k cos k

2-3-7

Derivation of a and k ab kb

a is the angular velocity of the aerodynamic reference frame Fa with respect to the bodyab xed reference frame Fb , expressed in Fa . The transformation from Fa to Fb is performed using two consecutive rotations (see section 2-2-9, p. 52): 1. Rotation a aerodynamic angle of attack about the Yb -axis. 2. Rotation a aerodynamic sideslip angle about the Za -axis (= Zb axis). The derivative of these angles are:
d dt d dt

(a ) = b b b (a ) =
b ab

y z

= a = a
Flight Dynamics

70

Reference Frames

Using equation (2-65) one can derive the following expression for the angular velocity: a = Tab b b + b ab b ab cos a sin a 0 0 = sin a cos a 0 a + 0 0 1 0 a sin a = a cos a a

0 0 a

The derivation of the angular velocity vector k is analogue to the derivation of a . The kb ab nal expression is given by: k sin k k = k cos k kb k

2-4

Quaternions, a brief introduction

In this chapter the principle of Euler angle rotations has been explained as it will be used throughout the book. The great drawback of using Euler angle rotations is the chance of getting stuck in the singularity (as mentioned in section 2-2-1). Luckily, the singularity is positioned at +90 and 90 pitch angle, a ight condition which is normally not reached for non-acrobatic aircraft. For simulating aircraft motion one can thus use Euler angles. However, for spacecraft the condition of the singularity will almost certainly be reached, i.e. there is no real up and down in space. The existence of the singularity will be explained in section 2-4-1. To remove the singularity we will introduce the concept of quaternions in section 2-4-2. The transformation between two reference frames expressed in the transformation matrix will be derived in the same section. When using quaternions during simulations, one integrates the time derivatives of the quaternions. How to determine the derivative is explained in section 2-4-3.

2-4-1

Euler angle singularity

The presence of a singularity at = 90 is best explained by looking at the Euler angle values for the 3-2-1 rotation sequence, i.e. . Imagine that the aircraft is in a vertical climb, see gure 2-29. If we try to determine the set of Euler angles for the particular case of vertical climb, it become clear that innitely many sets satisfy this ight condition! One set is: 90 90 0 . Another set is: 0 90 90 . The yaw and roll angle are now both dened about the same axis vertical axis. The singularity at = 90 can be avoided by choosing a dierent rotation sequence. However, another singularity will still exist for that new rotation sequence! For example, when the rotation sequence is y x z the singularity will lie at x = 90 .

2-4 Quaternions, a brief introduction

71

Xb

Zb

Yb

North XO YO Earth surface ZO

Figure 2-29: Euler angle singularity: Aircraft in vertical climb

For a-symmetrical rotation sequences (see 2-2-1) the singularity is always at 2 = 90 , i.e. 90 of the second rotation. Due to this rotation the axis for the third rotation will be aligned with the axis of the rst rotation. It will therefore not matter how much the body is rotated in the rst and third rotation, as long as the subtraction (for 2 = +90 ) or summation (for 2 = 90 ) of the two angles remains xed! This can be shown mathematically by looking at the transformation matrix for the 3-2-1 rotation sequence (similar to TbO , section 2-2-6): T321 = Tx Ty Tz 1 0 0 cos y 0 sin y = 0 cos x sin x 0 1 0 0 sin x cos x sin y 0 cos y cos z sin z 0 sin z cos z 0 0 0 1 cos y cos z cos y sin z sin x sin y cos z sin x sin y sin z cos sin + cos x cos z = x z cos x sin y sin z cos x sin y cos z + sin x sin z sin x cos z

(2-66) sin y

sin x cos y cos x cos y

Flight Dynamics

72 Substituting y = 90 will lead to: T321 =

Reference Frames

0 sin x cos z cos x sin z sin x cos z cos x sin z

0 sin x sin z + cos x cos z cos x sin z sin x cos z

Using the trigonometric rules of: sin (x z ) = sin x cos z cos x sin z sin (z x ) = cos x sin z sin x cos z cos (x z ) = sin x sin z + cos x cos z

0 0

This means that the transformation matrix will be similar for a constant A, meaning that the entire collection of x , z satisfying A = x z can describe the same rotation. For y = 90 , the collection of x , z satisfying A = x + z is obtained. The possibilities are innite, therefore indicating the singularity. For symmetric rotation sequences the singularity is at 0 and 180 for the second rotation. In both cases the rotation axis for the third rotation is aligned with of the rst rotation, creating a singularity.

one can rewrite the transformation matrix like: 0 0 1 0 0 1 T321 = sin (x z ) cos (x z ) 0 = sin A cos A 0 cos A sin A 0 cos (x z ) sin (x z ) 0

2-4-2

Quaternion transformation matrix

The solution to the problem is to use one rotation instead of three. This single rotation is taken about the so-called Euler axis or eigenaxis. Such an axis of rotation is xed to the reference frame which is rotated and has a xed orientation to this frame and to the inertial reference frame during the rotation. Consider two reference frames Fa and Fb respectively. The orientation of Fb with respect Fa can be characterized by a single rotation about the unit vector e along the Euler axis, e = e1 Xa + e2 Ya + e3 Za = e1 Xb + e2 Yb + e3 Zb (2-67)

where (e1 , e2 , e3 ) are the direction cosines of the Euler axis relative to both Fa and Fb (see gure 2-30). The fact that the Euler axis is stationary to the body-xed reference frame is shown clearly in equation (2-67), i.e. relation between Fb and e is equal to the Fa , e relation. It is graphically presented in gure 2-31 Let us take a look at the rotation about the Euler axis regarding an arbitrary vector v (see gure 2-32). We want to nd the relation between the vector expressed in Fa and Fb . From vector algebra one knows that: vb = OA + AB + BC

2-4 Quaternions, a brief introduction

73

Za

Zb

e 3 e3 = cos 3 1 2 Ya e1 = cos 1 e2 = cos 2 Xa Xb e2 = cos 2 1 3 2

e e3 = cos 3

e1 = cos 1

Yb

(a) Direction cosines for Fa

(b) Direction cosines for Fb

Figure 2-30: Direction cosines equality for Euler axis rotations

Za

Zb e

Xb

Yb Xa

Ya

Figure 2-31: Rotating a reference frame about the Euler axis

Flight Dynamics

74

Reference Frames

e A A B D vA e C D vb C

B O va

Figure 2-32: Rotating an arbitrary vector about the Euler axis

The rst term can be expressed in the Euler axis unity vector e and the vector v (see gure 2-32): OA = e (e va ) The second term can be dened as:

AB = AD cos = [va OA] cos = [va e (e va )] cos Finally, third term is expressed as: (2-68)

BC = [va e] sin where [va e] denotes the direction of BC with magnitude equal to the radius of the rotation circle while sin will scale the magnitude to the correct value. Combining the relations will lead to:

v2 = e (e va ) + [va e (e va )] cos + [va e] sin = va cos + e (e va ) (1 cos ) + [va e] sin

(2-69)

2-4 Quaternions, a brief introduction Elaborating equation (2-69): b a a vx vx vx e1 b a a vy = vy cos + e2 e1 e2 e3 vy (1 cos ) b a a e3 vz vz vz a vx e1 a vy e2 sin + a e3 vz a a a a e1 e1 vx + e2 vy + c3 vz (1 cos ) vx cos a a a a = vy cos + e2 e1 vx + e2 vy + c3 vz (1 cos ) a a a a vz cos e3 e1 vx + e2 vy + c3 vz (1 cos ) a a e3 vy e2 vz sin a e v a ) sin + (e1 vz 3 x a a e2 vx e1 vy sin

75

One can group the terms such that the relation can be characterized by a transformation matrix: vb = Tba va If this relation holds for an arbitrary vector v, then it will also hold for the unity axes of any reference frame. Therefore one can also write: Fb = Tba Fa After grouping the transformation becomes: e3 sin + cos + e1 e2 (1 cos ) e2 (1 cos ) 1 e3 sin + cos + Tba = 2 (1 cos ) e1 e2 (1 cos ) e2 e2 sin + e1 sin + e1 e3 (1 cos ) e2 e3 (1 cos )

e3 sin + e1 e3 (1 cos ) e1 sin + e2 e3 (1 cos ) cos + e2 (1 cos ) 3

(2-70)

Now that we have dened the transformation matrix between reference Fa and Fb in terms of the Euler axis e and the rotation angle it is time to introduce the Euler parameters, also called Quaternions. Euler parameter set consists of four parameters: q1 = e1 sin 2 q2 = e2 sin 2 q3 = e3 sin 2 q4 = cos 2 (2-71) (2-72) (2-73) (2-74)

The set is usually described as q q4 with the vector q dened as [q1 , q2 , q3 ]T . The Euler parameters are not independent of each other. Taking the square of each quaternion and
Flight Dynamics

76 summing all four yields:

2 2 2 2 q1 + q2 + q3 + q4 = e2 + e2 + e2 sin2 1 2 3

Reference Frames

+ cos2 2 2

= sin2 =1

+ cos2 2 2 (2-75)

where use is made that the length of the rotation axis is unity, i.e. e2 + e2 + e2 = 1. A 1 2 3 short-hand notation of the condition is:
2 qT q + q4 = 1

(2-76)

The transition matrix in equation (2-70) can be described in terms of the Euler parameters. When one uses the following trigonometric relations: cos 2 2 2 cos = cos sin2 2 2 = 2 cos2 1 2 = 1 2 sin2 2 sin = 2 sin one will come to the following transition 2 2 1 2 q2 + q3 Tba = 2 (q1 q2 q3 q4 ) 2 (q1 q3 + q2 q4 ) matrix: 2 (q1 q2 + q3 q4 ) 2 (q1 q3 q2 q4 ) 2 2 2 (q2 q3 + q1 q4 ) 1 2 q1 + q3 2 2 2 (q2 q3 q1 q4 ) 1 2 q1 + q2 (2-81) (2-77) (2-78) (2-79) (2-80)

To clarify this result two examples are given. The relation between matrix in (2-81) and the one in equation (2-70) will be determined for the rst and second entry one the rst row using the given relations7 . Substituting the expressions for the quaternions into the rst entry yields:
2 2 1 2 q2 + q3 = 1 2 e2 sin2 2

+ e2 sin2 3 2 2 = 1 2 1 cos2 e2 sin2 1 2 2 = 2 cos2 1 + e2 + 2e2 sin2 e2 1 1 1 2 2 (2-82)

= cos + e2 (1 cos ) 1 where use has been made in the rst line of the relation, e2 + e2 + e2 sin2 1 2 3
7

+ cos2 = 1 2 2

The other derivations are similar.

2-4 Quaternions, a brief introduction

77

and where e2 e2 has been added in the third line. Substituting the expressions for the 1 1 quaternions into the second entry of the rst row in (2-81) yields: 2 (q1 q2 + q3 q4 ) = 2e1 sin e2 sin + 2e3 sin cos 2 2 2 2 = e3 sin + e1 e2 2 sin2 e1 e2 + e1 e2 2 = e3 sin + e1 e2 (1 cos )

(2-83)

2-4-3

Quaternion propagation through time

What is the new set of quaternions for two reference frames when those frame are rotating with respect to another? We could determine the set of quaternions at each time step by determining the transformation matrix using Euler angle and then use that matrix to determine the values of the quaternions. However, we are using quaternions such that we do not have to use Euler angles! Another method of deriving the new set of quaternions is to integrate its derivatives. To derive the expression for the derivative a quaternion, we will have to enter the domain of vector algebra. Up till now we have dened vectors as components based parameters, i.e.: a vx a a vy v = a vz

a a a where vx , vy , and vz are the components of vector v expressed along the X, Y, and Z-axis of reference frame Fa respectively. Such a notation is used in the eld of linear algebra. Linear algebra is a sub-domain of vector algebra. The previous relation can also be expressed as: va = vx aa + vy aa + vz aa x y z

where ax , ay , and az are the unity vectors dening the orientation of the Xa , Ya , and Za -axis respectively. Expressing these vector in reference frame Fa will lead to: 1 0 0 vx va = vx 0 + vy 1 + vz 0 = vy 0 0 1 vz which is thus equal to the vector expressed as before in case of linear algebra. In the eld of vector algebra the unity vectors dening the reference frame axes are denoted as simply ax , ay , az or more commonly a1 , a2 , a3 where the letter denotes which frame the vectors span and in which frame the vectors are expressed (thus a1 is used instead of aa ). 1 Another important aspect of vector algebra is the fact that vector expressions are combined in one expression: b a b a v1 = Tba v1 v1 v1 a b = T va b v2 = Tba v2 v2 ba 2 a b = T va b v3 v3 v3 ba 3
Flight Dynamics

78

Reference Frames

Note that the entries in the column vectors are vectors themselves and not components (scalars) of one vector! Using this notation one can write the angular velocity vector b in ba terms of the unity vectors of reference frame Fb : b = 1 b1 + 2 b2 + 3 b3 ba (2-84)

where (1 , 2 , 3 ) are the vector components of b . ba Now, the derivation of the quaternion derivatives starts with the following relation: a1 b1 a2 = Tt b2 ba a3 b3 Taking the derivative with respect to reference frame Fa will lead to the expression of: a1 d a2 dt a3 Using the chain rule one obtains:
a

b1 o d Tt b2 = o = dt ba b3 o

o d Tt ba o = dt o

b1 b d 1 b2 + Tt b2 ba dt b3 b3 a

(2-85)
a

As for the case of linear algebra the vector dierentiation rules also applies for vector algebra, i.e.: b b b b1 d 1 d 1 b2 b2 = + b b2 ba dt dt b3 b3 b3 a b where ba is a vector based angular velocity vector (as dened in equation (2-84)). The rst term in the previous relation is equal to zero since the vectors b1 , b2 , b3 span frame Fb , therefore xating them to Fb . Collecting terms, relation (2-85) can be written as: d Tt ba 0= dt b1 b1 b2 + Tt b b2 ba ba b3 b3 a

(2-86)

The cross product can be written in terms of the components of the angular velocity vector.

2-4 Quaternions, a brief introduction The exact formulation of the cross-product term is: b1 b1 b b2 = (1 b1 + 2 b2 + 3 b3 ) b2 ba b3 b3 1 b1 b1 + 2 b2 b1 + 3 b3 b1 = 1 b1 b2 + 2 b2 b2 + 3 b3 b2 1 b1 b3 + 2 b2 b3 + 3 b3 b3 2 b2 b1 + 3 b3 b1 = 1 b1 b2 + 3 b3 b2 1 b1 b3 + 2 b2 b3 2 b3 + 3 b2 = 1 b3 3 b1 1 b2 + 2 b1 where the following relations where used: b1 = b2 b3 b2 = b3 b1 b3 = b1 b2 Relation (2-87) can be rewritten in terms of the skew-symmetric matrix M : b1 0 3 2 b1 b1 0 1 b2 = Mb b2 b b2 = 3 ba ba b3 2 1 0 b3 b3 Using the skew-symmetric matrix relation (2-86) can be written as: 0 = Tt Tt Mb ba ba b1 b2 b3

79

(2-87)

ba

Since the vectors b1 , b2 , b3 are nonzero the term between brackets must be zero, i.e.: Tt Tt Mb = 0 ba ba
ba

Taking the transpose of this relation and using the property of the angular velocity vector Mt b = Mb one can rewrite the relation as:
ba ba

Tba + Mb Tba = 0
ba

(2-88)

Equation (2-89) is called the kinematic dierential equation for the transition matrix Tab . It is valid for any transition matrix provided that the correct angular velocity skew matrix is taken. The dierential equation for each element of an arbitrary T (with appropriate M) can
Flight Dynamics

80 be written as: T11 = 3 T21 2 T31 T12 = 3 T22 2 T32 T13 = 3 T23 2 T33 T21 = 1 T31 3 T11 T22 = 1 T32 3 T12 T23 = 1 T33 3 T13 T31 = 2 T11 1 T21 T32 = 2 T12 1 T22 T33 = 2 T13 1 T23

Reference Frames

These terms can be rewritten to constitute relations which express the angular velocity components in terms of transition matrix entries: 1 = T21 T31 + T22 T32 + T23 T33 2 = T31 T11 + T32 T12 + T33 T13 3 = T11 T21 + T12 T22 + T13 T23 (2-89) (2-90) (2-91)

By substituting the transformation matrix entries in relation (2-81) into the previous expression one obtains: 1 = 2 ( q1 q4 + q2 q3 q3 q2 q4 q1 ) 2 = 2 ( q2 q4 + q3 q1 q1 q3 q4 q2 ) (2-92) 3 = 2 ( q3 q4 + q1 q2 q2 q1 q4 q3 ) Dierentiating equation (2-76) yields the forth equation: 0 = 2 ( q1 q1 + q2 q2 + q3 q3 + q4 q4 ) These four relations can be q1 q2 q3 q4 rewritten in matrix form as: 0 3 2 1 3 0 1 = 2 2 1 0 1 2 3 (2-93)

1 q1 2 q2 3 q3 0 q4

(2-94)

Relation (2-94) can be used to determine the derivatives of the quaternions provide that one know the angular velocity components and the previous set of quaternions. By numeric integration one can determine the quaternion set for the next time step. In practice one would only need to determine the set of quaternions at t = 0. This set is usually derived using the relations between the quaternions and the direction cosine or transformation matrix, i.e. using equation (2-81). The initial transformation matrix can be determined using Euler angles. As for the angular velocity components, one can typically measure angular acceleration onboard the vehicle. Integrating the angular acceleration will yield the angular velocity.

2-4 Quaternions, a brief introduction

81

The conclusion is that there is no longer any need for Euler angles (except at t = 0). The complete equations of motion could be derived using quaternions. Since quaternions have no physical interpretation, the Euler angles will still be presented to the pilot. As for the singularity encountered with Euler angles, one simply creates a small oset when the orientation of the vehicle is close to the singularity which avoids any numerical problems. This book still uses Euler angles instead of quaternions because they have physical meaning! The aspects of ight dynamics can be better explained when the reader can make the mental coupling between expressions in Euler angles and the physical world.

Flight Dynamics

82

Reference Frames

Chapter 3 Derivation of the Equations of Motion

To describe the dynamics of a vehicle we need a model of that vehicle. The model we will use in subsequent chapters is governed by dierential equations also known as the Equations of Motion (EOM). These EOM will be derived in this chapter. During the derivation the eects of rotating masses is not taken into account. It is assumed that the body is rigid. The eects will be appended to the EOM in section 3-6. The chapter will begin with a review of Newtons law in section 3-1. Two fundamental equations, which are the base on which the EOM are build, are derived in the same section. The four components in these equations will be investigated in section 3-2, 3-3, and 3-4. In section 3-2 and 3-3, an explicit expression will be derived for the translational acceleration and for the rotational acceleration respectively. Thereafter, in section 3-4, de external forces and moments are identied. After this section all necessary components for constructing the EOM are available. The EOM are constructed in section 3-5. A total of 12 equations in 3 sets of equations constitute the EOM: 3 force equations, 3 moment equations, and 6 kinematic relations. Having dened the EOM in the most general manner, simplications are applied in section 3-7. At the end of section 3-7 the EOM for at, non-rotating Earth are derived. This section of EOM will be used in subsequent chapter to explain aircraft dynamics. A simple set of EOM is used to ensure a clear explanation.

3-1

Newtons laws of physics

The equations of motion are based on Newtons laws and are stated as follows (see [112] and [153]): Newtons 1st law: A body of constant mass is at rest or moves at a constant velocity, unless a force acts on the body.
Flight Dynamics

84

Derivation of the Equations of Motion

Newtons 2nd law: If a force acts on a body, the force is equal to the time derivative of the momentum, i.e. the product of mass and velocity of the body. Newtons 3rd law: If two bodies at rest, or moving at constant velocity, exert forces upon one another, the force on the rst body is equal in magnitude but opposite in direction to the force of the second body. Newtons laws are only valid when the motion of a body is expressed in an inertial reference frame, i.e. a frame with absolute translational velocity and absolute angular velocity equal to zero, or when it is expressed in a Galilean frame. A Galilean frame is one which is in rectilinear translation with respect to an inertial reference frame. This means that the angular velocity of such a frame is equal to zero and the translational velocity of the origin is constant. Normally, a reference frame based on stellar references is chosen as inertial reference frame. In this book, however, it is assumed that the absolute translational velocity of the Earth centered inertial reference frame FI (see section 2-1-1) is negligible such that FI can be seen as an inertial reference frame. The equations of motion will be derived with respect to the inertial reference frame FI , but as explained in the previous chapter they can be expressed in any reference frame. In this section we will only discuss derivatives with respect the inertial reference frame. We will therefore drop the derivative notation and we will write da da instead of dt dt

Only once will we use another reference frame (equation (3-3)). We will then use the ocial notation.

Translational acceleration Consider a moving body. This body can be seen as a collection of point masses. If a point mass P with mass dm moves with time varying inertial velocity VP under the inuence of a force dFP , Newtons second law is expressed by: d (VP dm) dt

dFP =

(3-1)

The motion of the complete collection of point masses is then characterized by the forces acting on each point mass and subsequent accelerations of each point mass. When the body is rigid, the motions of the masses in the body are coherent, i.e. the body will not deform. In this book it is assumed that the vehicle under consideration is indeed a rigid body. Since the accelerations of the point masses in a rigid body are coherent, one can state, using Newtons third law, that the internal forces the masses exert on each other within the body cancel. This means that the motion of such a body is only inuenced by external forces. The summation of the external forces exerted on the body is equal to the integration of the momentum of

3-1 Newtons laws of physics

85

P r G

rP ZI rG

YI

XI

Figure 3-1: Vector denition for point mass P in a body.

each point mass in the body, so: d dt

dFextP =
P

We can write the velocity vector of point mass P can be expressed as (see gure 3-1): drP drG dr dr = + VP = VG + dt dt dt dt So, equation (3-2) can be written as: dFextP =
P

VP dm

(3-2)

(3-3)

The integral can be spilt up into two parts:

d dt

VG dm +

dr dm dt

(3-4)

VG dm +
m m

dr dm dt

(3-5)

Using the relation for continuous, dierential functions, d d (. . .) = dt dt the second part can be rewritten as: dr d dm = dt dt r dm = 0
m Flight Dynamics

(. . .)

(3-6)

86
dFP

Derivation of the Equations of Motion

dm

dMG,P

Figure 3-2: Moment about center of mass due to force on point mass

since the denition of the center of mass is that the integral of the distance between mass points and the center of mass is zero. The rst part of equation (3-5) can be written as, VG dm = VG
m m

dm = VG m

(3-7)

where m is the total mass of the body. So equation (3-2) can be rewritten as: dFextP =
P

d (VG m) dt

(3-8)

When the rigid body is assumed to have a constant mass, the previous relation can be rewritten as: dVG = m AG (3-9) dt which states that the translational acceleration of the center of mass with respect to the inertial space AI,G times the total mass m is equal to the resultant external force Fext . Fext = m

Angular acceleration Next to a translational acceleration, a body can also have an angular acceleration. From equation (3-1) (Newtons second law) follows that the moment of a force about a reference point is equal to the cross product of the force vector and the relative distance vector (see gure 3-2): d (VP dm) (3-10) dt The point G is chosen in the center of mass in order to simplify the following derivations. Equation (3-10) can be rewritten in terms of the angular momentum about G which is dened for a mass element in point P as: dMG,P = r dFP = r dBG,P = r (VP dm) (3-11)

3-1 Newtons laws of physics

87

Taking the time derivative of the angular momentum with respect to the inertial reference frame yields: d (dBG,P ) d (r (VP dm)) = dt dt d (VP dm) dr = (VP dm) + r dt dt The vector r can be written as (see gure 3-1): r = rP rG where A is the center of the spherical Earth and by denition the origin of the inertial reference frame. Substituting this relation in equation 3-13 will lead to: d (dBG,P ) = dt drP drG dt dt VP dm d (VP dm) dt d (VP dm) +r dt d (VP dm) +r dt +r (3-12) (3-13)

= (VP VG ) VP dm = VG VP dm Combining relation (3-14) with equation (3-10) leads to: dMG,P =

(3-14)

d (dBG,P ) + VG VP dm dt

(3-15)

When the body is assumed to be rigid the forces between the point masses in the body cancel. So, only the external applied forces will aect the motion of the body, in this case the angular motion. The total external moment is equal to the integration of all external moments on the body, i.e.: d (dBG,P ) + VG VP dm dt

dMG,P =
m m

(3-16)

Which is equal to: d dt

MG,ext =

dBG = + dt

dBG,P +
m

VG VP dm (3-17)

VG VP dm

Since VG is independent of mass element P the second term on the right hand side in equation
Flight Dynamics

88 (3-17) can be written as: VG VP dm= VG = VG = VG VI,P dm

m

Derivation of the Equations of Motion

VG +
m

dr dt

dm d dt rdm

VG dm +

(3-18)

Again, for the center of mass it holds that the integral of the distance between mass points in the body and the center of mass is zero: r dm = 0
m

(3-19)

So: VG VP dm= VG d VG dm + VG dt dm
m

= VG VG =0

r dm

(3-20)

Using this result, we state that the total external moment about the center of mass is equal to the time derivative of the angular momentum of the body about the center of mass with respect tot the inertial reference frame: MG,ext = dBG dt (3-21)

The relation given by equation (3-21) can be reformulated by rewriting the term for the angular momentum BG . The denition of the angular momentum about point G is: BG =
m

r VP dm

(3-22)

The velocity vector of the mass dm in point P can be written as: VP = VG + dr dt dr = VG + dt

+ bI r

(3-23)

Since the body is rigid the distance between the point of reference G and the location P of the mass does not vary over time the second term on the right hand side is equal to zero. This means we have: VP = VG + bI r (3-24)

3-1 Newtons laws of physics Substituting this result into equation (3-22) gives: BG =
m

89

r (VG + bI r) dm r VG dm + r (bI r) dm (3-25)

=
m

Since the point of reference G is chosen in the center of mass and because VG is independent of mass element P , the rst term on the right hand side of the previous equation is equal to zero: r VG dm = VG r dm r dm
m

(3-26) (3-27) (3-28)

= VG =0 Thus the following relation remains: BG =

m

r (bI r) dm

(3-29)

With the cross product rule: a (b c) = b (a c) c (a b) equation (3-29) can be rewritten as: BG =
m

(3-30)

bI (r r) r (r bI ) dm bIX xr bIY (xr )2 + (yr )2 + (zr )2 yr (xr bIX + yr bIY + zr bIZ ) dm bIZ zr bIX (yr )2 + (zr )2 bIY xr yr bIZ xr zr 2 2 bIY (xr ) + (zr ) bIX xr yr bIZ yr zr dm bIZ (xr )2 + (yr )2 bIX xr zr bIY yr zr (yr )2 + (zr )2 xr yr xr zr bIX bIY dm xr yr (xr )2 + (zr )2 yr zr 2 2 bIZ xr zr yr zr (xr ) + (yr ) 2 2 xr yr xr zr bIX (yr ) + (zr ) dm bIY xr yr (xr )2 + (zr )2 yr zr 2 2 bIZ xr zr yr zr (xr ) + (yr )

=
m

=
m

=
m

=
m

= IG bI

(3-31)

Flight Dynamics

90

Derivation of the Equations of Motion

Where IG is the matrix of inertia of the body about point G. A general denition of the matrix of inertia is: Ixx Jxy Jxz I = Jxy Iyy Jyz (3-32) Jxz Jyz Izz with the moments of inertia dened as: Ixx =
m

y 2 + z 2 dm x2 + z 2 dm
m

(3-33) (3-34) (3-35)

Iyy = Izz =
m

x2 + y 2 dm

and the products of inertia as: Jyz =

m

yz dm xz dm
m

(3-36) (3-37) (3-38)

Jxz = Jxy =
m

xy dm

This means that equation (3-21) can be rewritten as: dBG d (IG bI ) = (3-39) dt dt which states that the resultant external moment (MG,ext ) about point G is equal to the derivative of the angular velocity (bI ) of the body times the matrix of inertia about point G, i.e the derivative of body angular momentum about point G. MG,ext = Formal notation In the previous paragraphs, the formal notation has been dropped to make the text clearer. We will now give the formal notation for the two most important relations: Fext =m
dVG dt I

= m AG =
d(IG bI ) dt I

MG,ext =

dBG dt I

Equations of Motion derivation The force equation (3-9) and the moment equation (3-39) form the basis for the derivation of the equations of motion. There are three main parts to be considered: 1. Derivation of body translational acceleration

3-2 Derivation of body translational acceleration 2. Derivation of body angular momentum 3. Resultant external force and moment on the body In the next three sections these parts will be discussed individually.

91

3-2

Derivation of body translational acceleration

The goal in this section is to nd the expression for the inertial translational acceleration of a body in case of a spherical, rotating Earth. The body is again assumed to be a rigid so that all mass points have the same translational acceleration. Our task is now to nd the inertial translational acceleration of the center of mass, i.e. AG . The derivation starts with the notion that the inertial acceleration is equal to the derivative of the inertial velocity which in turn is equal to the derivative of the inertial position: AG = dVG dt =
I

d dt

drG dt

(3-40)
I I

Using the rule for vector dierentiation (see section 2-3) we can also write AG = dVG dt =
I

dVG dt

+ OI VG

(3-41)

The inertial velocity of the body VI,G can be expressed as the derivative of the inertial position of the body center of mass: VG = Which is equivalent to: VG = drG dt + EI rG (3-43) drG dt (3-42)
I

The position vector of the center of mass rG can be composed of two other vectors: the position vector for the origin of FO , AO, plus the position vector for the center of mass seen from the origin of FO , OG. Using this in the previous equation: VG = dAO dt +
E

dOG dt

+ EI rG

(3-44)

The position vector AO expressed in the normal Earth xed reference frame is always directed E along the ZE -axis and stays constant through time. The derivative dOG is by denition equal dt to the kinematic velocity Vk,G . This leads to following expression for VG : VG = Vk,G + EI rG (3-45)
Flight Dynamics

92

Derivation of the Equations of Motion

Substituting this relation into equation (3-41) will lead to: AG = dVG + OI VG dt O d (Vk,G + EI rG ) = + OI (Vk,G + EI rG ) dt O dVk,G dEI drG = + rG + EI + dt O dt O dt O OI Vk,G + OI (EI rG )

(3-46)

Relation (3-46) is the general expression of the inertial translational acceleration of a body, it is valid in any reference frame. An alternative equation can be dened to be able to indicate what consequences particular assumptions have on the formulation of the inertial translational acceleration. The alternative equation is: AG = Ar + AS + AR (3-47)

Where Ar is the relative acceleration of the aircraft with respect to the vehicle carried normal Earth reference frame. It is dened as: Ar = dVk,G dt (3-48)
O

AS is the acceleration due to the sphericity of the Earth (note that OI = OE + EI , used to split the term OI Vk,G in equation 3-46). It is dened as: AS = OE Vk,G Finally, AR is the acceleration due to the rotation of the Earth. It is dened as: AR = dEI dt rG + EI drG dt + EI Vk,G + OI (EI rG ) (3-50) (3-49)

For now, the inertial translational acceleration of the body will be expressed in the vehicle carried normal Earth reference frame FO . By selecting this reference frame the notation of the nal expression for the inertial translation acceleration will be simpler then for the case of any other previously dened reference frame. Each term in equation (3-46), expressed in FO , will be expanded before grouping the results. During the expansion of each term two relations will be used. They are given by: LtG = VN Rt + h VE + t (Rt + h) cos LtG (3-51) (3-52)

tG = LgG t =

Starting with the rst term, the derivative of the kinematic velocity expressed in FO . The kinematic velocity expressed in FO has the components: velocity due local North VN , velocity

3-2 Derivation of body translational acceleration

93

due local East VE , and velocity in local vertical direction VZ (see section 2-1-7). The derivative then simply becomes:

dVk,G dt

O O

VN = VE VZ

(3-53)

Moving on to the second term on the right in equation (3-46):

dEI dt

O O

rO = G

d TOI I EI rO G dt O cos LtG 0 d t 0 = 0 dt (Rt + h) sin LtG sin LtG 0 0 0 = t LtG cos LtG (Rt + h) 0 = t LtG (Rt + h) sin LtG 0 = t VN sin LtG

(3-54)

The third term becomes:

O EI

drG dt

O O

cos LtG 0 d 0 = t 0 dt (Rt + h) sin LtG 0 cos LtG 0 = t 0 sin LtG h 0 = t h cos LtG 0 0 = t VZ cos LtG 0

(3-55)

Flight Dynamics

94 The fourth term becomes: O OI

Derivation of the Equations of Motion

And the fth and nal term becomes: O O rO OI EI G

tG cos LtG VN O VE Vk = LtG VZ tG sin LtG VE VN (Rt +h) + t cos LtG (RVN = VE t +h) E VZ (RV+h) tan LtG t sin LtG t 2 VN VZ +VE tan LtG + t VE sin LtG Rt +h = VE VZ VE VN tan LtG + t VZ cos LtG t VN sin LtG Rt +h
2 2 VE +VN Rt +h

(3-56)

+ t VE cos LtG

= O O rO rO O O EI OI G G OI EI T tG cos LtG 0 t cos LtG 0 0 = LtG t sin LtG (Rt + h) tG sin LtG T tG cos LtG t cos LtG 0 G 0 0 Lt (Rt + h) t sin LtG tG sin LtG

t cos LtG (VE tan LtG + t (Rt + h) sin LtG ) = 0 t sin LtG 0 VE 0 t cos LtG + 2 cos2 LtG t Rt + h (Rt + h) VE t tan LtG sin LtG + 2 sin2 LtG t Rt + h VE t sin LtG + 2 (Rt + h) sin LtG cos LtG t = 0 2 (R + h) cos2 Lt VE t cos LtG + t t G +

(3-57)

Now that all the terms from equation (3-46) are known, they can be combined to form the complete expression. After grouping the terms the relation for the translational acceleration of a body expressed in the vehicle carried normal Earth reference frame is given by: V V +V 2 tan LtG VN + N Z Rt E + 2t VE sin LtG + 2 (Rt + h) sin LtG cos LtG t +h VE VZ VE VN tan LtG O AG = VE + + 2t VZ cos LtG 2t VN sin LtG (3-58) Rt +h 2 +V 2 VE N VZ + Rt +h + 2t VE cos LtG + 2 (Rt + h) cos2 LtG t

3-3 Derivation of body angular momentum derivative

95

The inertial translation acceleration is built up by three terms, see equation (3-47). Expressed in the vehicle carried normal Earth reference frame they are:
O O

AO = r

dVk dt

2 VN VZ + VE tan LtG 1 O VE VZ VE VN tan LtG AO = O Vk = S OE Rt + h 2 2 VE + VN 2VE sin LtG + t (Rt + h) sin LtG cos LtG 2VZ cos LtG 2VN sin LtG AO = t R 2VE cos LtG + t (Rt + h) cos2 Lt
G

VN = VE VZ

(3-59)

(3-60)

(3-61)

3-3

Derivation of body angular momentum derivative

The goal in this section is to nd an expression for the derivative of the body angular momentum. The body angular momentum is equal to the multiplication of the matrix of inertia IG and the angular velocity of the body-xed reference system with respect to the inertial reference system bI . It is common practice to dene the matrix of inertia using the bodyxed reference system, i.e. Ib . In most cases, a specic orientation of the Xb and Zb -axis can G be chosen such that some components in the matrix of inertia become zero. If it is assumed that the vehicle mass does not change over time and that the body is rigid, the derivative of the matrix of inertia with respect to the body-xed reference frame will be equal to zero. This is a very useful aspect which simplies the equations of motion. The derivation of the expression for the derivative of the body angular momentum starts with the relation given by equation (3-62). MG,ext = dBG dt (3-62)
I

To use the properties of the matrix of inertia, equation (3-62) can be written as: dBG dt dBG dt dBG = dt =
I

+ bI Bb G + bE Bb + EI Bb G G (3-63)

When using the following relation, BG = IG bI = IG bE + IG EI (3-64)

Flight Dynamics

96 equation (3-63) can be written as: dBG dt

Derivation of the Equations of Motion

d (IG bE ) + bE (IG bE ) + EI (IG bE ) + dt b d (IG EI ) + bE (IG EI ) + EI (IG EI ) dt b dbE =IG + bE IG b + EI IG b + bE bE dt b dEI b + bE IG b IG EI + EI IG EI dt b =

(3-65)

Equation (3-65) is a general expression for the derivative of the angular momentum about point G. The expression is valid in any reference frame. Equation (3-65) can be divided into three groups, the rst group containing all terms consisting of only b and b terms, bE bE b b terms, and the third group consisting of the second group only consisting of EI and EI multiplications of terms from b and b . Equation (3-65) can be written as: EI bE dBG dt With: 1 MG = IG dbE dt + b IG b bE bE (3-67) (3-68) (3-69) = 1 MG + 2 MG + 3 MG
I

(3-66)

2 MG = b IG b EI EI 3 MG = IG dEI dt
b b b + b IG b EI + EI IG bE bE

Since the matrix of inertia is commonly dened in the body-xed reference frame, the derivative of the body angular momentum will also be expressed in the body-xed reference frame. Before expanding equation (3-65), several subcomponents are formulated explicitly. The rst is the matrix of inertia expressed in the body-xed reference system: Ixx Jxy Jxz A F E Ib = Jxy Iyy Jyz = F B D (3-70) G Jxz Jyz Izz E D C The second component is the angular velocity of the body-xed reference frame with respect to the normal Earth-xed reference frame expressed in the body-xed reference frame: p b = q = b + TbO O bE bO OE r LgG cos LtG 1 0 sin = 0 cos cos sin + TbO (3-71) LtG G sin LtG 0 sin cos cos Lg

3-3 Derivation of body angular momentum derivative

97

The third and last component is the angular velocity of the normal Earth-xed reference frame with respect to the inertial reference frame expressed in the body-xed reference frame:

b EI

pb pt t b = qt = qt = TbO O EI b rt rt cos cos cos sin sin sin sin sin t cos LtG sin sin cos sin cos = cos sin + cos cos 0 cos sin cos t sin LtG cos sin sin cos cos + sin sin sin cos cos LtG cos cos + sin LtG sin sin sin cos sin LtG sin cos cos LtG cos sin = t (3-72) cos sin cos cos LtG sin LtG cos cos + sin sin

Each term in equation (3-65) will be expanded in the subsequent text. Thereafter the terms will be grouped where possible. The rst term can be written as:

Ib G

db bE dt

A F E p = F B D q E D C r Ap F q E r = B q F p Dr C r E p Dq

(3-73)

The second term can be written as: p A F E p Ib b = q F B D q G bE r E D C r p Ap F q Er = q Bq F p Dr r Cr Ep Dq (C B) qr Epq + F pr + D r 2 q 2 = (A C) pr F qr + Dpq + E p2 r 2 (B A) pq Dpr + Eqr + F q 2 p2

b bE

(3-74)

Flight Dynamics

98

Derivation of the Equations of Motion

The third term can be written as: A F E p pt b Ib b = qt F B D q EI G bE rt E D C r Ap F q Er pt = qt Bq F p Dr Cr Ep Dq rt Crqt Epqt Bqrt + F prt + D (rrt qqt ) = Aprt F qrt Crpt + Dqpt + E (ppt rrt ) Bqpt Drpt Apqt + Erqt + F (qqt ppt ) The fourth term can be written as: Ib G dEI dt
b

(3-75)

= Ib G
b

dEI dt dEI dt
b E

b E

+ b b EI Eb

= Ib G = Ib G

+ Ib b b G EI Eb

The fth term is similar to the third term: p A F E pt b Ib b = q F B D qt G EI bE r E D C rt Cqrt Eqpt Brqt + F rpt + D (rrt qqt ) = Arpt F rqt Cprt + Dpqt + E (ppt rrt ) Bpqt Dprt Aqpt + Eqrt + F (qqt ppt ) And nally, the sixth term is equal to the second term: pt A F E pt b Ib b = qt F B D qt EI G EI rt E D C rt (C B) qt rt Ept qt + F pt rt + D = (A C) pt rt F qt rt + Dpt qt + E (B A) pt qt Dpt rt + Eqt rt + F

A F E p pt = F B D q qt E D C r rt A (rqt qrt ) F (prt rpt ) E (qpt pqt ) = F (rqt qrt ) + B (prt rpt) D (qpt pqt) E (rqt qrt ) D (prt rpt ) + C (qpt pqt )

b bE

b EI

(3-76)

(3-77)

These six terms can be grouped together to form the general expression for the derivative of angular momentum about point G expressed in the body-xed reference frame. The grouping

2 2 rt qt 2 p2 rt t 2 q t p2 t

(3-78)

3-4 External forces and moments

99

is done according to equations (3-66) through (3-69) (expressed in the body-xed reference frame): dBG dt With: Ap + (C B) qr E (pq + r) + F (pr q) + D r 2 q 2 1 Mb = B q + (A C) pr F (qr + p) + D (pq r) + E p2 r 2 G C r + (B A) pq D (pr + q) + E (qr p) + F q 2 p2 2 2 (C B) qt rt Ept qt + F pt rt + D rt qt 2 2 Mb = (A C) pt rt F qt rt + Dpt qt + E p2 rt G t 2 p2 (B A) pt qt Dpt rt + Eqt rt + F qt t (C A B) qrt + (C + A B) rqt 3 Mb = (A + B C) prt + (A B C) rpt G (B + C A) qpt + (B C A) pqt pt (F r Eq) + D (rrt qqt ) + 2 qt (Dp F r) + E (ppt rrt ) rt (Eq Dp) + F (qqt ppt ) (3-80)
b

= 1 Mb + 2 Mb + 3 Mb G G G
I

(3-79)

(3-81)

(3-82)

3-4

External forces and moments

One could dene two types of forces acting on an aircraft: 1. Gravity 2. Aerodynamic forces (includes thrust) These forces are by nature distributed forces. To make the equations of motion more easy to read, these distributed forces are replaced by point forces acting at particular points on the body. Summation of all point forces will result in the total resultant force. In general these forces will generate a moment about the point of reference, which on the aircraft is chosen at the center of gravity. This choice is made for two reasons. Firstly, the force due to gravity will not cause a moment since it acts at the center of gravity. Secondly, the center of gravity is also the origin of the body-xed reference system such that no additional reference frames are needed. External forces will be discussed rst , next to be followed by the external moments.

3-4-1

External forces

We rst consider the gravity vector.

Flight Dynamics

100

Derivation of the Equations of Motion

Gravity Gravity is composed of two parts: the gravitational pull towards the center of the Earth gr,G and the centripetal acceleration due to the spinning of the Earth AI,G (see gure 3-3). In mathematical terms the relation is expressed as: mgG = m (gr,G AI,G ) (3-83)

However, the situation in gure 3-3 is not the one assumed in this book. The hypothesis of a spherical Earth will lead to an erroneous gravity vector if the centripetal force is taken into account. Because the gravitation vector gr,G points to the center of the Earth it will be perpendicular to the Earth surface when a spherical Earth is assumed. When the centripetal force is subtracted from the gravitation vector the resulting gravity vector gG is no longer perpendicular to the Earth surface. To keep the gravity vector perpendicular to the Earth surface, within the framework of the assumption of a spherical Earth, the centripetal acceleration is omitted, i.e.: (3-84) mgG = mgr,G Using the assumption of a spherical Earth, the gravity vector is expressed in the vehicle carried normal Earth reference frame FO as: 0 O O 0 gG = gr,G = (3-85)
G mt (Rt +h)2

Where:

G : the constant of gravitation mt : Earth total mass Rt : mean Earth radius

6.664 10 - 11 SI 5.983 1024 kg 6.368 106 m

The gravity vector can be rewritten as a function of the gravity at zero altitude gr , 0: gr,0 = gr = Gmt 2 Rt gr,0 1+
h Rt 2

(3-86) (3-87)

When investigating equation (3-87) one can see that the gravity does not vary much with altitude. To have a dierence of 1% in gravity one would have to travel 32 km in height: gr,2 gr,0 = 0.01 = gr,0 1+ h2 Rt
2

1 1+
h2 Rt 2

(3-88)

1 0.99 1 1 0.99

(3-89) (3-90) (3-91)

h2 = Rt 32 km

3-4 External forces and moments

101

Ellipsoid Earth G gr,G gG A AI,G

Spherical Earth G gr,G gG A AI,G

Figure 3-3: Earth gravity vector denition

This means that for normal aircraft the gravity vector magnitude can be assumed to be constant. For spaceight, however, the distances traveled in height are much larger creating larger changes in gravity, and the variation of the magnitude of gravity can no longer be neglected. Another issue is the fact that the gravity experienced by a body traveling over the Earth surface changes with latitude. This eect is caused by the attening of the Earth. A simple model for the gravity at zero altitude as a function of latitude was developed by Somigliana in 1929. The numerical version of that model is given by equation (3-92) (Ref. [30]). gr,0 = 9.8703 1.00335 + 0.00195 sin 2 LtG (3-92) where gr,0 is the magnitude of the gravity. To determine the gravity force one has to multiply the gravity acceleration with the total mass of the body. This force is also known as the weight of the body, i.e. WG = mgG (3-93)

Aerodynamic forces The ow of particles around a body creates pressure dierences that cause the body to experience a force. For ight dynamics the considered particles are air particles, hence the forces are called aerodynamic forces. The overall resultant aerodynamic force which acts on the body can be composed by a summation of aerodynamic forces which act on each part of the aircraft. This includes the forces generated by the aerodynamic surfaces and by the engines, i.e. thrust. The thrust can be seen as an aerodynamic force since an engine creates disturbances in the airow such that the aircraft is pulled forward. The resultant aerodynamic force is usually decomposed into either the body-xed reference frame or the aerodynamic reference frame. The decomposition into the body-xed reference frame is denoted as: b X b Yb Faero = (3-94) Zb
Flight Dynamics

102

Derivation of the Equations of Motion

The two decompositions of the resultant aerodynamic force are related as follows:
b Fa aero = Tab Faero a b X cos a cos a sin a cos a sin a X Y b = sin a cos a cos a sin a sin a Y a Zb sin a 0 cos a Za

The decomposition into the aerodynamic reference frame Fa is denoted as: a X Fa = Y a aero Za

(3-95)

(3-96)

3-4-2

External moments

The external moments are usually dened about the center of gravity G. Of course the point where the external forces act on the body can dier from the center of gravity, and they will then cause a moment about the center of gravity. In this section the external forces are reviewed to nd out whether they produce additional moments about the center of gravity. Gravity Since the gravity force mgG acts in the center of gravity, it will not cause a moment about the center of gravity. Aerodynamic forces The resultant aerodynamic force typically does not act in the center of gravity. It therefore causes a moment about the center of gravity. The resultant aerodynamic moment is dened in the body-xed reference frame as: b L Mb = Mb (3-97) aero,G Nb The resultant aerodynamic moment is dened in the aerodynamic reference frame as: a L Ma = Ma (3-98) aero,G a N Further details about the aerodynamic moment about the center of gravity will be given in chapter 4.

3-5

Composition of equations of motion

The components necessary to build the equation of motion have been dened in the previous sections. The complete set of equations of motions consists of fteen scalar equations: three

3-5 Composition of equations of motion

103

force equations, three moment equations, and nine kinematic relations1 . The equations of motion will be expressed in the body-xed reference frame with the origin in the center of mass. Three assumptions are made:

Constant gravity eld, i.e. center of mass is equal to the center of gravity (= G) Vehicle is a rigid body Vehicle has constant mass

First the force equations will be derived (section 3-5-1), then the moment equations (in section 3-5-2). The linear and angular velocities used will be stated in the kinematics section, section 3-5-3. It is important to note that the eects of rotating masses on the rotational accelerations is not taken into account here, i.e. the moment equations are not complete. These eects will be appended to the equations of motion in section 3-6.

3-5-1

Force equations

The derivation of the force equations starts with the basic formula of Newtons second law for rigid, constant mass body: mAb = Fb I,G ext (3-99)

Gathering the information obtained in previous sections, the force equation can be written as:
b mAb + mAb + mAb = mgG + Fb r S R aero

(3-100)

Which is equal to:

O mTbO AO + AO + AO = mTbO gG + Fb r S R aero

(3-101)

In literature one often considers twelve scalar equations. In this book, however, three additional (rotational) kinematic relations are added to make the equations of motion easier to read.
Flight Dynamics

104 with: cos cos sin sin cos cos sin

Derivation of the Equations of Motion

cos sin sin sin sin + cos cos

TbO

cos sin cos cos sin sin + sin sin sin cos VN AO = VE r VZ 2 VN VZ + VE tan LtG 1 VE VZ VE VN tan LtG AO = S Rt + h 2 2 VE + VN 2VE sin LtG + t (Rt + h) sin LtG cos LtG AO =t 2VZ cos LtG 2VN sin LtG R
O gG =

sin cos cos cos

sin

0 0

gr,O

2VE cos LtG + t (Rt + h) cos2 LtG

Fb aero

Xb = Y b Yb

The denition of the aerodynamic forces expressed in the body-xed reference frame will be given in chapter 4.

3-5-2

Moment equations

The derivation of the moment equations starts with the modied form of the basic formula of Newtons second law for rigid, constant mass body: dBG dt
b

= Mb G,ext
I

(3-102)

Gathering the information obtained in previous sections, the force equation can be written as: 1 Mb + 2 Mb + 3 Mb = Mb G G G aero,G (3-103)

3-5 Composition of equations of motion With: Ap + (C B) qr E (pq + r) + F (pr q) + D r 2 q 2 1 Mb = B q + (A C) pr F (qr + p) + D (pq r) + E p2 r 2 G C r + (B A) pq D (pr + q) + E (qr p) + F q 2 p2 2 q2 (C B) qt rt Ept qt + F pt rt + D rt t 2 2 Mb = (A C) pt rt F qt rt + Dpt qt + E p2 rt t G 2 p2 (B A) pt qt Dpt rt + Eqt rt + F qt t (C A B) qrt + (C + A B) rqt 3 Mb = (A + B C) prt + (A B C) rpt G (B + C A) qpt + (B C A) pqt pt (F r Eq) + D (rrt qqt ) + 2 qt (Dp F r) + E (ppt rrt ) rt (Eq Dp) + F (qqt ppt ) b L = Mb Mb aero,G Nb

105

3-5-3

Kinematic relations

The linear velocity scalars are related to the changes in latitude and longitude in the following manner: Vk,G = drG dt + OE rG (3-104)

Where:

Expressed in the vehicle carried normal Earth reference frame, this relation can be written as: 0 LgG cos LtG 0 VN VE = 0 + 0 Lt G sin LtG (Rt + h) VZ h Lg Lt (Rt + h) G cos LtG (Rt + h) = Lg (3-105) h LgG =tG tO = tG t Lt =LtG LtO = LtG (3-106) (3-107)

The angular velocity scalars are related to the changes in latitude and longitude and the changes in vehicle orientation with respect to the Earth surface, i.e. with respect to the vehicle carried normal Earth reference frame. The rst set of angular velocity scalars can be represented by: bE = bO + OE (3-108)
Flight Dynamics

106

Derivation of the Equations of Motion

Or, likewise: 1 sin tan cos tan LgG cos LtG p = 0 cos sin q TbO LtG sin cos r 0 LgG sin LtG cos cos

Expressed in the body-xed reference frame, this relation can be written as: p b = q = b + TbO O bE bO OE r LgG cos LtG 1 0 sin = 0 cos cos sin + TbO LtG 0 sin cos cos LgG sin LtG

(3-109)

(3-110)

The second set of angular velocity scalars denote the angular relation between the normal Earth xed reference frame and the inertial reference frame expressed in the body-xed reference frame: b pt pt b b = qt = qt = TbO O EI EI b rt rt cos LtG cos cos + sin LtG sin sin sin cos sin LtG sin cos cos LtG = t cos sin (3-111) cos sin cos cos LtG sin LtG cos cos + sin sin

3-6

Rotating masses

When deriving the angular momentum of the aircraft, the assumption was made that the coordinates xb , yb and zb of all mass elements are constant as the aircraft was considered to be a rigid body. In many cases the aircraft has one or more rotating masses, such as propellers or rotors, either helicopters rotors or the rotating masses of jet engines. Due to their own rotation, these rotors may provide an appreciable contribution to the angular momentum. In the following, this contribution is derived. The derivation starts with the assumption that the axis of rotation of the rotating mass does not change its attitude relative to the body-xed reference frame. In the following the case of an engine containing rotors is considered2 . The angular velocity r about this axis of rotation is then dependent on the commanded thrust level. The situation to be considered is shown in gure 3-4. There are two rotations which transform the Xb -axis to the rotation-axis orientation. The rst rotation is about the Yb -axis, rotation angle m . The second rotation is about the Zb (intermediate axis), rotation angle m .
2

The derivation stated in this section is valid for any rotating mass.

3-6 Rotating masses

G

107

Yb Xb

rgp rgq

rpq
Q

Zb

VP r m

// to Xb

Rotation axis

Figure 3-4: Engine rotation axis denition

The position of an arbitrary mass element dm, P , with respect to the center of mass G of the aircraft is, GP = GQ + QP (3-112)

where Q is the center of mass of the rotating mass (in this case an engine rotor). The velocity of dm now is: V P = V P0 + r QP (3-113)

where V P0 , is the velocity of dm if the engine rotor does not rotate at angular velocity r . The contribution of dm to the total angular momentum of the aircraft about its center of mass is:

dB G = dm GP V P = dB 1 + dB 2

= dm GP V P0 + dm GP (r QP )

(3-114)

Here dB 1 is the contribution to the angular momentum, as derived in section 3-3. The additional part, due to the angular velocity of the engine rotor, is:

dB 2 = dm GP (r QP )

= dm GQ (r QP ) + dm QP (r QP )

(3-115)
Flight Dynamics

108

Derivation of the Equations of Motion

In the integration across the entire rotor GQ and r in the rst term at the right hand of the latter expression are constants, hence,

mrotor

dm GQ (r QP ) = GQ r

mrotor

dm QP = 0

(3-116)

This integral is equal to zero, as QP denotes the position of the mass element dm relative to the center of mass Q of the engine rotor. As a result, dB 2 may now be written as,

dB 2 = dm QP (r QP )

2 2 2 = dm r Rx + Ry + Rz QP rx Rx + ry Ry + rz Rz

= dm {r (QP QP ) QP (r QP )}

(3-117)

where Rx , Ry , Rz are the components of the vector QP along the X, Y , and Z-axis of the chosen reference frame respectively. The required addition to the angular momentum of the aircraft is,

B G =
mrotor

dB 2

(3-118)

Expansion results in the components of B G , Bx = rx

2 2 Ry + Rz dm ry

mrotor mrotor mrotor

By = rx Bz = rx

Rx Ry dm + ry Rx Rz dm ry

mrotor mrotor

mrotor 2 Rx

Rx Ry dm rz

Rx Rz dm
mrotor

2 + Rz dm rz

Ry Rz dm
mrotor 2 Rx + 2 Ry dm

Ry Rz dm + rz

mrotor

In these expressions the inertial parameters of the rotor are taken about axes parallel to the body-xed reference frame axes, as it applies to the inertial parameters of the aircraft itself. The moments of inertia are,

Ixxr Iyyr Izzr

=
mrotor

2 2 Ry + Rz dm

(3-119) (3-120) (3-121)

=
mrotor

2 2 Rx + Rz dm

=
mrotor

2 2 Rx + Ry dm

3-6 Rotating masses and the products of inertia,

109

Jyzr Jxzr Jxyr

=
mrotor

Ry Rz dm Rx Rz dm
mrotor

(3-122) (3-123) (3-124) (3-125)

= =
mrotor

Rx Ry dm

and hence the addition to the angular momentum of the rotor expressed in the body-xed reference frame can be dened as: Bb = Ib b G rotor r
b Bx

(3-126)

b By = Jxyr b Jxzr Bz

Ixxr

Jxyr Iyyr Jyzr

Jxzr Jyzr Izzr

b x r

b r y b z r

(3-127)

The rotor angular velocity expressed in the body-xed reference frame is dened as: b = Tbr r r r (3-128)

where Tbr is the transition matrix from the reference frame of the engine3 to the body-xed reference frame. The relation can be written as: b x cos m cos m sin m cos m sin m r r b = 0 sin m cos m 0 ry b cos m sin m sin m sin m cos m 0 rz cos m cos m = r sin m cos m sin m pr = qr rr B G provides an extra contribution in the moment equation:
Which is dened as, Xr -axis along rotor angular velocity vector, Zr -axis is aircraft symmetry plane, and Yr -axis to complement the right-handed reference frame
Flight Dynamics
3

(3-129)

110

Derivation of the Equations of Motion

4 Mb = G

dB G dt

b I b b

dB G = dt

+ b B b G bI
b b

dIrotor r = dt

+ b B b + b B b bE G EI G

(3-130)

It is assumed that the rotor will stay intact during the operation of the engine, such that the moment of inertia of the rotor will stay constant with respect to the body-xed reference frame, i.e.: 4 Mb =Ib G rotor dr dt
b b

+ b B b + b B b bE G EI G

(3-131)

The derivative of the rotor angular velocity vector with respect to the body-xed reference frame is simply: dr dt
b b

cos m cos m pr = qr = r sin m sin m sin m rr

(3-132)

The denitions for the rotation vectors b and b are (see section 3-3, p. 96): EI bE p pt = q , b = qt EI rt r

b bE

(3-133)

The expression for the additional moment due to rotating masses expressed in the body-xed reference frame can be written as:
b b b b 4 MG = 4a MG + 4b MG + 4c MG

(3-134)

with:
b 4a MG = Ib rotor b 4b MG

dr dt

b b

b bE

Ib b rotor r

b 4c MG = b Ib b EI rotor r

Analogous to the derivation of 3 (M)b on page 96 till 99 the expressions for the three G

3-7 Simplications of the equations of motion components are: pr Ar Fr Er b 4a MG = Fr Br Dr qr Er Dr Cr rr Ar pr Fr qr Er rr Br qr Fr pr Dr rr = Cr rr Er pr Dr qr pr p Ar Fr Er b 4b MG = q Fr Br Dr qr Er Dr Cr rr r Cqrr Eqpr Brqr + F rpr + D (rrr qqr ) = Arpr F rqr Cprr + Dpqr + E (ppr rrr ) Bpqr Dprr Aqpr + Eqrr + F (qqr ppr ) pt Ar Fr Er pr b 4c MG = qt Fr Br Dr qr rt Er Dr Cr rr Cqt rr Eqt pr Brt qr + F rt pr + D (rt rr qt qr ) = Art pr F rt qr Cpt rr + Dpt qr + E (pt pr rt rr ) Bpt qr Dpt rr Aqt pr + Eqt rr + F (qt qr pt pr )

111

(3-135)

(3-136)

(3-137)

where the subscript r denotes that the moments of inertia (A, B, C) and the products of inertia (D, E, F ) are the properties of the rotor. The total moment equation dened in section 3-5-2 (equation (3-103), page 104) has to be extended to the form: 1 Mb + 2 Mb + 3 Mb + 4 Mb = Mb G G G G aero,G with 4 Mb consisting of the three components dened previously. G (3-138)

3-7

Simplications of the equations of motion

The set of equations of motion derived in section 3-6 are for a spherical, rotating Earth considering a rigid vehicle with constant mass moving in a constant gravity eld. In this section, the inuence of two commonly made assumptions on the equations of motion will be investigated. They are: Non-rotating Earth This assumption greatly reduces the complexity of the equations of motion. However, depending on the study case, it can introduce large errors during computations. Flat Earth If the distance traveled during the study case is small, then the curvature of the Earth will not play an important role in the changes of vehicle state. One could therefore assume that the Earth is at, which simplies the equations of motion. For larger distances, this assumption will cause signicant errors.
Flight Dynamics

112

Derivation of the Equations of Motion

3-7-1

Non-rotating Earth

If the Earth is assumed to be non-rotating then the angular velocity of the normal Earth xed reference frame with respect to the inertial reference frame will be zero, i.e. t = 0. This will simplify the equations of motions since in the force equations the term due to the rotation of the Earth AR will be zero. The force equations thus become (see equation (3-101), p. 103):
O mTbO AO + AO = mTbO gG + Fb r S aero

(3-139)

In the moment equations the terms 2 MG , 3 MG , and 4c MG will be zero (since EI = 0). The moment equations thus become (see equation (3-138), p. 111): 1 Mb + 4a Mb + 4b Mb = Mb G G G aero,G The kinematic relations which contain terms with t will become zero. (3-140)

3-7-2

Flat Earth

If one assumes that the Earth is at, then the normal Earth xed reference frame will coincide with the vehicle carried normal Earth reference frame. Therefore the angular velocity between those two reference frames will be zero, i.e. OE = 0. Using this in equation (3-49), one can see that the term AS in the force equations will become zero. Also, the term AR changes. The relation given by equation (3-50) will become: AR = dEI dt dEI dt rG + EI drG dt drG dt + EI Vk,G

+ (OE + EI ) (EI rG ) =
O

rG + EI

+ EI Vk,G (3-141)

+ EI (EI rG )

Only the last term of the relation has changed. Expressing the last term in the vehicle carried normal Earth reference frame will lead to: O O rO EI EI G

= O O rO rO O O EI EI G G EI EI T t cos LtG t cos LtG 0 = 0 0 0 t sin LtG t sin LtG (Rt + h) T 0 t cos LtG t cos LtG 0 0 0 (Rt + h) t sin LtG t sin LtG

(3-142)

3-7 Simplications of the equations of motion t cos LtG (t (Rt + h) sin LtG ) 0 t sin LtG 0 2 cos2 LtG + 2 sin2 LtG 0 t t (Rt + h) 2 (Rt + h) sin LtG cos LtG t 0 2 (R + h) cos2 Lt t t G

113

The relations given by the moment equation remain the same. Only the denition of the term b will change: bE b bE p = q = b + TbO O = b OE bO bO r 1 0 sin = 0 cos cos sin 0 sin cos cos

(3-143)

(3-144)

The consequences for the kinematic relations is that all terms which contain OE will become zero.

3-7-3

Flat and Non-rotating Earth

Combining the assumptions of a at Earth and a non-rotating Earth will result in the following force equations:
O mTbO AO = mTbO gG + Fb r aero

(3-145)

and the following moment equations: 1 Mb + 4a Mb + 4b Mb = Mb G G G aero,G with: b bE p 1 0 sin = q = 0 cos cos sin r 0 sin cos cos (3-146)

The relations given in this section can directly be used to obtain the scalar form of the equations of motion for a at, non-rotating Earth expressed in the body-xed reference frame. To show that the relations in this section are correct, another derivation is performed. The following tutorial will show this second derivation of the equations of motion for a at, nonrotating Earth and no wind. The derivation is similar to one used to obtain the general equations of motion but it is much more simpler. The simplied set of equations will be used in subsequent chapters to demonstrate the principles of ight dynamics.
Flight Dynamics

114

Derivation of the Equations of Motion

Mini-tutorial: Derivation of the equations of motion for a at, non-rotating Earth, no wind and no rotating masses. A simplied set of equations of motion is derived that will be used in the remainder of this book. The assumptions made here are: Vehicle is a rigid body. Vehicle mass is constant. Earth is at. Earth is non-rotating. Body-xed reference frame is chosen such that Jxy and Jyz are zero. Eects of rotating masses are neglected. Since the Earth is assumed to be at the vehicle carried normal earth reference frame and the normal Earth xed reference frame coincide. Because the Earth is assumed to be non-rotating, those two reference frames can also be seen as inertial reference frames, so: AE,G = AO,G = AG VE,G = VO,G = VG rE,G = rO,G = rG Here the normal Earth xed reference frame will be selected as inertial reference frame. The equations of motion will be expressed in the body-xed reference frame (with origin in the center of gravity G). The derivation starts with the force equations:
b mAb = Fext G

(3-147)

The inertial acceleration in the vehicle carried normal Earth reference frame can be rewritten as: Ab = G dVG dt
b E

The external forces on the vehicle remain the same as given in section 3-4 (keeping in mind that TbE = TbO in case of a at earth): b 0 X b 0 + Yb Fext = TbO Zb mgr,0 sin X = mgr,0 sin cos + Y (3-148) cos cos Z

(3-149)

3-7 Simplications of the equations of motion

115

Since there is no wind, the kinematic velocity is equal to the aerodynamic velocity. We can therefore drop the sub-scripts and dene the vector [u, v, w] as the inertial velocity vector expressed in the body-xed reference frame, i.e.: b b uk ua u b v b = va = v (3-151) k b b wk wa w

dVG b b + b VG bE dt b b b p uk uk b = vk + q vk b b wk b r wk =

(3-150)

Similar to the force equations, one can write for the moment equations: dBG dt =
I

Substituting relation (3-150) into equation (3-150) and substituting the result in equation (3-147) along with equation (3-148) will lead to the following set of force equations: sin X u + qw rv (3-152) m v + ru pw = mgr,0 sin cos + Y cos cos w + pv qu Z dBG dt dBG dt

=
E

= Mext
O

(3-153)

When neglecting the eect of rotating masses we can write the derivative of the inertial angular momentum as: dBG dt
b E

dBG = dt = Ib G

b b

+ b Bb G bO
b b

dbE dt

+ b Ib b G bE bE

(3-154)

By substituting: p b = q bE r A F E Ixx 0 Jxz Ib = F B D = 0 Iyy 0 G E D C Jxz 0 Izz (3-155)

(3-156)

Flight Dynamics

116

Derivation of the Equations of Motion

into equation (3-154) and substituting the result into equation (3-153) will lead to the following moment equations: Ixx p + (Izz Iyy ) qr Jxz (pq + r) L Iyy q + (Ixx Izz ) pr + Jxz p2 r 2 = M (3-157) Izz r + (Iyy Ixx ) pq + Jxz (qr p) N

The kinematic relations describe the relation of angular velocities between reference frames and the relation of translational velocities between reference frames. First the angular velocities. Only two reference frames are used, the body-xed reference frame and the normal earth-xed reference frame. The angular velocity expressed in Fb is related to the angular velocity expressed in FE by the following expression: p 1 0 sin q = 0 cos cos sin (3-158) r 0 sin cos cos For the relation between velocities, the following expression holds: VN u u u VE = TEb v = TOb v = Tt v bO w w w VZ sin sin cos cos sin cos cos cos cos sin + sin sin = sin sin sin cos sin sin cos sin + cos cos sin cos sin sin cos cos cos [u cos + (v sin + w cos ) sin ] cos (v cos w sin ) sin [u cos + (v sin + w cos ) sin ] sin = + (v cos w sin ) cos u sin (v sin + w cos ) cos

u v w (3-159)

The displacement of the aircrafts center of gravity relative to the earth is obtained by integrating the three components of the ground speed with respect to time, hence,
t

x(t) =
0 t

VN dt

(3-160)

y(t) =
0 t

VE dt

(3-161)

h(t) =
0

VZ dt

(3-162)

Chapter 4 Linearized Equations of Motion

The equations of motion of an aircraft may be used in several dierent ways. Most commonly, a disturbance function is given, such as a time-dependent control surface deection or a gust velocity, and the response of the aircraft is then calculated. An inverse situation of this problem also exists, when a certain non-steady motion of the aircraft is given and the control surface deection as a function of time causing this motion is to be found. In a third application of the equations of motions both the disturbing control surface deection and the resulting aircraft motion have been measured in ight. The purpose is then to nd expressions for the aerodynamic forces and moments in the equations of motions, as functions of the components of the motion and the disturbances, or stated dierently, to dins a mathematical model of the aircraft. The problems described above can be solved by a variety of numerical algorithms. For example the numerical integration routines provided by MatLab c may readily be applied to solve the set of nonlinear ordinary dierential equations derived in chapter 3, even in real time as needed for ight simulation. The present chapter focusses on the linearization of these nonlinear dierential equations, resulting in a set of linear dierential equations describing small deviations from a chosen reference condition. Performing simulations with a set of linearized equations will always lead to less accurate results. So why should we derive linearized versions of the equations of motion? The answer is that these linearized equations of motion enable us to: study the so-called characteristic modes of aircraft, derive important handling quality and stability characteristic parameters design manual or automatic ight control systems using well established linear control theories. The goal of this book is to provide insight into the characteristics of aircraft, presented in two ways: through determination of the characteristics modes on the one hand and of
Flight Dynamics

118

Linearized Equations of Motion

stability and control derivatives on the other. Characteristic modes are described by the impulse response of the aircraft and are completely determined by the eigenvalues of the system. Stability and control derivatives describe the relation between the variation of the aircraft state due to variation of that state and due to variation of control input respectively. The knowledge about the characteristics of the aircraft can be used to design control systems, such as autonomous control systems, like an autopilot, or performance increasing control systems to be used during manual control. The set of equations of motion used in this chapter is obtained by applying the following assumptions: Vehicle is a rigid body. Vehicles mass is constant. Earth is at. Earth is non-rotating. Gravity eld is constant. Aircraft has a plane of symmetry and the body-xed reference frame is chosen such that Jxy and Jyz are zero. Eects of rotating masses are neglected. The resultant thrust vector lies in the symmetry plane and thus only aects the aerodynamic forces X, Z and the aerodynamic moment M . Moreover we assume that the aircraft has a normal conguration with respect to control surfaces, i.e. it has ailerons (deection a ), an elevator (deection e ), a rudder (deection r ), and engines (setting t ). The complete set of equations can be described by (see section 3-7-3): Fx = Fy = Fz = Mx = My = Mz = W sin + X =m (u + qw rv)

+W cos sin + Y

+W cos cos + Z L M N

=m (v + ru pw)

=m (w + pv qu) =Ix p + (Iz Iy ) qr Jxz (r + pq)

(4-1)

=Iy q + (Ix Iz ) rp + Jxz p2 r 2 =Iz r + (Iy Ix ) pq Jxz (p rq)

where W = mgr,0 has been used. are: = = =

The kinematic relations accompanying this set of equations p + q sin tan + r cos tan q cos r sin q sin + r cos cos cos

Linearization of the equations of motion can be performed about any given ight condition. The results generated with the linearized equations are valid for that ight condition and

4-1 Linearization about arbitrary ight condition in arbitrary body-xed reference frame

119

for the conditions close to that particular point (see gure 4-1). This means that there is a certain validity-region about the initial state, i.e. the state about which the equations are linearized. It depends on the chosen ight condition and on the desired equation of motion output accuracy whether or not the aircraft moves quickly out of this region or not. If, for instance, a steady state ight condition is chosen, the aircraft will remain close to this condition if no changes are made to the inputs. On the other hand, if we select a instable or non-steady ight condition (with the same desired accuracy), the aircraft will deviate quickly from this point leaving us with a set of equations incapable of describing the motion accurately. If we want to simulate the response of the aircraft along a particular trajectory which consists of multiple ight conditions, we could linearize the equations for multiple conditions and switch between sets as we move along the trajectory (see gure 4-1). Especially when we want to design a control system for the entire trajectory (for instance, to keep the aircraft on that trajectory) we can design a controller for each ight condition and switch between controllers. Also, linearizing the equations for multiple ight conditions gives us insight into the variation of the aircraft characteristics.

Linearization about one ight condition or about an entire ight trajectory boils down to the same thing. The technique of linearization about an arbitrary ight condition is described in section 4-1. No specic body-xed reference frame is selected. In practice a steady state ight condition is usually chosen for the linearization. A steady state, straight, symmetric ight condition is chosen in section 4-2 for the linearization. At rst no choices are made regarding the type of body-xed reference frame. Thereafter we do make a choice and derive the linearized equation set for the same ight condition. The non-dimensional form the latter set of equations is derived in section 4-3. The set is made non-dimensional to decouple the aerodynamic properties of the aircraft from its dimensions and mass, creating the opportunity to compare dierent aircraft types.

4-1

Linearization about arbitrary ight condition in arbitrary bodyxed reference frame

Linearization can be performed on any function which is dierentiable. Consider a general function: Y = f (X) which states that the output Y is a function of the state X which can have any dimension. This function can also be written in a Taylor expansion form. If we consider the state to be one dimensional and the function to be real then the Taylor expansion about state x0 is given by: f (x0 ) y =f (x0 ) + f (x0 ) (x x0 ) + (x x0 )2 2! f (3) (x0 ) f (n) (x0 ) + (x x0 )3 + . . . + (x x0 )n + . . . 3! n!

(4-2)
Flight Dynamics

120

Linearized Equations of Motion

x2 condence region x20

x10

x1

(a) Point-linearization

x2

x2

x1

x1

(b) Trajectory-linearization

Figure 4-1: Linearization about a single point and a trajectory

4-1 Linearization about arbitrary ight condition in arbitrary body-xed reference frame

121

where f denotes the rst derivative of function f with respect to x, and f denotes the second derivative etc. If the state X is two dimensional, e.g. X = (x, u), the Taylor expansion about state (x0 , u0 ) of a real function can be written as: y =f (x0 , u0 ) + [fx (x0 , u0 ) x + fu (x0 , u0 ) u] 1 fxx (x0 , u0 ) x2 + 2fxu (x0 , u0 ) xu + fuu (x0 , u0 ) u2 + 2! 1 + fxxx (x0 , u0 ) x3 + 3fxxu (x0 , u0 ) x2 u 3! + 3fxuu (x0 , u0 ) xu2 + fuuu (x0 , u0 ) u3 + ... (4-3)

where fx denotes the partial derivative of f with respect to x and fu denotes the partial derivative of f with respect to u, etc. Of course the Taylor expansion can become rather large when the dimension of the state grows, however, the principle remains the same. Linearization of a function about a point (known as the initial point of linearization or simply the initial point) is dened as expanding the function in a Taylor series and taking only the rst term and the terms with a rst (partial) derivative from that expansion. In other words disregarding terms which contain (partial) derivatives of order higher than 1. For a function with an n-dimensional state the linearized function about point X0 becomes: Y = f (X0 ) + fx1 (X0 ) x1 + fx2 (X0 ) x2 + . . . + fxn (X0 ) xn
n

= f (X0 ) +
i=1

fxi (X0 ) xi

(4-4)

The function can be replaced by the relations given by the equations of motion. In the next section the linearization of the states, i.e. the right hand side of the equations, is performed while in section 4-1-2 the forces and moments, i.e. the left hand side of the equations, are linearized. In section 4-1-3 the kinematic relations are linearized.

4-1-1

Linearization of states

The equations of motion can be written as: Fx =f (u, v, w, q, r) Fy =f (v, u, w, p, r) Fz =f (w, u, v, p, q) Mx =f (p, r, p, q, r) My =f (q, p, r) Mz =f (p, r, p, q, r) (4-5)
Flight Dynamics

122

Linearized Equations of Motion

Applying the Taylor expansion and neglecting the higher order terms will give: Fx =f (X0 ) + fu (X0 ) u + fv (X0 ) v + fw (X0 ) w + fq (X0 ) q + fr (X0 ) r Fy =f (X0 ) + fv (X0 ) v + fu (X0 ) u + fw (X0 ) w + fp (X0 ) p + fr (X0 ) r Fz =f (X0 ) + fw (X0 ) w + fu (X0 ) u + fv (X0 ) v + fp (X0 ) p + fq (X0 ) q Mx =f (X0 ) + fp (X0 ) p + fr (X0 ) r + fp (X0 ) p + fq (X0 ) q + fr (X0 ) r My =f (X0 ) + fq (X0 ) q + fp (X0 ) p + fr (X0 ) r Mz =f (X0 ) + fp (X0 ) p + fr (X0 ) r + fp (X0 ) p + fq (X0 ) q + fr (X0 ) r (4-6)

where X0 denotes the state about which the linearization is performed. Taking the specic partial derivatives of the function using the equations in (4-1) and substituting the initial state for linearization results in: Fx = Fy = Fz = Mx = My = Mz = m (u0 + q0 w0 r0 v0 ) + m (u r0 v + q0 w + w0 q v0 r) m (v0 + r0 u0 p0 w0 ) + m (v + r0 u p0 w w0 p + u0 r) m (w0 + p0 v0 q0 u0 ) + m (w q0 u + p0 v + v0 p u0 q) Ix p0 + (Iz Iy ) q0 r0 Jxz (r0 + p0 q0 ) + Ix p Jxz r Jxz q0 p + [(Iz Iy ) r0 Jxz p0 ] q + (Iz Iy ) q0 r Iy q0 + (Ix Iz ) p0 r0 + Jxz p2 0 +
2 r0

(4-7)

+Iy q + [(Ix Iz ) r0 + Jxz p0 ] p + [(Ix Iz ) p0 + Jxz r0 ] r Ix r0 + (Iy Ix ) p0 q0 Jxz (p0 q0 r0 ) Jxz p + Ix r + (Iy Ix ) q0 p + [(Iy Ix ) p0 + Jxz r0 ] q + Jxz q0 r

4-1-2

Linearization of forces and moments

The left hand side of the equations of motion can be written as: Fx = f (, X) Fy = f (, , Y ) Fz = f (, , Z) Mx = f (L) f (M ) My = Mz = f (N )

(4-8)

where X, Y, Z are the aerodynamic forces and L, M, N are the aerodynamic moments, they are all dependent on the entire history of the aircraft state in time! The question is which states inuence which forces and moments. The most general framework to be chosen is to assume that the aerodynamic forces and moments are functions of the components of the motion and control surface deections, i.e.: X, Y, Z, L, M, N f (u(.), v(.), w(.), p(.), q(.), r(.), a (.), e (.), r (.), t (.)) (4-9)

where the notation (.) denote the dependence on the time-history of the parameters. The (.) function contains information regarding the entire time history of a specic parameter. The aerodynamic forces and moments are thus a function of time-functions.

4-1 Linearization about arbitrary ight condition in arbitrary body-xed reference frame

123

u(.)

u(t)

Figure 4-2: Example of a time-function: u(.)

Let us take a look at the time-functions, taking for example u(.), i.e. the time-function of the velocity along the Xb -axis (see gure 4-2). The information regarding the entire time history must now be molded into a form t for the equations of motion. The method adopted here is to replace the time-function u(.) by the following Taylor series: u (.) = u (t) + u (t) t +

1 1 ... u (t) t2 + u (t) t3 + . . . 2! 3! (4-10)

= u (t) +
i=1

1 di u i t i! dti

It is assumed that if we take into account all possible derivatives of u(t), that we have captured all information regarding the parameter history. In theory all the derivatives have an inuence on the value of u(.) on time t. However, it has been observed in practice that for most parameters the inuences of the derivatives is limited and can be neglected in the relation (i.e. neglecting the summation term in equation (4-10)). A few exceptions are the derivatives Y N Z v and v with respect to the acceleration v along the YB -axis and the derivatives w and M w with respect to the acceleration w along the ZB -axis. Of these four partial derivatives, the two relating the moments N and M (and of these two again M ) are by far the most v w w important. In the following, all derivatives (with the exception of the previously mentioned four derivatives) will be neglected. By neglecting the inuence of the derivatives we are in fact neglecting the inuence of parameter variation through time. These inuences arise from non-stationary wing-fuselage and tail interference. An example is given in the next mini-tutorial.

Flight Dynamics

124

Linearized Equations of Motion

Mini-tutorial: Inuence of non-stationary wing-fuselage and tail interference on aerodynamic forces and moments. We look here at the wing-fuselage and tail interference in the case of cross-wind, i.e. v = 0. The discussion in this tutorial is hypothetical and is not proven here, it is only meant to present some insight into why the aerodynamic forces and moments are dependent on the history of all motion parameters. Consider an aircraft ying at a velocity V which experiences at a certain moment a sudden cross-wind at time t (see gure 4-3(a)). Let us focus momentarily on the airow over the wing and fuselage due to the cross-wind only (see gure 4-3(b)). At the root of the wing the airow encounters the fuselage. Since the airow cannot penetrate the fuselage it needs to follow a path around the fuselage. The airow will be split in two, one part will travel over the fuselage and one will go underneath. The airow underneath the fuselage will travel around the wing downwards, causing a decrease in the angle of attack for that section of the wing which leads to a decrease in lift. On the other side of the fuselage the opposite happens. So the lift on one side decreases and on the other side increases. This will cause a moment about the Xb -axisa Now consider the lift generating mechanism. The magnitude of lift is directly related with the strength of the vortices generated by the wings. If the lift decreases then the vortex strength will also decrease and vice versa. So, in our situation, we have a local decrease in vortex strength on one side and an increase on the other. This can be represented by a placing a imaginary vortex with its centerline on the Xb -axis (see gure 4-3(c)). After time t the vortex hits the tail, therefore the tail will experienced a dierent airow. The forces generated by this change in airow will cause a rolling and yawing moment. This example shows that events on time t will still have eect at time t + t, showing the time-dependency of the aerodynamic force and moments on the motion parameters.
a

Under the assumption that the Xb -axis is aligned with the fuselage.

By neglecting the time-dependency, the relations given by equation (4-9) can thus be redened as:

X, Y, Z, L, M, N f (u(t), v(t), v(t), w(t), w(t), p(t), q(t), r(t), a (t), e (t), r (t), t (t)) (4-11) When dropping the time notation (t), the relations given equation (4-8) can be rewritten as:

Fx Fy Fz Mx My Mz

= = = = = =

f (, u, v, w, p, q, r, a , e , r , t ) f (, , u, v, w, v, p, q, r, a , e , r ) f (, , u, v, w, w, p, q, r, a , e , r , t ) f (u, v, w, p, q, r, a , e , r ) f (u, v, w, w, p, q, r, a , e , r , t ) f (u, v, w, v, p, q, r, a , e , r )

(4-12)

4-1 Linearization about arbitrary ight condition in arbitrary body-xed reference frame

125

Airow direction

(a) Cross-wind relative to the aircraft

Line of stagnation points

Streamlines

Increased Angle of attack Decreased Angle of attack Airow direction

(b) Streamlines over wing and fuselage due to cross-wind

F

(c) Additional forces on tail due to time-delayed vortices

Figure 4-3: Airow tutorial: example of time-dependent aerodynamic characteristics

Flight Dynamics

126 Linearizing these equations leads to: Fx Fy Fz Mx My Mz with Fx (X0 ) = Fy (X0 ) = Fz (X0 ) = Mx (X0 ) = My (X0 ) = Mz (X0 ) = = = = = = = Fx (X0 ) + Fx (X) Fy (X0 ) + Fy (X) Fz (X0 ) + Fz (X) Mx (X0 ) + Mx (X) My (X0 ) + My (X) Mz (X0 ) + Mz (X)

Linearized Equations of Motion

(4-13)

W sin 0 + X0 W cos 0 sin 0 + Y0 W cos 0 cos 0 + Z0 L0 M0 N0

(4-14)

where X0 , Y0 , and Z0 are the aerodynamic forces and L0 , M0 , and N0 are the aerodynamic moments for the initial state of linearization, i.e. for X0 = [u0 , v0 , w0 , p0 , q0 , r0 , a0 , e0 , r0 ]. The second part of equation 4-13 can be written as: Fx (X) = W cos 0 + Xu u + Xv v + Xw w + Xp p W sin 0 sin 0 + W cos 0 cos 0 +Ya a + Ye e + Yr r W sin 0 cos 0 W cos 0 sin 0 +Za a + Ze e + Zr r + Zt t Lu u + Lv v + Lw w +Lp p + Lq q + Lr r + La a + Le e + Lr r Mu u + Mv v + Mw w + Mw w + Mp p + Mq q +Mr r + Ma a + Me e + Mr r + Mt t Nu u + Nv v + Nw w + Nv v +Np p + Nq q + Nr r + Na a + Ne e + Nr r

+Xq q + Xr r + Xa a + Xe e + Xr r + Xt t +Yu u + Yv v + Yw w + Yv v + Yp p + Yq q + Yr r

Fy (X) =

Fz (X) =

+Zu u + Zv v + Zw w + Zw w + Zp p + Zq q + Zr r

(4-15)

Mx (X) = My (X) = Mz (X) =

As a consequence of the assumed symmetry of the aircraft and the linearization of the equations of motion a further simplication is possible. Suppose, a change in velocity +v along the YB -axis causes an extra force Xdv = Xv v = dX. Then, since the derivative is a constant, a change in velocity v will generate an extra force Xdv = Xv v = dX. Reecting the situation with v in the plane of symmetry must result in the situation with +v, see gure 4-4. This means, however, that dX must be equal to zero and, as a consequence, X cannot vary linearly with v. Only a variation with v 2 , v 4 , v 6 etcetera is possible. Such terms of higher degree however, have already been neglected during the

4-1 Linearization about arbitrary ight condition in arbitrary body-xed reference frame XB

127

+ dX due to +dv - dv + dv

YB

- dX due to dv - dX = + dX = 0

Figure 4-4: The impossibility of a force in the XB -direction arising from a variation in velocity dv along the YB -axis

linearization. This leads to the conclusion that the asymmetric deviation v has no inuence on the symmetric aerodynamic force X. A similar argument can be given for any of other combinations of symmetric and asymmetric variables. Evidently, the following rules apply,

Small asymmetric deviations and disturbances have no inuence on the symmetric forces X and Z or on the symmetric moment M Small symmetric deviations and disturbances have no inuence on the asymmetric force Y or on the asymmetric moments L and N

Expressed in a dierent way,

No aerodynamic coupling exists between the symmetric and the asymmetric degrees of freedom, as long as the deviations and disturbances remain small!
2

A single exception to this rule is the contribution M v 2 to dM , which has in many v2 practical situations a non-negligible magnitude at values of dv occurring in ight conditions where no other signicant non-linearities are noticeable. In the following, however, this contribution of dv 2 to dM will not be further considered.

Flight Dynamics

128

Linearized Equations of Motion

Under the previous assumptions, the linearized set of aerodynamic forces and moments and the weight can be written as follows: Fx = Fx (X0 ) W cos 0 + Xu u + Xw w + Xq q + Xe e + Xt t W sin 0 sin 0 + W cos 0 cos 0 Fy = Fz (X0 ) + +Yv v + Yv v + Yp p + Yr r + Ya a + Yr r Fz = Fz (X0 ) + +Zu u + Zw w + Zw w + Zq q + Ze e + Zt t W sin 0 cos 0 W cos 0 sin 0 (4-16)

Mx = Mx (X0 ) + Lv v + Lp p + Lr r + La a + Lr r My = My (X0 ) + Mu u + Mw w + Mw w + Mq q + Me e + Mt t Mz = Mz (X0 ) + Nv v + Nv v + Np p + Nr r + Na a + Nr r where Fx (X0 ) , ..., Mx (X0 ) , ... remain the same (see equation 4-14).

4-1-3

Linearization of kinematic relations

The kinematic relations for a non-rotating at Earth are functions of the Euler angles , and of the angular velocities p, q, r: = f (, , p, q, r) = f (, , q, r) = f (, , q, r) Linearization of these relations results in: = (X0 ) + (X) = (X0 ) + (X) = (X0 ) + (X) with (X0 ) = p0 + q0 sin 0 tan 0 + r0 cos 0 tan 0 (X0 ) = q0 cos 0 r0 sin 0 (X0 ) = q0 sin 0 + r cos 0 cos 0 cos 0 p + sin 0 tan 0 q + q0 cos 0 tan 0 sin 0 +q0 2 + cos 0 tan 0 r (X) = cos 0 cos 0 r0 sin 0 tan 0 + r0 2 cos 0 (X) = cos 0 q q0 sin 0 sin 0 r r0 cos 0 sin cos 0 sin 0 0 q + q0 + q0 tan 0 + cos 0 cos 0 cos 0 (X) = sin 0 cos 0 cos 0 r r0 + r0 tan 0 cos 0 cos 0 cos 0 (4-18) (4-17)

(4-19)

and

(4-20)

Since the parameters of the linear velocity, i.e. VN , VE , VZ , are not included in the state vector, we do not have to linearize the equations relating these parameters with the state

4-1 Linearization about arbitrary ight condition in arbitrary body-xed reference frame

129

parameters u, v, w, , , . After we known the values for the state vector we can determine VN , VE , VZ using the relation: VN u VE = Tt v bO w VZ cos cos = cos sin sin

sin sin cos cos sin sin sin sin + cos cos sin cos

cos sin cos + sin sin cos sin sin sin cos cos cos

u v w

At this point we have linearized the equations of motion and the kinematic relations. With this set of equations we are able to simulate any motion about the initial state of linearization X0 . During simulation one uses a state vector consisting of nine parameters: x = [u, v, w, p, q, r, , , ] T and an input vector containing three input parameters: u = [a , e , r , t ]T After rearranging the terms in the linearized equations of motion and kinematic relations one can write the relations in the state-space format: x0 + x = A0 x0 + Ax + B0 u0 + Bu where A0 , A, B0 , B are constant matrices and the vector x0 is also constant. The simulation is initialized at the initial state of linearization X0 . Bear in mind that this state includes information about the initial state vector and the initial input vector. If the initial state of linearization is a steady state, then the term x0 will be zero meaning that the terms A0 x0 and B0 u0 cancel and the state-space format becomes: x = Ax + Bu The matrix A contains the stability derivatives and inertial properties and the B contains the control derivatives. The derivatives will be dened in section 4-3. The previous relation is used during simulations to determine the changes in state due to input changes. The most common ying condition is a steady-state ight. In particular the steady, straight, symmetric ight conditions are often encountered. Therefore most analysis of the characteristics of an aircraft are performed for such ight conditions. We will derive the linearized set of equations about a steady, straight, symmetric ight condition in the next section.
Flight Dynamics

130

Linearized Equations of Motion

4-2

Linearization about steady, straight, symmetric ight condition

A steady, straight (wing level), symmetric ight condition has the following properties: u=0 u=0 v=0 v=0 w=0 w=0 =0 =0 =0 =0 =0 =0 p=0 q=0 r=0 p=0 q=0 r=0

X=0 X=0 Y =0 Y =0 Z=0 Z=0

which in principle means that the state of the aircraft remains constant, i.e. all derivatives, the lateral velocity, and the angular velocities are zero. The conditions that u, w, X, Z are unequal to zero means that the aircraft has a velocity equal to u2 + w2 and that the resultant aerodynamic force R = X + Z is equal in magnitude but opposite in sign with respect to the gravity vector W . The set of linearized equations of motion is obtained by taking the ight condition described above as initial state for linearization. By substituting the properties in the general linearized equations of motion (equation (4-21)) we get: Fx = m (u + w0 q) Fy = m (v w0 p + u0 r) Fz = m (w u0 q)

Mx = Ix p Jxz r My = Iy q Mz = Ix r Jxz p Fx Fy Fz Mx My Mz = = = = = =

(4-21)

The linearized forces and moments equation becomes: W cos 0 + Xu u + Xw w + Xq q + Xe e + Xt t +W cos 0 + Yv v + Yv v + Yp p + Yr r + Ya a + Yr r W sin 0 + Zu u + Zw w + Zw w + Zq q + Ze e + Zt t Lv v + Lp p + Lr r + La a + Lr r Mu u + Mw w + Mw w + Mq q + Me e + Mt t Nv v + Nv v + Np p + Nr r + Na a + Nr r

(4-22)

here, we applied the knowledge that the aerodynamic forces are canceled by the weight terms for the initial state. Therefore they are removed from the equations. The linearized kinematic relations become (see equation 4-18 to 4-20): = p + tan 0 r = q 1 = cos 0 r (4-23)

The aircraft body axes (Fb ) used in the foregoing to describe the motions, have not yet been completely dened. The reference frame is xed relative to the aircraft, but the direction of

4-2 Linearization about steady, straight, symmetric ight condition

131

the Xb -axis in the plane of symmetry has not yet been dened. Often, the choice depends on the type of motion to be studied. If the aircraft motion about a steady, straight, symmetric ight are to be determined, for instance to investigate the stability of the steady ight conditions, it is advantageous to use the so-called stability reference frame. This is a system of body axes, as previously dened, of which the XS -axis is parallel to the direction of motion of the center of gravity in the steady initial ight condition. But during the disturbed motion of the aircraft the reference frame is xed to the aircraft. This particular choice of the XS -axis results in the equations of motion in, u0 = V w0 = 0 In addition, the angle of attack in the steady reference condition, measured relative to the XS -axis, is zero, 0 = 0 (4-24) 0 = 0 A disadvantage of the stability axes becomes apparent, if the disturbed motions about several dierent equilibrium conditions have to be studied in succession. During each of these disturbed motions the XS -axis and the ZS -axis have a dierent direction relative to the aircraft. This implies, that the aerodynamic variables, the components of the motion and also the moments and products of inertia in each new case refer to a dierent attitude of the reference frame relative to the aircraft, although that attitude remains invariant during each of the disturbed motions. In the following section the linearized set of equations for steady, straight, symmetric ight expressed in the stability reference frame is given. Thereafter the equations needed to determine the moment and products of inertia about the stability reference frame are given.

4-2-1

Linearized set of equations

In order to simplify the notation in the remainder of this chapter the changes in the variable angles relative to the reference condition are indicated without the letter , e.g. rather than , e instead of e 1 . The magnitude of the relevant variable in steady ight is indicated by the subscript 0 ( 0 ). Since the aerodynamics are considered to be uncoupled, the set of equations can be split up into two parts. One describes the motion in the symmetry plane of the aircraft, i.e. the symmetric motion, and the other part describes the asymmetric motion. If the stability reference frame is used the equations of motion can now be written as follows,

During simulations one should bear in mind that u does not represent the true velocity along the Xb -axis. One should add the initial condition u0 to obtain the true value for that velocity!
Flight Dynamics

132 I Symmetric motions W cos 0 W sin 0

Linearized Equations of Motion

+Xu u + Xw w + Xq q + Xe e + Xt t = mu +Zu u + Zw w + Zw w + Zq q + Ze e + Zt t = m (w q V ) Mu u + Mw w + Mw w + Mq q + Me e + Mt t = Iy q = q

(4-25)

II Asymmetric motions W cos 0 +Yv v + Yv v + Yp p + Yr r + Ya a + Yr r = m (v + r V ) Lv v + Lp p + Lr r + La a + Lr r = Ix p Jxz r Nv v + Nv v + Np p + Nr r + Na a + Nr r = Iz r Jxz p =

r cos 0

(4-26)

= p + r tg 0 The kinematic relations remain the same, i.e.: = p + tan 0 r = q 1 = cos 0 r (4-27)

4-2-2

Moments and products of inertia

The moments and products of inertia of the aircraft depend not only on the mass distribution of the aircraft, but also on the chosen directions of the axes of the reference frame relative to the aircraft (stability frame of reference FS versus body-xed frame of reference Fb ). Due to the symmetry of the aircraft two of the three products of inertia are zero, Jyz =
m

yz dm = 0

Jxy =
m

xy dm = 0

This result does not depend on the directions of the two sets of body axes (stability and body-xed frame of reference) in the plane of symmetry. Suppose now, that for an arbitrary direction of the Xb -axis the magnitudes of the three moments of inertia and the remaining product of inertia are given. The direction of the Xb -axis is given by the angle + 1 between the Xb -axis and the negative Xr -axis, see

4-2 Linearization about steady, straight, symmetric ight condition

133

Zr

X0a n1 X1a

n2

X2a

Z2

Z1 X2 O

X1

Xr

Figure 4-5: The angles between the principal inertial axes and the aircraft reference axes

gure 4-5. This xes also the direction of the Zb -axis. If the Xb -axis has a dierent angle + 2 with the negative Xr -axis, see gure 4-5, and the rst and second reference frame are indicated by the subscripts 1 and 2 respectively, the following relations hold,

x2 = x1 cos (2 1 ) + z1 sin (2 1 ) y2 = y1 z2 = x1 sin (2 1 ) + z1 cos (2 1 ) This leads to the following moments and products of inertia, relative to the new reference frame, Ix2 Iy2 Iz2 Jxz2 = Ix1 cos2 (2 1 ) + Iz1 sin2 (2 1 ) Jxz1 sin2 (2 1 ) = Iy1 (4-28) = Ix1 sin2 (2 1 ) + Iz1 cos2 (2 1 ) + Jxz1 sin2 (2 1 ) =
1 2

(Ix1 Iz1 ) sin (2 (2 1 )) + Jxz1 cos (2 (2 1 ))

As could be expected, the moment of inertia about the Yb -axis does not depend on the directions of the body axes in the plane of symmetry. For one particular direction of the
Flight Dynamics

134

Linearized Equations of Motion

Xb -axis, and correspondingly of the Zb -axis, the product of inertia Jxz is zero. The axes having these directions are called the principal inertial axes in the plane of symmetry. They are indicated as X0 - and Z0 -axes and have an angle with the negative Xr - and Zr -axis, see gure 4-5. As the products of inertia Jxy and Jyz are always zero, the Yb -axis is always a principal inertial axis. This follows also directly from the fact that the Yb -axis is perpendicular to the plane of symmetry. If for an arbitrary direction + of the Xb -axis the inertial parameters are known the angle of these axes with the principal inertial axes follow from equation (4-28) by equating Jxz2 and 2 to zero, tan 2 = 2 Jxz Ix Iz

Using the principal moments of inertia Ix0 and Iz0 , which can be calculated from equation (4-28) for a given value by letting 2 = 0, the inertial parameters for an arbitrary direction of the body axes are, Ix = Ix0 cos2 + Iz0 sin2 Iy = Iy0 Iz = Ix0 sin2 + Iz0 cos2 Jxz = 1 (Ix0 Iz0 ) sin 2 2

4-3

Equations of motion in non-dimensional form

The dierential equations derived in the foregoing can be applied directly to the calculation of the symmetric and asymmetric aircraft motions about a given condition of steady, straight, symmetric ight as a response to a given disturbance. Very often, however, the equations are used in a non-dimensional form. The reason for this practice is that the aerodynamic forces and moments in the equations are usually expressed in non-dimensional coecients2 . This leads to writing the entire equations in non-dimensional form. To this end, also the inertial characteristics and the time scale are made non-dimensional. One of the several ways in which this can be done, is discussed in the following. As a starting point, three independent units are chosen, i.e. length having the dimension [], velocity with dimension [][t]1 and mass with dimension [m]. The actual values of these units are, Symmetric
2

Asymmetric

With the non-dimensional form of the aerodynamic forces and moments, we can compare aircraft independent of aircraft dimensions and weight.

4-3 Equations of motion in non-dimensional form motions : length velocity mass c V S c motions : b V Sb

135

Evidently dierent units of length are used for the symmetric and the asymmetric motions, the mean aerodynamic chord c and the wing span b respectively. These are also the customary lengths used to make the symmetric moment M and the asymmetric moments L and N , respectively, non-dimensional. The unit of mass chosen for both types of motion is the mass of a certain volume of air surrounding the aircraft equal to the wing area multiplied by the respective unit of length. With these units all variables in the equations of motion can be made non-dimensional, according to the schemes in tables 4-1 and 4-2. Using these schemes, the way in which the equations of motion are made non-dimensional can now be discussed more in detail.

4-3-1

Symmetric Motions

For the symmetric motions the two force equations, in which each of the terms has the dimension of a force, i.e. X and Z, are made non-dimensional by a division by 1 V 2 S. The 2 moment equation, in which each of the terms has the dimension of a moment, i.e. M , is divided by 1 V 2 S. Starting from equations (4-25) the result is, c 2 Xu u + Xw w + Xq q + Xe e + Xt t mu W cos0 + = 1 2 1 1 2S 2S 2 V 2 V 2 V S Zu u + Zw w + Zw w + Zq q + Ze e + Zt t W sin0 m (w q V ) 1 2 + = 1 1 2S 2 2 V S 2 V 2 V S Iy q Mu u + Mw w + Mw w + Mq q + Me e + Mt t = 1 2 1 2 S c c 2 V 2 V S c q c = V V In elaborating the two force equations, use is made of the two equilibrium conditions, W sin 0 = X0 W cos 0 = Z0

(4-29) (4-30) (4-31) (4-32)

Using the previous scheme of table 4-1, these conditions are written as follows in equation (4-29), W cos 0 = CZ0 1 2 2 V S (4-33)
Flight Dynamics

136 and in equation (4-30), W sin 0 = CX0 1 2 2 V S

Linearized Equations of Motion

(4-34)

Next, the partial derivatives of the aerodynamic forces and the moment are further elaborated. As an example the contribution Xu u is considered. It is written as, Xu u 1 2 2 V S = Xu 1 2 V S u Xu = 1 u V 2 V S

As both the left and right hand sides of this latter expression are non-dimensional, as u is, Xu the quotient 1 V S must also be non-dimensional. This quotient is called a stability derivative written as CXu , CXu = and, Xu u 1 2 2 V S = CXu u Xu 1 2 V S
2

It should be emphasized that for the stability derivative CXu in particular, but also for CZu and Cmu , the order of dierentiation with respect to airspeed and non-dimensionalizing using the factor 1 V 2 S in which the airspeed gures as well, is relevant for the resulting values of 2 CXu , CZu and Cmu , CXu = In a similar manner it follows, Xw w Xw w = 1 = CX 1 2S V 2 V 2 V S where the stability derivative CX is, CX = Xw 1 2 V S = CX 1
1 2 V

X CX = u u

(4-35)

Now the order of dierentiation with respect to w and non-dimensionalizing using the factor 1 1 2 2 2 V S may be changed, as 2 V S does not contain the factor w. The expressions for the dierent stability derivatives for symmetric motions are collected in table 4-3. The non-dimensional control derivatives are given in table 4-4. See also table 4-5 for the denition of the non-dimensional moments and products of inertia.

4-3 Equations of motion in non-dimensional form

137

The non-dimensional equations for the symmetric motions now become, CZ0 + CXu u + CX + CXq Dc + CXe e + CXt t = 2c Dc u CX0 + CZu u + CZ + CZ Dc + CZq Dc + CZe e + CZt t = 2c (Dc Dc ) q c 2 Cmu u + Cm + Cm Dc + Cmq Dc + Cme e + Cmt t = 2c KY Dc V c q c = (4-36) V V The four equations (4-36) are now rewritten in an ordered form. In this way, the following c dierential equations in the variables u, , and q result in, V q c + CXe e + CXt t V q c + CZe e + CZt t CZu u + [CZ + (CZ + 2c ) Dc ] CX0 + CZq + 2c V q c Dc + V q c 2 Cmu u + (Cm + Cm Dc ) + Cmq 2c KY Dc + Cme e + Cmt t V (CXu 2c Dc ) u + CX + CZ0 + CXq or using matrix notation, CXu 2c Dc u CX CZ0 CXq CZu CZ + (CZ 2c ) Dc CX0 CZq + 2c 0 0 Dc 1 q c 2D Cmu Cm + Cm Dc 0 Cmq 2c KY c V CXe CZ e = 0 Cme CXt CZt 0 Cmt e t (4-38)

=0 =0 =0 =0 (4-37)

and with,

= + the left hand side of equation (4-38) can also be expressed in the variables u, , and equation (4-38) is written as P x = Q e with x = u, , , CXu 2c Dc CX CX + CZ0 CZu CZ (CZ 2c ) Dc CZ CX0 0 0 Dc Cmu Cm Cm Dc Cm
q T c , V q c V .

If

matix P is equal to, CXq CZ + CZq 1 2D + Cmq 2c KY c

Cm

The right hand side of equation (4-38), i.e. Q [e , t ]T , remains unchanged with matrix Q equal to,
Flight Dynamics

138

Linearized Equations of Motion

CXe CZ e Q= 0 Cme

CXt CZt 0 Cmt

4-3-2

Asymmetric Motions

The equations for the asymmetric motions are made non-dimensional in close analogy with the foregoing. The force equation is again divided by 1 V 2 S and the two moment equations 2 by 1 V 2 Sb. The result is, 2

W cos 0 Yv v + Yv v + Yp p + Yr r + Ya a + Yr r + 1 1 2 2 2 V S 2 V S Lv v + Lp p + Lr r + La a + Lr r 1 2 2 V Sb Nv v + Nv v + Np p + Nr r + Na a + Nr r 1 2 2 V Sb b V / b V

= = = = =

m (v + r V ) 1 2 2 V S Ix p Jxz r 1 2 Sb 2 V Iz r Jxz p 1 2 Sb 2 V b r V cos 0 pb rb + tg 0 V V

The component of the weight along the ZS -axis is now written as,

W cos 0 W cos 0 = 1 2 = CL 1 2S 2 V 2 V S Next, the partial derivatives of the aerodynamic force and the two moments are expressed using the non-dimensional stability derivatives according to the scheme of table 4-6. See also table 4-7 and 4-8 for the denition of the non-dimensional control derivatives and the moments and products of inertia respectively. If in addition 0 in the kinematic relations is replaced by 0 , using equation (4-24), the non-dimensional equations for the asymmetric motions are,

4-3 Equations of motion in non-dimensional form

139

pb rb CL + CY + CY Db + CYp 2V + CYr 2V + CYa a + CYr r = 2b Db + 2 pb pb rb 2 C + Cp 2V + Cr 2V + Ca a + Cr r = 4b KX Db 2V rb KXZ Db 2V pb rb rb 2 Cn + Cn Db + Cnp 2V + Cnr 2V + Cna a + Cnr r = 4b KZ Db 2V pb KXZ Db 2V 1 2 Db 1 2 Db

rb 2V

= =

1 rb cos 0 2V pb 2V

rb 2V

tg 0 (4-39)

In level ight (0 = 0) the kinematic relations are reduced to,

1 2 Db 1 2 Db

= =

rb 2V pb 2V

By adding the latter of these two relations, 1 pb Db + =0 2 2V the equations of motion (4-39) are changed into a set of four rst order dierential equations pb rb in , , 2V and 2V ,
h i pb CY + CY 2b Db + CL + CYp 2V + (CYr 4b )
rb 2V

+ CYa a + CYr r
1 2 Db + pb 2V

= = = =

0 0 0 0

2 C + Cp 4b KX Db ` Cn + Cn Db + Cnp + 4b KXZ Db

pb 2V pb 2V

+ (Cr + 4b KXZ Db ) ` 2 + Cnr 4b KZ Db

rb 2V

+ Ca a + Cr r + Cna a + Cnr r

rb 2V

(4-40) Using matrix notation, equations (4-40) are written as, CY + CY 2b Db Cn 0 C + Cn Db CL CYp 0 pb Cr + 4b KXZ Db 2V rb 2 Cnr 4b KZ Db 2V CYr 4b
Flight Dynamics

1 Db 1 2 2 0 Cp 4b KX Db 0 Cnp + 4b KXZ Db

140

Linearized Equations of Motion

At the end of this section it should be noted that,

CYa 0 = C a Cna

CYr 0 Cr Cnr

a r

(4-41)

Often, especially in the British literature, the equations of motion are made nonc c dimensional using a unit of time not equal to V , but c c V seconds for the symmetric b b motions and b b V seconds rather than V seconds for the asymmetric motions. The particular way of making the equations non-dimensional is determined in the rst place by the habits of the user. None of the various methods oers clear advantages or disadvantages in comparison with other possible methods. This explains why only one method has been described in detail here. In chapters 6 and 8 attention is given to the methods used to obtain the various stability and control derivatives for the symmetric and asymmetric motions respectively. Reliable quantitative data on moments and products of inertia of aircraft are rare. Quantitative calculations can be time-consuming and may produce rather inaccurate results. The most reliable inertial parameters of aircraft are those obtained experimentally, usually by oscillating the suspended aircraft on the ground. Details on this experimental technique may be found in references [40, 95, 164, 168]. Figure 4-6 shows the relation between the moment of inertia Iy and the mass for multiple types of aircraft. One can make educated guesses for the inertia properties of new types of aircraft using such gures. Appendix D lists symmetrical and asymmetrical inertial data, as well as stability and control derivatives for both the symmetric and asymmetric motions.

4-3 Equations of motion in non-dimensional form

141

[Moment of Inertia Iy] 1/5 vs [mass]1/3, inertia2002,12Jan2006 50 45 40 35 IY(1/5) [kgm2] 30 25 20 15 10 5 0 0 10 20 30 40 Mass

(1/3)

Best Fit border 50% of data

50 [kg]

60

70

80

90

Figure 4-6: Relation between moment of inertia Iy and the mass

Flight Dynamics

142

Linearized Equations of Motion

Dimensional parameter

Dimension

Divisor

Non-dimensional parameter

t
d dt d2 dt2

[t] [t]1 [t]2 [][t]1 [][t]1 [t]1 [][t]2 [][t]2 [t]2 [m] [m][]2 [] [m][][t]2 [m][][t]2 [m][]2 [t]2

c V V c V2 c2

sc = Dc =

V c

t =
d dsc

c d V dt

2 Dc =

c2 d2 V 2 dt2

d2 ds2 c

u w q u w q m IY ky X Z M

V V
V c V2 c V2 c V2 c2

u= =
q c V

u V w V

Dc u = Dc =

u c V V w c V V
2

= =

uc V c V

c Dc q = q V 2 c V

S c S c2 c c
1 c 2 S 1 c 2 S 1 c2 2 S

c =

m S c IY S3 c

2 c KY =

KY = c c
V2 c2 V2 c2 V2 c2

ky c X

CX = CZ = Cm =

1 V 2 S 2

1 V 2 S 2

1 V 2 S c 2

Table 4-1: Non-dimensional parameters in the equations of motion, symmetric motions

4-3 Equations of motion in non-dimensional form

143

Dimensional parameter

Dimension

Divisor

Non-dimensional parameter

t
d dt d2 dt2

[t] [t]1 [t]2 [][t]1 [][t]1 [t]1 [][t]2 [t]2 [t]2 [m] [m][]2 [m][]2 [m][]2 [] []

b V V b V2 b2

sb = Db =

V b

t =
d dsb

b d V dt

2 Db =

b2 d2 V 2 dt2

d2 ds2 b

v p r v p r m IX IZ JXZ kx kz

V
2V b 2V b V2 b 2V 2 b2 2V 2 b2

=
pb 2V rb 2V

v V

Db =

v b V V

b V

pb Db 2V = rb Db 2V =

pb2 2V 2 rb2 2V 2

Sb Sb b2 Sb b2 Sb b2 b b

b =

m Sb IX Sb3 IZ Sb3 JXZ Sb3

2 b KX = 2 b KZ =

b KXZ = KX = KZ =
kx b kz b

Table 4-2: Non-dimensional parameters in the equations of motion, asymmetric motions

Flight Dynamics

144

Linearized Equations of Motion

Dimensional parameter

Dimension

Divisor

Non-dimensional parameter

Y L N

[m][][t]2 [m][]2 [t]2 [m][]2 [t]2

1 2 Sb

V2 b2 V2 b2 V2 b2

CY = C = Cn =

1 V 2 S 2

1 2 2 Sb 1 2 2 Sb

b b

1 V 2 Sb 2

1 V 2 Sb 2

Table 4-2: (Continued) Non-dimensional parameters in the equations of motion, asymmetric motions

Denition

CXu = CZu = Cmu = CX = CZ = Cm = CZ = Cm = CXq = CZq = Cmq =

1 V 2

X u Z u M u X w Z w M w Z w M w X q Z q M q

= = = = = = = = = =

CX u CZ u Cm u CX CZ Cm

1 V 2

1 V 2

1 S c

1 V 2

1 V S 2

1 V 2

1 S c

1 S c 2

CZ c V Cm c V CX c q V CZ c q V Cm c q V

1 S2 c 2

1 V 2

1 S c

1 V 2

1 S c

1 V S2 c 2

Table 4-3: Stability derivatives, symmetric motions

4-3 Equations of motion in non-dimensional form

145

Denition

CXe = CZe = Cme = CXt = CZt = Cmt =

1 V 2 S 2

X e Z e M e X t X t X t

= = = = = =

CX e CZ e Cm e CX t CZ t Cm t

1 V 2 S 2

1 V 2 S c 2

1 V 2 S 2

1 V 2 S 2

1 V 2 S 2

Table 4-4: Control derivatives, symmetric motions

Denition

c =
2 KY =

m S c IY m2 c

Table 4-5: Moments and products of inertia, symmetric motions

Flight Dynamics

146

Linearized Equations of Motion

Denition

CY = C = Cn = CY = Cn = CYp = Cp = Cnp = CYr = Cr = Cnr =

1 V 2

Y v L v N v Y v N v Y p L p N p Y r L r N r

= = = = = = =

CY C Cn CY b V Cn b V CY pb 2V C pb 2V Cn pb 2V CY rb 2V C rb 2V Cn rb 2V

1 V Sb 2

1 V 2

1 Sb

1 V 2

1 Sb

1 Sb2 2

1 V 2

2 Sb

1 V 2

2 Sb2 2 Sb2

1 V 2

= = = =

1 V 2

2 Sb

1 V 2

2 Sb2

1 V 2

2 Sb2

Table 4-6: Stability derivatives, asymmetric motions

4-3 Equations of motion in non-dimensional form

147

Denition

CYa = Ca = Cna = CYr = Cr = Cnr =

1 V 2 S 2

Y a L a N a Y r L r N r

= = = = = =

CY a C a Cn a CY r C r Cn r

1 V 2 Sb 2

1 V 2 Sb 2

1 V 2 S 2

1 V 2 Sb 2

1 V 2 Sb 2

Table 4-7: Control derivatives, asymmetric motions

Denition

b =
2 KX = 2 KZ =

m Sb IX mb2 IZ mb2 IXZ mb2

JXZ =

Table 4-8: Moments and products of inertia, asymmetric motions

Flight Dynamics

148

Linearized Equations of Motion

4-4

Symmetric equations of motion in state-space Form

The linearized deviation equations for the symmetric motions, where CXq = 0, are, see equation (4-38), CX CZ0 0 CXu 2c Dc u CZu CZ + (CZ 2c ) Dc CX0 CZq + 2c 0 0 Dc 1 q c 2D Cmu Cm + Cm Dc 0 Cmq 2c KY c V CXe CZ e = 0 Cme CXt CZt 0 Cmt e t

Linear Time Invariant (LTI) systems are given in the so called state-space form, x =Ax+Bu (4-42)

y =Cx+D u

(4-43)

With A (n n) the state-matrix, B (n m) the input-matrix, C (r n) the output matrix and D (r m) the direct matrix. Furthermore, x (n 1) and y (r 1) are the state and output vector respectively, while u (m 1) is the input vector. Since the system is LTI, the matrices A, B, C, and D have constant elements. This form can be obtained by rearranging equation (4-38). First, all terms without the dierential operator Dc are placed on the right hand side of equation (4-38). Furtherc d more, the dierential operator Dc is replaced by V dt . The result is,
c d 2c V dt 0 0 0

0 c d (CZ 2c ) V dt 0 c d Cm V dt

0 0
c d V dt 0

0 0 0 2 2c KY

c d V dt

u =
q c V

This equation can be written in the form, Px = P with,

CXu CZu 0 Cmu

CX CZ0 CZ CX0 0 0 Cm 0

0 u CZq + 2c 1 q c Cmq V

CXe CZ e + 0 Cme

CXt CZt 0 Cmt

e t

dx =Qx+Ru dt

(4-44)

4-4 Symmetric equations of motion in state-space Form

149

c 2c V 0 P= 0 0

0 c (CZ 2c ) V 0 c Cm V CX CZ0 CZ CX0 0 0 Cm 0

0 0 c V 0

0 0 0 2 2c KY

c V

CXu CZu Q= 0 Cmu

0 CZq + 2c 1 Cmq

c and x = u q , while u = e . The nal state-space form is obtained by premultiplying V equation (4-44) by the inverse of matrix P,

CXe CZ e R= 0 Cme

CXt CZt 0 Cmt

x = P1 Q x + P1 R u = A x + B u In ight dynamics, equation (4-45) is usually written as, u xu zu = 0 mu x z 0 m x 0 u z zq V 0 c q c m mq V xe ze + 0 me xt zt 0 mt

(4-45)

e t

(4-46)

q c V

The denition of the newly introduced symbols are recapitulated in table 4-9. Dimensional eigenvalues are obtained using MATLABs routine eig.m (for the state matrix A).

Flight Dynamics

150

Linearized Equations of Motion

V CXu c 2c

CZu V c 2c CZ

V c

Cmu +CZu 2
2 2c KY

Cm c CZ

V CX c 2c

CZ V c 2c CZ CX0 V 2c CZ c V 2c +CZq c 2c CZ CZ V e c 2c CZ CZ V t c 2c CZ

V c

Cm +CZ 2
2 2c KY

Cm c CZ

V CZ0 c 2c

V c
V c

CX0 2

Cm c CZ 2 2c KY

V CXq c 2c

Cmq +Cm 2

2c +CZ q c CZ 2 2c KY Cm e 2c CZ 2 2c KY Cm t 2c CZ 2 2c KY

V CXe c 2c

V c

Cm +CZ
e

C V Xt c 2c

V c

Cm +CZ
t

Table 4-9: Symbols appearing in the general state-space representation of equation (4-46)

4-5 Asymmetric equations of motion in state-space form

151

4-5

Asymmetric equations of motion in state-space form

Similar to section 4-4 the linearized deviation equations for the asymmetric motions, are, see equation (4-41),

CY + CY 2b Db Cn 0 C + Cn Db

CL

CYp

1 Db 1 2 2 0 Cp 4b KX Db 0 Cnp + 4b KXZ Db CYa 0 C a Cna CYr 0 Cr Cnr

0 = pb Cr + 4b KXZ Db 2V rb 2 Cnr 4b KZ Db 2V CYr 4b

a r

These equations of motion are transformed to the state-space form as in, x =Ax+Bu (4-47)

This form can be obtained by rearranging equation (4-41). First, all terms without the dierential operator Db are placed on the right hand side of equation (4-41). Furthermore, b d the dierential operator Db is replaced by V dt . The result is, CY 2b 0 0
b Cn V d dt b d V dt

0
b 1 V 2 0 0 d dt

0 0 2 b d 4b KX V dt b d 4b KXZ V dt

0 0 b 4b KXZ V 2 b 4b KZ V
d dt d dt

pb = 2V
rb 2V

Again, this equation can be written in the form, Px = P with in this case, CY 2b 0 0
b Cn V b V

CY 0 C Cn

CL CYp 0 1 0 Cp 0 Cnp

(CYr 4b ) CYa 0 0 pb + C Cr a 2V rb Cna Cnr 2V dx =Qx+Ru dt

CYr 0 Cr Cnr

a r

(4-48)

0
b 1 V 2 0 0

0 0 2 b 4b KX V b 4b KXZ V

P=

0 b 4b KXZ V 2 b 4b KZ V
Flight Dynamics

152

Linearized Equations of Motion

y p r a r
V CY b 2b V CL b 2b V CYp b 2b V CYr 4b b 2b V CYa b 2b V CYr b 2b

l
2 V C KZ +Cn KXZ b 4b (K 2 K 2 K 2 ) X Z XZ

n
2 V C KXZ +Cn KX b 4b (K 2 K 2 K 2 ) X Z XZ

0
2 V Cp KZ +Cnp KXZ b 4b (K 2 K 2 K 2 ) X Z XZ 2 V Cr KZ +Cnr KXZ b 4b (K 2 K 2 K 2 ) X Z XZ 2 V Ca KZ +Cna KXZ b 4b (K 2 K 2 K 2 ) X Z XZ 2 V Cr KZ +Cnr KXZ b 4b (K 2 K 2 K 2 ) X Z XZ

0
2 V Cp KXZ +Cnp KX b 4b (K 2 K 2 K 2 ) X Z XZ 2 V Cr KXZ +Cnr KX b 4b (K 2 K 2 K 2 ) X Z XZ 2 V Ca KXZ +Cna KX b 4b (K 2 K 2 K 2 ) X Z XZ 2 V Cr KXZ +Cnr KX b 4b (K 2 K 2 K 2 ) X Z XZ

Table 4-10: Symbols appearing in the general state-space representation of equation (4-50)

CY 0 Q= C Cn

CL CYp 0 1 0 Cp 0 Cnp

(CYr 4b ) 0 Cr Cnr

pb rb and x = 2V 2V , while u = [a r ]T . The nal state-space form is obtained by pre-multiplying equation (4-48) by the inverse of matrix P,

CYa 0 R= C a Cna

CYr 0 Cr Cnr

x = P1 Q x + P1 R u = A x + B u In ight dynamics, equation (4-49) is usually written as, y 0 pb = l 2V rb n 2V y 0 0 0 yp 2V b lp np yr 0 0 0 + pb lr 2V la rb na nr 2V yr 0 lr nr

(4-49)

a r

(4-50)

The denition of the newly introduced symbols are recapitulated in table 4-10. Dimensional eigenvalues are obtained using MATLABs routine eig.m (for the state-matrix A).

Part II

Static Stability Analysis

Flight Dynamics

Chapter 5 Analysis of Steady Symmetric Flight

The analysis of both static and dynamic stability and control characteristics of aircraft requires a mathematical model describing the eects of aerodynamic ow parameters as airspeed, angle of attack, angle of sideslip and body rotation rates on the aerodynamic forces and aerodynamic moments acting on the aircraft, see also chapters 3 and 4. Current practice is to extensively rely on tools from Computational Fluid Dynamics (CFD) to derive these characteristics from. It was and and still is considered good practice, however, to augment and verify these results from CFD by using results from steady or rotary balance wind tunnel experiments with models of the complete aircraft or of isolated components. During the ight test campaign of the prototype aircraft these models can subsequently be improved by including ight test results using aerodynamic model identication techniques, see for example [94]. The disadvantage, however, of just using data from either CFD, wind tunnel experiments or ight tests on the complete model or actual aircraft is that it will not necessarily contribute much to the physical understanding of the resulting aerodynamic mechanisms leading to the observed aerodynamic force and moments characteristics. Also, there is still the problem of incorporating and interpreting the wind tunnel experimental results on isolated aircraft components. This is the rationale behind the approach taken in the present chapter to rst look at the characteristics of the isolated wing to which subsequently a fuselage and nacelles on the wing or fuselage are added. The nal step is to add the eects of horizontal and vertical tail planes to this wing-body-nacelles composite. The present chapter starts with an analysis of the aerodynamic forces and the aerodynamic moment acting on a symmetrical wing at some angle of attack. It is shown that the introduction of the concept of aerodynamic center leads to a surprisingly simple model for the force and moment characteristics. The concept of static stability is introduced and the necessary conditions to allow steady, stable ight. Next, it is shown how to derive a model of the force and moment characteristics of wings of arbitrary shape and twist from wing prole characteristics by applying classical 2-dimensional strip theory. It is shown in particular how to compute the location of the wing aerodynamic center as well as the aerodynamic moment
Flight Dynamics

156 CN CL XB V Cm CD CT CR

Analysis of Steady Symmetric Flight

ZB
Figure 5-1: The aerodynamic forces and the moment acting on a wing in symmetric ight

about it. Next, the isolated fuselage is discussed and the eect the fuselage has on the aerodynamic characteristics of the wing, and vice versa, the so called interference eects. In a similar way the eects of wing nacelles are discussed. The contribution of the horizontal tailplane to the aerodynamic forces and moment is discussed next. In classical aircraft congurations the horizontal tail plane is located somewhere behind the wing in the ow eld induced by the wing-body-nacelles composite. This has a signicant eect on the contribution of the tail plane to the aerodynamic forces in the plane of symmetry and on the pitching moment. The chapter proceeds with a general discussion on steady state ight conditions and is concluded with a discussion on the aerodynamic hinge moments of control surfaces.

5-1
5-1-1

Aerodynamic center
Aerodynamic forces and moments acting on a wing

A. Aerodynamic moment as a function of angle of attack Figure 5-1 shows the components CL and CD , and CN and CT , of the dimensionless total aerodynamic force vector C R and the dimensionless aerodynamic moment Cm about a given arbitrary reference point, acting on a wing in symmetric ight, with

CN = CL cos + CD sin CT = CD cos CL sin Figure 5-2 presents the familiar picture of CL and CD of a typical wing as a function of , as measured in the wind tunnel. In addition the corresponding CN - and CT - curves are

5-1 Aerodynamic center

157

shown. From these gures it follows that the CL and CN curves are almost identical while the CD and CT are much more dierent. While CL and CN vary almost linearly with over a large range of angles of attack, the relations betweenCD and CT and is clearly non-linear. Figure 5-3 shows the relation between CN and CT of this wing. Note that CT turns negative as increases! Similar results are obtained using linear-potential ow simulations. The conguration used for these simulations is depicted in gure 5-4 while the CN , CL versus , CT , CD versus , CN versus CT and CL versus CD curves are plotted in gures 5-5 and 5-6. The aerodynamic moment coecient Cm not only depends on but also on the position of the selected reference point. This holds true also for the slope of the Cm versus curve which, as we will show later is of crucial importance for longitudinal stability. It is easy to compute Cm for an arbitrary reference point from wind tunnel measurements about a xed model suspension point. Given the measured force and moment coecients CN and CT and Cm about the suspension point (x1 , z1 ) for some angle of attack, the moment about an arbitrary reference point (x2 , z2 ) in the plane of symmetry can be computed as, see gure 5-7, Cm(x2 ,z2 ) = Cm(x1 ,z1 ) + CN x2 x1 z2 z1 CT c c (5-1)

where x and z are coordinates in the aircraft reference frame Fr . Figure 5-9 shows the resulting curves of Cm as a function of for various positions of the reference point on the mac again for the case of a Fokker F-27 wing. From this gure it appears that, 1. The slope of the Cm curve strongly depends on the position of the reference point in the Xr -direction 2. I a large interval Cm varies almost linearly with for reference points close to the mac Figure 5-9 also shows the inuence on the Cm curve of a change in position of the reference point parallel to the Zr -axis. Now, the eect on the moment is much smaller. This is easily explained by equation (5-1), where CT is small relative to CN , see also gure 5-3. It appears that when the reference point is situated far below or above the mac, the moment is no longer a linear function of which is due the non-linear variation of CT with , see gure 5-2. Similar results from linear potential ow simulations are given in gure 5-10. For this case use has been made of the wing conguration as depicted in gure 5-4 (the data apply to the same wing as used in gures 5-5 and 5-6. The dependence of the Cm curve the selected reference point is a general phenomenon which does not apply to only wings but also to complete aircraft congurations. The relevant reference point in that case is the aircraft center of gravity, as we will see later.

Flight Dynamics

158

Analysis of Steady Symmetric Flight

CN
1.0

CL

CL CN

0.8

0.6

0.4

0.2

-4

-2

10

-0.2

[o ]

(A) CT
0.08

CD CD
0.04

-4

-2

10

-0.04

[o ]

-0.08

CT
-0.12

-0.14

(B)
Figure 5-2: CN , CL and CT , CD as functions of for the case of a Fokker F-27 wing (from reference [149])

5-1 Aerodynamic center

159

CN
1.0

0.8

0.6

0.4

0.2

-0.14

-0.12

-0.10

-0.08

-0.06

-0.04

-0.02

-0.2

CT

Figure 5-3: CN as a function of CT for a Fokker F-27 wing (from reference [149])

Figure 5-4: Rectangular wing conguration with zero-twist, = simulations

1 3,

as used for linear potential ow

Flight Dynamics

160

Analysis of Steady Symmetric Flight

0.5

CN , CL

CN CL

0.5 5

10

0.04 0.02 0 0.02

CT CD

CT , CD

0.04 0.06 0.08 0.1 0.12 0.14 5

10

Figure 5-5: Simulated CN , CL and CT , CD as functions of for the wing conguration depicted in gure 5-4

5-1 Aerodynamic center

161

0.5

CN
0 0.5 0.14

0.12

0.1

0.08

0.06 CT

0.04

0.02

0.02

0.5

CL
0 0.5 0

0.005

0.01

0.015

0.02 CD

0.025

0.03

0.035

0.04

Figure 5-6: Simulated CN as a function of CT and CL as a function of CD for the wing conguration depicted in gure 5-4

Flight Dynamics

162

Analysis of Steady Symmetric Flight

Zr

CN CR

CN CR

z2

z1
1 2

CT x1 Cm(x1 ,z1 ) x2

CT Cm(x2 ,z2 )

Xr

Cm(x2 ,z2 ) = Cm(x1 ,z1 ) + CN x2 x1 CT z2 z1 c c

Figure 5-7: The variation of the moment with changes in the reference point

z x

mac

c
Figure 5-8: Location of point (x, z) with respect to the position of the mean aerodynamic chord

5-1 Aerodynamic center

163

It is possible to interpret the bundle of Cm curves in terms of some new concepts: the line of action of C R , the center of pressure, the aerodynamic center (ac) and the aerodynamic moment about the ac. B. The line of action of C R and the center of pressure The aerodynamic forces as well as the aerodynamic moment acting on the wing may also be obtained from the line of action along of the resulting aerodynamic force vector C R . The magnitude of the vector C R is, CR =
2 2 CL + CD = 2 2 CN + CT

The direction of C R is follows from the angle 1 , see gure 5-11, of the vector with the CN -axis, 1 = arctan CT CN

The position of the line of action of C R in the plane of symmetry, may be dened by its intersection with the (extended line through the ) mac. This point is called the center of pressure. The position of the center of pressure (xd , z0 ) for an arbitrary moment reference point (x1 , z1 ) can be derived with equation (5-1) from Cm(x1 ,z1 ) , Cm(xd ,z0 ) = 0 = Cm(x1 ,z1 ) + CN or simply 0 = Cm(x0 ,z0 ) + CN xd x0 c xd x1 z0 z1 CT c c (5-2)

if the leading edge of the mac (x0 , z0 ) is taken as the moment reference point. Now the center of pressure follows as, Cm(x0 ,z0 ) e xd x0 = = c c CN (5-3)

It was shown above that for moment reference points on the mac, the wing aerodynamic moment Cm varies more or less linearly with the angle of attack, and thus also with CN . The relation between Cm(x0 ,z0 ) and CN can then be written as, Cm(x0 ,z0 ) = Cm0 + dCm(x0 ,z0 ) dCN CN (5-4)

Cm0 denoting the value of Cm at CN = 0. We note that all moment lines intercept at one single point at CN =0 with one value for Cm0 for all moment reference points. We note also
Flight Dynamics

164

Analysis of Steady Symmetric Flight

Cm
0.8

z/ = 0 c
0.6

x/ = 1.00 c

0.4

x/ = 0.75 c
0.2

CN =0
-4 -2 0 2 4 6 8

x/ = 0.50 c
10

Cm0 = Cmac

x/ = 0.25 c
-0.2

[o ]

Cmx0 ,z0
-0.4

x/ = 0.00 c

(A)

x c

variable

Cm
0.2

x/ = 0.25 c
-4 -2 0 2 4 6 8 10

z/ = 1.00 c z/ = 0.50 c z/ = 0.00 c [o ] z/ = 0.50 c z/ = 1.00 c

-0.2

-0.4

(B)

z c

variable

Figure 5-9: Moment curves for various positions of the reference point for the Fokker F-27 wing (from reference [149])

5-1 Aerodynamic center

165

x/ = 1.00 c
0.5

x/ = 0.75 c x/ = 0.50 c x/ = 0.25 c x/ = 0.00 c

Cm
0 0.5 4

(a) Moment curve for varying

0.4

4
x z , c c

10

=0

0.3

0.2

0.1

z/ = 1.00 c z/ = 0.50 c z/ = 0.00 c z/ = 1.00 c z/ = 0.50 c

Cm

0.1

0.2

0.3

0.4 4

(b) Moment curve for varying

4
z x , c c

10

= 0.25

Figure 5-10: Simulated moment curves for various positions of the reference point for the wing conguration of gure 5-4

Flight Dynamics

166 CN

Analysis of Steady Symmetric Flight

CR

CT

Figure 5-11: Denition of the angle 1

that this and all other moment lines are close to straight up to an angle of attack of almost 10 The center of pressure may now be written as, dCm(x0 ,z0 ) e Cm0 = c CN dCN
dCm

(5-5)

(x0 ,z0 ) is negative, the center of pressure lies behind the leading edge of the mac for As dCN positive values of CN .

Figure 5-13 shows the total aerodynamic force vector on the Fokker F-27 wing for a range of angles of attack. Figure 5-14 gives the corresponding position of the center of pressure. The dierent ways to describe the aerodynamic moment characteristics of the wing, i.e. the Cm -curve and the line of action of C R with the center of pressure, both have their disadvantages. The Cm -curve for example is not unique but depends on the location of the reference point. In the other case of the line of action and the center of pressure both C R as well as the position of the center of pressure depend on . As a consequence the resulting picture of the aerodynamic moment character is not simple to interpret, and easily leads to mistakes. There, however, is an alternative and more attractive way to describe the aerodynamic moments which is based on the concept of the so called aerodynamic center. The aerodynamic center (ac) is a very special moment reference point with the important property that the aerodynamic moment about it is invariant, i.e. it does not change with the angle of attack! In many cases of practical interest also the position of the aerodynamic center (ac) is inde-

5-1 Aerodynamic center

167

Zr

CN line of action of C R

(x0 , z0 ) V Cm(x0 ,z0 )

(xd , z0 ) CT e Cm(xd ,z0 ) = 0 Xr mac

Figure 5-12: Denition of the center of pressure relative on the mean aerodynamic chord (mac)

= 18.95o = 8.97o = 12.94o = 6.87o = 2.65o mac = 0.56o

= 1.62o = 4.81o

Figure 5-13: The magnitude of C R and the position of the line of action of C R as a function of the angle of attack for Fokker F-27 wing (from reference [149])

Flight Dynamics

168

Analysis of Steady Symmetric Flight

16

[o ]

12

-1.2

-0.8

-0.4

0.4

0.8
e c

CN =0
-4

Figure 5-14: The position of the center of pressure as a function of the angle of attack for Fokker F-27 wing (from reference [149])

5-1 Aerodynamic center CN + CN C R + C R CN CN C R CT CT CT + CT Cm + Cm Cm

169

CR

Figure 5-15: The aerodynamic forces and moment about an arbitrary point at two adjacent values of the angle of attack

pendent of the angle of attack. In the next section the aerodynamic center will be discussed in more detail.

5-1-2

A compact description for the aerodynamic moment characteristics

The following discussion it is shown that a moment reference point called the aerodynamic center (ac), about which aerodynamic moment (Cmac ) is constant, does indeed exist. In this section the simplifying assumption is made that also the position of the ac does not vary with angle of attack.

A. The rst metacenter, the neutral line and the neutral point Figure 5-15 shows the forces and the moment about an arbitrary reference point at two adjacent angles of attack, and + . Now the line of action of C R in Figure 5-16, may readily be constructed as the vector dierence of the aerodynamic forces corresponding to and + . About any point on the lines of action of C R and C R + C R the aerodynamic moment at the respective angle of attack, i.e. and + , is zero. Consequently, the moment about the point of intersection of the two lines, M1 , equals zero both at and + . In the limiting case, where 0, this point of intersection of the two adjacent lines of action is well known as the rst metacenter, M1 , in naval architecture.
Flight Dynamics

170

Analysis of Steady Symmetric Flight

Zr centers of pressure C R (x0 , z0 ) CT 2 (< 0) CR mac Xr neutral point (xn , z0 ) C R

CN

C R + C R

M1 , rst metacenter

neutral line
Figure 5-16: The line of action of the dierence between aerodynamic force vectors at adjacent values of the angle of attack

5-1 Aerodynamic center We may conclude that in the rst metacenter the following two conditions hold, Cm
dCm d

171

= 0 (5-6) = 0

The physical meaning of the line of action of C R is that for each reference point on the line the change of aerodynamic moment will be zero. Consequently, for innitely small change of alpha the analytical expression for all points on the line of action of dC R is, dCm =0 d (5-7)

The line of action of dC R is called the neutral line. The rst metacenter turns out to be a special point on the neutral line where Cm = 0 as it is the intersection of the line of action of C R and the neutral line. If the aerodynamic forces and the moment about the leading edge of the mac (x0 , z0 ) are given the expression for the neutral line can be extracted from equations (5-1) and (5-7), dCm(x0 ,z0 ) d dCN x x0 dCT z z0 =0 d c d c

(5-8)

The intersection of the neutral line with the mac, where z = z0 , is called the neutral point of the wing. The distance of this neutral point (xn , z0 ) behind the leading edge of the mac follows from equation (5-8), dCm(x0 ,z0 ) xn x0 = c dCN with, zn = z0 The orientation of the neutral line follows from (see gure 5-16), 2 = arctan dCT dCN (5-10)

(5-9)

B. The second metacenter and the aerodynamic center Next, we consider two vectors dC R at two values of the angle of attack, 1 = and 2 = + . It follows that now, we have two neutral lines, one for 1 and for 2 . The point where these two neutral lines intersect satises the following conditions:
Flight Dynamics

172

Analysis of Steady Symmetric Flight neutral lines

C R2

C R1

( 0) M2

second metacenter = aerodynamic center

Figure 5-17: The position of the aerodynamic center

dCm d dCm d

=0
1

=0
2

see gure 5-17. In the limiting case where 0 the point of intersection follows from, dCm =0 d d2 Cm =0 d2

(5-11) (5-12)

The point of intersection of two neutral lines belonging to two adjacent angles of attack is called the second metacenter, M2 in naval architecture. In aeronautical engineering this point is referred to as the aerodynamic center (ac). In theoretical aerodynamics it is possible to prove that thin airfoils in a potential ow have an ac which is exactly invariant with the angle of attack. According to this thin airfoil theory, the ac of all thin airfoils is located at 25 % of the airfoil chord. In experimental aerodynamics, measurements on two-dimensional wing airfoils show that in the range of angles of attack where no ow separation occurs there is indeed an ac position which is very nearly independent of the angle of attack. Also in the case of a complete wing in three-dimensional ow the ac turns out to have a constant position over a wide range of angles of attack in many cases for moderate to large aspect ratios and small to moderate sweep angles.

5-1 Aerodynamic center

173

The . The location of the ac with coordinates xac and zac with respect to the leading edge of the mac follows from the conditions (5-11, 5-12), dCmac =0 d or, dCm(x0 ,z0 ) d and, dCN xac x0 dCT zac z0 =0 d c d c

(5-13)

d2 Cmac d2 2C d m(x0 ,z0 ) d2

= 0 + d2 CN xac x0 d2 CT zac z0 =0 d2 c d2 c

(5-14) (5-15)

in which it was assumed that xac and zac are independent of . After computing the two coordinates xac and zac of the ac by solving these equations, the aerodynamic moment about the ac, Cmac , follows by substitution into (5-1). This leads to the following expression for Cmac : Cmac = Cm(x0 ,z0 ) + CN xac x0 zac z0 CT c c (5-16)

If CN , CT and Cm(x0 ,z0 ) are known from measurements as functions of the angle of attack, it is possible in principle to obtain the coordinates xac and zac using the equations above (5-13). This would require, however, to dierentiate the experimental data on CN , CT and Cm two times with respect to alpha, which maybe dicult to do. An alternative way is to use the expressions for the neutral line (5-7), dCmac =0 d for two adjacent angles of attack. The coordinates xac and zac follow from their intersection .

With the concept of the aerodynamic center, the model of the aerodynamic forces and moment acting on a wing can be written in a very simple way. Each neutral line for any angle of attack must pass through the ac . As by denition the moment Cmac is constant, the model of the aerodynamic forces and moment acting on the wing thus consists of just a constant Cmac about the ac of the wing and the total aerodynamic force C R through the ac for all angles of attack (see gure 5-18).
Flight Dynamics

174 Zr

Analysis of Steady Symmetric Flight

C R + C R zac C R aerodynamic center xac Cmac mac

Xr

Figure 5-18: The aerodynamic forces and moment, using the ac as the moment reference point

C. Replacing the ac by the neutral point We have seen that the aerodynamic center is the moment reference point for which Cm is constant. It follows from practice that the position of the ac is indeed invariant for a lange of angles of attack. Furthermore, it appears that the aerodynamic center is usually located very close to the wings mac. This follows also from equation 5-15 if we take account of the virtually perfect linear relationships between Cm(x0 ,z0 ) and alpha, and CN and alpha. So, their second derivatives will vanish, resulting in zac = z0 , see again gure 5-9. By assuming that the ac is located exactly on the mac, the ac coincides with the neutral point. The following shows also that this simplication is usually quite acceptable. As Cmac is constant: dCmac =0 d it follows from equation (5-1) that, dCm(x0 ,z0 ) dCmac dCN xac x0 dCT zac z0 = + d d d c d c

(5-17)

In this expression dCT is relatively small compared to dCN . Moreover, the distance from the d d ac to the mac is small (a few percent of c at most). This means that the last term in equation ac eq:eq2p14 may be neglected and the x-coordinate of the ac xc , follows from equation (5-17) as, dCm(x0 ,z0 ) xac x0 = c dCN

(5-18)

5-1 Aerodynamic center

175

This result is identical to the expression for the location of the neutral point, equation (5-9), meaning that if the contribution of the tangential force is neglected in (5-17), the ac will indeed coincide with the neutral point. With zac = z0 , Cmac follows, with (5-1), from, Cmac = Cm(x0 ,z0 ) + CN and, clearly, Cmac = Cm(x0 ,z0 )C = Cm0 (5-20) xac x0 c (5-19)

N =0

As has been shown, the ac and Cmac can be obtained simply from experimental data on the aerodynamic moment about the leading edge of the mac. Next, using equation (5-18), xac is obtained from the slope of the Cm CN -curve. According to equation (5-20), Cmac is equal to Cm0 at CN = 0. In many cases in the literature this simplied and approximated position of the ac is used. See for instance reference [49].
ac Finally if xc and Cmac are known, the aerodynamic moment coecient about an arbitrary reference point on the mac is given by the fo0llowing very simple expression:

Cm (x) = Cmac + CN

x xac c

The description of Cm using the ac and Cmac derives its practical value from the fact that Cmac is constant and the position of the ac is independent of the angle of attack . However, for wings with a low aspect ratios and/or a large sweep angle the variation of the ac with angle of attack may not always be negligible, not even at those angles of attack occurring in normal ight. In order to enlarge the applicability of the ac and Cmac , a more general denition may be used in which Cmac is kept constant but the ac is allowed to shift position depending on alpha.

5-1-3

The role of the aerodynamic center and Cmac in static stability

If the Cmac and the position of the ac (xac , zac ) of a wing are known, the moment about an arbitrary point (x.z) is given by, see gure 5-19, Cm (x, z) = Cmac + CN x xac z zac CT c c (5-21)

For qualitative discussions it usually is permissable to simplify equation (5-21), as was already indicated in section 5-1-2, by neglecting the contribution of the tangential force, Cm = Cmac + CN x xac c (5-22)
Flight Dynamics

176 Zr

Analysis of Steady Symmetric Flight

CN zac xac x CT Or z Cm(x,z)

aerodynamic center

Cmac Xr

Figure 5-19: The moment about an arbitrary point (x, z) if the position of the ac and the Cmac are known

This simplied expression for Cm will often be used in the following discussions. Suppose a model of a wing has been mounted in a windtunnel so that the model is free to rotate about an axis through the moment reference point, serving now as the suspension point of the model. Furthermore, it is supposed that the center of gravity of the model has been made to coincide with this axis of rotation. Now a necessary condition for a steady equilibrium condition is Cm = 0. Small disturbances will induce model attitude (i.e. angle of attack) deviations from the nominal equilibrium condition. If such a deviation causes an aerodynamic moment acting against the change in angle of attack the original equilibrium situation was a stable one. This means that an increase in angle of attack must give rise to a negative, nose-down, change in aerodynamic moment. The condition for stability may thus be written as: dCm < 0 at Cm = 0 d (5-23)

It is easy to see that if dCm > 0 at Cm = 0, the equilibrium is unstable. If a change in angle of d attack causes no change in the aerodynamic moment, the equilibrium is said to be indierent or neutrally stable. In that case Cm = 0 at all angles of attack and there will be equilibrium at any value of . The importance of the position of the suspension point relative to the ac of the wing is illustrated by considering the equilibrium of the moment and the stability of the equilibrium of the wing for dierent positions of the suspension point. From equation (5-22) follows for the equilibrium (Cm = 0), Cm = Cmac + CN x xac =0 c (5-24)

5-1 Aerodynamic center and for the change of the moment with angle of attack, dCm dCN x xac = d d c where dCN in the range of linear variation of CN with is always positive. d Three cases are considered, 1. suspension point ahead of the ac,
xxac c

177

(5-25)

<0

According to equation (5-24), equilibrium is possible for either positive or negative values of CN depending on the magnitude and the sign of Cmac , see gures 5-20 and 5-21. From equation (5-25) follows that dCm < 0. This means that the equilibrium is d stable, as can be seen from gures 5-20 and 5-21. 2. suspension point in the ac,
xxac c

=0

In this situation, equilibrium is possible only if Cmac = 0, see equation (5-24). If this is the case, equilibrium exists at any value of CN , or (see gure 5-22). The equilibrium is indierent ( dCm = 0). d If Cmac = 0 no equilibrium situation is possible and as a consequence there can be no discussion of stability either (see gure 5-23). 3. suspension point behind the ac,
xxac c

>0

if Cmac < 0 equilibrium will exist at a certain positive CN and for Cmac > 0 at a negative value of CN (see gures 5-24 and 5-25). In both cases the equilibrium is unstable, dCm > 0. d

In summary it can be said that in the general case, i.e. where Cmac = 0, a condition of equilibrium (Cm = 0) is possible for any position of the suspension point ahead or behind the ac. The value of or CN for which Cm = 0 does depend on Cmac , see equation (5-24), but has by itself nothing to do with the stability of the equilibrium. At a given CN the stability follows only from the position of the suspension point relative to the ac, see equation (5-25). Summarizing, If the suspension point is situated ahead of the ac the model is always stable, dCm < 0. d If the suspension point moves back, the model becomes less stable as dCm becomes less d negative. If the suspension point coincides with the ac, the slope of the moment curve is zero. In that case equilibrium is possible only if Cmac = 0. This equilibrium is said to be indierent or neutrally stable. The equilibrium is always unstable if the suspension point lies behind the ac as then dCm d > 0.
Flight Dynamics

178

Analysis of Steady Symmetric Flight

CN Cm equilibrium at 1 1 2 Cm Cmac Cmac < 0; Cm (a)

dCm dCN

CN

<0

stable equilibrium

V 1 ac suspension point (b) CN CN1

Cmac Cm = 0

Cm 2 V

Cmac CN2 ac Cm < 0

suspension point (c)

CN1

Figure 5-20: Equilibrium and stability of a wing, suspension point ahead of the ac, Cmac negative

5-1 Aerodynamic center

179

CN Cm CN

Cmac

1 2 Cm

equilibrium at 1

Cmac > 0; Cm
dCm dCN

<0

stable equilibrium

(a) CN1

V ac suspension point

Cmac Cm = 0

(b)

CN2 CN1 2 V ac suspension point Cm < 0 Cm CN Cmac

(c)
Figure 5-21: Equilibrium and stability of a wing, suspension point ahead of the ac, Cmac positive

Flight Dynamics

180

Analysis of Steady Symmetric Flight

CN Cm CN

1 Cmac = 0 2 Cm

equilibrium at any

Cmac = 0;
dCm dCN

=0

neutrally stable equilibrium (a) CN1

V suspension point ac CN2 CN1 Cm = 0

(b)

2 V suspension point (c)

CN

ac Cm = 0

Figure 5-22: Equilibrium and stability of a wing, suspension point in the aerodynamic center, Cmac equals 0

5-1 Aerodynamic center

181

CN Cm CN

1 Cmac Cm1 Cm2 Cmac < 0; 2 Cm

dCm dCN

=0

no equilibrium (a) CN1 Cmac V suspension point (b) CN2 CN1 2 V suspension point (c)
Figure 5-23: Equilibrium and stability of a wing, suspension point in the aerodynamic center, Cmac negative

ac

Cm < 0

CN

Cmac

ac Cm = 0, Cm < 0

Flight Dynamics

182

Analysis of Steady Symmetric Flight

CN Cm equilibrium at 1 Cm 1 Cmac Cmac < 0; Cm 2

dCm dCN

CN

>0

unstable equilibrium (a) CN1 Cmac V ac suspension point CN2 CN1 2 V ac (c)
Figure 5-24: Equilibrium and stability of a wing, suspension point behind the ac, Cmac negative

Cm = 0

(b)

CN

Cmac Cm

suspension point Cm > 0

5-1 Aerodynamic center

183

CN Cm equilibrium at 1 Cm 1 2 Cmac

CN

Cm

Cmac > 0;
dCm dCN

>0

unstable equilibrium (a)

Cmac 1 V ac CN1 (b) CN Cmac 2 V ac suspension point CN2 CN1 Cm > 0 (c)
Figure 5-25: Equilibrium and stability of a wing, suspension point behind the ac, Cmac positive

suspension point Cm = 0

Cm

Flight Dynamics

184

Analysis of Steady Symmetric Flight

In the foregoing the inuence of a change in the x-coordinate of the moment reference point (the suspension point and rotation axis) was considered. In a similar way the inuence of a change in the position of the ac at a constant position of the suspension point can be studied. Such a shift in ac location may be caused for instance by the inuence of the fuselage on Cm as will be discussed later. From equation (5-25) it follows, that a forward shift of the ac (xac decreases) always has a destabilizing inuence and a rearward shift of the ac a stabilizing eect. In chapter 8 it is shown that the discussion given here for a wing, in a more extended form applies to the complete aircraft as well. The center of gravity of an aircraft has in this respect the same role as the axis of rotation of the wing. The slope of the moment curve (Cm ), with the center of gravity as the suspension point, is a measure of the static stability of the aircraft, i.e. its tendency to return the original angle of attack following a hypothetical disturbance.

5-1-4

The position of the aerodynamic center and Cmac of a wing

In the foregoing sections an elegant and compact model was derived for the aerodynamic moment acting on a wing. Section 5-1-3 discussed how to use that model for analysis of static stability. The present section describes a method to derive such a model for a wing of arbitrary shape and twist from strip theory rather than from wind tunnel measurements. The main motive to discus this theory here is to allow the reader to understand the relationships between the parameters dening wing shape and the aerodynamic center ac and Cmac The basic idea behind strip theory is to derive the aerodynamic characteristics of wings of arbitrary geometry, dihedral and twist from the aerodynamic characteristics of two dimensional wing proles. A two-dimensional dimensionless lift coecient may be dened as c =

denotes lift force per unit of length. In a similar way the two dimensional normal force coecient cn , drag coecient cd , tangential force coecient ct and moment coefcient cm may be dened. Just for simplicity, below, the dierence between the two dimensional lift and normal force is neglected, as well as the contributions of the drag or tangential force to the overall aerodynamic moment. The latter implies that eects of dihedral are neglected too. All formulas, however, may easily be extended to include these eects. The simple model of the aerodynamic forces and moments developed in part c of section 5-1-2 is now used as the model for the aerodynamic forces and moment acting on a wing strip at a given span-wise location. The lift distribution across the span of a wing can be divided into two parts. One part is the basic lift distribution cb (y), which is the lift distribution at CL =0 . This lift distribution depends on only on the geometry of the wing and the wing proles applied. The other part is

1 V 2 c 2

, in which

5-1 Aerodynamic center

185

cb c

b 2

b +2

(A) Basic lift distribution ca c

increasing

b 2

b +2

(B) Additional lift distribution c c

ca c
b 2 b +2

(C) Total lift distribution

Figure 5-26: The basic, additional and total lift distribution of a wing with negative twist

Flight Dynamics

186

Analysis of Steady Symmetric Flight

the additional lift distribution ca (y) which depends on the wing geometry and wing proles as well but in addition depends on the angle of attack, see gure 5-26. Several examples of basic, additional and total lift distributions using a source/doublet singularity panel method (based on linearized potential ow) are given in gures 5-27 to 5-32. The gures clearly illustrate the eects of wing-twist, taper ratio and wing sweep angle. We note that Cmac is the aerodynamic moment when CL = 0. This must imply that Cmac is caused by the torque resulting from the basic lift distribution and the cmac s of the wing strip airfoils. From the characteristics of the applied wing airfoils as well as the shape of the wing we now continue by computing the position of the ac, the magnitude of Cmac and lift gradient CL . The lift acting on a wing strip dy is, dL = dLa + dLb = (ca + cb ) 1 2 V c dy 2

This local lift force is assumed to act through the local ac, the constant moment cmac must be added, see gure 5-33. The moment due to the additional lift distribution about x0 , the leading edge of the mac, follows as,
b 2

Ma = 2

ca
0

1 2 V c 2

(x x0 ) cos dy

The resultant force of the additional lift distribution of the wing may be assumed to act through xac the ac of the wing on the mac. So, the moment about the leading edge x0 of the mac due to the additional lift distribution may be written as, Ma = La (xac x0 ) Combining the above two expressions results in,
b 2

La (xac x0 ) = 2 Using,

ca
0

1 2 V c 2

(x x0 ) cos dy

L = La = CL

1 2 V S 2

and assuming cos 1, results in the required position of the ac of the wing, xac x0 1 2 = c CL S c
b 2

ca c (x x0 ) dy

(5-26)

Z [m]

0 5 0 0 2 4 6 8 5

Z [m]

0 5 0 0 2 4 6 8 5

5 2

5 2

Y [m]

Y [m]

X [m] without wing-twist

1.4 1.2 1 0.8 1.4

X [m] including wing-twist c c ca c cb c

1.2 1 0.8

c c ca c cb c

c c

0.6 0.4

c c

0.6 0.4 0.2 0

5-1 Aerodynamic center

0.2 0 0.2 0 0.2 0.4 0.6 0.8 1

0.2

b y/( 2 )

0.2

0.4

b y/( 2 )

0.6

0.8

Figure 5-27: Numerical simulation of the basic lift distribution (cb c at = CL =0 ), additional lift distribution (ca c) and total lift distribution (c c) for a wing without wing-twist (left, = 0o ) and a wing including wing-twist (right, = 3o ); NACA 0012 airfoil, b = 16 m, S = 32 m2 , c = 2 m, = 0o , = 0o , = 1, r = 0o , = 3o , = 5o . Results from a source/doublet singularity panel-method, linearized potential ow.

Flight Dynamics

187

Analysis of Steady Symmetric Flight

Z [m]

0 5 0 0 2 4 6 8 5

Z [m]

0 5 0 0 2 4 6 8 5

5 2

5 2

Y [m]

Y [m]

X [m] without wing-twist

1.4 1.2 1 0.8 1.4

X [m] including wing-twist c c ca c cb c

1.2 1 0.8

c c ca c cb c

c c

0.6 0.4 0.2 0 0.2 0 0.2 0.4 0.6 0.8 1

c c

0.6 0.4 0.2 0 0.2 0 0.2 0.4 0.6 0.8 1

b y/( 2 )

b y/( 2 )

Figure 5-28: Numerical simulation of the basic lift distribution (cb c at = CL =0 ), additional lift distribution (ca c) and total lift distribution (c c) for a wing without wing-twist (left, = 0o ) and a wing including wing-twist (right, = 3o ); NACA 0012 airfoil, b = 16 m, S = 32 m2 , c = 2 m, = 0o , = 37o , = 1, r = 0o , = 3o , = 5o . Results from a source/doublet singularity panel-method, linearized potential ow.

188

Z [m]

0 5 0 0 2 4 6 8 5

Z [m]

0 5 0 0 2 4 6 8 5

5 2

5 2

Y [m]

Y [m]

X [m] without wing-twist

1.4 1.2 1 0.8 1.4

X [m] including wing-twist c c ca c cb c

1.2 1 0.8

c c ca c cb c

c c

0.6 0.4

c c

0.6 0.4 0.2 0

5-1 Aerodynamic center

0.2 0 0.2 0 0.2 0.4 0.6 0.8 1

0.2

b y/( 2 )

0.2

0.4

b y/( 2 )

0.6

0.8

Figure 5-29: Numerical simulation of the basic lift distribution (cb c at = CL =0 ), additional lift distribution (ca c) and total lift distribution (c c) for a wing without wing-twist (left, = 0o ) and a wing including wing-twist (right, = 3o ); NACA 0012 airfoil, b = 16 m, S = 32 m2 , c = 2 m, = 0o , = 0o , = 1 , r = 0o , = 3o , = 5o . Results from a source/doublet singularity panel-method, linearized potential 3 ow.

Flight Dynamics

189

Analysis of Steady Symmetric Flight

Z [m]

0 5 0 0 2 4 6 8 5

Z [m]

0 5 0 0 2 4 6 8 5

5 2

5 2

Y [m]

Y [m]

X [m] without wing-twist

1.4 1.2 1 0.8 1.4

X [m] including wing-twist c c ca c cb c

1.2 1 0.8

c c ca c cb c

c c

0.6 0.4 0.2 0 0.2 0 0.2 0.4 0.6 0.8 1

c c

0.6 0.4 0.2 0 0.2 0 0.2 0.4 0.6 0.8 1

b y/( 2 )

b y/( 2 )

Figure 5-30: Numerical simulation of the basic lift distribution (cb c at = CL =0 ), additional lift distribution (ca c) and total lift distribution (c c) for a wing without wing-twist (left, = 0o ) and a wing including wing-twist (right, = 3o ); NACA 0012 airfoil, b = 16 m, S = 32 m2 , c = 2 m, = 0o , = 37o , = 1 , r = 0o , = 3o , = 5o . Results from a source/doublet singularity panel-method, linearized potential 3 ow.

190

Z [m]

0 5 0 0 2 4 6 8 5

Z [m]

0 5 0 0 2 4 6 8 5

5 2

5 2

Y [m]

Y [m]

X [m] without wing-twist

1.4 1.2 1 0.8 1.4

X [m] including wing-twist c c ca c cb c

1.2 1 0.8

c c ca c cb c

c c

0.6 0.4

c c

0.6 0.4 0.2 0

5-1 Aerodynamic center

0.2 0 0.2 0 0.2 0.4 0.6 0.8 1

0.2

b y/( 2 )

0.2

0.4

b y/( 2 )

0.6

0.8

Figure 5-31: Numerical simulation of the basic lift distribution (cb c at = CL =0 ), additional lift distribution (ca c) and total lift distribution (c c) for a wing without wing-twist (left, = 0o ) and a wing including wing-twist (right, = 3o ); NACA 0012 airfoil, b = 16 m, S = 32 m2 , c = 2 m, = 5o , = 0o , = 1 , r = 0o , = 3o , = 5o . Results from a source/doublet singularity panel-method, linearized potential 3 ow.

Flight Dynamics

191

Analysis of Steady Symmetric Flight

Z [m]

0 5 0 0 2 4 6 8 5

Z [m]

0 5 0 0 2 4 6 8 5

5 2

5 2

Y [m]

Y [m]

X [m] without wing-twist

1.4 1.2 1 0.8 1.4

X [m] including wing-twist c c ca c cb c

1.2 1 0.8

c c ca c cb c

c c

0.6 0.4 0.2 0 0.2 0 0.2 0.4 0.6 0.8 1

c c

0.6 0.4 0.2 0 0.2 0 0.2 0.4 0.6 0.8 1

b y/( 2 )

b y/( 2 )

Figure 5-32: Numerical simulation of the basic lift distribution (cb c at = CL =0 ), additional lift distribution (ca c) and total lift distribution (c c) for a wing without wing-twist (left, = 0o ) and a wing including wing-twist (right, = 3o ); NACA 0012 airfoil, b = 16 m, S = 32 m2 , c = 2 m, = 5o , = 37o , = 1 , r = 0o , = 3o , = 5o . Results from a source/doublet singularity panel-method, linearized potential 3 ow.

192

5-1 Aerodynamic center

193

Yr x0

x line of local acs

mac

dy

Xr

Zr (ca + cb ) 1 V 2 cdy 2

x0 x ac

cmac 1 V 2 c2 dy 2

Xr
Figure 5-33: The calculation of the position of the ac and Cmac for a wing with arbitrary geometry

Flight Dynamics

194

Analysis of Steady Symmetric Flight

The contribution of a wing strip dy to the total moment about x0 , the leading edge of the mac, follows from gure 5-33, 1 dM = {cmac c c (x x0 ) cos } V 2 c dy 2 The non-dimensional moment of the entire wing about x0 , if again cos 1, is obtained by integration along the wing span and division by 1 V 2 S c 2 2 = S c
b 2 b 2

Cm or since,

cmac c2 dy

c c (x x0 ) dy

(5-27)

c = ca + cb equation (5-27) may be written as, 2 = S c

b 2 b 2 b 2

Cm

cmac c2 dy

cb c (x x0 ) dy

ca c (x x0 ) dy

(5-28)

According to equation (5-22), Cm about x0 ), the leading edge of the mac, assuming again CN CL , can be written as, Cm = Cmac + CL x0 xac c (5-29)

Substituting equation (5-26) in equation (5-29) and equating equation (5-28) to equation (5-29) results in the expression for Cmac , 2 = S c
b 2 b 2

Cmac

cmac c2 dy

cb c (x x0 ) dy

(5-30)

We now have expressed the position of the ac of the wing as in equation (5-26) and the value of Cmac in equation (5-30). These results are subject to the same simplifying assumptions as were made in part c of section 5-1-2, 1. The contribution of the tangential force to the moment is neglected and as a consequence the ac is located on the mac 2. The position of the ac is assumed to be independent of Reference [5] provides for a large number of wing shapes the values of xac according to equation (5-26), in addition to the magnitudes of the rst and second integral in equation (5-30). Data on the position of the ac and of Cmac can be found in references [74, 152, 172].

5-1 Aerodynamic center

195

5-1-5

Inuence of wing shape on the aerodynamic center and Cmac

It is quite dicult to give general rules for the inuence of the various wing geometry parameters describing the wing shape, such as aspect ratio and wing sweep on the aerodynamic moment curve of wings. The reason is, that the inuence of one parameter strongly depends on the value of the others. Not too many experimental (wind tunnel) results are available in which those parameters were systematically varied. Most of these experimental results apply to wing-fuselage combinations, so variations of the inuence of the fuselage with wing sweep, wing taper ratio, etcetera, are included in the experimental results. For these reasons it is not feasible to aim for completeness and a description of only a few of the more general eects of wing shape on the aerodynamic moment has to suce. We have seen that the aerodynamic moment characteristics of a wing can be expressed in terms of a constant value of Cmac and an approximated position of the ac on the wing mac. Relative to an arbitrary reference point (x1 , z0 ) on the mac the aerodynamic moment may then be written as see equation (5-21), Cm(x1 ,z0 ) = Cmac + CN x1 xac c

This equation is valid as far as linear wing strip theory is valid, so up to only moderate values of the angle of attack, wing sweep angle and slender wings with moderate to large wing aspect ratios. For such wings, Cm and CN vary more or less linearly with if the same holds true for the individual wing strip proles. The Cmac of a complete wing could be written as as, see equation (5-30), 2 = S c
b 2 b 2

Cmac

cmac c2 dy

cb c (x x0 ) dy

From this expression we see that Cmac is partly determined by the cmac s in the rst integral, i.e. by the camber of the individual wing airfoils. However, from the second integral in equation (5-30) it appears that Cmac also depends on the wing twist in case the wings are swept , due to the contribution made of the basic lift distribution. Straight wings have a constant factor (x x0 ) in the second integral in equation (5-30), reducing it to zero, and Cmac now only results from the contribution of the camber of the wing airfoils to the local cmac . Increasing the camber leads to a more negative cmac , see for example the wind tunnel results of gure 5-34. If a straight wing has a constant cmac along the wing span, the result is Cmac = cmac thanks to the denition of the mac. For sweptback wings having a negative wing twist the positive cb at the wing root and the negative cb at the tip will result in a positive (tail-heavy) contribution to Cmac , see gure 5-35a. This contribution increases with wing sweep and aspect ratio, see gure 5-35b. ct Decreasing the taper ratio ( = cr ) decreases again this eect.

Flight Dynamics

196

Analysis of Steady Symmetric Flight

Cm0.25

0.1

NACA Airfoil 63A010

-0.4 -0.2 0 0.2 0.4 0.6 0.8

63A210

1.0

CL

-0.1

63A410
t c

(A) = 1, = 0o , A = 4, M = 0.6, Cm0.25

0.1

= 0.10

NACA Airfoil 63A010

-0.4 -0.2 0 0.2 0.4 0.6 0.8 1.0

CL

63A210 63A410
-0.1

(B) = 1, = 0o , A = 2, M = 0.6, Cm0.25

0.1

t c

= 0.10

NACA Airfoil 63A004

-0.4 -0.2 0 0.2 0.4 0.6 0.8 1.0

CL

63A204 63A404
-0.1

(C) = 1, = 0o , A = 4, M = 0.6, Cm0.25

0.1

t c

= 0.04

NACA Airfoil 63A004

-0.4 -0.2 0 0.2 0.4 0.6 0.8 1.0

63A204 63A404
-0.1

CL

(D) = 1, = 0o , A = 2, M = 0.6,

t c

= 0.04

Figure 5-34: The inuence of camber on the moment curves (from references [119, 118])

5-1 Aerodynamic center

197

YB cb c XB

ZB (A) The moment due to basic lift distribution (left wing)

Cmac (per o )

0.008

A = 6.0 A = 3.5 A = 1.5 (o )

0.004

-60

-40

-20

20

40

60

-0.004

(B) Cmac as a function of A and , = 0.5

Figure 5-35: The variation of Cmac with wing sweep and angle of twist (from reference [49])

Flight Dynamics

198

Analysis of Steady Symmetric Flight acs of the local airfoils

local acs ac 1 4c

ac

1 4c

Figure 5-36: The positions of the local acs of swept wings and delta wings

The deection of landing aps may cause a signicant increase of the camber of the wing cross-section over a part of the wing span. For straight wings the result is a more negative Cmac for the entire wing. Flap deection also causes a change of the basic lift distribution similar to the eect an increased negative twist at the wing tips. If the wing has positive (backward) sweep back, this change in the basic lift distribution induces itself a positive change in Cmac . As a consequence, the total change in Cmac due to ap deection may even be positive. Calculations of this eect have been made for instance in references [121, 47]. The position of the ac is, (5-26),
b 2

xac x0 2 = c CL S c

ca c (x x0 ) dy

The position of local acs of wing airfoils of straight wings with moderate to large aspect ratios correspond are close to the acs in two-dimensional ow. As a consequence, the ac of such wings will lie around the 25 % position on the mac of the wing. Straight wings with low aspect ratio (A < 3), however, show an appreciable forward shift with decreasing aspect ratio. For these straight wings, ap deection causes no signicant shift of the ac. The lift distribution in chordwise direction of swept wings and delta wings is strongly inuenced by wing sweep. The result is that the local acs of the wing no longer agree with the acs of the local wing airfoils in two-dimensional ow, see gure 5-36. The local acs at the tips of swept back wings lie ahead of the acs of the corresponding wing

5-1 Aerodynamic center

199

Cm0.25

= 60o
0.28

0.24

= 60o

0.20

= 45o

0.16

0.12

= 45o
0.08

= 32.6o

= 32.6o

0.04

-0.2

0.2

0.4

0.6

0.8

CL
-0.04

= 3.6o

= 3.6o

Figure 5-37: The inuence of wing sweep on the moment curves of wings of aspect ratio A = 4, = 0.6, t c = 0.06, M = 0.40 (from reference [169])

Flight Dynamics

200

Analysis of Steady Symmetric Flight

airfoils. Since for swept back wings the additional lift distribution is concentrated more near the wing tips, the ac of the complete wing will be situated ahead of the 25 % mac position. This forward shift increases with increasing sweepback angle and we know that this has a destabilizing inuence, see also [5]. However, the combination of a swept back wing with a fuselage often shows a rearward shift of the ac with increasing sweepback angle, see the linear part of the moment curves up to CL 0.4 in gure 5-37. This is due to the inuence of wing sweepback on the wing-fuselage interference, see also section 5-1-6.
ct If the taper ratio ( = cr ) decreases, the contribution of the wing tips to the aerodynamic moment decreases. As the ac at the wing tip is shifted forward due to wing sweep, see 5-36, decreasing the taper-ratio of swept back wings moves the ac backwards, thus has a stabilizing eect, see the linear part of the moment curves up to CL 0.4 in gure 5-38.

Wing aspect ratio as such only has a relatively minor inuence on the ac. Figure 5-39 for example shows the position of the ac as a function of aspect ratio for wings with 45o of sweep back angle . Delta wings have local acs situated behind the quarter-chord position over a large part of the wing span, see gure 5-36. In addition, the lift distribution of delta wings shows a concentration more towards the center parts of the wing. This causes the wing ac to be situated behind the 25 % mac position. Wing aspect ratio and sweep back of delta wings cannot be varied independently. Increasing sweepback decreases aspect ratio, leading to a rearward shift of the ac, see gure 5-39. Similar to the case of swept wings, a fuselage added to a delta wing has a stabilizing eect (i.e. rearward shift of the ac), which increases with decreasing wing aspect ratio, and thus with increasing sweepback angle. At suciently large angles of attack in particular the eects of ow separation result in nonlinear moment curves for straight wings, wings with sweep back angle as well as for delta wings. Now, for these large angles of attack, the concepts of ac and Cmac start loosing their signicance as in order to match the slope of the moment curve, not only the location of the ac, but also the value of the Cmac must be adapted. As also argued in reference [1] the characteristics of wings in such conditions may best be judged from aerodynamic moment curves for some suitable reference point as stability characteristics may still be deduced from the slope of the moment curves, as discussed earlier. Straight wings usually exhibit large negative changes in dCm dCm at lift coecients near dCL dCN CLmax due to ow separation. Figures 5-41 and 5-34 give examples of this phenomenon. As in ight such a negative (nose-down) change in Cm would help to decrease the angle of attack this is a favourable characteristic of straight wings. Swept wings,however, with moderate to large aspect ratios usually suer from a large positive slope change increasing angle of attack. The change may start already at relatively small

5-1 Aerodynamic center

201

Cm0.25

0.28

0.24

0.20

= 1.0

0.16

= 1.0
0.12

= 0.6
0.08

0.04

= 0.6
-0.2 0 0.2 0.4 0.6 0.8

CL = 0.3
-0.04

= 0.3

Figure 5-38: The inuence of taper ratio on the moment curves of swept wings, A = 4, = 45o , t c = 0.06, M = 0.40 (from reference [87])

Flight Dynamics

202
xac c
0.37

Analysis of Steady Symmetric Flight

0.33

Delta wings

xac

1 4c

ac
0.29

0.25

A xac ac 1 4c c

0.21

0.17

Swept wings

Figure 5-39: The position of the ac of delta wings and concept wings as a function of aspect ratio, = 45o (from reference [159])

values of . Figures 5-37, 5-38, 5-40 and 5-41 give examples. These changes are caused by premature ow separation at the outer wings, predominantly caused by the high tip loading and the cross ow towards the wing tips. The result is an increasing thickness of the boundary layer and early ow separation. This eect occurs at lower CL values, and the eect on Cm is stronger as sweep back angle and aspect ratio are larger. Figure 5-42 gives an indication of the combinations of aspect ratio and wing sweep angle for which destabilizing changes in the slope of the moment curves of the wing may be expected. This sudden change of slope of the moment curve is called pitch up. If it would occur in ight its overwhelming eect is to increase the angle of attack which evidently is highly undesirable and potentially very dangerous. A horizontal tailplane has an important eect on the aerodynamic pitching moment and may be designed as to compensate for the adverse eect of pitch up as we will see later. However, even then pitch up remains a very undesirable wing characteristic which should best be avoided through design. To avoid too much lift concentration near the wing tips (and the pitch up phenomenon) negative wing twist is often used. Camber and other shape parameters of the wing airfoils may be adjusted to obtain desired values of local cmax without undue drag increases in cruising ight. To prevent cross-ow towards the wing tips (inducing early ow separation), several means may be applied such as boundary layer fences or a saw tooth leading edge. Reference [68] provides an extensive compilation of the various methods and tools which may be employed for this purpose.

5-1 Aerodynamic center

203

Cm0.25

0.28

A=6
0.24

0.20

A=6

0.16

0.12

A=4
0.08

A=4
0.04

-0.2

0.2

0.4

0.6

0.8

CL
-0.04

A=2

A=2

Figure 5-40: The inuence of aspect ratio on the moment curve of swept wings, = 45o , = 0.6, t c = 0.06, M = 0.40 (from reference [97])

Flight Dynamics

204

Analysis of Steady Symmetric Flight

Cm0.25
0.12

0.08

Swept wing = 45o

0.04

Straight wing = 0o

12

16

20

[o ]

-0.04

-0.08

Figure 5-41: The moment curve of a straight wing and a swept wing, A = 5, = 1, Re = 1 106 (DUT measurements)

5-1 Aerodynamic center

205

A
14

12

=1

10

Ref. [58] =0

Ref. [75] unstable

6

stable
2

10

20

30

40

50

60

70

1 c [o ]
4

Figure 5-42: Boundaries for the combinations of aspect ratio and sweep angle for which destabilizing changes in the moment curves may be expected (from references [58, 75])

Flight Dynamics

206

Analysis of Steady Symmetric Flight

5-1-6

Characteristics of wing-fuselage-nacelle congurations

The model of the aerodynamic moment of the combination of a wing with fuselage and nacelles attached to the wing is in principle the same as for the isolated wing. At small angles of attack Cm varies linearly with CN , as it does for the isolated wing. As a consequence, changes in the longitudinal moment caused by the fuselage and the nacelles are commonly expressed as a shift in the position of the ac and a change in Cmac . Neglecting again the contribution of the tangential force, the moment coecient of the wing was written as, see equation (5-22), Cmw = Cmacw + CN x xacw c

In a similar way the moment of the combination of the wing with fuselage and nacelles can be expressed as, Cmw+f +n = Cmacw+f +n + CN At constant CN the change in the moment is, Cm = Cmw+f +n Cmw = Cmac CN xac c (5-31) x xacw+f +n c

Similar to Cmacw , also Cmacw+f +n is dened as constant and independent of the angle of attack. This means that Cmac is independent of as well. Cmac is the change of Cm at CN = 0 due to the addition of a fuselage (and nacelles). The shift in the ac position then follows from equation (5-31), d (Cm ) xac 1 d (Cm ) = = c dCN CN d and the change in Cmac is, Cmac = (Cm )CN =0 (5-33) (5-32)

The change in the longitudinal moment due to the presence of the fuselage and the wings nacelles can be thought of as consisting of the contributions of the fuselage and the nacelles separately, the contribution of the wing-fuselage interference and the contribution of the wing-nacelle interference. Nacelles mounted at the rear fuselage are commonly regarded as part of the fuselage. Wing mounted nacelles exert an inuence on the longitudinal moment similar to the fuselage. For these reasons only the eect of adding a fuselage to the wing is discussed in the following. The change in the moment due to the presence of the fuselage, Cm , can be written in the following expression, Cmw+f = Cmw + Cm

5-1 Aerodynamic center where, Cm = Cmf + Cmi

207

Here Cmf denotes the longitudinal moment of the free fuselage in undisturbed ow while Cmi denotes the wing-fuselage interference. From windtunnel tests the total Cm can be simply obtained by comparing the moment curves of the isolated wing and the wing-fuselage combination at equal values of CN . If, in addition, measurements have been made on the isolated fuselage, i.e. to determine Cmf , Cmi can be found as well. For theoretical studies it appears to be useful to divide the total interference eect Cmi into two parts, 1. Cmf.i. ; the change in the fuselage moment caused by situating the fuselage in the eld of ow around the wing. 2. Cmw.i. ; the change in the wing moment caused by placing the wing in the eld of ow around the fuselage. So, Cm = Cmf + Cmf.i. + Cmw.i. Usually, the total moment Cmf.i. contributed by the fuselage in the eld of ow around the wing is calculated as one eect, Cmf.i. = Cmf + Cmf.i. This results in, Cm = Cmf.i. + Cmw.i. In the following the latter two contributions to the change in Cm and the subsequent changes in the position of the ac and the Cmac are discussed. A. The moment on the fuselage positioned in the eld of ow induced by the wing, the magnitude of Cmf.i. For a fuselage placed under a non-zero angle of attack in an inviscid and incompressible ow it is possible to prove that it will experience just an aerodynamic moment while the aerodynamic force will be zero. The mechanism behind this moment is illustrated in gure 5-43 (see also gure 5-44 for a numerical simulation of particle lines and pressure distribution over a slender fuselage, linearized potential ow). The magnitude of this moment acting on
Flight Dynamics

208
- + + ++ - -

Analysis of Steady Symmetric Flight

+ + + + - - - - +

Figure 5-43: Pressure distribution over a fuselage in inviscid ow

a slender fuselage of circular crosssection in these conditions of incompressible, inviscid ow has been derived by Munk (reference [115]), Mf = V 2 f (Volume of Fuselage) From this expression follows,
lf

Cmf Where, bf (x) lf f S, c

f = 2S c
0

b2 (x) dx f

(5-34)

local fuselage width fuselage length fuselage angle of attack, [Rad.] characteristic parameters of the wing to which Cmf is referred

According to equation (5-34) the moment is positive (tail-heavy) at positive angles of attack and it increases with increasing angles of attack. The implication is that the idealized isolated fuselage is statically unstable at f = 0o and cannot be in equilibrium at f = 0o . If the wing and the fuselage are combined, the fuselage will experience the eect of air velocity variations as induced by the wing. This causes the angle of attack along the fuselage axis to be dierent from that of an isolated fuselage. As a consequence, also the pressure distribution over the fuselage is dierent from that of an isolated fuselage. The wing induced variation of the local angle of attack and the corresponding local normal force along the fuselage axis is shown schematically in gure 5-45. Ahead of the wing, for positive values of CL , the upwash causes an increase of f (x) up to the wing leading edge. Across the wing chord the airow is guided parallel to the wing chord resulting in f (x) being zero. Behind the wing the local angle of attack along the fuselage is reduced due to the downwash induced by the wing. Results of some numerical simulations of the oweld of the wing-fuselage combination of gure 5-46 are depicted in gures 5-47 and 5-48. The result of the induced upwash in front of the wing is an increase in tail-down moment with increasing angle of attack, which as we know, is destabilizing. The wing induced downwash has the opposite eect, however, usually not enough to compensate fully the destabilizing

5-1 Aerodynamic center

209

Pressure distribution Cp

0.6

0.4

0.2

0.2

0.4

0.6

0.8

1.2

Figure 5-44: Numerical simulation of particle lines (top) and pressure distribution over a fuselage in inviscid ow, = 10o , linearized potential ow

Flight Dynamics

210

Analysis of Steady Symmetric Flight

eect of the wing induced ow eld. Now it is not dicult to see that a long fuselage in front of the wing, as needed for transport aircraft with rear mounted engines, amplies this eect. As shown in the classical reference [114], the aerodynamic moment acting on a slender fuselage in an inviscid, incompressible ow eld with induced velocities can be written as,
lf

Cmf.i.

= 2S c
0

b2 (x) f (x) dx f

(5-35)

Note that by substituting f (x) for the case of an isolated fuselage in (5-35 this equation should yield the same result as equation (5-34) ! f (x) can be expressed in terms of the wing angle of attack as, f (x) = f0 + df (x) ( 0 ) d (5-36)

where 0 denotes CN =0 and f0 is the fuselage angle of attack at = 0 . With equation (5-36) follows for Cmf.i. ,
lf lf

Cmf.i.

f0 = 2S c
0

( 0 ) b2 (x) dx + f 2S c
0

b2 (x) f

d f (x) dx d

(5-37)

in which the rst term on the right hand side represents the fuselage moment at CN = 0, and so Cmac due to the fuselage is,
lf

Cmac = (Cm )CN =0

f0 = 2S c
0

b2 (x) dx f

(5-38)

With equation (5-32) the shift in ac position due to the fuselage is, xac c 1 dCmf.i. 1 = = CN d CN 2S c
lf

b2 (x) f
0

f.i.

d f (x) dx d

(5-39)

Reference [66] presents some classical semi-empirical correction formulae for Cmac and xac accounting for friction eects of fuselages of nite length. c f.i.
f Computing Cmf.i. with equation (5-37)is possible if the variation of d along the fuselage axis is known. Some classical references are [114], [165] and [160]. Nowadays one would no doubt rather rely on some numerical method as used to compute the numerical simulations in the gures above.

Some nice qualitative results are shown in Figures 5-49 and 5-50. It turns out that both d f a reduction of aspect ratio and an increase of sweepback angles result in decreasing d in

5-1 Aerodynamic center

211

f (x)

df d

upwash downwash
1

1 f = 0 and
df d

d d

dh d

=0

dNf dx

X
Figure 5-45: The variation of the angle of attack and the normal force along the fuselage axis in a wing induced ow eld .

Flight Dynamics

212

Analysis of Steady Symmetric Flight

Figure 5-46: Wing-fuselage conguration as used in the ow eld simulations depicted in gures 5-47 and 5-48 .

Figure 5-47: Numerical simulation of the velocity eld of a wing-fuselage conguration, = 10o .

5-1 Aerodynamic center

213

Figure 5-48: Numerical simulation of the velocity eld of a wing-fuselage conguration, = 10o , detail of gure 5-47 .
f.i. front of the wing. This means that the contribution of the fuselage nose to d becomes less stabilizing, see equation (5-39). The downwash angle behind the wing increases with d f d decreasing wing aspect ratio and increasing sweepback angle, d increases and d is closer to zero. In this way the destabilizing contribution to Cmf.i. of the rear part of the fuselage behind the wing decreases as well. We conclude that both decreasing the wing aspect ratio and increasing the sweepback angle reduce the destabilizing eect of the fuselage to Cm .

d Cm

B. The inuence of the fuselage on the wing moment, the magnitude of Cmw.i. The inuence of the fuselage on the characteristics of the wing itself consists of two parts, 1. The extra upwash induced by the presence of the fuselage over the wing root next to the fuselage, see gures 5-47, 5-48 and 5-51. Results of numerical simulations are shown in gure 5-52. These simulations were performed using the wing-fuselage conguration as of gure 5-46. 2. The change in lift distribution over that part of the wing covered by the fuselage, see gure 5-53. A further look at the fuselage induced upwash learns that its eect on the wing moment is relatively small and can be usually be neglected, see reference [143]. The induced upwash of wing-mounted nacelles, however, may have a non-negligible eect, see reference [114].

Flight Dynamics

214
df d

Analysis of Steady Symmetric Flight

A= increasing A A=9

A=3 increasing A A= A=9 A=3 A=0

-2 -1 0 1 2 3
x cr

A=0

Figure 5-49: The inuence of aspect ratio on the variation of reference [143])
df d

df d

along the fuselage axis ( = 0) (from

= 45o = 0o increasing

= 45o increasing
1

0 -2 -1 0 1 2 3

= 45o = 0o = 45o

x cr

Figure 5-50: The inuence of sweep angle of wings with A = on reference [143])

df d

along the fuselage () (from

5-1 Aerodynamic center Z

215

(A) Flow around the fuselage inuence of the fuselage (y)

y (B) The variation of the angle of attack along the wingspan

Figure 5-51: Fuselage induced variations of the local angle of attack along the wing span

Left wing

Figure 5-52: Numerical simulation of the fuselage induced upwash along the wing span, = 5o
Flight Dynamics

216 c c
0.8

Analysis of Steady Symmetric Flight

wing
0.6

0.4

wing with fuselage

0.2
1 2 bf

0 0 0.2 0.4 0.6 0.8 1.0

b y/ 2

Figure 5-53: The c c-distribution along the span of a wing with and without a fuselage (from reference [54])

Figure 5-53 shows measurements of the lift distribution across the span of a wing with and without fuselage. It appears that the fuselage aects lift only close to the fuselage. For straight wings this fuselage eect on the Cmac of the wing can safely be neglected. For swept back wings, and also for delta wings, the ac of the entire wing lies behind the local acs of the center wing proles , see gure 5-54. This means that the lift reduction CLw of the center part of the wing will induce in a nose-down change in moment, i.e. Cmw.i. is negative. As CLw is more or less proportional to CL this eect is stabilizing , shifting ac of the wing rearward. For swept forward wings, however, the same eect causes a tail-down change in the wing moment and is destabilizing ,shifting the ac forward . Reference [8] allows one to estimate Cmw.i. with a simple method. The inuence of the sweep angle on the ac as computed with, xac c = 1 dCmw.i. CN d (5-40)

w.i.

is shown in gure 5-55 for a set of idealized wing-body congurations. The gure shows the contribution to xac caused by the fuselage in the wing induced eld of ow Cmf.i. as derived from equation (5-37), the eect of the fuselage on aerodynamic moment of the wing Cmw.i. and the resulting shift of ac position caused by the fuselage, both calculated and measured.

5-1 Aerodynamic center

217

line of acs

ac of the wing center part

ac of the wing

ac of the wing center part

ac of the wing
Figure 5-54: The change in wing moment due to loss in lift over the wing center part

Flight Dynamics

218
xac c

Analysis of Steady Symmetric Flight

-0.18

measurements
-0.12

fuselage in wing induced eld of ow Cmf.i.

-0.06

total fuselage inuence

-30 -20 -10 0 10 20 30 40

[o ]

0.06

eect of fuselage on the wing Cmw.i.

A = 5, = 0.2

Xr
Figure 5-55: Shift of ac due to fuselage eects, as a function of wing sweep angle(from reference [141])

5-1 Aerodynamic center

219

Figure 5-56: Flow simulations over a nacelle, = 15o , computed with a source/doublet singularity panelmethod, linearized potential ow

From this gure it appears once again that the fuselage has a destabilizing inuence, whereas a wing with positive sweepback angle has a stabilizing eect. As has already been noted in 5-1-5 this stabilizing inuence of wing sweepback may compensate the destabilizing eect of the fuselage. The resulting shift of the ac may turn out to be stabilizing.

5-1-7

Aerodynamic eects of nacelles

Analysis of the aerodynamic eects of wing- or fuselage-mounted nacelles on the aerodynamic characteristics of a wing-fuselage conguration starts with the isolated nacelle . In gure 5-56 a generic conguration of a engine nacelle including simulated particle lines is given for the case of an empty nacelle (no engine thrust eects ). Considering this nacelle at an angle-of-attack of = 15o it is obvious that the local airow will conform to the exterior of the nacelle, the nacelle can be seen as a ring shaped wing .

Example: aircraft with wing-mounted nacelles We rst consider a transport aircraft model with swept back wings without its nacelles, see gure 5-57. The local airow over the wing at the nacelle position is shown in gure 5-58. In gure 5-58 simulated particle lines are shown for an angle-of-attack = 10o as well as the non-dimensional pressure distribution coecient Cp . The aerodynamic eect of the sweep
Flight Dynamics

220

Analysis of Steady Symmetric Flight

Figure 5-57: Generic Large Transport Aircraft (GLTA), no nacelles

angle on the non-dimensional pressure distribution Cp of the wing sections is evident from the increasing pressures moving to the wing tip. Next we add nacelles to the wing . In gure 5-59 results are shown for the case of one nacelle per wing including simulated particle traces and non-dimensional sectional pressure distribution coecients Cp . It is clear that the nacelle has a signicant eect on the local pressure distribution resulting in a reduction of the local angle-of-attack as the airow conforms to the external shape of the nacelle. This is conrmed by gure 5-60 showing the spanwise lift distribution expressed in terms of c c for an angle-of-attack = 10o with and without nacelles. Similar results are shown for the case of a four-engined model in gures 5-61 and 5-62. The numerical examples shown above on the example case of a transport aircraft conguration show the potential benets of numerical ow simulation tools as the source/doublet singularity panel-method, linearized potential ow applied in our case. We may use these methods in several ways to compute ow characteristics including pressure distributions and lift distributions, to analyze the eect of aircraft conguration changes including the eect of components as nacelles and, by computing results for a range of angle of attach values, computing the location of the aerodynamic center xac and the Cmac , and their variations for dierent congurations xac and Cmac . This includes the computation of the corresponding important eects on the downwash characteristics behind the wing, as we will see below.

5-2

Equilibrium in steady, straight, symmetric Flight

The present section discusses steady, symmetric, straight ight conditions. The stability of these conditions is discussed in subsequent chapters.

5-2 Equilibrium in steady, straight, symmetric Flight

221

Figure 5-58: Particle lines in an XB OZB -plane (top) and non-dimensional sectional pressure distribution Cp (bottom) for the left wing of a generic transport aircraft model, no nacelles, = 10o , computed with a source/doublet singularity panel-method, linearized potential ow
Flight Dynamics

222

Analysis of Steady Symmetric Flight

Figure 5-59: Particle lines in an XB OZB -plane (top) and non-dimensional sectional pressure distribution Cp (bottom) for the left wing of a twin-engined transport aircraft , = 10o , computed with a source/doublet singularity panel-method, linearized potential ow

5-2 Equilibrium in steady, straight, symmetric Flight

223

3 No nacelles 1 nacelle per wing

2.5

c c

1.5

0.5

0 0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

b y/( 2 )

Figure 5-60: Spanwise lift distribution c c at an angle-of-attack = 10o twin-engined transport aircraft conguration , source/doublet singularity panel-method, linearized potential ow

Flight Dynamics

224

Analysis of Steady Symmetric Flight

Figure 5-61: Particle lines in an XB OZB -plane and non-dimensional sectional pressure distribution coefcients Cp for the left wing of a four-engined transport aircraft , = 10o , source/doublet singularity panel-method, linearized potential ow

5-2 Equilibrium in steady, straight, symmetric Flight

225

3 No nacelles 2 nacelles per wing

2.5

c c

1.5

0.5

0 0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

b y/( 2 )

Figure 5-62: Four-engined transport aircraft with spanwise lift distribution c c at an angle-of-attack = 10o , source/doublet singularity panel-method, linearized potential ow

Flight Dynamics

226

Analysis of Steady Symmetric Flight

For an aircraft it should be possible to obtain equilibrium of the (longitudinal) aerodynamic moment at any angle of attack and center of gravity position at which equilibrium of the forces is possible, i.e. at all feasible ight conditions, see reference [123]. To achieve this a horizontal tailplane with an elevator is used. The horizontal tailplane serves a dual purpose of generation a longitudinal moment about the center of gravity such that the resulting aerodynamic moment is zero and of changing the aerodynamic moment characteristics in a stable sense, as we will discuss in more detail later. The following sections discuss the equilibrium of the forces and moments in steady, symmetric, straight ight conditions. In particular the contribution of the horizontal tailplane and the elevator to the longitudinal equilibrium of moments is considered. The stability of these conditions of equilibrium is the subject of next chapters.

5-2-1

Conditions for equilibrium

Since only steady ight is considered here, forces and moments due to mass and rotational inertia of the aircraft are omitted. The aircraft weight and the aerodynamic forces and moments are in equilibrium so,

W +R = 0 M = 0 The forces and the moment are resolved in components along the aircraft body axes. In this way the conditions for equilibrium may be written as,

Wx + X = 0 Wy + Y = 0 Wz + Z = 0 L = 0 M N = 0 = 0

In the following, it is always assumed that the aircraft is symmetric and that in the considered ight conditions the ow around the aircraft has a plane of symmetry coinciding with that of the aircraft. In these symmetric ight conditions the resultant aerodynamic force R and the weight vector W lie in the plane of symmetry of the aircraft, so,

5-2 Equilibrium in steady, straight, symmetric Flight R XB W sin Z

227

horizontal c.g. W cos

W ZB
Figure 5-63: The equilibrium in steady, symmetric ight

Wy L

= =

Y N

= 0 = 0

The three conditions for equilibrium of these symmetric ight conditions are , see gure 5-63,

W sin + X = 0 W cos + Z = 0 M denoting the pitch angle. These equations may also be written in terms of normal and tangential force N and T , T = W sin N = W cos M =0 (5-41) (5-42) (5-43) = 0

with T = X and N = Z. For simplicity , the aircraft body axes are chosen parallel to the aircraft reference axes, origin of the body axes remaining in the aircraft center of gravity. In the following, these conditions are further analyzed.

Flight Dynamics

228

Analysis of Steady Symmetric Flight

The resultant aerodynamic force R is built up from the contributions of the dierent parts of the aircraft. These parts also provide the elements of the resultant aerodynamic moment M . Three parts of the aircraft may be distinguished. Wing with fuselage and nacelles The contributions of these parts of the aircraft usually are taken as one entity, the index w refering to the combination of the wing with fuselage and nacelles. The model of the forces acting on this part of the aircraft is given by the components Nw and Tw of the total aerodynamic force Rw , acting in the aerodynamic center (ac) of the wing with fuselage and nacelles (acw ), and the aerodynamic moment Macw acting about the acw . Horizontal tailplane Just as for the wing with fuselage and nacelles, the model of the forces acting on the horizontal tailplane is given by the components Nh and Th , acting in the ac of the tailplane (ach ), and the moment Mach . The propulsion unit The contribution of the propeller, or the jet engine, consists of the thrust Tp along the propeller or engine-axis and the normal force Np acting in the plane of the propeller or in the plane of the engine inlet depending on the local angle of attack . Figure 5-64 depicts the contributions of the various parts of the aircraft. The symmetric conditions of equilibrium, equations (5-41), (5-42) and (5-43) can now be written as follows, Forces along the XB -axis T = +Tw + Th Tp cos ip + Np sin ip = W sin Forces along the ZB -axis N = +Nw + Nh + Np cos ip + Tp sin ip = +W cos Moments about the YB -axis M = +Macw + Nw (xc.g. xw ) Tw (zc.g. zw ) + +Mach + Nh (xc.g. xh ) Th (zc.g. zh ) + + (Np cos ip + Tp sin ip ) (xc.g. xp ) + + (Tp cos ip Np sin ip ) (zc.g. zp ) = 0 (5-46) (5-45) (5-44)

5-2 Equilibrium in steady, straight, symmetric Flight

229

Zr

in plane of propeller ip Np

Nw

Tp

ach acw iw zp Mac zw W xp xw xc.g. xh Xr zc.g. zh c.g. Tw ih Nh Th

Mach

Figure 5-64: The forces and moments in steady, straight, symmetric ight

Flight Dynamics

230

Analysis of Steady Symmetric Flight

Before bringing the expressions (5-44), (5-45) and (5-46) in a non-dimensional form, some new non-dimensional coecients have to be dened. The non-dimensional coecients of normal force and tangential force on the tailplan, and the moment about the ac of the tailplane are now dened by, Nh Th
2 = CNh 1 Vh Sh 2 2 = CTh 1 Vh Sh 2 1 2 2 Vh Sh ch

(5-47)

Mach = Cmach

The non-dimensional coecients CNh , CTh and Cmach are thus referred to the surface area Sh and the mac ch of the horizontal tailplane (see gure 5-65a) and the average local 2 dynamic pressure 1 Vh at the horizontal tailplane. In general, this dynamic pressure is dif2 ferent from that of the undisturbed ow (at innity). This is further discussed in section 5-2-4. The thrust of the propeller and the normal force in the plane of the propeller may be written as, Tp = Tc V 2 D2 Np = CNp 1 2 V Sp 2 (5-48) (5-49)

where D is the propeller diameter and Sp = D2 is the area of the propeller disc. The 4 equations of equilibrium are made non-dimensional in the usual way by dividing equations (5-44) and (5-45) by 1 V 2 S and equation (5-46) by 1 V 2 S. Using , c 2 2 Tp cos ip + Np sin ip Tp the result is, Vh V
2

CT = +CTw + CTh

Sh 2D2 W Tc = 1 2 sin S S 2 V S

(5-50)

CN = +CNw + CNh

Vh V

Sp Sh 2D2 W + Tc sin ip + CNp = 1 2 cos S S S 2 V S Vh V

2

(5-51)

Cm = +Cmacw + CNw +CNh

xc.g. xw zc.g. zw CTw + Cmach c c

2

Sh ch + S c

Vh V

Sh xc.g. xh CTh S c

Vh V

Sh zc.g. zh + S c

5-2 Equilibrium in steady, straight, symmetric Flight

231

Sh

Se

ce

St

ch

hinge line of elevator

hinge line of tab (A) Geometry parameters of the horizontal tailplane, elevator and trim tab

h Vh e te

(B) Positive deections of h , e and te

Figure 5-65: Geometry parameters of the horizontal tailplane, elevator and trim tab, and their positive deections

Flight Dynamics

232

Analysis of Steady Symmetric Flight

+ Tc

Sp 2D2 sin ip + CNp S S

xc.g. xp 2D2 zc.g. zp + Tc =0 c S c

(5-52)

These equilibrium equations may be used for dierent purposes such as to explore the set of feasible steady state ight conditions. In the following, however, the purpose is an analysis of stability and ying qualities parameters. In that case it is permissable and even desirable to simplify the equations of equilibrium in such a way that only the most important contributions remain, resulting in the following simplications , 1. Propulsion eects, i.e. the terms with Tc and CNp are neglected. 2. The contribution of CTh to the force in XB -direction in equation (5-50) and to the moment in equation (5-52) is neglected. This simplication is nearly always permissable. 3. The contribution of CTw to the moment in equation (5-52) is omitted. For large angles of attack, or in cases when the center of gravity is situated far above or below the acw , this simplication may not always be permissable. 4. The contribution of Cmach is relatively small, even zero in case of symmetric airfoils, and so Cmach in equation (5-52) is neglected too. The index w of Cmacw may therefore be omitted as well without raising confusion. The resulting equations of equilibrium now read , CT = CTw = 1 CN = CNw + CNh Vh V W sin 2 2 V S
2

(5-53)

Sh W = 1 2 cos S 2 V S Vh V
2

(5-54)

Cm = Cmac

xc.g. xw + CNw + CNh c

Sh xc.g. xh =0 S c

(5-55)

The assumptions made in the model of the equilibrium lead to a very simple model, see gure 5-66. The distance between the ac of the horizontal tailplane and the ac of the wing with fuselage and nacelles, (xh xw ), is called the tail length lh . For conventional aircraft, lh = xh xw xh xcg With (5-56) equation (5-55) may be written as, Cm = Cmac + CNw xc.g. xw CNh c Vh V
2

(5-56)

Sh lh =0 S c

(5-57)

5-2 Equilibrium in steady, straight, symmetric Flight or, Cm = Cmw + Cmh in which
Sh lh S c

233

is called the tailplane volume.

The tailplane serves to compensate for Cmw through an appropriate deection of the elevator or by adjusting the angle of incidence of the entire horizontal tailplane. In this way the contribution Cmh from the horizontal tailplane to the moment as generated by CNh , is equal and opposite to the contribution Cmw of the wing with its fuselage and nacelles. In the following sections the characteristics of the horizontal tailplane and the eects of the elevator are discussed in more detail. Also, expressions will be derived for the elevator angle and the pilots control force necessary to maintain equilibrium for given ight conditions. The equations of equilibrium (5-53), (5-54) and (5-55) apply to the conventional aircraft congurations where the horizontal tailplane is placed behind the wing. In the given form the expressions are valid also for aircraft where the horizontal tailplane is located in front of the wing. In that case lh in equation (5-57) is negative , see also equation (5-56). For tailless aircraft the contribution of the horizontal tailplane obviously does not exist. In that case equation (5-57) reduces to, Cm = Cmac + CNw see also gure 5-21. The elevator of tailless aircraft is situated at the trailing edge of the wing. Deection of the elevator of such an aircraft causes a change of Cmac . The elevator deection also has an inuence on CN which may be non-negligible in many cases. For tailless aircraft, equilibrium for some value of CN is possible if Cmac in equation (5-58) is made equal and opposite in sign to the moment of CN about the center of gravity. xc.g. xw =0 c (5-58)

5-2-2

Normal force on the horizontal tailplane

The horizontal tailplane usually consists of a stabilizer and a control surface. In many cases a small auxiliary control surface is added to the control surface. This may be a trim tab, a balance tab or a servo tab. The non-dimensional pressure distribution over the horizontal tailplane is completely determined by three deection angles, 1. The angle of attack of the horizontal tailplane, h 2. The control deection angle, e (index e stands for elevator)
Flight Dynamics

234

Analysis of Steady Symmetric Flight

CNh Zr CNw ach

Vh V

Sh S

c.g. Cmac acw CTw

W cos 1 V 2 S 2

W sin 1 V 2 S 2

xw xc.g. lh xh

Xr

Figure 5-66: Simplied version of the equilibrium of forces and moments

5-2 Equilibrium in steady, straight, symmetric Flight

235

Normal force derivatives

Derivative

CNh h

CNh e

CNh te

Shorthand notation

CNh

CNh

CNh

Table 5-1: Shorthand notation for the normal force derivatives on a tailplane

3. The tab angle, te (index t stands for tab) Figure 5-65b shows the positive sense of the three angles h , e and te . Figure 5-67 shows the changes in the chordwise pressure distribution caused by changes in the local angle of attack, the control surface deection and the tab angle deection. The non-dimensional normal force CNh as dened in equation (5-47) is determined by the local angle of attack and two deection angles , CNh = CNh (h , e , te ) In gure 5-68 results are given of measurements of CNh as a function of h , e and te . It follows that in selected small intervals of h , e and te , CNh may be assumed to vary approximately linearly with these parameters. In the range of those small angle excursions a reasonable approximation of CNh is obtained by, CNh = CNh0 + CNh CNh CNh h + e + te h e te (5-59)

A convenient shorthand notations is used for the aerodynamic partial derivatives introduced in equation (5-59), see table 5-1. The expression (5-59) can be simplied a little. First we note that the inuence of te on CNh is usually small, see for example gure 5-68 and may be neglected. Furthermore, many tailplanes have a symmetric or close to symmetric airfoils, so CNh0 = 0 . Using the shorthand notation (which corresponds to the American notation) this results in, CNh = CNh h + CNh e which serves as an approximation of the much more complex nonlinear relation,
Flight Dynamics

(5-60)

236

Analysis of Steady Symmetric Flight

p q

e = te = 0 h2 h1 x/ c

h1 h2
p q

h = te = 0

e2 e1 x/ c e1

e2

p q

h = e = 0

te2 te1 x/ c te1

te2

Figure 5-67: Pressure distributions over the tailplane chord due to h , e and te

5-2 Equilibrium in steady, straight, symmetric Flight CNh e [o ]

0.8 -5 20 15 10 5 0 -10 -15 -20 0.4 0.4 -15 -25 -16 -8 0 8 0.8 15 0

237 CNh te [o ]

te =

0o

e = 0o

-16

-8

-0.4

-0.4

-0.8

-0.8

Figure 5-68: The normal force coecient CNh as a function of h , e and te for the tailplane of the Fokker F-27 (Wind tunnel measurements from reference [17])

CNh = CNh (h , e , te ) as may be obtained from windtunnel measurements or ight test. Both positive changes in the angle of attack h and the control deection e cause positive changes in CNh which means, CNh > 0 and CNh > 0

With references [71, 157, 4, 101, 61, 21], based on empirical data, it is possible to estimate CNh and CNh . The normal force gradient CNh can be obtained in principle using one of the available wing theories. The presence of the fuselage and the existence of an open gap between the stabilizer and the control surface must be taken into account, reducing the value of CNh . From measurements on two- and three-dimensional models it was shown that CNh depends almost linearly on CNh . This is reected in gure 5-69 showing
ce ch CNh CNh

to depend on the

ratio of the control surface chord to the tailplane chord, and the trailing edge angle of the airfoil. This result holds for aerodynamically unbalanced control surfaces extending over the entire span of the horizontal tailplane with a closed gap between the horizontal tailplane (stabilizer) and the control surface. However, CNh is aected by several other factors such as cut-outs in the control surface at the fuselage and the vertical tailplane, a gap between
Flight Dynamics

238
CNh CNh

Analysis of Steady Symmetric Flight 0o

0.7

10o 20o

0.6

0.5

0.4

0.3

0.2

Closed gap Full span elevator No Aerodynamic balance

0.1

0.1

0.2

0.3

0.4

0.5

ce ch

Figure 5-69: The quotient [21])

CNh CNh

as a function of

ce ch

for various trailing edge angles (from reference

the stabilizer and the aerodynamic shape of the nose of control surfaces (used to balance or compensate hinge moments), see references [101, 61].

5-2-3

Hinge moment of the elevator

The aerodynamic hinge moment about the elevator hinge line, He , is determined by the pressure distribution over the elevator. This hinge moment is expressed by a non-dimensional coecient, Che , Che = He 1 2 2 Vh Se ce (5-61)

The hinge moment coecient is thus referred to the average local dynamic pressure at the horizontal tailplane and to the surface Se and the mac ce of the part of the control surface

5-2 Equilibrium in steady, straight, symmetric Flight Che te =

0.08

239 Che e [o ]
-12 -9 -6

0o

h = 0o

0.04

e
-5

[o ]
-10

0.04 -15 -3

0 0 8

-16

-8

-16

-8

e [o ]

-0.04 -0.04 5

6 -0.08 -0.08 9 12

te [o ]

Figure 5-70: The hinge moment coecient Che as a function of h , e and te for a tailplane of the Fokker F-27 (wind tunnel measurements from reference [17])

behind the hinge line, see gure 5-65a. The hinge moment is dened positive in the same direction as the control surface deection. Similar to CNh , also Che depends on the three angles h , e and te , see the pressure distributions in gure 5-67, Che = Che (h , e , te ) Measurements of Che as a function of h , e and te are shown in gure 5-70. We now note that te has an important eect on Che . Again for a limited interval of h , e and te variations a linear approximation for Che may be written as, Che = Ch0 + Che Che Che h + e + t h e te e (5-62)

If a symmetric airfoil is used for the tailplane, Ch0 = 0. Using in addition the shorthand notation of table 5-2, Che can be written as, Che = Ch h + Ch e + Cht te (5-63)

Note that in equation (5-63), Cht is the derivative of the moment about the elevator hinge axis with respect to the tab deection. Table 5-2 also denes the derivatives Cht , Cht and Cht which refer to the hinge moment about the tab hinge axis. These latter derivatives are t made non-dimensional using St and ct of the tab.
Flight Dynamics

240

Analysis of Steady Symmetric Flight

Hinge moment derivatives

Derivative

Che h

Che e

Che te

Chte h

Chte e

Chte te

American notation

Ch

Ch

Cht

Cht

Cht

Cht

Table 5-2: Abbreviated notation for the hinge moment derivatives

Computing the hinge moment derivatives of a control surface of given shape and dimensions is not always possible with sucient accuracy. The shape of the tailplane, the airfoil and minute details of the form of the gap between stabilizer and control surface have an eect on the derivatives. The main problem, however, is the relatively large inuence of viscous eects which are dicult to account for quantitatively. Best is to obtain data on Che from windtunnel measurements, using as large a model of the tailplane as possible. Depending on the location of the hinge line of the control surface the change in hinge moment, with increasing angle of attack and control surface deection, will be negative in most cases, see also the pressure distributions given in gure 5-67. In that case both derivatives are negative,

Ch < 0

and

Ch < 0

If the hinge line would be moved backwards, however, in most cases it is the derivative Ch which rst changes sign to positive, see again 5-67, and would result in unacceptable ying qualities as discussed later. It follows from equation (5-61) that the hinge moment varies with the square of the airspeed and the cube of the dimensions of the aircraft. Fast and large aircraft would soon be confronted with very large hinge moments which would be dicult for pilots to handle by conventional mechanical control systems . This led early mechanical ight control system designers to carefully aerodynamically balance control surfaces such that Ch and Ch stay small, while keeping Ch negative. As long as both Ch and Ch are negative, the control surface is called aerodynamically underbalanced. Shifting the hinge-line rearward will result in the control surface to become overbalanced with respect to the angle of attack, or with respect to the control deection, or both. Ch and Ch will then be positive see gure 5-71. One of the means to change Ch and Ch in the positive sense, is to employ a rearward shift of the hinge axis, or a forward extension of the nose of the control surface in front of the hinge line. As long as both Ch and Ch are negative, the control surface is called aerodynamically underbalanced.

5-2 Equilibrium in steady, straight, symmetric Flight

241

5-2-4

Flow direction and dynamic pressure at the horizontal tailplane

A horizontal tailplane located behind the wing experiences the eld of ow disturbed by the wing, the fuselage, the nacelles and propellers. Mainly due to the downwash behind the wing, the average angle of attack h of the horizontal tailplane is decreased. If the horizontal tailplane lies in the wake produced by the wing, or in the boundary layer of the fuselage, the local average dynamic pressure at the horizontal tailplane is reduced. This is expressed in a value of
Vh V 2

< 1 in equation (5-57).

Both eects cause a change in the contribution of the horizontal tailplane to the moment about the aircrafts lateral axis. The slipstream of propeller driven aircraft and the exhaust stream of jet-propelled aircraft may also cause changes in h and Vh . V If the average downwash angle at the location of the horizontal tailplane is designated , and the tailplane incidence setting relative to the Xr -axis is ih , the average angle of attack of the horizontal tailplane follows from (see gure 5-72), h = + ih (5-64)
2

The downwash angle is approximately proportional to CL , in the range where CL varies linearly with , = ( 0 ) where 0 = CL =0 and
d d

d d

is approximately constant. Thus h can be written as, h = ( 0 ) 1 d d + (0 + ih ) (5-65)

From (5-65) the derivative of h with respect to is, dh = d 1 d d (5-66)

We will see below that this derivative is always smaller than one, resulting in a reduction of the positive contribution of the horizontal tailplane to the static stability of conventional aircraft with rear mounted tail planes. The downwash angle and the factor Vh depend in the rst place on the shape of the wing V and on the location of the horizontal tailplane relative to the wing. The downwash behind the wing may be thought of as to be induced by a system of nested U shaped vortices. Part of each vortex is bound to the wing, inducing an upwash in front of the wing and a downwash behind the wing. The free, remaining parts of each vortex (the two antisymmetrical tails) leave the wing at the trailing edge of the wing in a vortex sheet of free vortices, reinforcing the downwash and reduce the upwash, see gure 5-73. The vertical velocities induced by these vortices result in the vortex sheet itself being pushed downwards, assuming the form
Flight Dynamics 2

242

Analysis of Steady Symmetric Flight

p q

due to h

x/ c

Che > 0

(A)

Che h

> 0 Ch > 0

p q

due to e

Che > 0

x/ c

e (B)
Che e

> 0 Ch > 0

Figure 5-71: Overbalance with respect to angle of attack and elevator angle.

5-2 Equilibrium in steady, straight, symmetric Flight V Vh Xr Xr h = + ih

Figure 5-72: The angle of attack h of the horizontal tailplane, h = + ih

243

ih

as depicted in gure 5-74. While the inclination of the vortex sheet conforms to the upper and lower surfaces of the wing at the trailing edge (Joukowski condition), downstream the inclination decreases and the vortex sheet is bent towards the direction of the undisturbed ow. Upwash, downwash and the vertical displacement of the vortex sheet are approximately proportional to CL for a given wing geometry. The boundary layers over the wing result in a wake of reduced dynamic pressure which is assumed to coincide with the the vortex sheet. The vertical displacement of the wing wake and vortex sheet increases with angle of attack but at a smaller rate than that of the horizontal tailplane , see gure 5-75. It is good practice to keep the horizontal tailplane out of the wing wake at all ight conditions. We conclude that this can indeed be assured by mounting it either high, keeping it above, or mounting it low, keeping it below the wake of the wing. The tendency of a vortex plane is to roll-up, changing its shape until it is transformed into two tip vortices, see gure 5-76. This process depends on the lift distribution over the wing and it is intensied as the lift coecients are larger and the wing aspect ratio smaller, see references [151, 171]. For wings with a large aspect ratio and a small sweep angle the roll-up of the vortex plane is of little importance in the range of practical angles of attack. For these wings we may compute the downwash at the horizontal tailplane by assuming a at vortex plane, only slightly curled at the edges. For small aspect ratio wings , or wings with a large sweep angle , or both, however, this would be too rough an approximation. One would have to take account in these cases of the intricate deformed shape of the vortex sheet. The same holds true in cases where at some values of the angle of attack local ow separation occurs, often resulting in the functional form of the wing to be dierent from what one would expect by considering just its geometry. The main eect of ow separation is to reduce downwash rather than to aect the shape (rolling-up) of the vortex sheet, as discussed below. According to reference [117], the inuence of the deformation and the rolling-up of the vortex sheet on the downwash close to the wing plane of symmetry may as well be neglected at low values of CL . d Figure 5-77 gives the theoretical value for d in the core of the vortex plane, as a function of aspect ratio and distance behind the wing, for wings having an elliptical lift distribution and
Flight Dynamics

244

Analysis of Steady Symmetric Flight

x x

due to free vortex

total induced vertical velocity

due to lifting vortex

Figure 5-73: The contribution of the lifting and free vortices to the induced vertical velocities in front of and behind a wing

depression of the vortex plane V V Xr

vortex plane
Figure 5-74: The position of the vortex plane behind the wing

5-2 Equilibrium in steady, straight, symmetric Flight position of the vortex plane 1 2 xw V zwake zh

245

Xr at 1

xh Xr at 2

Figure 5-75: The vertical displacement of the wake and the tailplane relative to the ow eld due to an inxh xh

crease in angle of attack ; zh =

xw

dx = (xh xw ), zwake =

(x) dx,
xw

zwake < zh as (x) is smaller than . With increasing angle of attack the horizontal stabilizer moves downwards through the wake

attached (non separated) ow. The dominant inuence of the wing aspect ratio is apparent.
d A rough estimate of d at suciently large distances behind the wing is possible for straight wings for not too small aspect ratios. The classical formula to be used is,

d CL 2 A 4 =2 2 = d A A A+2 A+2

(5-67)

d This expression results in values for d for A > 5 which are in very good agreement with the more exact values given in gure 5-77 for an elliptical wing.

Figure 5-78 shows the variation in spanwise direction of the lift distribution c c for wings of 2b ct dierent taper ratios. Note that for wings with small values of = cr the average downwash angle near the plane of symmetry is larger than for wings with larger values of at the same values of CL . This is caused by the higher lift loading near the wing center of former wings. d As a consequence d at the horizontal tailplane increases with decreasing . Sweepback of an d untwisted wing causes a lower loading near the wing center. As a consequence, d decreases with increasing sweep angle , see also gure 5-79. The downwash angle decreases with increasing distances above and below the vortex plane. Figure 5-80 shows the results of numerical simulations using a source and doublet singularity method (also known as panel methods). Measurements of the downwash angle in the plane of symmetry behind a tapered wing can be found in reference [146]. From gure 5-80 it is clear that the downwash in a certain distance above the wake is larger than in a point at the same distance below the wake. This is due to the fact that as a result of the depression of the wake and of the deformation of the vortex plane the free and the lifting vortices contribute more to the downwash above the wake than below the wake.

Flight Dynamics

246

Analysis of Steady Symmetric Flight

Figure 5-76: Numerical simulation of the deformation and wake roll-up of the vortex sheet behind a wing

5-2 Equilibrium in steady, straight, symmetric Flight

d d
1.2

247

1.0

0.8

x b/2
0.6

= 0.5

1.0
0.4

1.5

2.0 and 2.5

0.2

0 0 2 4 6 8 10 12 14

A V
b 2

X
d Figure 5-77: Computed values of d behind a wing of elliptical lift distribution, as a function of aspect ratio and distance behind the wing

Flight Dynamics

Analysis of Steady Symmetric Flight

3 2 1 Z [m] Z [m] 0 1 2 3 2 0 2 X [m] 0 2 Y [m] 2

3 2 1 0 1 2 3 2 0 2 X [m] 0 2 Y [m] 2 Z [m]

3 2 1 0 1 2 3 2 0 2 X [m] 0 2 Y [m] 2

0.15

0.15

0.15

c c/(2b)

c c/(2b)

0.05

0.05

c c/(2b)
0 0.5 1

0.1

0.1

0.1

0.05

0.5

248

y b/2

y b/2

0.5

y b/2

c Figure 5-78: Numerical simulation of c2b for wings with varying taper ratio (CL = 1.0, = 0o , A = 6), computed with a source/doublet singularity method, also see reference [143]

3 2 1 Z [m] Z [m] 0 1 2 3 2 0 2 X [m] 0 2 Y [m] 2

3 2 1 0 1 2 3 2 0 2 X [m] 0 2 Y [m] 2 Z [m]

3 2 1 0 1 2 3 2 0 2 X [m] 0 2 Y [m] 2

5-2 Equilibrium in steady, straight, symmetric Flight

0.15

0.15

0.15

c c/(2b)

c c/(2b)

0.05

0.05

c c/(2b)

0.1

0.1

0.1

0.05

0.5

y b/2

0.5

y b/2

0.5

y b/2

c Figure 5-79: Numerical simulation of c2b for wings with varying sweep angle (CL = 1.0, = 1, A = 6), computed with a source/doublet singularity method, also see reference [143]

Flight Dynamics

249

250

Analysis of Steady Symmetric Flight

Z [m]
3 2 1 0 1 2 1 0 1 2 3

= 4o

= 2o

= 8o

= 6o

X [m]

Figure 5-80: Numerical simulation of downwash angles in the plane of symmetry behind a wing ( = 13o , NACA 23014 airfoil, A = 6, = 0.5 and CL = 1.1615, , computed with a source/doublet singularity method, see also reference [146])

2 3 4 5 6

[o ]

10

11

5-2 Equilibrium in steady, straight, symmetric Flight

251

The eld of ow behind wings with separated ow is signicantly dierent from the case of d attached ow. Flow separation will cause appreciable changes in d with increasing angle of attack. A detailed discussion of these changes related to ow separation can be found in reference [117]. For straight wings with just little taper (so a large ), ow separation starts d near the wing root. The reduction in lift over the center wing results in a decrease of d . In addition a broad and deep wake is formed, causing a strong inow into the wake. Flow separation near the wing tips, occurring with highly tapered and swept back wings, causes an increase in downwash at the tailplane due to a concentration of lift force to the center of the wing caused by ow separation. Swept wings and delta wings with thin airfoils often exhibit ow separation near the wing leading edge, already at small angles of attack. This leads to characteristic conical vortices above the wing. Such separation vortices may considerably increase the downwash. The camber and nose shape of airfoils, wing twist, boundary layer fences, nose aps and slats, all serve to inuence the ow separation characteristics. It will be clear that these wing geometry parameters will also aect the downwash characteristics at the horizontal tailplane. The fuselage causes a change in the ow behind the wing. This is due to the ow around the fuselage itself and due to the change in the wing lift distribution caused by the fuselage. The resultant inuence on the average downwash at the horizontal tailplane is relatively small for slender wings. For wings of small aspect ratio the inuence of the fuselage can be considerable and is strongly dependent of the position of the horizontal tailplane relative to the fuselage. If the tailplane is mounted on the fuselage a decrease in the average value of
Vh V 2

occurs since part of the tailplane is in the boundary layer of the fuselage. In such
Vh V 2

cases a reduction in

of some 5 to 10% is often taken into account.

Example: aircraft with fuselage-mounted nacelles The aerodynamic eects of fuselage-mounted nacelles are shown next on two aircraft models depicted in gures 5-81. These models represent a Cessna Ce550 Citation II with and without nacelles, respectively the top and bottom gure in gures 5-81.
d In gures 5-82 the numerical simulation of d (for X = 5.0 m, Z = 1.5 m) is given along the wing span (left wing) of the aircraft models presented in gures 5-81. Note that the top gure of gure 5-82 is a 2-dimensional representation (aft view) of the bottom gure. From gures d 5-82 it follows that in the vicinity of the horizontal stabilizer the parameter d is increased due to the presence of nacelles. This eect cannot be attributed to the induced velocities of free and bounded vortices but rather to the local airow conforming to the external shape of the nacelles.

Flap deection causes a concentration of lift at the center parts of the wing. This leads to d an increased downwash and d .

Flight Dynamics

252

Analysis of Steady Symmetric Flight

Figure 5-81: Cessna Ce550 Citation II conguration with and without nacelles

5-2 Equilibrium in steady, straight, symmetric Flight

253

4 No nacelles, X = 5.0 m, Z = 1.5 m Including nacelles, X = 5.0 m, Z = 1.5 m 3

Z [m]

LEFT WING
0

d d ,

3 10

Y [m]

Z [m]
d d ,

2 0 2 4 6

No nacelles, X = 5.0 m, Z = 1.5 m Including nacelles, X = 5.0 m, Z = 1.5 m 5 6 7

Y [m]

X [m]

d Figure 5-82: The eect of nacelles on d (at X = 5.0 m, Z = 1.5 m) aft of the wing of a Cessna Ce550 Citation II, computed with a source/doublet singularity panel-method, linearized potential ow

Flight Dynamics

254

Analysis of Steady Symmetric Flight

Real aircraft

Ground surface (mirror plane)

Reected aircraft

Figure 5-83: Calculation of ground eect

In take-o and landing the eld of ow around and in particular behind the wing is strongly inuenced by the proximity of the ground. Ground eect can be modeled by introducing additional mirror lifting vortices and trailing vortices, see gure 5-83. The mirror vortices cause a decrease in the induced angles of attack and a corresponding decrease in the eective angle of attack, proportional to the strength of the mirror vortices and thus to CL . The result of ground eect is an increase in CNw and CNh . As CNh is proportional to CNh , CNh increases as well. The downwash angle behind the wing decreases due to the mirror vortices. This eect can also be understood directly, as the presence of the ground plane inhibits the downward ow behind the wing. Figure 5-84 gives an example of the change in at the horizontal tailplane d caused by the ground eect. This gure also demonstrates that the reduction in d due to ground eect can be quite important.

An additional consequence of the reduced downwash angles behind the wing is a smaller downward displacement of the wake, as compared with the situation without ground eect, see gure 5-84b . When choosing the vertical position of the horizontal tailplane in the

5-2 Equilibrium in steady, straight, symmetric Flight course of the design process, this eect has to be taken into account.

255

For slender, straight wings without and with deected aps references [157, 145] shows how to compute the average downwash angle and the average dynamic pressure at the horizontal tailplane, see also reference [138]. The downwash behind swept wings can be determined with references [73, 52, 48, 50, 158]. An extensive discussion of ground eect is given in reference [59]. The ground eect on the downwash and the wake can be calculated with reference [92]. For wings with a small aspect ratio or large sweep angle or both, the fuselage has a large inuence on the ow at the horizontal tailplane and in case of ow separation , the ow eld at the horizontal tailplane should be derived from measurements in the wind tunnel or by CFD simulations. An extensive summary and analysis of experimental data is given in reference [117], where also a method is described to estimate the characteristics of the horizontal tailplane.

5-2-5

Eect of airspeed and center of gravity on tail load

As discussed in section 5-2-1, a normal force acts on the horizontal tailplane to produce equilibrium of the longitudinal moment in an equilibrium situation. This normal force is called the tail load. The magnitude of the tail load Nh follows from equation (5-57) multiplied by 1 V 2 S, c 2 M = 0 = Cmac or, Nh = 1 lh Cmac 1 2 V S + Nw (xc.g. xw ) c 2 (5-69) 1 2 V S + Nw (xc.g. xw ) Nh lh c 2 (5-68)

At small values of the angle of pitch , and if in addition Nh << Nw , it follows with equation (5-54), Nw W The result becomes, Nh 1 lh Cmac 1 2 V S + W (xc.g. xw ) c 2 (5-70)

This also follows directly from gure 5-85, where the model of equilibrium of the forces and moment has been given in a strongly simplied form. It can be seen from equation (5-70) that the tail load Nh is needed, 1. to compensate for the moment caused by the coecient Cmac ; this moment depends on the airspeed but not on center of gravity position; in addition it varies with ap deection f
Flight Dynamics

256 [o ]
6

Analysis of Steady Symmetric Flight

f = 0o
4

without ground eect

2

0 0 2 4 6 8

with ground eect z0.25 zground = 0.19 b/2 c

10 12

[o ]

(A) The downwash angle as a function of

-0.4

0 0 2 4 6 8 10 12

0.4

with ground eect z0.25 zground = 0.19 b/2 c f = 0o without ground eect

0.8

1.2

h b/2

[o ]

tailplane h wing center line of wake (B) The distance from the wake center line to the horizontal tailplane
Figure 5-84: The ground eect on the downwash angle and the location of the wake relative to the horizontal tailplane of a Siebel 204 D-1 aircraft

5-2 Equilibrium in steady, straight, symmetric Flight W cos + Nh W

257

c Mac = Cmac 1 V 2 S 2

Nh

acw

c.g. ach

lh

xw xc.g. W cos
Figure 5-85: Simplied picture of the equilibrium of moments

2. to balance the moment of Nw ( W ) about the center of gravity; this moment is independent of airspeed, but varies with the center of gravity position From equation (5-70) it can be seen that Nh varies quadratically with V and increases in the upward sense with a rearward shift of the center of gravity. Figure 5-86 gives a sketch of the variations of the tail load with airspeed for various c.g. positions and signs of Cmac . The point of intersection of the parabola with the line V = 0, in actual ight the airspeed can, of course, not drop below Vmin , follows from equation (5-70), Nh = W xc.g. xw lh

Figure 5-86 illustrates that due to the usually negative sign of Cmac of the wing, fuselage and nacelles, the tail load reaches its largest positive value at the rearmost c.g. position and at minimum airspeed. The largest negative, downward, tail load occurs for a given negative Cmac at the most forward c.g. position and at a maximum airspeed. Figure 5-87 presents tail loads measured in ight on a DeHavilland Mosquito. The inuence of airspeed on Nh as well as of c.g. position agrees in principle with equation (5-70). From
Flight Dynamics

258 Nh

Analysis of Steady Symmetric Flight

Cmac > 0 xc.g. > xw c.g. behind ac Cmac = 0

xc.g. xw lh

Cmac < 0 Vmin Cmac > 0 V

c.g. forward of ac xc.g. < xw Cmac = 0

Cmac < 0
Figure 5-86: The variation of the tail toad with airspeed at dierent c.g. positions and values of Cmac

this gure it can be derived that for the range of airspeeds 350 < Ve < 500 km/h : xw = 0.08 c and Cmac = 0.021.

5-2-6

Elevator deection required for moment equilibrium

The way in which the elevator is used to produce moment equilibrium follows from expression (5-57), Cm = Cmac xc.g. xw + CNw CNh c Vh V
2

Sh lh =0 S c

CNw can be written in the linear range as, CNw = CNw ( 0 ) In section 5-2-2, CNh has been expressed by, see equation (5-60), CNh = CNh h + CNh e while in section 5-2-4 the expression for the angle of attack of the horizontal tailplane h was found to be, see equation (5-65), h = ( 0 ) 1 d d + (0 + ih ) (5-71)

5-2 Equilibrium in steady, straight, symmetric Flight

259

Nh
700

xc.g.1 = 0.341 c
600

500

c xc.g.2 = 0.292
400

300

200

100

0 100 200 300 400 500 600

Ve [km/h]

-100

-200

Figure 5-87: The tail load as a function of airspeed at two c.g. positions for De Havilland Mosquito II F (from reference [26]), W = 6800 kg, S = 41.8 m2 , lh = 8.0 m and c = 2.81 m

Flight Dynamics

260

Analysis of Steady Symmetric Flight

Substituting equations (5-71), (5-60) and (5-65) in equation (5-57) results in the condition for moment equilibrium, Cm = Cmac + CNw ( 0 ) CNh ( 0 ) 1 d d xc.g. xw c Vh V
2

+ (0 + ih ) + CNh e = 0

Sh lh S c (5-72)

or, in an ordered form, Cm = Cm0 + Cm ( 0 ) + Cme e = 0 with, a constant Cm0 , Cm0 = Cmac CNh (0 + ih ) the static longitudinal stability, stick xed Cm , Cm = CNw xc.g. xw CNh c 1 d d Vh V
2

(5-73)

Vh V

Sh lh S c

(5-74)

Sh lh S c

(5-75)

the elevator eectivity Cme , Cme = CNh Vh V

2

Sh lh S c

(5-76)

For a given aircraft the pilot can obtain equilibrium of the moment at a certain angle of attack, or ( 0 ), and c.g. position, or Cm , see equation (5-75), by deecting the elevator, i.e. by choosing e such that the resultant Cm is equal to zero. If the total moment Cm is written as the sum of the moment at e = 0 and the moment due to the elevator deection, Cm = (Cm )e =0 + Cme e (5-77)

the condition for equilibrium of the total moment, Cm = 0, results in the required elevator deection, e = 1 (Cm )e =0 Cme (5-78)

5-2 Equilibrium in steady, straight, symmetric Flight Since, according to equation (5-73), (Cm )e =0 = Cm0 + Cm ( 0 ) the elevator angle e becomes, e = 1 Cme {Cm0 + Cm ( 0 )}

261

(5-79)

(5-80)

The above expressions are based on the simple model as derived in section 5-2-1. In addition, use has been made of several linear approximations. As a consequence, the above expressions are valid, strictly speaking, only for those angles of attack and elevator angles where CNw and CNh are proportional to , or h and e respectively. For more accurate calculations it will be better to rely on wind tunnel measurements of Cm for a range of values of interest of , ih and e . The above expressions apply to the case of an aircraft with a xed stabilizer and a movable elevator. In many cases, however, the available power Cme e max is insucient to obtain moment equilibrium at all desired c.g. positions over the entire range of airspeeds. Often an adjustable stabilizer with a variable ih is employed, and set such that e = 0. In that case Cm0 in equation (5-74) is of course no longer a constant . Some aircraft types have a movable horizontal tailplane with even no elevator at all, the so-called ying-tail. Expressions can be derived quite similarly for the value of ih required for equilibrium in such cases. We conclude this section on equilibrium and tail load with the following comments on equations (5-72) through (5-76), Cmac , CNw and 0 are xed by the aerodynamic design of the wing, the fuselage and the nacelles.
hl The required tailvolume SSh follows primarily from the requirement that the aircraft c shall be statically stable (i.e. Cm < 0)at the rearmost permissable c.g. position, the derivative CNh following from the the shape and size of the horizontal tailplane, see section 5-2-2.

If the angle of incidence of the horizontal tailplane, ih , is not adjustable, ih is generally set such that the elevator angle is zero, minimizing the parasite tailplane drag .
d Both d and Vh follow from the characteristics of the wing and the location of the V horizontal tailplane relative to the wing and the fuselage, see section 5-2-4. 2

For tail planes with xed ih , (Cm )e =0 depends on angle of attack and c.g. position, and has to be compensated for by the moment generated by the elevator Cme e , see equation (5-77).
Flight Dynamics

262

Analysis of Steady Symmetric Flight For most aircraft the maximum, positive value of Cme e (elevator up) required in any steady ight condition is larger than the minimum, negative (elevator down) value. The largest positive values of Cme e will be needed to compensate for the largest value of (Cm )e =0 . The latter occurs at the most forward c.g. position and at the maximum angle of attack.

Ground eect usually produces a further negative contribution to (Cm )e =0 ; as a consequence for many aircraft the required size of the elevator is determined by the requirement that the aircraft can take-o and land at the most forward c.g. position.

5-2-7

Stick forces in steady ight

To obtain equilibrium of the aerodynamic moment moment about the lateral axis through the center of gravity, the elevator must be deected by applying a force on the control stick or wheel to compensate for the corresponding aerodynamic hinge moment. In many modern aircraft the connection between the pilots control manipulator and the elevator is such that this eort provides only part of the total required hinge moment. Sometimes there is even no direct mechanical connection at all between the control stick or wheel and the control surface, such as in case of a Fly by Wire ight control system. these classes of control systems will not be discussed now, however, the concepts developed here and in subsequent chapters for mechanical ight control systems do in fact apply also to these types of ight control systems. In gure 5-88 a classical manual control system is depicted. The positive directions of the control deections s, control forces F , surface angles and hinge moments H are as indicated. In the following the longitudinal control force required for equilibrium about the lateral axis, Fe , will be considered in more detail. We now derive a simple model for the control force on the elevator control stick or wheel. Similar models can be derived for the aileron control force, Fa , and for the control force on the rudder pedals, Fr . If the control stick is displaced over a small distance dse , the work done by the control force Fe is Fe dse . Neglecting exibility, the elevator deects over an angle de and the work done by the hinge moment He is He de . If the internal friction in the control system, as well as cable stretch, is neglected, the control system absorbs no work internally and the total amount of external work has to be zero if the stick and the elevator are again at rest after the displacement. This implies that, Fe dse + He de = 0 or, see also gure 5-89, Fe = de He dse (5-82) (5-81)

5-2 Equilibrium in steady, straight, symmetric Flight

263

Figure 5-88: The positive direction of control deections, control forces, control surface deections and hinge moments

Figure 5-89: The relation between the control force and the hinge moment

Flight Dynamics

264

Analysis of Steady Symmetric Flight

Figure 5-90: Contributions to the hinge moment

where the gear ratio

de dse

is positive , see gure 5-88.

The total hinge moment to be balanced by the control force, see gure 5-90, consists of several components, i.e. the aerodynamic hinge moment (Hea ), a hinge moment caused by friction in the control system (Hef ) and, if the control surface is not statically balanced, a hinge moment (Hew ) caused by the static unbalance. In many aircraft a spring is applied in the control system, giving rise to yet another component of the total hinge moment, Hes . The total hinge moment then is, H e = H ea + H ef + H ew + H es For simplicity , the hinge moments due to friction, static unbalance and the spring will be omitted in the following. This results in, H e = H ea Substituting equation (5-61) in (5-82) the longitudinal control force can then be written as, Fe = de 1 2 V Se ce Che dse 2 h

and, using the linearized expression as in equation (5-63) for Che , Fe = de 1 2 V Se ce dse 2 h Ch h + Ch e + Cht te (5-83)

Suppose now, that for a given c.g. position and angle of attack, or airspeed, the elevator angle, e , required for equilibrium of the moment about the lateral axis has been determined

5-2 Equilibrium in steady, straight, symmetric Flight

265

according to equation (5-80). Now the longitudinal control force required to maintain this elevator angle, for a given tab angle, te , follows from equation (5-83). Chapter 7 deals in more detail with this control force as a function of airspeed, c.g. position and tab angle. We conclude with some remarks on the required signs and magnitudes of the hinge moment derivatives Cht , Ch and Ch . The required magnitude of Cht (this is the case of a classical tail plane with xed incidence of the horizontal tailplane, ih ) follows from the requirement that in steady ight conditions it must be possible to trim the control forces down to zero by adjusting the trim tab angle. A more detailed denition of aircraft congurations and ight conditions for which this requirement holds can be found in the airworthiness requirements, see references [11, 19, 6, 20, 12, 13, 7]. The required magnitude of Cht also depends of course on the allowable tab deecting angles, te , and on the values of Ch and Ch . As has been mentioned before, in order to keep the required control forces within reasonable limits when changing either the aircraft conguration, indicated airspeed the steady ight condition or when executing turns or pull up manoeuvres, Ch and Ch should be small in an absolute sense. As shown in chapter 7 and mentioned above, Ch must always be negative. Usually Ch is negative as well, but small positive values of Ch are permissable, even desirable and occur indeed in practice. When studying the control forces in more detail in chapter 7, this will be discussed in more detail.

Flight Dynamics

266

Analysis of Steady Symmetric Flight

Chapter 6 Longitudinal Stability and Control Derivatives

In the following sections attention is given to the methods used to obtain the various stability and control derivatives for the symmetric motions. The corresponding methods used for the asymmetric motions will be discussed in chapter 7. Evidently, the accuracy to which the derivatives can be determined has a direct inuence on the accuracy of the calculated non-steady aircraft motions. The stability derivatives can be found not only by calculation from generalized data, but also from measurements. Flight tests with the actual aircraft can be made, or wind tunnel tests on a model. From comparisons of measured and calculated stability derivatives obtained in dierent ways, it appears that none of these various ways to obtain the various derivatives gives satisfactory results in all cases of practical interest. However, recent research suggests that Computational Fluid Dynamics (CFD) techniques can also be used to estimate the stability and control derivatives with an accuracy comparable to results obtained by ight test techniques. Calculated derivatives may be expected to show an acceptable correspondence with experimentally obtained values, only for conventional aircraft congurations. For less conventional congurations, calculated derivatives may be considered only as a very rst approximation. In such situations measured stability derivatives are indispensable for calculations of stability and control characteristics on which some reliance has to be placed. In the following the various stability derivatives are discussed in turn. When considering the forces and moments in the initial equilibrium condition, the contributions from the propulsive system are included; when discussing the stability derivatives these contributions are not considered.

Flight Dynamics

268 Zr

Longitudinal Stability and Control Derivatives

L Xr

XS

0 V D

Tp

ip

c.g. Xr ZS Xr

Figure 6-1: The attitude of the stability reference frame relative to the Xr - and Zr -axis, after a disturbance from the equilibrium ight condition

Various calculation methods are mentioned, for more detailed discussions and quantitative data reference is made to the literature. All derivatives, with the exception of CZ , Cm and Cn , are calculated under the assumption that the airow is always stationary.

6-1

Aerodynamic forces in the nominal ight condition

For the following discussions the direction of the XS -axis of the system of stability axes is dened relative to the aircraft, using the xed and invariable system of the aircraft reference axes, see gure 6-1 and chapter 2. The angle between the negative Xr -axis and the positive XS -axis is 0 . Note: This is not the 0 meant in equation (4-24). This angle remains constant during the disturbed motion and varies only if one other initial, steady ight condition is chosen. The total aerodynamic force acting on the aircraft in symmetric ight is divided in, 1. The thrust of the propulsive system, Tp 2. The remaining aerodynamic force on the aircraft, consisting of the components, (a) lift, L, perpendicular to the direction of the undisturbed ow (b) drag, D, parallel to the direction of the undisturbed ow

6-1 Aerodynamic forces in the nominal ight condition

269

These three forces are depicted in gure 6-1. The angle between Tp and the negative Xr -axis is ip . In the following it is assumed that ip is independent of the ight condition and does not change during the disturbed motion, as is the case with 0 . Furthermore, 0 and ip are assumed to be small, so,

cos (0 + ip ) 1 sin (0 + ip ) 0 + ip The total aerodynamic moment acting on the aircraft in steady ight is not divided in contributions explicitly assigned to thrust, lift and drag. The components X and Z of the total aerodynamic force along the XS -axis and the ZS -axis of the system of stability axes can now be calculated. From gure 6-1 it follows, X = Z L sin D cos + Tp

= L cos D sin Tp (0 + ip )

In non-dimensional form, after division by 1 V 2 S, it follows, 2 CX CZ = CL sin CD cos + Tc Tc (6-1) (0 + ip )

= CL cos CD sin

where is now the change in angle of attack relative to the value in the equilibrium situation. In equation (6-1) is, Tc = Tp 2D2 = Tc 1 2 S 2 V S

In the initial, steady ight condition (the equilibrium situation) = 0, therefore,

CX0 CZ0

= CD + Tc = CL Tc (0 + ip )

In addition, the following relations hold because of equations (4-34) and (4-33) and = 0, CX0 CZ0 =
1 V 2 S 2

sin 0 (6-2)

W = 1 V 2 S cos 0
2

Flight Dynamics

270 or,

Longitudinal Stability and Control Derivatives

CX0 = CZ0 tg 0 +CL tg 0 If the steady, initial condition is one of horizontal steady ight, then according to equation (6-2), CX0 = 0 and, by consequence, CD = Tc For the calculation of the moments reference is made to previous sections of this chapter. For the equations of motion it is important that in the initial steady ight condition the moments are balanced and therefore, Cm0 = 0

6-2

Derivatives with respect to airspeed

Experimentally obtained values of the stability derivatives CXu , CZu and Cmu are rare. Nearly all measurements of stability derivatives in ight are restricted to the motions at constant airspeed. Measurements in wind tunnels are also invariably made at constant speed. Due to these circumstances it is usually not possible to compare calculated with measured values of CXu , CZu and Cmu . The expressions for the derivatives with respect to airspeed to be discussed in the following, should be used with care, since they have been calibrated against experimental data to an insucient degree.

6-2-1

Stability derivative CXu

According to table 4-3, this derivative is equal to, CXu = Also, X = CX From this follows, X CX 1 2 = CX V S + V S V V 2 1 2 V S 2 1
1 2 V

X S V

6-2 Derivatives with respect to airspeed and, by consequence, CXu = 2 CX + where, CX CX V = V u CX V V

271

(6-3)

See also equation (4-35). The partial derivative with respect to airspeed is determined for deviations from the steady, initial ight condition in airspeed only, i.e. at = 0. This means that in equation (6-3), CX = CX0 and, CXu = 2 CX0 + CX V V
CX V

(6-4) is the expression for

The starting point for the determination of the partial derivative CX at = 0, CX = CD + Tc Dierentiating this expression with respect to V yields, CX Tc CD = V V V

(6-5)

6-2-2

Stability derivative CZu

In analogy with equation (6-4) it follows for the derivative CZu , CZu = 2 CZ0 + In the steady, initial ight condition is, CZ = CL Tc (0 + ip ) After dierentiating it follows, in analogy with equation (6-5), CZ CL Tc = (0 + ip ) V V V (6-7)
Flight Dynamics

CZ V V

(6-6)

272

Longitudinal Stability and Control Derivatives

6-2-3

Stability derivative Cmu

Following equations (6-4) and (6-6), Cmu is written as, Cmu = 2 Cm0 + Cm V V

Since Cm = 0 in the steady, initial condition, Cmu is reduced to, Cmu = The derivatives
CD CL V , V

Cm V V

(6-8)

and

Cm V

may dier from zero due to various causes,

1. The variation of Mach number with airspeed. Due to the eects of compressibility, the aerodynamic coecients vary with airspeed at Mach numbers higher than 0.6 to 0.7. 2. The variation of the aircrafts aeroelastic deformation with airspeed, or rather with dynamic pressure. This eect is not further considered here. 3. The variation of the Reynolds number with airspeed. This inuence is entirely neglected for the small variations of airspeed considered here. 4. For a propeller-driven aircraft the contribution to the coecients CL , CD and Cm made by those parts of the wing and the tailplane submerged in the slipstream, varies with airspeed via the thrust coecient Tc . For CD and CL the inuences of compressibility V V and slipstream eects are taken separately. The same applies to Cm . The partial V derivatives with respect to V then are,
CD V CL V Cm V

= = =

CD M M V CL M M V Cm M M V

+ + +

CD Tc Tc V CL Tc Tc V Cm Tc Tc V

Expressions (6-4), (6-6) and (6-8) can now be written as, CXu = 2 CX0 + 1 CD Tc dTc CD V M dV M dTc CL V M dV M (6-9)

CZu = 2 CZ0 (0 + ip ) +

CL Tc

(6-10)

Cmu =

Cm dTc Cm V + M dV Tc M

(6-11)

6-2 Derivatives with respect to airspeed

273

In equations (6-9), (6-10) and (6-11) the partial derivatives with respect to Mach number are not further considered here. The eects of compressibility, expressed here through the presence of the partial derivatives CD , CL and Cm , are not the subject of this book. The M M M slipstream eect, expressed through CD , CL and the various contributions to Cm , are not Tc Tc Tc discussed in this book either. The derivative Tc is now briey discussed. The basic assumption is always that during the V disturbed aircraft motions to be studied, the power setting of the engine(s), i.e. the throttle position and the setting of any other possible control lever, remains unchanged. Another assumption to be made when determining the variation of Tc is that changes in airspeed always occur in a quasi-steady manner, i.e. so slowly that Tc varies with airspeed just as in a series of steady ight conditions. If these assumptions are met, Tc can be calculated in a V fairly simple manner. In the following Tc is derived for various types of propulsive systems. V 1. Jet turbines and rocket motors In studies on dynamic stability characteristics the usual assumption is that the thrust of these engines is independent of airspeed at constant throttle setting. This means, Tp = Tc This results in, dTc 2T = c dV V At supersonic speeds in particular the thrust of a jet turbine does vary with airspeed. The correct value dTc then has to be obtained from more detailed calculations or from dV measurements. 2. Piston or turbine engines driving constant speed propellers In this case, again at constant power setting, the power produced by the engine is approximately independent of airspeed. If the propeller eciency is also assumed to be constant for relatively small variations in airspeed, then, Tp V = constant or, Tc V 3 = constant This results in, dTc T = 3 c dV V
Flight Dynamics

1 2 V S = constant 2

274 3. Gliding ight

Longitudinal Stability and Control Derivatives

Finally, it should be noted that for any propulsive system in gliding ight (Tc = 0) the following relation holds, dTc =0 dV The foregoing expressions for
dTc dV

may be summarized as follows, dTc T = k c dV V

where k is, 1. in gliding ight, k = 0 2. jet turbines at subsonic speeds, rocket motors, k = 2 3. piston and turbine engines driving constant speed propellers, k = 3 As derived in the previous subsection,

CX0 CZ0

= CD + Tc = CL Tc (0 + ip )

Using these relations, equations (6-9), (6-10) and (6-11) can nally be written as, CXu = 2 CD + Tc 2k 1 CD Tc CL Tc CD M M CL M M (6-12)

CZu = 2 CL + Tc

(2 + k) (0 + ip ) + k

(6-13)

Cmu = k Tc

Cm Cm + M Tc M

(6-14)

If CX0 CL tg 0 is used, equation (6-12) becomes, CXu = 2 CL tg 0 k Tc A few specic cases are, 1 CD Tc CD M M (6-15)

6-2 Derivatives with respect to airspeed 1. Gliding ight at subsonic speed Or, with k = 0, and, CD CL Cm = = =0 M M M yields, CXu CZu Cmu = 2 CL tg 0 = 2 CD = 2 CL = 0

275

2. Level ight at subsonic speed for a jet- or rocket-propelled aircraft Or, with 0 = 0, k = 2, Tc = CD and, CL Cm CD = = =0 M M M yields, CXu CZu Cmu = 2 CD = 2 CL = 2 CD
Cm Tc

3. Level ight at subsonic speed for a propeller-driven aircraft having constant speed propellers Or, with 0 = 0, k = 3, Tc = CD and, CD CL Cm = = =0 M M M yields, CXu CZu Cmu = 3 CD 1
CD Tc CD Tc

= 2 CL + CD (0 + ip ) + 3 = 3 CD
Cm Tc

Flight Dynamics

276

Longitudinal Stability and Control Derivatives

6-3

Derivatives with respect to angle of attack

A change in angle of attack only is obtained by varying the speed w along the ZS -axis. A pure -motion is thus seen to be a steady motion where the aircraft performs a translation along the ZS -axis (plunging motion). The stability derivatives CX , CZ and Cm pertaining to this type of motion may be obtained in a relatively simple way from wind tunnel measurements on a model of the aircraft. CX and CZ are derived from a measured or calculated polar curve corresponding to the specic aircraft conguration, whereas Cm follows from a measured or calculated moment curve Cm . A discussion of each of the three derivatives is given below.

6-3-1

Stability derivative CX

According to table 4-3, this derivative is equal to, CX = From equation (6-1) follows that, CX = CL sin CD cos + Tc Taking the derivative with respect to , leaving Tc constant, results in, CX = CL cos + CL sin + CD sin CD cos In the steady, initial condition = 0, leading to, CX = CL CD if CX is expressed using CN and CT , CX can be written as, CX = CT cos 0 CN sin 0 If a parabolic polar curve is assumed, then, CD = CD0 +
2 CL Ae

1
1 2 V

X CX = S w

(6-16)

where CD0 is the drag coecient at CL = 0. Then, CD is, CD = 2 and CX can be written as, CL CL Ae

6-3 Derivatives with respect to angle of attack

Figure 6-2: The denition of aircraft parts wing, horizontal stabilizer, pylon, nacelles, vertical n and fuselage for the Cessna Ce550 Citation II model (starting from the left top gure, in clockwise direction)

Flight Dynamics

277

278
0.16

Longitudinal Stability and Control Derivatives

0.14

0.12

wing horizontal stabilizer pylons fuselage vertical fin nacelles total

0.1

CX

0.08

0.06

0.04

0.02

0.02 10

[Deg.]

10

Figure 6-3: Calculated contributions of various aircraft parts to the force curve CX versus for the Cessna Ce550 Citation II

CX = CL 1

2CL Ae

Contrary to most stability derivatives, CX is normally positive. Using linearized potential ow theory (panel methods), absolute values of CX as a function of have been calculated for the Cessna Ce550 Citation II, see gure 6-3. The denition of the aircraft parts wing, horizontal stabilizer, pylons, fuselage, vertical n and nacelles is given in gure 6-2. The main contribution to CX is from the wing (induced drag).

6-3-2

Stability derivative CZ

According to table 4-3, this derivative is equal to, CZ = 1

1 2 V

Z CZ = S w

CZ has been written in equation (6-1) as, CZ = CL cos CD sin Tc (0 + ip )

6-3 Derivatives with respect to angle of attack This results in CZ , CZ = CL sin CL cos CD cos CD sin which reduces for = 0 to, CZ = CL CD As normally CD << CL , a simplied expression for CZ is, CZ CL

279

The dominant contribution to CZ is, of course, provided by the wing. But in some cases the contribution made by the horizontal tailplane is not negligible. Then it may be practical to express CZ , CZ = CNw CNh 1 d d Vh V
2

Sh S

(6-17)

Data for a quantitative determination of CZ using equation (6-17) can be found in various places in the literature, see references [72] and [9]. Using linearized potential ow theory (panel methods), absolute values of CZ as a function of have been calculated for the Cessna Ce550 Citation II, see gure 6-4. The denition of the aircraft parts wing, horizontal stabilizer, pylons, fuselage, vertical n and nacelles is given in gure 6-2. The main contribution to CZ is from the wing.

6-3-3

Stability derivative Cm

According to table 4-3, this derivative is equal to, Cm = 1

1 2 V

M Cm = S w c

In this chapter the simplied case is considered where the contributions of the tangential forces to Cm are neglected and the contributions of the thrust and slipstream eects are also not considered. It was found in chapter 5 for Cm , Cm = CNw xc.g. xw CNh c 1 d d Vh V
2

Sh lh S c

(6-18)

More general. Cm is the slope of the moment curve Cm = Cm () at the respective angle of attack at which according to section 6-1, Cm0 = 0. A calculation of Cm is possible using the references of chapter 5.
Flight Dynamics

280
1

Longitudinal Stability and Control Derivatives

0.5

wing horizontal stabilizer pylons fuselage vertical fin nacelles total

CZ
0.5 1 1.5 10

[Deg.]

10

Figure 6-4: Calculated contributions of various aircraft parts to the force curve CZ versus for the Cessna Ce550 Citation II

Using linearized potential ow theory (panel methods), absolute values of Cm as a function of have been calculated for the Cessna Ce550 Citation II, see gure 6-5. The denition of the aircraft parts wing, horizontal stabilizer, pylons, fuselage, vertical n and nacelles is given in gure 6-2. The main contributions to Cm are from the fuselage (destabilizing) and horizontal stabilizer (stabilizing).

6-4

Derivatives with respect to pitching velocity

Before discussing the derivatives CXq , CZq and Cmq , the angular motion under consideration will be studied more closely. During the turning, symmetric motion, indicated here as qmotion, the center of rotation lies on the top axis through the aircrafts center of gravity, see gure 6-6. The distance R to the c.g. is, V q

R=

At a positive pitching velocity the center of rotation lies on the negative ZS -axis, i.e. above the aircraft. In all points of the aircraft, with the exception of the c.g. and all other points on the ZS -axis, the geometric angle of attack is proportional to the angular velocity and to the distance in

6-4 Derivatives with respect to pitching velocity

0.25 wing horizontal stabilizer pylons fuselage vertical fin nacelles total

281

0.2

0.15

0.1

Cm

0.05

0.05

0.1

0.15

0.2 10

[Deg.]

10

Figure 6-5: Calculated contributions of various aircraft parts to the moment curve Cm versus for the Cessna Ce550 Citation II

XS -direction to the center of gravity, = See also gure 6-6. The magnitude of the local airspeed varies (in principle) also with the angular velocity. This is because the local velocity in a given point varies proportional with the distance of that point to the center of rotation. Only in points at distance R of the center of rotation, i.e. also in the c.g., the magnitude of the local velocity does not vary with the rate of pitch. The aircrafts c.g. is the only point where during this q-motion neither the magnitude of the velocity, nor the direction, i.e. the geometric angle of attack, vary. Changes in the magnitude of the local airspeed and the geometric angle of attack at the aircrafts c.g. are described as the u-motion and the - or plunging-motion respectively. As far as the airow at the center of gravity is concerned, the q-motion causes only a curvature of the streamlines. If the angular velocity during a q-motion is not constant, but varies harmonically with time, the aircraft describes due to this q-motion a sinusoidal trajectory, as indicated in gure 6-7. The most characteristic feature of this trajectory is, that the angle of attack at the aircrafts c.g. remains constant during this motion, i.e. = 0. The center of gravity moves along an undulating trajectory, such that,
Flight Dynamics

x xc.g. x xc.g. q c = R c V

(6-19)

282

Longitudinal Stability and Control Derivatives

Zr

xc.g. x

Xr
Figure 6-6: The pure q-motion, with =
xxc.g. R

xxc.g. q c c V

6-4 Derivatives with respect to pitching velocity

283

XS XS XS ZS XS ZS ZS ZS ZS XS

Figure 6-7: Harmonic q-motion, = constant, =

Flight Dynamics

284

Longitudinal Stability and Control Derivatives

XS ZS XS

XS XS ZS ZS

XS ZS ZS

Figure 6-8: Harmonic -motion, q = 0

6-4 Derivatives with respect to pitching velocity

285

XS XS ZS XS ZS XS ZS ZS

Figure 6-9: Combined harmonic q- and -motion, q = , = constant

=q =+ = Both the angle of pitch and the ight path angle will vary harmonically. In gure 6-8 the case for harmonic plunge is presented, i.e. varies harmonically, however, now q = 0. An entirely dierent motion, to be distinguished very clearly from the q-motion, results if the center of rotation lies in the aircrafts c.g., see gure 6-9. Then the angle of attack varies also in the center of gravity and, =q= and, by consequence, ==0 (6-21) (6-20)

According to equation (6-21) the center of gravity moves along a straight line, if the aircraft rotates while the c.g. is the center of rotation. This motion may be considered a superposition of a q-motion as dened above and a variable -motion, such that the condition (6-20) is satised.
Flight Dynamics

286

Longitudinal Stability and Control Derivatives

center of gravity

XS

ZS
Figure 6-10: Aircraft in a curved ow eld

Zr (xc.g. , zc.g. )

xc.g. zc.g. x

(x, z)

Xr
Figure 6-11: Equivalent curved aircraft in a straight ow eld with aircraft frame of reference, Fr

6-4 Derivatives with respect to pitching velocity

287

Figure 6-12: Equivalent curved aircraft in a straight ow eld

Figure 6-13: Equivalent curved aircraft in a straight ow eld, 3-dimensional view

Flight Dynamics

288

Longitudinal Stability and Control Derivatives

This q-motion is here the subject of discussion, the inuence on the aerodynamic forces and moments due to an acceleration along the top axis during a non-steady -motion is discussed in the next section. The variation of the magnitude of the local airspeed with the distance to the center of rotation during the q-motion is usually neglected because of the small dimensions of the aircraft in the ZS -direction if compared with common values of R. The only eect of the rotation is then a variation of the geometric angle of attack in the XS -direction, expressed by equation (6-19), see gure 6-10. A eld of ow equivalent to that of the aircraft in curved ow can be obtained by placing a suitable curved aircraft in a eld of parallel ow, see gures 6-11, 6-12 and 6-13. The curvature of the aircraft or the wind tunnel model must be such that, = see equation (6-19), or, z zc.g. 1 = c 2 x xc.g. c
2

x xc.g. q dz c =+ dx c V

q c V

c The curvature is parabolic in the XS -direction and proportional to q . For theoretical calV culations it may be advantageous to use the curved aircraft in parallel ow. If, however, the stability derivatives would have to be measured using a curved model in parallel ow, it would c be necessary to use a separate model for each value of q . V

The changes in the forces X and Z and in the moment M caused by the q-motion are discussed in the following. The change in X, expressed by the derivative CXq , CXq = 1
1 2 V

X CX = q S q c Vc

is always neglected. The assumption thus is, CXq = 0

Using linearized potential ow theory (panel methods), absolute values of CX as a function of q c V have been calculated for the Cessna Ce550 Citation II, see gure 6-14. The denition of the aircraft parts wing, horizontal stabilizer, pylons, fuselage, vertical n and nacelles is given in gure 6-2. All dened aircraft parts have a neglible contribution to CX (as compared to the force curve CX versus , see gure 6-3).

6-4 Derivatives with respect to pitching velocity

0.03 wing horizontal stabilizer pylons fuselage vertical fin nacelles total

289

0.025

0.02

0.015

CX

0.01

0.005

0.005

0.01 0.05

0.04

0.03

0.02

0.01

0.01 q 0 c V [Rad.]

0.02

0.03

0.04

0.05

Figure 6-14: Calculated contributions of various aircraft parts to the force curve CX versus Cessna Ce550 Citation II

q c V

for the

The two remaining derivatives are CZq and Cmq , CZq = 1

1 2 V

Z CZ = q q S c Vc

Cmq =

1 2 V

1 M Cm = q 2 q S c Vc

Of these two derivatives Cmq is the most important. For an aircraft having a horizontal tailplane, the largest contribution to Cmq is provided by this tailplane. The change in wing lift caused by the angular velocity, i.e. by the curvature of the ow eld, is small. Therefore, no change in downwash, as a reaction to a change in wing lift, is accounted for either. The contribution from the tailplane is calculated as follows. According to equation (6-19) the change in angle of attack of the horizontal tailplane due to the rotation is, h = xh xc.g. q lh q c c c V c V

This change in angle of attack causes a normal force at the tailplane, CNh = CNh Vh V
2

Sh h S
Flight Dynamics

290

Longitudinal Stability and Control Derivatives

and this generates a moment about the aircrafts c.g., Vh V

2

Cm = CNh h

Sh lh = CNh S c

Vh V

2 Sh lh q c 2 V S c

As a consequence the contribution of the horizontal tailplane to Cmq is, Cmq and the contribution to CZq is, CZq = CNh Vh V
2

= CNh

Vh V

2 Sh lh S2 c

Sh lh S c

A rough estimate of CZq for the complete aircraft is sometimes taken as, CZq = 2 CZq = 2 CNh Vh V
2

Sh lh S c

For aircraft having a straight, slender wing Cmq of the entire aircraft is usually approximated by multiplying the contribution of the tailplane with a factor 1.1 to 1.2, in order to account for the inuence of the wing. For these conventional aircraft Cmq can be expressed by, Cmq = (1.1) CNh Vh V
2 2 Sh lh S2 c

Finally, a practical remark should be made. Often, quantitative data for the determination of CZq and Cmq are taken from publications of the American National Aeronautics and Space Administration (NASA). It is of interest to realize that NASA, and some other organizations q c c as well, commonly relate the stability derivatives CZq and Cmq to 2V rather than q . V In contrast to a u- or -motion, the q-motion changes with a shift in the c.g. position. This can be seen as follows. In the new c.g. the change in angle of attack due to a rotation about the new center of rotation is, of course, again zero. This change in angle of attack at the new c.g. position c.g.2 , c.g.2 (where c.g.2 = 0), is the sum of a contribution from a rotation about the old c.g. (c.g.1 ) and a change in angle of attack due to an -motion along the ZS -axis. From this consideration Cmq can be calculated in a straightforward manner. In table 6-2 the expressions for the variations of CZq and Cmq with the c.g. position are given, based on the above discussion. Using linearized potential ow theory (panel methods), absolute values of CZ and Cm as c a function of q have been calculated for the Cessna Ce550 Citation II, see gure 6-15 V and gure 6-16. The denition of the aircraft parts wing, horizontal stabilizer, pylons, fuselage, vertical n and nacelles is given in gure 6-2. The main contributions to CZ

6-5 Derivatives with respect to the acceleration along the top axis
0.15

291

0.1

0.05

0.05

CZ
0.1 wing horizontal stabilizer pylons fuselage vertical fin nacelles total 0.15 0.2 0.25 0.3 0.05

0.04

0.03

0.02

0.01

0.01 q 0 c V [Rad.]

0.02

0.03

0.04

0.05

Figure 6-15: Calculated contributions of various aircraft parts to the force curve CZ versus Cessna Ce550 Citation II

q c V

for the

are from the wing and the horizontal stabilizer. The main contribution to Cm is from the c c horizontal stabilizer. Note that both the slope of CZ q and Cm q are representative for V V respectively the stability derivatives CZq and Cmq . E.g. from gure 6-15 it is obvious that CZq indeed is approximately, CZq = 2 CZq

Furthermore, from gure 6-16 it also holds that Cmq is indeed approximately, Cmq = 1.1 CZq lh c

6-5

Derivatives with respect to the acceleration along the top axis

If the components of airspeed u or w or the rate of pitch q of an aircraft experience a sudden change, a certain time interval passes before the pressure distribution over the entire aircraft has adjusted to the new ow condition. Usually, changes in the airspeed, i.e. in the component u, occur so slowly that this delayed adjustment is not noticeable. Changes in u are, therefore, always assumed to occur in a quasi-steady fashion. On the other hand, changes in angle of attack, i.e. in the speed component w along the ZS -axis, may occur much more quickly. Such changes are discussed in the following.
Flight Dynamics

292
0.4

Longitudinal Stability and Control Derivatives

0.3

0.2

wing horizontal stabilizer pylons fuselage vertical fin nacelles total

0.1

Cm

0.1

0.2

0.3

0.4 0.05

0.04

0.03

0.02

0.01

0.01 q 0 c V [Rad.]

0.02

0.03

0.04

0.05

Figure 6-16: Calculated contributions of various aircraft parts to the moment curve Cm versus Cessna Ce550 Citation II

q c V

for the

In reference [38] Cowley and Glauert were the rst to point to the fact that an acceleration w of the aircraft in the direction of the top axis has a non-negligible inuence on the longitudinal moment Cm . This eect has an important inuence on the damping of the symmetric motions. Following Cowley and Glauert, this eect is usually expressed through a stability derivative Cm . In order to account for the inuence of the c.g. on this derivative Cm , it is necessary in principle to know as well the derivative CZ which is in itself quite important. As described earlier the relation between the aerodynamic coecient Cm and the component of the motion may be expressed in a Taylor series, Cm Cm 1 c + + c V 2! V 2 Cm
2 c V

Cm =

2 Cm 2 Cm c ()2 + 2 + c 2 V V
2

c V

1 { } + 3!
3

(6-22)

c c In principle, the derivatives with respect to 2 , 3 , etcetera, could also be included in this V V series expansion. Experimental evidence shows, however, that the changes in angle of attack occur slowly enough to neglect any inuences of the higher time derivatives than . As usual, in the series (6-22) only the linear terms are maintained under the assumption that the values c of and remain suciently small. V

6-5 Derivatives with respect to the acceleration along the top axis The change in the moment coecient, Cm , can then be written as, Cm = Cm Cm c + c V V

293

or, in the common notation used in stability analyses, Cm = Cm + Cm c V (6-23)

c The term Cm is then supposed to express the inuence on Cm of the delayed adjustment V of the airow to changes in .

For aircraft having a horizontal tailplane, this tailplane usually provides the most important contribution to Cm . A simplied explanation of this eect is the following. In chapter 5 the angle of attack of the horizontal tailplane in steady, straight ight was expressed by equation (5-64), h = + ih During the accelerated translation along the ZS -axis here considered, ih remains constant. In accordance with the conventions made in the present chapter, the explicit indication that a change in a variable is considered, is omitted. As a consequence h (t) can be written here as, h (t) = (t) (t) The horizontal tailplane is always situated in an air mass having passed the wing a brief time interval t, t lh V

earlier. As a consequence, the downwash angle at the horizontal tailplane is at any time proportional to the wing angle of attack that brief time interval earlier, (t) = d (t t) d (6-24)

The angle of attack of the wing at the time t t can be written as, (t t) = (t) t + 1 1 (t)2 2! 3!

In this expression the time derivative of higher than the rst are omitted, as in equation (6-22). The result is,
Flight Dynamics

294

Longitudinal Stability and Control Derivatives

(t t) = (t) t In equation (6-24) this leads to, d d lh (t) d d V

(t) = and,

h (t) = (t)

d d

d lh d V

The term in the change in h proportional to is then, h (t) = + This h produces a force, Vh V
2

d lh c d c V

(6-25)

CZ = CNh h and a longitudinal moment,

Sh S

Cm = CNh h

Vh V

Sh lh S c

Substituting equation (6-25) in the latter two expressions results in the desired, approximated, expressions for the stability derivatives, Vh V Vh V
2

CZ = CNh Cm = CNh

d Sh lh d S c
2 d Sh lh d S2 c

The variation of Cm with the c.g. position is calculated by, Cm = CZ xc.g. c

In the calculation of CZ and Cm attention should be given, as with CZq and Cmq , to the c c fact that in the literature these derivatives occur referenced to 2V rather than to . V

6-6 Derivatives with respect to the elevator angle

295

6-6
6-6-1

Derivatives with respect to the elevator angle

Control derivative CXe

According to table 4-3, this derivative is equal to, CXe = 1

1 2 2 V S

X CX = e e

This derivative expresses the variation of aircraft drag with the elevator angle. Although the total drag caused by the elevator deection may be non-negligible, especially at supersonic speed (trim drag), the variation in CX with small variations of e is commonly neglected. Or, CXe = 0

6-6-2

Control derivative CZe

According to table 4-3, this derivative is equal to, CZe = 1

1 2 2 V S

Z CZ = e e

If the aircraft has a horizontal tailplane, CZe can also be written more explicitly as, CZe = CNh Vh V
2

Sh S

(6-26)

References [9, 102, 42, 45, 85, 98, 37] provide data to derive CNh if the form and the e dimensions of the stabilizer and the elevator are known. For tailless aircraft CZe should preferably be obtained from measurements on a wind tunnel model, see also references [42, 45, 98, 37].

6-6-3

Control derivative Cme

According to table 4-3, this derivative is equal to, Cme = 1

1 2 c 2 V S

M Cm = e e

If the aircraft has a horizontal tailplane, Cme can be simply expressed using CZe , Cme = CZe or,
Flight Dynamics

xh xc.g. c

296

Longitudinal Stability and Control Derivatives

Cme = CNe An approximation for Cme is,

Vh V

Sh xh xc.g. S c

Cme = CNe

Vh V

Sh lh S c

This results in Cme , if CZe has been previously obtained using equation (6-26). Of the three control derivatives Cme has the largest eect on the aircraft motions. For aircraft having a horizontal tailplane, the inuence of Cme dominates to such an extent that very often CZe for these aircraft is neglected. This approximation is less applicable to tailless aircraft. The derivative Cme for tailless aircraft is derived by the same calculation methods or wind tunnel measurements yielding also CZe . Table 6-1 summarizes the expressions for the stability and control derivatives discussed in the previous sections.

6-7

Symmetric inertial parameters

Reliable quantitative data on moments and products of inertia of aircraft are rare. Quantitative calculations are very time-consuming and may produce rather inaccurate results. The most reliable inertial parameters of aircraft are those obtained experimentally, usually by oscillating the suspended aircraft on the ground. Details on this experimental technique can be found in references [167, 163, 41, 96].

6-7 Symmetric inertial parameters

297

CX0 = CZ0 = Cm0 =

CD + Tc cos (0 + ip ) =

1 V 2 S 2

sin0 CD + Tc

W CL Tc sin (0 + ip ) = 1 V 2 S cos0 CL
2

CXu = CZu = Cmu =

2 CL tg 0 2 CL 0

CX =

CL CD = CT cos0 CN sin0 ; if CD = CD0 + CX = CL 1

2CL Ae

2 CL Ae

then

CZ = Cm =

CL CD CL CN CNw
xc.g. xw c

CNh 1

d d

Vh V

Sh lh S c

CXq = CZq = Cmq =

0 2 CZq = 2 CNh
h Vh V 2 Sh lh S c Vh V 2 S l2 h h S2 c

(1.1 to 1.2) Cmq

= (1.1 to 1.2) CNh

CX = CZ = Cm =

0 CNh CNh
Vh V Vh V 2 d Sh lh d S c
2 d Sh lh d S2 c

Table 6-1: Simplied formulae for the calculation of stability and control derivatives and coecients in the initial, steady ight condition for the symmetric motions (without the eects of propellers and jets, aeroelasticity and compressibility of the air).

Flight Dynamics

298

Longitudinal Stability and Control Derivatives

CXe = CZe = Cme =

0 CNh CNe
Vh V Vh V 2 2 Sh S Sh lh S c

Table 6-1: (Continued) Simplied formulae for the calculation of stability and control derivatives and coecients in the initial, steady ight condition for the symmetric motions (without the eects of propellers and jets, aeroelasticity and compressibility of the air).

Cm2 =

Cm1 CZ CZq1 CZ

xc.g.2 xc.g.1 c

CZq2 = Cmq2 = =

xc.g.2 xc.g.1 c xc.g.2 xc.g.1 c

Cmq1 CZq2

Cm1

xc.g.2 xc.g.1 c

=
xc.g.2 xc.g.1 c 2

Cmq1 CZq1 + Cm1

xc.g.2 xc.g.1 c

xc.g.2 xc.g.1 c

+ CZ

Cm2 =

Cm1 CZ

Table 6-2: Inuence of the center of gravity position on some stability derivatives

Chapter 7 Lateral Stability and Control Derivatives

In the previous chapter the equilibrium about the lateral axis and longitudinal control in steady, symmetric ight was studied. In addition, static longitudinal stability was considered since this is usually the most critical condition for dynamic longitudinal stability. Studies in dynamic lateral stability relate to the asymmetric motions relative to a steady, symmetric ight condition. There is, however, not one specic static lateral stability characteristic for the asymmetric disturbed motions. These disturbed asymmetric motions appear to be determined by more than just one stability derivative, while the inertial parameters also have an important inuence. In this chapter the lateral stability and control derivatives are discussed, whereas in chapter 9 the lateral equilibrium and lateral control in steady, asymmetric ight conditions will be studied, similar to chapter 8.

7-1

Aerodynamic force and moments due to side slipping, rolling and yawing

The asymmetric degrees of freedom of an aircraft are the following, see gure 7-1. The translation of the aircraft center of gravity along the YB -axis of the aircraft body axes, positive in the direction of this Y -axis. The component of airspeed along the YB -axis is indicated as v. If such a component is present the aircraft is said to be in sideslipping ight. The angle of sideslip is the angle between the velocity vector V of the center of gravity and the plane of symmetry, see gure 7-2,
Flight Dynamics

300 XB

Lateral Stability and Control Derivatives q u v YB

c.g. w

ZB
Figure 7-1: The six degrees of freedom of a rigid aircraft

= arcsin or, at suciently small angles of sideslip, v V

v V

(Radians)

(7-1)

The sideslip angle is positive if the aircraft moves in the direction of the positive YB -axis, so to the right as seen by the pilot. The angular velocity about the XB -axis, positive if the right wing moves down. The angular velocity about the XB -axis, the rolling velocity, is indicated as p. Often the pb non-dimensional form 2V is used. The angular velocity about the ZB -axis, positive if the aircrafts nose moves to the right as seen by the pilot. The angular velocity about the ZB -axis, the yawing velocity, is pb indicated as r. The non-dimensional yaw velocity about this top axis is 2V . The asymmetric motions in general cause an aerodynamic force Y along the YB -axis, a moment L about the XB -axis and a moment N about the ZB -axis, see gure 7-3. As already noted in section 4-1, asymmetric motions generally do not produce symmetric aerodynamic forces and moments. As a consequence the symmetric aerodynamic forces X and Z and the moment M , with a single exception, do not have to be considered. The rolling moment L is commonly indicated as L in situations where confusion with the lift L is possible. The non-dimensional asymmetric forces and moments are dened as,

7-1 Aerodynamic force and moments due to side slipping, rolling and yawing XB u c.g. w V v YB

301

ZB
Figure 7-2: Denition of the angle of attack and sideslip angle

CY =

Y
1 2 2 V S

C =

L
1 2 2 V Sb

(7-2)

Cn =

N
1 2 2 V Sb

where C and Cn are referenced to the wing span and not, like Cm , to the wing m.a.c., c. The non-dimensional aerodynamic force and moments resulting from side slipping, rolling and yawing are determined by the partial derivatives with respect to the non-dimensional pb rb asymmetric components of the moment , 2V and 2V P. These derivatives are, For sideslipping ight, CY = For rolling ight, CYp = For yawing ight, CYr = CY rb 2V Cr = C rb 2V Cnr = Cn rb 2V
Flight Dynamics

CY

C =

Cn =

Cn

CY
pb 2V

Cp =

C
pb 2V

Cnp =

Cn
pb 2V

302 XB L = C 1 V 2 Sb 2 V
pb 2V

Lateral Stability and Control Derivatives YB

v c.g.

Y = CY

1 2 2 V S

rb 2V

N = Cn 1 V 2 Sb 2 ZB
Figure 7-3: Asymmetric force and moments
pb rb The derivatives with respect to 2V and 2V are indicated only with the subscript p and r respectively, to simplify the notation. The above mentioned partial derivatives are called the stability derivatives with respect to angle of sideslip, non-dimensional rolling velocity and non-dimensional yawing velocity, as these derivatives are used primarily in relation to the stability of aircraft ight. When using the stability derivatives it is assumed that the forces pb rb and moments vary linearly with , 2V and 2V . The additional assumption is made, that in symmetric ight CY , C and Cn are zero. As a result, for instance in steady, straight sideslipping ight (p = r = 0), it can be written,

CY = CY

C = C

Cn = Cn

For small deviations from the symmetric equilibrium condition to which the study of stability refers, this latter assumption is generally acceptable. Commonly, this assumption also holds pb rb when considering steady, asymmetric ight at values of , 2V and 2V which are not too large. During the non-steady motion of aircraft caused by a disturbance from the symmetric equilibrium condition the accelerations also generate aerodynamic forces and moments. These are usually small enough to be negligible. An exception is only made for the moment about the top axis caused by an acceleration along the YB -axis. The stability derivative describing this moment, indicated in accordance with the above as Cn , can be compared to the stability derivative Cm for the symmetric motions already discussed in section 6-5. A deviation from the rule that asymmetric motions cause no symmetric forces and moments, is the change in pitching moment caused by a velocity about the lateral axis, v. In the study of stability, where small deviations from the symmetric equilibrium situation are considered, this change in Cm is commonly negligible. In steady asymmetric ight where large angles of sideslip occur, but also during take-o runs and during landing with cross wind, this pitching moment due to sideslip certainly has to be considered. The eect is expressed quantitatively

7-2 Stability derivatives with respect to the sideslip angle by the derivative Cm2 .

303

Summarizing the above, the non-dimensional asymmetric aerodynamic forces and moments caused by the lateral motions are written as, CY = CY + CYp pb rb + CYr 2V 2V pb rb + Cr 2V 2V (7-3)

C = C + Cp

Cn = Cn + Cn

b pb rb + Cnp + Cnr V 2V 2V

In the following the underlying mechanisms, magnitudes and signs of the stability derivatives are discussed. Figures 7-50 give a summary of the forces and moments acting on the aircraft pb rb due to , 2V and 2V . The usual signs of the corresponding stability derivatives are shown as well. Quantitative values of the stability derivatives have been compiled in appendix B for a number of aircraft in various ight conditions, see also reference [18]. When it comes to the actual calculation of stability derivatives, reference will be made to relevant references. An extensive list of references on the calculation of stability derivatives can be found in reference [154]. The calculation methods indicated in the references often rely on empirical data obtained from similar aircraft congurations. Calculation methods based entirely on theory result in most instances in insuciently accurate results. Wind tunnel measurements remain necessary in later design stages to arrive at suciently accurate data. This applies in particular to the stability derivatives with respect to angle of sideslip, , which have a predominant inuence on lateral stability and lateral control characteristics of aircraft.

7-2
7-2-1

Stability derivatives with respect to the sideslip angle

Stability derivative CY

Sideslipping motion generates an aerodynamic lateral force, usually negative at positive angles of sideslip, Y = CY 1 2 V S 2 (7-4)

where CY is negative. The dominant contributions to CY are caused by the fuselage and the vertical tailplane, see gure 7-4. The contribution of the wing is generally negligible unless the wing has a large angle of sweep. The propulsion system may also give a contribution to CY that is not negligible.

Flight Dynamics

304

Lateral Stability and Control Derivatives

CY
0.3

f = 0o complete model
0.2

0.1

wing+fuselage fuselage
-16 -12 -8 -4 0 4 8 12

[o ]

wing
-0.1

-0.2

-0.3

Figure 7-4: CY as a function of measured on a model of the Fokker F-27 in gliding ight (from references [148, 150, 24])

7-2 Stability derivatives with respect to the sideslip angle

305

The small contribution of the wing may, if necessary, be obtained for low airspeeds using references [10, 70, 34, 122, 156, 63]. The inuence of the compressibility of the air can be accounted for using reference [56]. Calculation methods applicable to supersonic speeds are given in references [10, 70, 34, 3, 139, 2, 76]. The contribution of the fuselage arises in the same way as the normal force on a symmetric fuselage at a non-zero angle of attack. Calculation methods can be found in references [10, 70, 34, 3, 139, 2, 76]. The contribution of the vertical tailplane can, in analogy with the gradient of the normal force on the horizontal tailplane in the symmetric case, be written as, CY = CYv dv d Vv V
2

Sv S

(7-5)

Compare this equation to the corresponding expression for the horizontal tailplane, (CN )h = CNh
C

dh d

Vh V

Sh S

Y In equation (7-5) CYv = vv is the gradient of the normal force on the isolated vertical tailplane referenced to the area Sv of the vertical tailplane and the average local dynamic pressure 1 Vv2 at the vertical tailplane. The angle of attack v is the local angle of attack 2 of the vertical tailplane measured in the XB OYB -plane and counted as positive if the airow hits the tailplane from the left, see gure 7-5.

Even when disregarding the sign, v is generally not equal to . Due to the presence of the fuselage and the wing-fuselage interaction a change in the component of the local velocity of the air parallel to the YB -axis occurs. The resulting change in the direction of the airow is described using the sidewash angle , comparable to the downwash angle in symmetric ow. Consulting gure 7-5 it follows, v = ( ) The corresponding expression for the horizontal tailplane was, see equation (5-64), h = + ih From equation (7-6) follows, dv d = 1 d d Substituting equation (7-7) in equation (7-5) results in, CY = CYv 1 d d Vv V
2

(7-6)

(7-7)

Sv S

(7-8)
Flight Dynamics

306

Lateral Stability and Control Derivatives

(> 0)

v (< 0) (> 0)

v = ( )

Figure 7-5: Relation between the sideslip angle , the sidewash angle and the angle of attack v at the vertical tailplane

As a result of the sign convention for v , CYv is positive like any normal force gradient. In the quantitative calculation of CYv the presence of the horizontal tailplane, sometimes acting as an end plate to the vertical tail plane, has to be taken into account. Due to this eect the eective aspect ratio of the vertical tailplane may be considerably larger than its geometric aspect ratio. For subsonic airspeeds the calculation of CYv can be made using references [10, 70, 103, 91, 116, 135, 131, 31]. For supersonic speeds references [109, 29, 106, 107] provide calculation methods. If the angle of attack is not too large, ) for the horizontal tailplane. The dealing with Cn .
Vv 2 V

may usually be taken as 1 if the tailplane is not

located in a slipstream. In the contrary case,

2 Vv 2 is determined in the same way as Vh V V magnitude of d will be discussed in more detail when d

The propulsion system provides a negative contribution to CY caused by a lateral force in the propeller plane or in the plane of the engine inlet in the case of jet propulsion. This eect is a direct analogy with the normal force generated by a propeller or the engine air inlet if placed under a non-zero angle of attack. Calculation methods to determine this contribution are given in references [133, 134]. Using linearized potential ow theory (panel methods), CY as a function of has been calculated for the Cessna Ce550 Citation II, see gure 7-6. The denition of the aircraft

7-2 Stability derivatives with respect to the sideslip angle

0.1

307

0.08

0.06

0.04

wing horizontal stabilizer pylons fuselage vertical fin nacelles total

0.02

CY

0.02

0.04

0.06

0.08

0.1 10

[Deg.]

10

Figure 7-6: Aerodynamic force coecient CY as a function of for the Cessna Ce550 Citation II, = 0o

parts wing, horizontal stabilizer, pylons, fuselage, vertical n and nacelles is given in gure 6-2. The main contributions to CY are from the vertical n and the fuselage.

7-2-2

Stability derivative C

The stability derivative C is one of the primary lateral stability derivatives, usually indicated as the eective dihedral of the aircraft. The explanation of this name lies in the fact that for conventional aircraft without wing sweep C can usually be varied primarily changing the dihedral of the wing. To obtain desirable lateral control charactersitics it is required that C is negative. This can be illustrated as follows. Suppose a disturbance causes an angle of roll to the right. Under the inuence of the component of gravity along the YB -axis, the aircraft starts sideslipping to the right. A negative C then causes a rolling moment trying to return the aircraft to an even keel, without interference by the pilot. In a qualitative sense this explains the desirability of a negative C . Apart from wing dihedral, wing sweep and the wing-fuselage interaction have a large inuence on C . The slipstream inuence on C of propeller driven aircraft can be considerable. The contributions of the fuselage and the tailplane are generally small. The contribution of wing-dihedral to C is caused by the dierence in geometric angle of attack of the left and right hand halves of the wing in sideslipping ight. From subgure b of gure 7-7 follows for the right wing without sweep and having not too large a dihedral,
Flight Dynamics

308

Lateral Stability and Control Derivatives

Vnr = w cos + v sin w + v where for small values of and , see subgure a of gure 7-7, w V sin V and, v V sin V resulting in, Vnr = V ( + ) The geometric angle of attack of the right wing then is, see subgure c of gure 7-7, wr = arctan Vnr u + (7-10) (7-9)

In a similar manner it can be derived for the left wing, Vnl V ( ) hence, wl (7-12) (7-11)

The increase in lift on the right wing due to wr = + and the decrease in lift on the left wing due to wl = are thus proportional to the geometric dihedral . From the above it can be concluded that C w will also be approximately proportional to . Swept wings experience an additional dierence in lift between the two halves of the wing, independent of dihedral. This dierence in lift is caused by the dierence in the components of airspeed perpendicular to the wing leading edge on the 1 -chord line. It can be seen from 4 gure 7-8 that the dierence in lift L can be roughly approximated as, L = CL 1 2 S V 2 2 cos2 ( ) cos2 ( + ) (7-13)

Further analysis results for small values of in, L = CL 1 2 V S sin 2 2 (7-14)

7-2 Stability derivatives with respect to the sideslip angle w v V u XB YB

309

ZB (A) Wing with dihedral in sideslipping ight

w Vn v v

Vnr

YB

b 2

b +2

ZB (B) The normal velocities Vn and Vnr at two sides of the wing Vn w Vnr u u (C) The angles of attack at the left and right wing wr

Figure 7-7: The origin of the dierence in angles of attack at the left and right wing for a wing with dihedral in sideslipping ight
Flight Dynamics

310

Lateral Stability and Control Derivatives

Vnr = V cos( )

Vnl = V cos( + )

+ V

Figure 7-8: The origin of the dierence in the velocities over the left and right wing for a swept wing in sideslipping ight

<0 CL () =0

C due to > 0

>0
Figure 7-9: The variation of C with CL for swept wings

7-2 Stability derivatives with respect to the sideslip angle

311

As the rolling moment of the wing C w will be proportional to L, to a rst approximation, this simple calculation shows that C w will be approximately proportional to CL and to sin 2, see gure 7-9. Flap deection causes, at constant lift on the entire wing, a concentration of the lift at the center parts of the wing. Swept wings will have a less negative C due to ap deection. The same eect is obtained by applying negative wing-twist to swept wings. The inuence of dihedral and sweep on C w are additive to a rst order approximation. In summary it is seen from the foregoing that for a swept wing the contribution C w is composed of one part proportional to the wing dihedral (independent of CL and wing sweep ) and of one part independent of and proportional to CL and sin 2. A consequence of the variation of C with CL is that swept wings often have a zero or even negative geometric dihedral, to avoid the unfavourable eects of too large and eective dihedral at large angles of attack. This will be further discussed when studying the dynamic lateral stability. Calculation methods to determine C w are presented in references [10, 70, 122, 44, 130, 127]. The contribution of the fuselage itself to C is very small. The wing fuselage interaction in sideslipping ight, however, can cause signicant dierences in the angles of attack of the two halves of the wing. This again has a strong inuence on the rolling moment due to side-slip. Important for the way in which the rolling moment varies with angle of sideslip is the vertical position of the wing relative to the fuselage, see gure 7-13. The forward half wing of a low wing aircraft experiences a decrease in angle of attack due to sideslip. This produces a decrease in lift whereas the receding half wing experiences an increase in angle of attack and thus an increase in lift, due to the presence of the fuselage. The situation is the reverse for a high wing aircraft. A positive angle of sideslip causes for a low wing aircraft an extra positive rolling moment, C becomes less negative. On the other hand, C of a high wing aircraft becomes more negative. The magnitude of the contribution is strongly inuenced by the shape of the fuselage and by its size compared to that of the wing. A detailed description of the inuence of the wing-fuselage interaction on C is given in reference [79]. A determination of C of aircraft without tailplanes is possible using references [122, 156, 25, 44, 130, 127] for subsonic speeds and references [129, 84, 83, 108, 144] for supersonic speeds. For example, see also the following illustrations; for the case of a high-wing-fuselage conguration in side-slipping ight, the conguration and velocity eld are plotted in gure 7-10, while for the case of a low-wing-fuselage conguration in side-slipping ight, the conguration and velocity eld are plotted in gure 7-11. Figure 7-12 presents the airow around the Cessna Ce550 Citation II in side-slipping ight. Particle traces depict the ow around the bottom side of the fuselage; as is obvious, the right wing induces a downwash in front of it, i.e. the particle traces follow a path below the wing and fuselage, while the left wing induces an upwash in front of it, i.e. the particle lines follow a path over the wing. In this gure a speedbar is given as well, depicting a measure for the local airspeeds magnitude along particle traces. The results of these simulations were obtained using a panel method (linearized potential ow).
Flight Dynamics

312

Lateral Stability and Control Derivatives

The contribution of the vertical tailplane to C is usually not large for conventional aircraft, but may not be neglected for aircraft having a relatively large vertical tailplane. The magnitude of C v follows from the magnitude and the position of the point of action of the lateral force CY v . In the calculation of C v the choice of the reference frame is important, as will be shown below. In the study of stability, the so-called stability reference frame is commonly used. This is a system of aircraft body axes of which the XS -axis in the plane of symmetry is chosen in the direction of the undisturbed ow. In this situation the rolling moment due to the side force on the vertical tailplane depends on the angle of attack of the aircraft. From gure 7-15 follows for C C = CY , xv xc.g. zv zc.g. cos 0 sin 0 b b (7-15)

the ordinates xv and zv can be determined using the same references as indicated for the determination of CYv . The eective dihedral of a propeller driven aircraft having a propeller in front of the wing can experience an important change due to the interaction between the slipstream and the wing. Due to the increased dynamic pressure in the slipstream an extra rolling moment is generated in sideslipping ight, see gure 7-16. At a positive side-slip angle this rolling moment acts in the right wing down sense, thus giving a positive contribution to C , The contribution increases with Tc and CL . Deection of the landing aps will increase the eective angle of attack over the center part of the wing. It will further increase the eect of the slipstream on C . Using linearized potential ow theory (panel methods), C as a function of has been calculated for the Cessna Ce550 Citation II, see gure 7-14. The denition of the aircraft parts wing, horizontal stabilizer, pylons, fuselage, vertical n and nacelles is given in gure 6-2. The main contributions to C are from the vertical n and the wing.

7-2-3

Stability derivative Cn

The stability derivative Cn is called the static directional, or Weathercock, stability, as this derivative is geometrically directly comparable to the static longitudinal stability Cm For good control characteristics it is desirable that the angle of sideslip causes a moment about the top axis trying to reduce the sideslip. The desired sign of Cn is thus positive. The magnitude of Cn is determined by a small, positive contribution of the wing, an important destabilizing, negative contribution of the fuselage and a usually large stabilizing, positive contribution of the vertical tailplane, see gure 7-17. The contribution of the propulsion

7-2 Stability derivatives with respect to the sideslip angle

313

Right wing

Figure 7-10: High-wing-fuselage conguration with the calculated velocity eld in sideslipping ight ( = 10o , = 0o )

Flight Dynamics

314

Lateral Stability and Control Derivatives

Right wing

Figure 7-11: Low-wing-fuselage conguration with the calculated velocity eld in sideslipping ight ( = 10o , = 0o )

7-2 Stability derivatives with respect to the sideslip angle

315

55.0 m/s 54.5 m/s 54.0 m/s 53.5 m/s 53.0 m/s 52.5 m/s 52.0 m/s 51.5 m/s 51.0 m/s 50.5 m/s 50.0 m/s 55.0 m/s 54.5 m/s 54.0 m/s 53.5 m/s 53.0 m/s 52.5 m/s 52.0 m/s 51.5 m/s 51.0 m/s 50.5 m/s 50.0 m/s
Figure 7-12: Calculated particle traces around the bottom half of the fuselage of the Cessna Ce550 Citation II, sideslipping ight = 10o , = 0o (note: the speedbar is related to the magnitude of the airows velocity along the calculated particle traces)

Flight Dynamics

316

Lateral Stability and Control Derivatives

c c left wing right wing V sin

b 2 b +2

c (< 0) c c

(A) High-wing aircraft

V sin
b 2 b +2

left wing

right wing

c (> 0)

(B) Low-wing aircraft

Figure 7-13: Origin of a rolling moment caused by wing-fuselage interactions in sideslipping ight

0.02

0.015

0.01

0.005

wing horizontal stabilizer pylons fuselage vertical fin nacelles total

0.005

0.01

0.015

0.02 10

[Deg.]

10

Figure 7-14: Aerodynamic moment coecient C as a function of for the Cessna Ce550 Citation II, = 0o

7-2 Stability derivatives with respect to the sideslip angle

317

Zr

YB,S

(CY )v XB,S 0 c.g.

zc.g. xc.g.

zv

Xr xv

ZB,S

(zv zc.g. ) sin0 (CY )v (xv xc.g. ) cos0 c.g. 0 (zv zc.g. ) cos0 (xv xc.g. ) sin0 xv xc.g. (zv zc.g. )
Figure 7-15: The position of the point of action of (CYv )v relative to the X- and Z-axis in the stability Flight Dynamics reference frame

a.c.v

318 system may be either positive or negative.

Lateral Stability and Control Derivatives

The contribution to Cn of a wing without sweep is very small. The contribution of a swept back wing may be important especially at high CL values due to the sideforce acting relatively far behind the center of gravity. This wing contribution may be calculated using the same references as for the determination of the contribution of the wing to CY . The negative destabilizing contribution of the fuselage arises in a similar manner as the corresponding distribution of the fuselage to Cm in symmetric ight, see references [114, 2, 76, 3, 139, 22, 128]. The contribution of the vertical tailplane to Cn , expressed in the stability reference frame, follows from gure 7-15, Cn = CY
v

xv xc.g. zv zc.g. sin 0 + cos 0 b b

(7-16)

In this expression CY

is given by equation (7-8).

At small angles of attack with xv xc.g. = v follows from equations (7-8) and (7-16), Cn
v

= CYv

d d

Vv V

Sv lv Sb

(7-17)

The corresponding expression for the horizontal tailplane is, see equation (8-7), (Cm )h = CNh 1 d d Vh V
2

Sh lh S c

For the sake of simplicity lv is often taken as the distance between the 1 -chord points on the 4 m.a.c.s of the wing and the vertical tailplane. Calculation methods to determine CYv have been mentioned already in section 7-2. The calculation of the sidewash gives rise to consirable diculties in practice as the ow is strongly determined by the interaction between ow around the fuselage, the wing and the tailplane. The ow around an isolated fuselage can be presented schematically as in gure 7-18. A positive angle of sideslip causes a negative induced sidewash, or d < 0. d Another numerical example of the ow around a complete aircraft conguration in side-slip is given in gure 7-19. It is obvious from this gure that the determination of at the vertical tailplane will be dicult because of strong interactions between the fuselage, vertical tailplane and nacelles. Also in this gure speedbars have been added to give an impression of the magnitude of the local airspeed along the calculated particle traces. Note that the characteristic ow over the top of a fuselage in sideslip is given in gure 7-18, see also the examples given in gures 7-20 where particle traces have been given over the top of a fuselage of a large 4-engined transport aircraft in sideslip.

7-2 Stability derivatives with respect to the sideslip angle

319

V XB

YB

(A) The part of the wing submerged in the slipstream c c

propeller on

propeller o y

b 2

c (> 0) (B) The change in lift distribution

Figure 7-16: The contribution of the slipstream to C

b +2

Flight Dynamics

320

Lateral Stability and Control Derivatives

Cn
0.06

f = 0o
0.04

= 2.5o

0.02

wing+fuselage fuselage
-16 -12 -8 -4 0 4 8 12

[o ]

wing
-0.02

complete model

-0.04

-0.06

Figure 7-17: Cn as a function of measured on a model of the Fokker F-27 in gliding ight (from references [148, 150, 24])

7-2 Stability derivatives with respect to the sideslip angle V sin ( ) V sin

321

(A) Velocity components in a plane perpendicular to the fuselage axes

(> 0) V cos V sin (< 0) V cos V sin ( )

(B) The ow around the fuselage and the vertical tailplane

Figure 7-18: The sidewash induced by the fuselage at the vertical tailplane
Flight Dynamics

322

Lateral Stability and Control Derivatives

55.0 m/s 54.5 m/s 54.0 m/s 53.5 m/s 53.0 m/s 52.5 m/s 52.0 m/s 51.5 m/s 51.0 m/s 50.5 m/s 50.0 m/s 55.0 m/s 54.5 m/s 54.0 m/s 53.5 m/s 53.0 m/s 52.5 m/s 52.0 m/s 51.5 m/s 51.0 m/s 50.5 m/s 50.0 m/s
Figure 7-19: Calculated particle traces around the fuselage of the Cessna Ce550 Citation II, sideslipping ight = 10o , = 0o (note: the speedbar is related to the magnitude of the airows velocity along the calculated particle traces)

7-2 Stability derivatives with respect to the sideslip angle

323

55.0 m/s 54.5 m/s 54.0 m/s 53.5 m/s 53.0 m/s 52.5 m/s 52.0 m/s 51.5 m/s 51.0 m/s 50.5 m/s 50.0 m/s 55.0 m/s 54.5 m/s 54.0 m/s 53.5 m/s 53.0 m/s 52.5 m/s 52.0 m/s 51.5 m/s 51.0 m/s 50.5 m/s 50.0 m/s
Flight sideslipFigure 7-20: Calculated particle traces around the fuselage of a large 4-engined transport aircraft, Dynamics ping ight = 10o , = 0o (note: the speedbar is related to the magnitude of the airows velocity along the calculated particle traces)

324

Lateral Stability and Control Derivatives

Earlier in section 7-2 the change in lift distribution in sideslip due to the presence of the fuselage was discussed. The extra lift distribution caused by the interaction also induces a side-wash at the location of the vertical tailplane. The dierence in lift on the two halves of the wing causes a dierence between the downwash at either side of the fuselage, see subgures a and b of gure 7-21. The downwash pattern is displaced laterally along with the main ow, over an angle with the X-axis, see subgure b of gure 7-21. The main eect, however, is a circulation about the fuselage. If a low wing aircraft sideslips to the right, the circulation is counter-clockwise, subgure c of gure 7-21. This circulation generates at the vertical tailplane an extra negative sidewash. This induced cross-ow is stabilizing, as can also be seen from equation (7-17), d < 0. d For a high wing aircraft in sideslip to the right, the sidewash is positive, the inuence of the wing fuselage interaction on Cn v is thus seen to be destabilizing, d > 0. d Figure 7-22 shows the results of measurements of the sidewash and the static directional stability illustrating the above. From this gure it can also be seen that the sidewash is greatly inuenced by the presence of the horizontal tailplane. A low horizontal tailplane acts as an end plate to the vertical tailplane. This reduces the sidewash considerably. An accurate calculation of the average sidewash , or d , at the vertical tailplane is not well d possible, although references [78, 80, 81, 82, 90, 28] give calculation methods. In many cases results of systematic measurements on similar congurations are used to determine d . d The contribution of the propulsion system to Cn is caused by the lateral force acting on the propeller or the engine inlet in cross-ow, see gure 7-23. If the propeller or the engine inlet is situated forward of the center of gravity this contribution is destabilizing. For propeller-driven aircraft the increased dynamic pressure in the slipstream causes a higher static directional stability, if the vertical tailplane is placed in the slipstream. Using linearized potential ow theory (panel methods), Cn as a function of has been calculated for the Cessna Ce550 Citation II, see gure 7-24. The denition of the aircraft parts wing, horizontal stabilizer, pylons, fuselage, vertical n and nacelles is given in gure 6-2. The main contributions to Cn are from the fuselage (destabilizing contribution) and the vertical n (stabilizing contribution).

7-2-4

Stability derivative Cn

The stability derivative Cn is a measure of the moment about the top axis caused during a change in the angle of sideslip when the aircraft executes an accelerated motion along the Y -axis. The derivative Cn corresponds entirely to the stability derivative Cm for the symmetric motions. After a sudden change in the angle of sideslip a short interval passes before the changed sidewash, caused by the wing-fuselage interaction, has reached the vertical tailplane. As a

7-2 Stability derivatives with respect to the sideslip angle

325

V sin YB

ZB > 0o c c = 0o

b 2

b +2

(A) The lift distribution of a low-wing aircraft sideslipping to the right

b 2

> 0o

b +2

= 0o

(B) The change in downwash behind the wing and fuselage Due to

Due to

V sin ( )

Circulation (C) The change in sidewash at the vertical tailplane

Figure 7-21: The eect of wing-fuselage interactions on the sidewash at the vertical tailplane of a low-wing aircraft in sideslipping ight
Flight Dynamics

326

Lateral Stability and Control Derivatives

Figure 7-22: The eect of wing-fuselage interactions on the sidewash at the tailplanes and the derivative Cn (from reference [142])

7-2 Stability derivatives with respect to the sideslip angle XB

327

CY (< 0) YB c.g. Cn (< 0)

Figure 7-23: The contribution of the sideforce on the propeller to Cn

0.025

0.02

0.015

0.01

wing horizontal stabilizer pylons fuselage vertical fin nacelles total

0.005

Cn

0.005

0.01

0.015

0.02

0.025 10

[Deg.]

10

Figure 7-24: Aerodynamic moment coecient Cn as a function of for the Cessna Ce550 Citation II, = 0o

Flight Dynamics

328

Lateral Stability and Control Derivatives

Figure 7-25: The pitching moment as a function of sideslip angle (from reference [126])

consequence, the change of Cn with does not occur as sudden as the change in . This phenomenon is described using the stability derivative Cn . For aircraft with a straight wing of suciently large aspect ratio (e.g. A > 4 to 5), Cn is usually neglected. Also in the discussion of the dynamic lateral stability, Cn will not be considered. For aircraft having swept wings of low aspect ratio, Cn has to be determined using the results of systematic measurements on oscillating models, see for instance reference [55].

7-2-5

Stability derivative Cm 2

A sideslip can cause a pitching moment that may be considerable, especially at large angles of sideslip, see gure 7-25. It can be seen from this gure, as follows also from considerations of symmetry, that the change in pitching moment has the same sign for both positive and

7-3 Stability derivatives with respect to roll rate

329

negative angles of sideslip. The stability derivative expressing the pitching moment due to sideslip is, therefore, written as Cm 2 . This moment is caused by the wing-fuselage interaction and the interaction between the fuselage and the tailplanes already discussed in relation to Cn . Reference [79] discusses the inuence of wing-fuselage interactions on the pitching moment for combinations of a fuselage with a straight wing and a swept wing in a sideslip. Here also the vertical position of the wing relative to the fuselage is important. Especially at large angles of sideslip the change in pitching moment is determined to a large extent by the contribution of the horizontal tailplane, see references [126] and gure 7-25. The tailplane is located in the downwash eld behind the wing, modied by the wing-fuselage interaction. If the tailplane is mounted low on the fuselage the eld of ow is modied additionally by the fuselage and the vertical tailplane. The inuence of the vertical position of the horizontal tailplane on Cm in sideslip is presented in gure 7-26. For a quantitative determination of Cm 2 the designer has to rely on wind tunnel measurements and/or CFD simulations.

7-3

Stability derivatives with respect to roll rate

If the aircraft rolls about the XB -axis, the geometric angle of attack of the various wing and tailplane chords varies proportional to the rolling-velocity p and to the distance of the chords to the XB -axis, see gure 7-27. For a wing chord at a distance y from the plane of symmetry, the change in the geometric angle of attack is, = pb y py = V 2V b/2 (7-18)

Here y is measured in the system of aircraft body axes. For the sake of simplicity the aircraft pb c.g. is assumed to lie in the plane of symmetry. The non-dimensional rolling velocity 2V is thus equal to the change in geometric angle of attack at the wing tip where y = b/2. This is also the helix angle of the helix described by the wing tip in rolling ight, which is why the b rolling velocity is made non-dimensional by multiplying with 2V .

7-3-1

Stability derivative CYp

The derivative CYp is always relatively small and is very often neglected. Only if the wing has a large sweep angle, or if a relatively large vertical tailplane is used, this derivative has to be taken into account. The origin of the contribution from a swept wing to CYp will be explained when discussing Cnp w . For swept back wings Cnp w is negative. The contribution of the vertical tailplane to CYp is caused by the change in angle of attack experienced by the vertical tailplane in rolling ight. A positive rolling velocity gives rise to a negative lateral force, Yv = CYp pb 1 2 V S 2V 2 (7-19)
Flight Dynamics

330

Lateral Stability and Control Derivatives

Figure 7-26: The inuence of the vertical position of the horizontal tailplane on the variation of Cm in sideslipping ight for a fuselage with tailplanes (from reference [126])

Figure 7-27: The variation of the local geometric angle of attack at the wing and the tailplane of a rolling aircraft

7-3 Stability derivatives with respect to roll rate

0.03

331

0.025

0.02

0.015

wing horizontal stabilizer pylons fuselage vertical fin nacelles total

0.01

CY

0.005

0.005

0.01

0.015

0.02 0.2

0.15

0.1

0.05

0pb

2V

[Rad.]

0.05

0.1

0.15

0.2

Figure 7-28: Aerodynamic force coecient CY as a function of = 0o

pb 2V

for the Cessna Ce550 Citation II,

Where CYp v is thus negative. In the determination of CYp v the sidewash at the vertical tailplane, induced by the rolling velocity has to be taken into account. This subject will be further discussed when dealing with Cp . Calculation methods can be found in references [89, 31, 126, 106, 107].
pb Using linearized potential ow theory (panel methods), CY as a function of 2V has been calculated for the Cessna Ce550 Citation II, see gure 7-28. The denition of the aircraft parts wing, horizontal stabilizer, pylons, fuselage, vertical n and nacelles is given in gure 6-2. The main contributions to CYp are from the wing, vertical n and horizontal tailplane. The stability derivative CYp is usually very snall.

7-3-2

Stability derivative Cp

The stability derivative Cp is a measure of the moment about the XB -axis due to rolling about this axis, L = Cp pb 1 2 V Sb 2V 2 (7-20)

In all normal ight conditions this moment opposes the rolling, trying to slow the motion down. Cp is normally negative and a measure for the roll damping.

Flight Dynamics

332

Lateral Stability and Control Derivatives

pb + 2V b +2

b 2

y
pb 2V

(A) The variation in geometric angle of attack e p

pb + 2V b +2

b 2

y
pb 2V

(B) The variation in eective angle of attack

Figure 7-29: The variation of the local geometric and eective angle of attack along the span of a rolling wing

7-3 Stability derivatives with respect to roll rate

333

Cp
0 2 4 6 8 10

A 60o

-0.4

45o 30o 0o

-0.8

Cp

(A) taper ratio = 1.0 A

0 2 4 6 8 10

60o
-0.4

45o 30o 0o

-0.8

Cp

(B) taper ratio = 0.5 A

0 2 4 6 8 10

60o
-0.4

45o 30o 0o

-0.8

(C) taper ratio = 0

Figure 7-30: The roll-damping as a function of aspect ratio and sweep for various wing taper ratios (from reference [34])

Flight Dynamics

334

Lateral Stability and Control Derivatives

The dominant contribution to Cp is generated by the wing. In some cases, the contribution from the tailplanes has to be taken into account as well. The contribution due to the fuselage, the wing-fuselage interaction and the propeller are usually negligible. The contribution of the wing is caused by the change in geometric angle of attack and the resulting change in eective angle of attack along the wing span, see gure 7-29. An additional anti-symmetric lift distribution is generated causing a rolling moment opposite to the sense of the rotation. This has a damping eect on the rotation. Calculated values of Cp for various wing planforms have been collected in gure 7-30. It can be seen that the roll damping of the wing increases in the absolute sense with the taper ratio, , and aspect ratio, A. The damping decreases with increasing wing sweep. If the airow over the wing remains attached, the wing contribution to Cp is independent of CL but proportional to the lift gradient CL . If the lift gradient decreases at large angles of attack due to ow separation, Cp decreases in the absolute sense as well. At very large angles of attack the down going wing may stall, causing a considerable reduction in wing damping. It is even possible that the inuence of the loss in lift caused by the stall of the down going wing dominates the inuence of the decrease in lift due to smaller angle of attack of the up going wing. In that case Cp may become positive. The lift distribution in rolling ight can be calculated using references [25, 44]. At subsonic speeds the wing contribution to Cp can be determined with references [10, 70, 156, 124, 125, 62]. Reference [110] enables the wing contribution to be determined for angles of attack where the relation between c and of airfoils is non-linear. For supersonic speeds references [10, 70, 84, 83, 67] are available. The contribution of the horizontal and the vertical tailplane to Cp , caused by the same mechanism as the wing contribution, see gure 7-27, is generally much smaller than the wing contribution. Due to the smaller span of the tailplanes the maximum change in angle of attack is smaller than that of the wing. A further decrease in the change of angle of attack at the horizontal and vertical tailplanes is caused by the change in downwash behind the rolling wing. Behind the downgoing wing the downwash increases, it decreases behind the upgoing wing, relative to non-rolling ight, see gure 7-31. This extra downwash distribution causes a circulation as depicted in subgure c of gure 7-31. This downwash decreases the change in angle of attack along the span of the tailplanes. A further study of this phenomenon can be found in reference [89]. For conventional, subsonic aircraft the contributions of the tailplanes to Cp can usually be neglected. Modern ghter aircraft and V/STOL-aircraft usually have relatively large tailplanes. Calculation of the tailplane contributions to Cp for such aircraft is possible using references [131, 31, 29, 106, 107].
pb Using linearized potential ow theory (panel methods), C as a function of 2V has been calculated for the Cessna Ce550 Citation II, see gure 7-32. The denition of the aircraft parts wing, horizontal stabilizer, pylons, fuselage, vertical n and nacelles is given in

7-3 Stability derivatives with respect to roll rate

335

YB

ZB resultant c c-distribution c c

due to due to
b 2 pb 2V

b +2

(A) The change in lift distribution on a rolling wing wi pb due to 2V

b 2

b +2

due to resultant downwash

(B) The change in downwash behind a rolling wing due to

pb 2V

due to

due to due to p due to

pb 2V pb 2V

circulation due to (C) The ow at the horizontal and vertical tailplanes due to p and
Figure 7-31: The ow at the tailplanes of a rolling aircraft
Flight Dynamics

336
0.1

Lateral Stability and Control Derivatives

0.08

0.06

0.04

wing horizontal stabilizer pylons fuselage vertical fin nacelles total

0.02

0.02

0.04

0.06

0.08

0.1 0.2

0.15

0.1

0.05

0pb

2V

[Rad.]

0.05

0.1

0.15

0.2

Figure 7-32: Aerodynamic moment coecient C as a function of = 0o

pb 2V

for the Cessna Ce550 Citation II,

gure 6-2. The main contributions to Cp is from the wing. The wing tries to damp the rolling motion.

7-3-3

Stability derivative Cnp

This derivative is determined mainly by the contribution of the wing. Only a relatively large vertical tailplane produces a non-negligible contribution to Cnp . Under normal conditions Cnp is negative. The generation of the wings contribution to Cnp is illustrated in gure 7-33. In rolling ight an extra force in XB -direction acts on each section of the wing. Since the local angle of attack varies along the wing span, it is essential to study the distribution of this extra force in the stability reference frame. The axes of the reference frame have a xed direction relative to the aircraft. In this reference frame the down going wing experiences an extra forward force due to the increased angle of attack. The up going wing experiences an extra longitudinal force pointing backwards. As a result, a positive rolling velocity causes a negative yawing moment, N = Cnp where Cnp
w

pb 1 2 V Sb 2V 2

(7-21)

< 0.

7-3 Stability derivatives with respect to roll rate

337

Figure 7-33: The origin of a negative yawing moment about the Z-axis of the stability reference frame for a wing having a positive rate of roll (attached ow)

Flight Dynamics

338 XB

Lateral Stability and Control Derivatives

CXn
CXn

CY

CY0 Cn

CYn

CX0
CX0

Figure 7-34: The side force and yawing moment on a rolling, swept back wing

If ow separation occurs on the down going wing, the drag increases considerably causing the extra force in XB - or XS -direction to become negative. This renders Cnp w less negative or even positive. This applies also to wings with a sharp leading edge. Due to the absence of the suction peak on the airfoils nose, the resultant aerodynamic force and the change in the resultant aerodynamic force act here approximately perpendicular to the wing chord. From subgure b of gure 7-33 it can be seen that for normal wings the extra force in XS direction on the down going wing is positive (pointing forwards) and the extra force on the up going wing acts in the direction of the negative XS direction. For normal wings Cnp w apparently is always negative. For swept wings the change in resultant aerodynamic force acting on a section of the rolling wing has a component in the YB - or YS -direction, see gure 7-34. This lateral force has pb the same direction on both wing halves. This causes a resultant lateral force CYp w 2V . This force generates a second, less important contribution to Cnp w , depending on the c.g. position. Calculation methods to determine CYp w and Cnp w for subsonic speeds are given in references [10, 70, 122, 156, 64]. For supersonic speeds, see references [84, 83]. The contribution of the vertical tailplane to Cnp follows directly from CYp and the position

7-4 Stability derivatives with respect to yaw rate

3

339

x 10

wing horizontal stabilizer pylons fuselage vertical fin nacelles total

Cn

4 0.2

0.15

0.1

0.05

0pb

2V

[Rad.]

0.05

0.1

0.15

0.2

Figure 7-35: Aerodynamic moment coecient Cn as a function of = 0o

pb 2V

for the Cessna Ce550 Citation II,

of the point of action of CYp v using an expression in analogy with equation (7-16). The position of the point of action CYp v can be determined using the same references as given for the determination of the magnitude of CYp w .
pb Using linearized potential ow theory (panel methods), Cn as a function of 2V has been calculated for the Cessna Ce550 Citation II, see gure 7-35. The denition of the aircraft parts wing, horizontal stabilizer, pylons, fuselage, vertical n and nacelles is given in gure 6-2. The main contributions to Cnp are from the vertical n, wing and horizontal stabilizer.

7-4

Stability derivatives with respect to yaw rate

Before going into a discussion of the derivatives with respect to yawing velocity, CYr , Cr and Cnr separately, the motion to be considered needs further description. The notion of a so called r-motion is introduced here, in analogy with the q-motion, see section 6-4. When performing such an r-motion, the aircraft moves along a curved trajectory in the XOY -plane, such that the velocity vector of the center of gravity remains in the plane of symmetry. By denition the angle of sideslip of the aircraft remains zero during aa r-motion, see gure 7-36. Apparently, the r-motion causes merely a curvature of the streamlines in the XOY -plane. At the position of the c.g. the radius of curvature is R,
Flight Dynamics

340

Lateral Stability and Control Derivatives

R=

V r

The center of rotation is situated at a distance R from the c.g. on the positive Y -axis if r is positive. The eect of this r-motion is twofold. In the rst place, and in analogy with the change in angle of attack caused by the q-motion, a change in the ow direction measured in planes parallel to the XOY -plane occurs at all points of the aircraft, see gure 7-37, rb x xc.g. 2V b/2

(7-22)

As a consequence the airow meets the vertical tailplane from the left if r is positive. In the second place, the local speed of the airow changes at all points of the aircraft by an amount, see gure 7-38, V rb y = V 2V b/2

(7-23)

As in equation (7-18), y is measured here in the stability reference frame. At positive r, the right wing has a lower airspeed and the left wing a larger airspeed than the aircraft center of gravity. Similar dierences in airspeed occur in the plane of symmetry during the q-motion, but they are negligible due to the relatively small dimensions of the aircraft in Z-direction.
rb According to equation (7-23), 2V is the non-dimensional change in airspeed occurring at the b wing tip. The choice of the factor 2V to make the angular velocity non-dimensional thus has a geometric interpretation.

7-4-1

Stability derivative CYr

This derivative is usually of minor importance. The dominant contribution to CYr is due to the vertical tailplane, CYr is commonly positive. The contribution of the wing is very small. Calculation methods are given in references [10, 70, 122, 156] for subsonic speeds and in references [84, 83] for supersonic speeds. The contributions of the fuselage and the wing-fuselage interaction are negligible, as is the contribution of the-horizontal tailplane. The lateral force on the vertical tailplane due to yawing can be written as, rb 1 2 V S 2V 2

Yv = (CYr )v

(7-24)

7-4 Stability derivatives with respect to yaw rate XB

V r

341

R=

V YB

c.g.

Figure 7-36: The pure r- motion

The lateral force is caused by the change in angle of attack v of the vertical tailplane, see gure 7-39. Accordingly, 1 2 V Sv 2 v

Yv = CYv v

(7-25)

The change in angle of attack v follows from equation (7-22) with xv xc.g. = lv , v = rb lv 2V b/2 (7-26)

From equations (7-24), (7-25) and (7-26) it follows after some reduction, Vv V
2

(CYr )v = 2 CYv

Sv lv Sb

(7-27)

From this expression it follows that (CYr )v is positive. The values of (CYr )v calculated using equation (7-27) may not be very accurate due to the unknown inuence of the wing-fuselage in the curved ow eld on the ow direction at the vertical tailplane. For these reasons it is commonly necessary to use the results of wind tunnel measurements on similar congurations to determine (CYr )v .

Flight Dynamics

342

Lateral Stability and Control Derivatives

Yr

XB

xc.g. r x c.g. Xr =
rb 2V

YB

xxc.g. b/2

Figure 7-37: The change in geometric ow direction at an arbitrary point of the aircraft due to an rmotion

XB

V V c.g. y

V r

YB

R=

V r

V V

rb = 2V

y b/2

Figure 7-38: The variation in airspeed in spanwise direction due to an r-motion

7-4 Stability derivatives with respect to yaw rate XB p (< 0)

343

Yp (< 0) xp xc.g.

V r c.g. YB

lv v (> 0) Yv (> 0)

Figure 7-39: The side forces on the vertical tailplane and the propeller due to an r-motion.

During the r-motion the propeller experiences an oblique inow also causing a contribution to CYr , see gure 7-39. In analogy with equation (7-27), (CYr )p can be written as, Vp V
2

(CYr )p = 2 CYp

Sv xp xc.g. S b
Vp V 2

(7-28)

The derivative (CYr )p usually is negative. CYp and [133, 134].

can be determined using references

rb Using linearized potential ow theory (panel methods), CY as a function of 2V has been calculated for the Cessna Ce550 Citation II, see gure 7-40. The denition of the aircraft parts wing, horizontal stabilizer, pylons, fuselage, vertical n and nacelles is given in gure 6-2. The main contributions to CYr are from the vertical n, the fuselage and the wing.

7-4-2

Stability derivative Cr

The stability derivative Cr is determined primarily by the contributions of the wing and the vertical tailplane. The contributions of the fuselage, the wing-fuselage interaction, the horizontal tailplane and the propulsion system may usually be neglected. Cr is always positive.

Flight Dynamics

344
0.06

Lateral Stability and Control Derivatives

0.04

0.02

wing horizontal stabilizer pylons fuselage vertical fin nacelles total

CY

0.02

0.04

0.06 0.2

0.15

0.1

0.05

rb 2V

[Rad.]

0.05

0.1

0.15

0.2

Figure 7-40: Aerodynamic force coecient CY as a function of = 0o

rb 2V

for the Cessna Ce550 Citation II,

The contribution of the wing arises due to the dierences in airspeed over the two wing halves during the r-motion, as described previously. At a positive yawing velocity, i.e. a rotation to the right, the left wing generates more lift than the right wing, causing a right wing down, positive, rolling moment, rb 1 2 V Sb 2V 2

L = (Cr )w

(7-29)

This rolling moment, and thus (Cr )w , is proportional to the lift coecient and at a given CL dependent on the lift distribution over the wing span. Tapered wings show a higher lift concentration near the center of the wing, causing an increase in speed over the outer wing to have less eect. As a consequence, (Cr )w decreases with at constant CL . The same applies to wings having negative twist and to wings with deected landing aps. Swept back wings have a higher lift concentration on the outer wings. This causes (Cr )w to increase somewhat with . The derivative (Cr )w can be calculated using references [10, 70, 122, 156, 33] for subsonic speeds and references [84, 83] for supersonic speeds. The contribution of the vertical tailplane to Cr is found from (CYr )v and the position of the point of action of (CYr )v , (Cr )v = (CYr )v zv zc.g. xv xc.g. cos 0 sin 0 b b (7-30)

7-4 Stability derivatives with respect to yaw rate

0.015

345

0.01

0.005

wing horizontal stabilizer pylons fuselage vertical fin nacelles total

0.005

0.01

0.015 0.2

0.15

0.1

0.05

rb 2V

[Rad.]
rb 2V

0.05

0.1

0.15

0.2

Figure 7-41: Aerodynamic moment coecient C as a function of = 0o

for the Cessna Ce550 Citation II,

As zv zc.g. is usually positive, (Cr )v is usually positive as well.

rb Using linearized potential ow theory (panel methods), C as a function of 2V has been calculated for the Cessna Ce550 Citation II, see gure 7-41. The denition of the aircraft parts wing, horizontal stabilizer, pylons, fuselage, vertical n and nacelles is given in gure 6-2. The main contributions to Cr are from the vertical n and the wing.

7-4-3

Stability derivative Cnr

Similar to the derivatives Cmq and Cp , Cnr is a measure of the aerodynamic moment about an axis caused by an angular velocity about the same axis. Under normal circumstances such a moment tries to slow down the motion, which implies that Cnr is normally negative, again similar to Cmq and Cp . The dominant contribution to Cnr is provided by the vertical tailplane. Additionally, only the wing, the fuselage and the propulsion system deliver non-negligible contributions. The part provided by the wing is generated by the dierences in drag between the two halves of the wing, caused by the dierences in local airspeed. Calculation methods for subsonic speeds are found in references [10, 70, 122, 156] and in references [84, 83] for supersonic speeds. The contribution due to the fuselage is small, but it may be non-negligible for modern congurations having a relatively large fuselage in comparison with the wing. A calculation
Flight Dynamics

346

Lateral Stability and Control Derivatives

Figure 7-42: The eects of the size of the vertical tailplane and the taillength on Cnr (from reference [88])

is possible using references [139, 88].

The contribution of the vertical tailplane follows directly from, see equation (7-27) and the taillength of the aircraft, see also gure 7-39,

(Cnr )v = (CYr )v

lv b

(7-31)

For the determination of (Cnr )v in the stability reference frame, at large angles of attack use has to be made of an expression in analogy with equation (7-16). v Figure 7-42 gives Cnr of a model for dierent values of lb and Sbv . The contribution of the propeller to Cnr follows in a similar manner from (CYr )p . If the propeller is situated in front of the c.g., (CYr )p is negative. The damping moment is then increased by the propeller, see gure 7-39. Additionally, the increased dynamic pressure in the slipstream may cause an increase in the contribution of the vertical tailplane.

rb Using linearized potential ow theory (panel methods), Cn as a function of 2V has been calculated for the Cessna Ce550 Citation II, see gure 7-43. The denition of the aircraft parts wing, horizontal stabilizer, pylons, fuselage, vertical n and nacelles is given in gure 6-2. The main contributions to Cnr are from the vertical n and the fuselage.

7-5 The forces and moments due to aileron, rudder and spoiler deections
0.025

347

0.02

0.015

0.01

wing horizontal stabilizer pylons fuselage vertical fin nacelles total

0.005

Cn

0.005

0.01

0.015

0.02

0.025 0.2

0.15

0.1

0.05

rb 2V

[Rad.]
rb 2V

0.05

0.1

0.15

0.2

Figure 7-43: Aerodynamic moment coecient Cn as a function of = 0o

for the Cessna Ce550 Citation II,

7-5

The forces and moments due to aileron, rudder and spoiler deections

For lateral control it is necessary to generate variable moments about the XB -axis and the ZB -axis. The moment about the XB -axis is usually obtained by deecting ailerons, at high airspeeds the use of spoilers is also quite common. The moment about the ZB -axis is provided by the rudder. The deection of the ailerons or the spoilers generates in addition to the rolling moment L also a lateral force Y and a yawing moment N . Equally, rudder deections cause not only a yawing moment, but also a lateral force and a rolling moment. As is the case with the lateral motions of the entire aircraft, in principle no symmetric forces and moments are generated. Measures of the non-dimensional force CY and the nondimensional moments C and Cn due to control surface deections are again the partial derivatives with respect to aileron deection a and the rudder deection r , these control derivatives are, 1. aileron deection, CYa = CY a Ca = C a Cna = Cn a
Flight Dynamics

348 2. rudder deection, CYr = CY r Cr =

Lateral Stability and Control Derivatives

C r

Cnr =

Cn r

When using the control deections as measures for the forces and moments, in analogy with the stability derivatives, the assumption is made that the forces and moments vary linearly with a and r . For not too large control surface deections, this assumption is usually quite acceptable. Appendix B provides, in addition to stability derivatives, also the control derivatives of some aircraft. Figure 7-44 shows once more the positive direction of the control surface deections. They are always measured in a plane perpendicular to the hinge axes of the surfaces. The deection a is dened as, a = aright alef t (7-32)

A positive deection of the control stick, or a positive deection of the control wheel (to the left), can be seen to produce a positive a . Equally, a positive deection of the rudder pedals (left pedal forward) causes a positive r . If spoilers are used, derivatives with respect to spoiler deection s are dened, in analogy with the above control deections, CY s C s Cn s

CYs =

Cs =

Cns =

A spoiler deection on the left wing is taken as positive. A positive s corresponds with a positive deection of the roll control manipulator. When using the derivatives with respect to s great care is required, as for many spoiler congurations the forces and moments vary strongly non-linear with the spoiler deection. The control derivatives are discussed in the following. The use of spoilers for roll control is discussed in section 7-7.

7-6
7-6-1

Aileron control derivatives

Control derivative CYa

The lateral force caused by an aileron deection may be taken as zero for straight wings. On swept wings a lateral force does arise as is the case in rolling ight. The derivative CYa is very small in magnitude and is usually neglected.

7-6 Aileron control derivatives

349

Figure 7-44: The positive direction of control deections, control forces, control surface deections and hinge moments

Flight Dynamics

350 c c

Lateral Stability and Control Derivatives

wing only

wing with deected ailerons

ailerons only
b 2

c (< 0)

b +2

ailerons only ba ba

Figure 7-45: The eect of an aileron deection on the lift distribution in spanwise direction

7-6-2

Control derivative Ca

This control derivative is, of course, the primary one for the ailerons. Due to a positive aileron deection the lift on the right wing increases, on the left wing the lift decreases. The result is a rolling moment to the left, which is counted as negative, see gure 7-45, 1 2 V Sb 2

L = Ca a

(7-33)

The derivative Ca is negative and is also called the aileron eectivity. If the aeroelastic deformation of the wing is neglected, the inuence of the Mach number is not considered and no ow separation on the wing in the area of the ailerons occurs, then Ca may be considered as independent of CL . The magnitude of Ca depends strongly on the dimensions and the location of the ailerons on the wing. Wing sweep and taper ratio also have an important inuence on the aileron eectivity. For a given location and dimension of the ailerons, Ca decreases in the absolute sense with increasing wing sweep and taper ratio (lower ), especially when the wing aspect ratio is large. The calculation of Ca is possible, using references [10, 70, 155, 44, 99] For supersonic speeds Ca can be determined with reference [99]. In many instances measured values of Ca obtained from ight tests turn out to be smaller in magnitude than was expected on the basis of calculations.

7-6 Aileron control derivatives

351

a < 0

a = 0

a > 0

Figure 7-46: The Frise aileron

7-6-3

Control derivative Cna

The yawing due to aileron deection arises because the drag of the wing having the downward deected aileron increases, while that of the other side decreases, or perhaps increases less. As a consequence, a positive aileron deection causes a positive yawing moment, N = Cna a 1 2 V Sb 2 (7-34)

where Cna is positive. To a rst approximation Cna is proportional to CL . Too large values of Cna are undesirable. If, for instance, the pilot wants to initiate a left turn, he deects the roll control (stick or wheel) to the left to obtain the required negative roll-angle (left wing down). While the aircraft must obtain a negative yaw-rate (to the left), the positive aileron deection generates a positive yawing moment, because Cna is positive. This yawing moment causes the aircraft to yaw initially to the right. This eect is amplied by the usually positive yawing moment caused by the negative rolling velocity of the aircraft, Cnp was seen to be negative. This initial yaw opposite to the desired direction is usually called the adverse yaw of the aircraft. The derivative Cna can be kept small in various ways. One of the means is the application of dierential deections of the ailerons when applying roll control. This implies that the
Flight Dynamics

352

Lateral Stability and Control Derivatives

downward deection of the aileron is made smaller than the upward deection of the other aileron. The dierence in drag is reduced in this manner. An other method uses so called Frise ailerons. The shape of the nose of these ailerons is such that the drag of the aileron strongly increases if it is deected upward, see gure 7-46 and reference [155]. A calculation of Cna having some accuracy is generally not possible. As a consequence it is always desirable to determine Cna on the basis of experimentally obtained data, see reference [70].

7-7

Spoiler control derivatives

At transonic and supersonic airspeeds roll control of the aircraft can often not be eected by ailerons alone. In such cases spoilers, or a combination of spoilers and ailerons are used. The latter solution is the more usual one. A spoiler is nothing but a ap, deected from the upper surface on one side of the wing to spoil the airow over that part of the wing, see subgure a of gure 7-47. The local disturbance causes a decrease in lift and thereby generates a rolling moment. A deection of the spoiler on the left wing is counted as positive. This causes, like an upward deection of the left aileron, a negative rolling moment, hence Cns is negative as Cna is. Spoilers are more eective if they are located closer to the wing leading edge. The response of the aircraft to spoiler deection becomes slightly more sluggish in this way. Therefore, in practice spoilers are located mostly on the rear part of the wing chord. Various causes make the use of spoilers on high speed aircraft more or less mandatory. In the rst place, aileron eectivity decreases due to compressibility eects with increasing Mach number. On the other hand, the eectiveness of spoilers slightly increases with increasing Mach number. A second, even more important, phenomenon is the decrease in aileron eectivity due to wing deformation caused by aileron deection. Figure 7-48 shows how a downward deection of the aileron causes the wing to twist such that the local angle of attack decreases. This counteracts the increase in lift due to the aileron deection. Conversely, an upward deection of the aileron causes an increase of the local angle of attack. Such a decrease in aileron eectivity may be appreciable especially at high airspeeds, both due to the high dynamic pressure and to the rearward shift of the point of action of the local change in lift caused by the aileron deection with increasing Mach number. In this way the resultant eect of an aileron deection on highly swept wings with thin wing sections may even be opposite to the desired one. The eect is called control reversal and the airspeed from which onward this occurs is the reversal speed. The point of action of the resultant change in lift for spoilers lies more forward on the chord than for ailerons. Hence, the wing twist is less and as a result the reversal speed is higher.

7-7 Spoiler control derivatives

353

(A) Plate type spoiler

(B) Plugtype spoiler

Figure 7-47: Spoiler types

Flight Dynamics

354

Lateral Stability and Control Derivatives

Figure 7-48: The local wing twist of a wing cross-section due to elastic deformation caused by an aileron deection

Finally and quite apart from the foregoing reasons, it is often not very well possible to use large ailerons on wings having a high sweep angle or a low aspect ratio. The trailing edges of such wings are taken up largely by landing aps to obtain an acceptable low landing speed. A spoiler deection causes not only a reduction in lift, but also causes a strong increase in drag. A secondary advantage of the use of spoilers is, that due to this drag increase, Cns is nearly always negative, contrary to Cna . The adverse yaw at the initiation of a turn is thereby decreased or even entirely avoided. The increase in drag created by the deployment of spoilers may cause the designer to use the spoilers not only for roll control but as speed brakes as well. The spoilers are deected symmetrically for this purpose. Symmetric deection of spoilers situated in front of the landing aps, when applied directly after touch down, provides the possibility to drastically decrease the wing lift during the roll out. A larger part of the weight acts on the undercarriage making the wheel brakes more eective, thus yielding an appreciable reduction in the stopping distance. A disadvantage of spoilers is their low eectivity at small deections, caused by re-attachment of the airow behind the spoiler. So-called plug type spoilers, see subgure b of gure 7-47 have better characteristics in this respect due to the open connection between lower and upper surface of the wing caused by the spoiler deection. Re-attachment of the airow is avoided in this manner. At large angles of attack the eectivity of the spoiler may also be low, especially if ow separation occurs at the trailing edge of the wing. A further disadvantage during the approach and landing phases is the decrease in total wing lift caused by the operation of the spoilers. This renders the precise control of the aircrafts ight path slightly more dicult. The hinge moment of a plain ap type spoiler, see subgure a of gure 7-47, may be

7-8 Rudder control derivatives

355

appreciable, and it usually changes non-linearly with the deection angle. On the other hand, the hinge moment of a spoiler in the shape of a half circular plate, see subgure b of gure 7-47, may be next to negligible. However, the total aerodynamic load on a spoiler of this type may be relatively high. Cs and Cns can be determined using references [70]. Reference [57] gives theoretical calculation methods to determine Cs while reference [100] contains an extensive list of references on the use of spoilers.

7-8
7-8-1

Rudder control derivatives

Control derivative CYr

The rudder deection causes a lateral force, Y = CYr r 1 2 V S 2 (7-35)

A positive r gives rise to a positive lateral force, see gure 7-49. The control derivative CYr is positive. The lateral force can be written as, Y = CYv r 1 2 V Sv 2 v (7-36)

where CYv is the gradient of the normal, here lateral, force on the vertical tailplane. From equations (7-35) and (7-36) it follows, CYr = CYv Vv V
2

Sv S

(7-37)

The derivative CYv is determined in a similar way as the derivative CNh of the horizontal tailplane, see references [155, 43, 46, 126]. For supersonic speeds this derivative can be calculated with reference [99].

7-8-2

Control derivative Cr

A positive lateral force acting on the vertical tailplane, caused by a positive rudder deection, causes a positive rolling moment, L = Cr r 1 2 V Sb 2 (7-38)
v

where Cr is positive. In analogy with the contribution of the vertical tailplane C Cp v and (Cr )v , the control derivative Cr can be written as, Cr = CYr zv zc.g. xv xc.g. cos 0 sin 0 b b

(7-39)
Flight Dynamics

356

Lateral Stability and Control Derivatives

XB

Cnr r (< 0) YB c.g. lv

CYr r (> 0)

r (> 0)
Figure 7-49: The side force and yawing moment due to a rudder deection

7-8 Rudder control derivatives

357

Large positive values of Cr could be detrimental to good lateral control characteristics. In ight, when accurate lateral control is required, small deviations from the desired heading may be corrected by using rudder deections, without recourse to the ailerons. This may give a result more quickly than changing the aircrafts heading via a change in angle of roll generated by the ailerons. If in such a situation a positive rudder deection is used to yaw the aircraft to the left, a large positive Cr causes an angle of roll to the right. Such a roll angle is associated with a turn to the right and opposes the desired yawing motion of the aircraft. The eect just mentioned occurs mainly with aircraft having a highly mounted vertical tailplane, where zv zc.g. in equation (7-39) is large. In these aircraft the rudder and ailerons may be coupled mechanically such that a positive aileron deection is generated by a positive rudder deection. The moment Ca a then opposes the moment Cr r .

7-8-3

Control derivative Cnr

The primary eect of a rudder deection is the yawing moment, N = Cnr r 1 2 V Sb 2 (7-40)

The control derivative Cnr follows from gure 7-49 at small 0 and taking xv xc.g. = lv , Cnr = CYr lv b (7-41)

At large angles of attack, Cnr follows from an expression in analogy with equation (7-16) for Cn v . The control derivative Cnr is negative.

Flight Dynamics

358

Lateral Stability and Control Derivatives

Figure 7-50: Asymmetric stability and control derivatives

Chapter 8 Longitudinal Stability and Control in Steady Flight

8-1

Stick Fixed Static Longitudinal Stability

In chapter 5 we discussed the equilibrium of the aerodynamic and gravitational forces in the plane of symmetry and the aerodynamic moment about the center of gravity. The also showed that the horizontal tailplane and its elevator play a crucial rle in the equilibrium o of the aerodynamic moment about the center of gravity. For the equilibrium, the following equations (5-53), (5-54) and (5-55), were derived, CT = CTw = 1 CN = CNw + CNh Vh V W sin 2 2 V S
2

Sh W = 1 2 cos S 2 V S Vh V
2

Cm = Cmac + CNw

xcg xw + CNh c

Sh xcg xh =0 S c

For the equilibrium of the moments equation (5-57) was introduced, xcg xw CNh c Vh V
2

Cm = Cmac + CNw

Sh lh =0 S c

Equilibrium of forces and moments acting on an aircraft in ight is a necessary, but certainly not a sucient condition for steady ight. A steady ight condition must also be stable as pointed out for example already in reference [123]. This means that after the occurrence of a small disturbance the aircraft should return to the original state of equilibrium. If this is
Flight Dynamics

360

Longitudinal Stability and Control in Steady Flight

the case, the equilibrium is called dynamically stable. Although, as remarked in reference [123], stability is strictly a characteristic of an equilibrium, it is common practice to call the aircraft itself dynamically stable if its equilibrium is stable. The motions of aircraft after a disturbance, and along with it the dynamic stability, are determined by the equations of motion of the aircraft. A discussion of aircraft motions, using these equations of motion (see chapters 3 and 4), is postponed at this stage. There is, as in reference [53], a simpler way to study stability. Experience has shown, that most critical requirements to be satised for dynamic stability can be expressed for most aircraft by the condition that the aircraft must be statically stable. The discussion of static stability is based on the equations for the equilibrium of the aircraft, see equations (5-53), (5-54) and (5-57). These equations are in fact special cases of the more general equations of motion. When we discuss static longitudinal stability we refer to pitch attitude stability. So we just need to consider the expression for the longitudinal moment, equations (5-57) or (5-55) allowing us to study the change of aerodynamic moment caused by a change in angle of attack in a similar way as we did with the two dimensional wing prole, wings and wing-fuselage composites in section 5-1-3. Here we analize the static stability of complete aircraft including horizontal tail surfaces. New is also that we must consider static stability for the case of stick xed (xed elevator deection) and the case of stick free (oating elevator). It will be shown that a close relationship exists between the way in which the stick position varies with airspeed, or angle of attack, in steady, straight ight on the one side and the static stability, stick xed on the other. In a similar way it will be shown that static stability with stick free, is closely related to the way in which the stick force varies with airspeed in steady, straight ight. For simplicity , propulsion eects will be neglected and the dimensionless aerodynamic coecients are assumed to be independent of airspeed. We could readily take these eects into account, at the cost, however, of much more cumbersome derivations. It is interesting to note that steady ight conditions as discussed in the present chapter, are in fact just a special case of dynamic ight conditions described by equations of motion. This fact allows us to present an alternative way to derive static stability and control characteristics directly from the linearized equations of motion. The resulting expressions contain stability and control derivatives which do take account of things as propulsion eects, and so, are in a sense more general than those which were derived for the more simple case of the model above. We present these derivations from the linearized equations of motion in boxes under the heading of Quick derivation of .... following the more elaborate analytical derivations based on the simplied model above.

8-1-1

Stick xed static longitudinal stability in gliding ight

In section 5-1-3 the concept of static stability was presented for two dimensional wing proles, wings and wing-fuselage composites. This discussion is now extended to a complete aircraft. Suppose the aircraft is in steady ight, so the moment about the center of gravity is zero, Cm = 0. If some external disturbance would cause an increase of the pitch angle and

8-1 Stick Fixed Static Longitudinal Stability

361

so of the angle of attack, we would like to see a negative, nose-down change in Cm , a nose down, restoring moment. The condition for static longitudinal stability then is, dCm = Cm < 0, at Cm = 0 d (8-1)

If Cm = 0 the static stability is called neutral. If Cm > 0 the aircraft is statically unstable. To study the various contributions to the derivative Cm of the complete aircraft, the expression for the moment coecient is reconsidered as in equation (5-72) Cm = Cmac + CNw ( 0 ) CNh ( 0 ) 1 d d xcg xw c Vh V
2

+ (0 + ih ) + CNh e
e

Sh lh S c

(8-2)

Equation (8-2) can be rearranged to combine the contributions of the wing with fuselage and nacelles (indicated by the index w) and the contribution of the horizontal tailplane (indicated by index h), Cm = Cmw + Cmh with, Cmw = Cmac + CNw ( 0 ) and, xcg xw c (8-4) (8-3)

Cmh = CNh

( 0 ) 1

d d
h CNh

+ (0 + ih ) +CNh e e

Vh V

Sh lh S c

(8-5)

Figures 8-1 schematically show the Cm -curves of the contributions according to equations (8-4) and (8-5) and of the complete aircraft conguration according to (8-2). Figure 8-2 presents the contributions of the various parts of the aircraft to the Cm -curve as measured in a wind tunnel on a model of the Fokker F-27. Figure 8-3 presents calculated moment curves for several parts of the Cessna Ce550 Citation II. This data was calculated using an inviscid panel method (linearized potential ow). The denition of the aircraft parts is presented in gure 6-2; these parts are: wing, horizontal stabilizer, pylon, nacelles, vertical n and fuselage (in clockwise direction, starting from the left top gure). The results presented
Flight Dynamics

362

Longitudinal Stability and Control in Steady Flight

Cmw

Cmh (A) Wing-fuselage contribution

Cm (B) Horizontal tailplane contribution

Equilibrium steady ight

(C) Cm = Cmh + Cmw

Figure 8-1: The contribution of various parts of the aircraft to the moment curve Cm

8-1 Stick Fixed Static Longitudinal Stability Cm

363

0.2

model without tailplane contribution of fuselage and nacelles

16

0 -4 0 4 8

wing
12

[o ]

complete model
-0.2

contribution of horizontal tailplane

-0.4

-0.6

Figure 8-2: Measured contributions of various aircraft parts to the moment curve of the Fokker F- 27, reference point at 0.346 c (from reference [147])

in gure 8-3 clearly show the destabilizing eect of the fuselage and the stabilizing eect of the horizontal tailplane, while the contribution of the wing and nacelles is almost indierent. The slopes of the moment curves follow from equations (8-4) and (8-5) by dierentiating with respect to the angle . Because the stick xed situation is considered, the elevator angle e is taken constant, Cmw = CNw xcg xw c d d Vh V
2

(8-6) Sh lh S c

Cmh = CNh

(8-7)

Because the cg usually lies behind the ac of the combination of a wing with fuselage and nacelles, the derivative Cmw is positive, see equation (8-6). This means, that even if the aircraft without the horizontal tailplane could be at equilibrium at all (Cmw = 0), the equilibrium would be unstable, see also section 5-1-3. The contribution of the tailplane has a stabilizing eect, according to equation (8-7), (Cm )h < 0. This contribution must be large enough, that Cm is, see equation (5-75), Cm = CNw is negative,
Flight Dynamics

xcg xw CNh c

d d

Vh V

Sh lh S c

364
0.25

Longitudinal Stability and Control in Steady Flight

0.2

0.15

0.1

wing horizontal stabilizer pylons fuselage vertical fin nacelles total

Cm

0.05

0.05

0.1

0.15

0.2 10

[Deg.]

10

Figure 8-3: Calculated contributions of various aircraft parts to the moment curve of the Cessna Ce550 Citation II

Cm < 0 The important role of the horizontal tailplane for obtaining static stability is obvious, as already discussed in section 5-2-1. The horizontal tailplane is also essential to obtain equilibrium of the longitudinal moment. As can be seen from a number of measured moment curves in gures 8-4 and 8-5, Cm varies almost linearly with over a large range of angles of attack. But at large values of , Cm is no longer constant. This is due to the contribution of CT to the longitudinal moment and to the non-linearity caused by ow separation. Strictly speaking, static stability is the slope of the moment curve Cm , only at Cm = 0. By varying ih and e the moment curve can be shifted up and down. If the angle of attack is not too large, the shifted moment curves are all parallel, see gure 8-4. A change in ih inuences Cm according to the expression in equation (5-73), Cm = Cm0 + Cm ( 0 ) + Cme e = 0 only by changes in Cm0 , which is independent of , see equation (5-74) Cm0 = Cmac CNh (0 + ih ) Vh V
2

Sh lh S c

8-1 Stick Fixed Static Longitudinal Stability Cm

365

0.2

ih = 2o ih = 0o

0 -4 0 4 8 12 16

[o ]

-0.2

ih = 2o
-0.4

-0.6

Figure 8-4: Inuence of the tailplane incidence on the moment curve of the Fokker F-27, reference point at 0.346 c (from reference [147])

A change in e inuences Cm via the term Cme e , in principle again independent of , see also equation (5-76), Cme e = CNh Vh V
2

Sh lh e S c

If Cm can be made zero at any angle of attack by a suitable choice of ih or e or both without inuencing Cm , the slope of the moment curve at an angle of attack where Cm = 0 may be considered the static stability at that particular angle of attack .

8-1-2

Neutral point, stick xed

It follows from equation (5-75) and gure 8-5 that the cg position has a strong inuence on the static stability. We will study this in more detail now. In order to obtain slightly more elegant expressions, we start with the expression for Cm obtained by dierentiating equation (5-55) with respect to , Cm = CNw xcg xw + CNh c 1 d d Vh V
2

Sh xcg xh S c

(8-8)

The only dierence between equation (8-8) and the expression in equation (5-75), following from equation (5-57), is, that lh has been replaced by xh xcg . If the shift in cg positions are small relative to the tail length they are typically about 20 % of the mean aerodynamic chord c and since lh is typically 3 to 4 times c, the contribution of the tailplane to Cm may
Flight Dynamics

366

Longitudinal Stability and Control in Steady Flight

be considered as constant to a rst approximation. The main eect of a change in cg position is a change in the contribution of the wing, fuselage and nacelles to Cm . Shifting the cg rearward (xcg increases) will change the contribution of Cmw , usually already positive, to grow further in the positive sense. The result is, that Cm of the complete aircraft becomes less negative. A rearward shift of the center gravity thus has a destabilizing eect, see also gure 8-5. At a suciently far aft cg position the positive contribution of the wing-fuselage combination just compensates the negative contribution of the horizontal tailplane. Then, Cm = 0 This center of gravity position, at which the equilibrium of the moment is neutrally stable, stick xed, is called the neutral point stick xed (n.p.x). This is the rst interpretation of the neutral point, stick xed. The abscissa of this point is xnf ix . If the center of gravity coincides with this neutral point, it follows from equation (8-8) that, Cm = 0 = CNw xnf ix xw + CNh c 1 d d Vh V
2

Sh xnf ix xh S c

(8-9)

From equation (8-9) the position of the neutral point is derived. To this end CNw is eliminated from equation (8-9) by multiplying the expression for the normal force gradient, CN = CNw + CNh by, xnf ix xw c The result is, xnf ix xw CNh = c CN 1 d d Vh V
2

d d

Vh V

Sh S

(8-10)

Sh lh S c

(8-11)

where, as in an equation similar to equation (5-56), lh = xh xw in an aircraft xed reference frame. Next, equation (8-9) is substracted from equation (8-8). Using equation (8-10), the resulting expression for Cm becomes, Cm = CN xcg xnf ix c (8-12)

8-1 Stick Fixed Static Longitudinal Stability

367

Cm

0.2

C
0 -4 0 4 8 12 16

[o ]

-0.2

D
-0.4

-0.6

Zr D 0.08 c A B C Xr

0.135 c

0.085 c

xcg / c A B C D 0.211 0.346 0.431 0.346

zcg / c 0.044 0.044 0.044 +0.036

Figure 8-5: Inuence of the position of the reference point (center of gravity) on the moment curves of the Fokker F-27, (from reference [147])

Flight Dynamics

368 Zr

Longitudinal Stability and Control in Steady Flight dCm = dCN

xcg xnf ix c

dCNw xw

dCN lh dCNh
Vh V 2 Sh S

acw

xcg xnf ix

cg

n.p.f ix xh

ach

Xr
Figure 8-6: Change in the moment dCm due to a change in the angle of attack

Consider now by means of equation (8-12) a change in the moment dCm due to a change in angle of attack d such that dCm = Cm d. Since the accompanying change in the normal force is dCN = CN d, equation (8-12) may be written as, dCm = dCN xcg xnf ix c (8-13)

According to equation (8-13), the change in the moment dCm may be interpreted as being xcg xnf ix caused by a change in the normal force dCN , acting a distance from the center of c gravity, see gure 8-6. From this follows the second interpretation of the neutral point with stick xed as the point of action on the m.ac of the normal force increment, caused by a step change of the angle of attack with xed elevator angle deection. In concurrence with this second interpretation of the neutral point, the non-dimensional xcg xnf ix distance is often called the stability margin, stick xed. c The present interpretation of the neutral point agrees entirely with its rst introduction in section 5-1-2as the point of intersection of the line of action of an increment in the aerodynamic force, dC R and the mean aerodynamic chord.

8-1-3

Elevator trim curve and elevator trim stability

A. Elevator trim curve In section 5-2-6 the elevator angle required for equilibrium, i.e. for Cm = 0, was expressed in equation (5-80):

8-1 Stick Fixed Static Longitudinal Stability

369

e =

1 Cme

{Cm0 + Cm ( 0 )}

where Cm is the static stability according to equations (5-75) or (8-8). The constant Cm0 and the elevator eciency Cme were already expressed in equations (5-74) and (5-76) respectively, see section 8-1-1. If Cme and Cm are constant, the elevator angle required for equilibrium varies linearly with , according to equation (5-80). The graphic representation of the elevator angle required for equilibrium as a function of or airspeed V is known as the elevator trim curve. The relation between the angle of attack and the airspeed V is derived from the equilibrium in Z-direction, or in a slightly modied form, CN CN ( 0 ) W
1 2 2 V S

(8-14)

From equation (5-80) follows with equation (8-14) the expression for the elevator trim curve e as a function of V , e = 1 Cme Cm0 + Cm CN W
1 2 2 V S

(8-15)

Figure 8-7 presents schematically some elevator trim curves and the corresponding moment curves for a statically stable (Cm < 0), a neutrally stable (Cm = 0) and a statically unstable (Cm > 0) aircraft. It is common practice to plot an elevator-up deection in the upward direction, implying that the negative e -axis points upwards. From these gures, and also from the expressions in equations (5-80) and (8-15), it can be seen that there is a close relation between the moment curve and the elevator trim curve. In particular the slope of the elevator-trim curve is directly related to the slope of the moment curve Cm , i.e. the static longitudinal stability, if Cme is exactly constant. B. Elevator stick position stability The following discussion shows how the slope of the elevator trim curve is important for the pilots opinion on the handling qualities of aircraft. The slope of the elevator trim curve e follows from equation (5-80) by dierentiating with respect to , de Cm = d Cme (8-16)

Considering that Cme is always negative, it follows that for a statically stable aircraft (Cm < 0) the slope of the trim curve e is always negative, i.e., de <0 d (8-17)
Flight Dynamics

370

Longitudinal Stability and Control in Steady Flight

(Cm )e =0

dCm d

>0

dCm d

=0

0 = 0o assumed

Cm0

dCm d

<0

e ()

e ()
dCm d dCm d

<0

<0 V
dCm d

e =

Cm Cm 0

dCm d

=0

=0

dCm d

>0
dCm d

>0

Figure 8-7: Moment curves and corresponding trim curves

8-1 Stick Fixed Static Longitudinal Stability

371

The slope de of the elevator trim curve e V follows from equation (8-15) by dierentiating dV with respect to V, de 4W 1 Cm = 3S C dV V me CN For a statically stable aircraft (Cm < 0) it follows from equation (8-18), de >0 dV (8-19) (8-18)

If the relation between the elevator angle and the airspeed in steady, straight ight is such that equation (8-19) is satised, the aircraft is said to have elevator trim stability, see also reference [161]. This concept of the elevator control position stability can be interpreted in two dierent ways, i.e., 1. According to the rst interpretation, trim stability (the slope de of the elevator trim dV curve e V ) provides an indication of the static longitudinal stability of the aircraft. Elevator trim stability is a convenient characteristic to obtain the static stability both qualitatively and quantitatively from measurements in actual ight, see section 8-1-6 for a more detailed discussion. 2. The second interpretation of trim stability lies in the fact that an aircraft possessing such stability is pleasant for the pilot and safe to y. This will be further illustrated in the following. Suppose that from a given steady ight condition pilot elects to transition to another steady ight condition at a slightly lower airspeed and a corresponding larger angle of attack. He must therefore initiate a nose-up rotation of the aircraft over the small angle , to obtain the required new angle of attack. This initial rotation is obtained through an initial moment Cmi about the lateral axis, generated by an initial elevator deection ei , see gure 8-8. If and are intended to increase and as a consequence V is intended to decrease, the initial elevator displacement must be upward and the corresponding stick displacement directed aft. For any aircraft at any cg position the following holds, ei >0 V and, sei >0 V Now it is known that it is pleasant for the pilot, because it eases the ying of the aircraft, if the initial control displacement sei and the nal control displacement seu are in the same direction, seu >0 sei eu >0 ei (8-20)
Flight Dynamics

372

Longitudinal Stability and Control in Steady Flight

Cmi i XB
Figure 8-8: Initial control surface deections

ei cg

with the subscript u as in seu and eu implying the nal, or ultimate, control (surface) displacement. There is no need for eu and ei to be of equal magnitude. In order to speed up the response of the aircraft, the pilot might well generate an initial control deection larger than the nal control displacement. This has been indicated by the dashed line in gure 8-9a. Evidently, the following holds, eu = ei
eu V ei V

e It was argued, that always Vi > 0. As a consequence equation (8-20) corresponds to the requirement eu > 0 or, for innitesimally small changes : de > 0. V dV

It is thus seen, that equation (8-20) leads to the requirement in equation (8-19) for elevator control position stability. Figure 8-10 may serve as a further illustration of the importance of this desired control characteristic. The gure shows in a highly schematic way how the pilot handles the elevator control, if he or she wants to change the aircraft from one steady ight condition into another steady condition. The case of a trim curve showing control position stability is considered rst. Suppose again the pilot wants to reduce airspeed, starting from a steady ight condition (indicated 1 in the gure). Pulling the control wheel aft, will cause and to increase, and soon after V to decrease. For an aircraft with positive elevator control position stability, the new elevator deection e2 corresponds to a new steady ight condition at a lower airspeed V2 (see the ight condition indicated as 2 in gure 8-10). The response of the aircraft to the initial control displacement is in the direction of the desired new state equilibrium. Figure 8-10 also shows the case of an aircraft with negative elevator control position stability. It can be seen, that the response of the aircraft to the initial control displacement is in the other direction as needed for the desired new steady ight condition. As a consequence a much more complicated time-history of the control displacement is needed to arrive at the steady ight condition 2. This characteristic where the initial control input and the nal control input have dierent signs is considered to be an objectionable characteristic which pilots are supposed not to appreciate. Worse is, of course, that in this case this control characteristic reects the fact that the aircraft is statically unstable, and would require a lot

8-1 Stick Fixed Static Longitudinal Stability

373

Figure 8-9: Initial and ultimate control displacement and control surface deection for the transition to a lower airspeed

Flight Dynamics

374

Longitudinal Stability and Control in Steady Flight

Figure 8-10: Aircraft responses to a control deection

8-1 Stick Fixed Static Longitudinal Stability (Cm )e =0

375

xcg = xnf ix 0 = 0o assumed cg moves rearward e () e ()

xcg = xnf ix xcg = xnf ix cg moves rearward

cg moves rearward

Figure 8-11: Inuence of cg position on the trim curve

of attention and eort by the pilot to keep under control, and it might well be an impossible task for even the most gifted pilot.

Flight Dynamics

376

Longitudinal Stability and Control in Steady Flight

A quick derivation of stick position stability We start with the symmetrical equations of motion, in which, for simplicity, the derivatives CZ and CZq were set to zero. u CX CZo 0 CXu 2c Dc CZu CZ 2c Dc 0 2c + 0 0 Dc 1 q c 2 Cmu Cm + Cm Dc 0 Cmq 2c KY Dc V 0 CZ 0 .e = 0 Cm For stationary straight ight reduce to: CXu CZu Cmu
c c conditions, where q = Dc = Dc u = Dc q = 0 these equations V V CX CZo u 0 CZ 0 . + CZ .e = 0 Cm 0 Cm

Now , we introduce a small stick deection e , in a trimmed ight condition with V0 , 0 , 0 . This results, after some time, in a new steady state ight condition with V = Vo + V ( = u So, V ), =o + , =o + Vo

or where Ass

CXu CZu Cmu

CX CZo u 0 CZ 0 . + CZ .e = 0 Cm 0 Cm Ass .x B.e = 0

After dividing by e we get:

x e

CXu = CZu Cmu

CX CZ Cm

CZo u 0 0 , x = and B = CZ 0 Cm
u e e e

CXu CZu Cmu now simply follows from

CX CZo CZ 0 . Cm 0

0 + CZ = 0 Cm

x = A1 .B ss e

8-1 Stick Fixed Static Longitudinal Stability By applying Cramers rule we nd: 0 CZ Cm CX CZ Cm |Ass | CZo 0 0

377

u = e

CZ Cm + Cm CZ CZu Cm Cmu CZ

If for simplicity the less usually important derivatives CZ and Cmu are set to zero, we get: u Cm CZ = e CZu Cm Substituting CZ = CN , and CZu 2CL results in u Cm CN Cm CN = . or, Vo = . . e 2CL Cm e 2CL Cm After introducing: W CL = 1 /2Vo2 S de 4W Cm 1 = . . dV Vo3 S CN Cm For a statically stable aeroplane where Cm < 0 and de > 0, which implies that for ight at dV a lower speed, an aft (negative) stick deection is required.
V

we arrive at the well known expression for stick displacement stability:

8-1-4

Inuence of various parameters on elevator trim curve

A. Inuence of the center of gravity position Using equation (8-12) the expression in equation (8-15) for the elevator trim curve can be rewritten as, 1 Cme W
1 2 2 V S

e =

Cm0 +

xcg xnf ix c

(8-21)

The slope of the trim curve, the trim stability, follows from equations (8-18) and (8-12), de 4W 1 xcg xnf ix = 3S C dV V c me (8-22)

Equations (8-21) and (8-22) reveal the direct inuence of the position of the center of gravity relative to the neutral point, stick-xed, on the elevator angle required for equilibrium and on the elevator control position stability. A rearward cg shift increases the elevator angle required for equilibrium, the elevator trim
Flight Dynamics

378 Cm

Longitudinal Stability and Control in Steady Flight

ih increases e () ih increases ih increases V e ()

Figure 8-12: Inuence of the tailplane angle of incidence on the trim curve (aircraft has control position stability)

curve e V in gure 8-11 shifts downward. In addition when the cg shifts aft the stability margin decreases. The elevator control position stability, i.e. the slope of the trim curve, decreases, see also gure 8-11. B. Inuence of the stabilizer setting In the expression for the trim curve the part of the elevator angle independent of airspeed is obtained by letting in equation (8-21) V = (and by consequence = 0 ). Using eqution (5-74), the resulting eV = is, Cm0 1 = Cme Cme Vh V
2

eV = =

Cmac CNh (0 + ih )
Cm0

Sh lh S c

(8-23)

For many aircraft the stabilizer setting ih can be varied in ight. From equation (8-23) it follows that such a change has an inuence on eV = . The stabilizer setting is adjusted in

8-1 Stick Fixed Static Longitudinal Stability

379

ight to reduce the elevator control force to zero in a given steady ight condition, see section 8-2. In such an aircraft the stabilizer thus has a function comparable to that of a trim tab. It follows from equation (8-23) that an increase of ih (the stabilizer leading edge moves up) causes an increase in the negative sense of the part of e independent of airspeed. The trailing edge of the elevator moves up as well. The simple explanation is that at constant airspeeds and cg positions the value of CN required for equilibrium of the moment remains constant. This has been indicated schematically in gure 8-12. C. Inuence of the trim tab angle Many aircraft, especially the smaller and slower ones, have a trim tab at the trailing edge of the elevator. If the tab angle te is changed, the trim curve shifts parallel to itself. Qualitatively, this shift is entirely comparable to the shift caused by a change in the stabilizer angle of incidence ih . Usually, the infuence of a change in te on the elevator trim curve is neglected, see also equation (5-60). Subgure a of gure 8-21 shows a number of trim curves measured at constant cg positions and dierent trim tab angles. D. Shape of the elevator trim curve The shape of the trim curve not only depends on the variables discussed so far: the cg position, the stabilizer angle of incidence ih and the trim tab angle te . In addition, the inuence of engine power setting can be appreciable, in particular for propeller-driven aircraft. Usually, an increase in engine power decreases the elevator control position stability. The Mach number eects and aeroelastic deformation can also have an important inuence on the elevator trim curves, because they make the aerodynamic coecient functions of airspeed. As noted in section 1-2 these phenomena will not be considered here.

8-1-5

Static longitudinal stability of tailless aircraft

The stability of tailless aircraft has already been discussed in section 5-1-3. It was shown, see gure 5-21, that equilibrium (Cm = 0) at positive values of CN is possible only if Cmac > 0. The equilibrium is stable, i.e. Cm < 0, if the center of gravity lies ahead of the ac, xcg < xac = xw By choosing a suciently forward cg position this condition can be satised. The requirement for a positive Cmac needs a further discussion. In section 5-1-4 the following expression was derived for the Cmac of a wing, see equation (5-30), 2 = S c
b 2 b 2

Cmac

cmac c2 dy

cb c (x x0 ) dy

Flight Dynamics

380 cb c

Longitudinal Stability and Control in Steady Flight

b 2

b +2

croot

<0 ctip

Figure 8-13: A positive Cma.c is obtained by combining sweepback and negative wing twist

It is possible to obtain a positive Cmac by choosing wing airfoil sections having a positive cmac . Airfoils showing an S-shaped camber line possess at c = 0 a positive value for the moment coecient. For wings having such airfoils the rst integral in equation (5-30) is positive. A second possibility to obtain a positive Cmac is oered by the second integral in equation (5-30). As was discussed in section 5-1-4, this integral represents the moment due to the basic lift distribution. Figure 8-13 shows that this moment is positive at CL = 0 for a wing having sweep back and negative wing twist. Although it is thus shown that static longitudinal stability can be obtained without resorting to a horizontal tailplane, most aircraft nevertheless have a horizontal tailplane. The elevator of a tailless aircraft is placed at the trailing edge of the wing. The two parts of the elevator are commonly used as ailerons (or named as elevons) as well. Due to the relatively small distance to the aircraft center of gravity, elevons are less eective for pitch control than an elevator mounted on a tailplane. When the aircraft is in a trimmed condition, in general e = 0. The increased drag due to the elevator deection, called trimdrag, of the tailless aircraft may be one of the factors in favour of a design with a tailplane. The horizontal tailplane also contributes to a considerable degree in the damping of the so-called short-period oscillation, occurring after a symmetric disturbance of the equilibrium, see chapter 10. This is an additional advantage of the horizontal tailplane. Summarizing, it can be said that the horizontal tailplane contributes to, the equilibrium of the moment about the aircraft cg

8-1 Stick Fixed Static Longitudinal Stability the static longitudinal stability the aerodynamic damping about the lateral axis The price to be paid is, however, an increased aerodynamic drag an increased aircraft weight

381

8-1-6

Determination Cme from measurements in ight

In section 5-2-6 the condition for equilibrium of the moment was written as, see equation (5-77), Cm = (Cm )e =0 + Cme e Suppose Cme were known. From a trim curve, e V , measured in ight the moment coecient (Cm )e =0 can be obtained as a function of V or . In many cases Cme varies markedly with ight condition, for instance under the inuence of the wing wake or the slipstream. It follows that it is necessary to know Cme in the ight conditions of interest, since a measured trim curve is not a proper reection of the moment curve (Cm )e =0 . Methods to determine the elevator eciency Cme from measurements ight are all based on the principle of applying a known moment about the lateral axis in the ight conditions to be investigated. The extra elevator deection needed to compensate for this moment is a measure of Cme in that ight condition. The simplest manner to generate a known moment employs a shift of the center of gravity in the X-direction, for instance by shifting a known amount of ballast over a known distance. In gure 8-14 cg positions (cg1 and cg2 ) have been indicated as well as the equilibrium at cg1 . At constant elevator angle e = e1 no equilibrium exists about cg2 . At constant and e , CN does not change and it still passes through cg1 . Equilibrium of the moment at cg2 at constant is obtained by changing the elevator angle. The change of the moment about cg2 caused by e2 e1 is, Cm = Cme (e2 e1 ) = Cme e This moment is balanced by the moment of CN about cg2 , i.e., CN xcg2 xcg1 + Cme (e2 e1 ) = 0 c (8-24)

The inuence of the elevator deection e on CN is neglected here, or, CN = CNh e Vh V

2

Sh << CN S
Flight Dynamics

382 XB

Longitudinal Stability and Control in Steady Flight R N W sin

cg1 W cos cg2 W ZB

Figure 8-14: The determination of elevator eciency from ight tests

e1 e2 e

From equation (8-24) it follows, Cme = 1 xcg CN e c (8-25)

The above means that if the trim curves have been measured at two cg positions diering in their x-coordinates, Cme can be obtained with equation (8-25). In the foregoing the inuence of the cg position on Cme has been neglected. Because the tail volume equation (5-76) gives,
Sh lh S c

is obtained directly from the dimensions of the aircraft, use of

CNh

Vh V

= Cme

S c Sh lh

(8-26)
2

The normal force gradient CNh can only be obtained from ight tests if in addition Vh V has been measured along the span of the horizontal tailplane. Once the elevator eciency Cme has been measured as just described, the stability margin (stick xed) can be derived from the slope of the trim curve, using equation (8-22). If the aircraft is equipped with an adjustable stabilizer, the derivative Cmih can be obtained as well, once Cme has been measured. To this end, the elevator angle required for equilibrium is measured in two ight conditions, diering only in the choice of the stabilizer angle of incidence but identical in all other respects. In section 5-2-4 the angle of attack h of the horizontal tailplane was derived as, see equation (5-64), h = + ih From this expression it follows that if ight conditions do not change (both and remain constant) a change in ih inuences only h , as expressed by,

8-1 Stick Fixed Static Longitudinal Stability

383

h = ih At constant and , CNh has to remain constant as well to maintain the same steady ight conditions, where Cm = 0, CNh = CNh h + CNh e = constant or, CNh h + CNh e = 0 The result is, CNh = As mentioned before, in general CNh
Vh V 2

e CN ih h
Vh V 2

will be determined, rather than CNh . If

Vh V 2

is not known, the procedure just described can only produce CNh CNh Vh V
2

e CNh ih

Vh V

By way of illustration, the following gives an example of the determination of stability characteristics from measured elevator trim curves. The aircraft in the example is a Fokker F-27, the measured trim curves were taken from reference [51]. Figure 8-15 shows two elevator trim curves e Ve . They were measured at an engine power of 420 hp (approximately gliding ight) and a trim tab angle of te = +4.5o . Apart from cg positions also the aircraft weight was dierent. For this reason, the trim curve for xcg = 0.227 c, pertaining to an aircraft weight of 14,500 kg, was corrected to a weight of 13,000 kg. At a constant ight condition, i.e. at constant CN and e , the following holds, Ve2 = Ve1 W2 W1

The trim curve thus corrected is also shown in gure 8-15. In the rst place Cme is derived from the two trim curves belonging to the same aircraft weight of 13,000 kg. At Ve = 240 x km/h, gure 8-15 shows that e = 1.4o . The shift in cg position is ccg = 0.058. At an aircraft weight of 13,000 kg the airspeed of Ve = 240 km/h corresponds to a normal force coecient of CN = 0.681. From equation (8-25) it then follows, Cme = with e in degrees. 1 0.681 0.058 = 0.0282 1.4

Flight Dynamics

384

Longitudinal Stability and Control in Steady Flight

Figure 8-15: Trim curves for two cg positions of the Fokker F-27 (from reference [51])

8-2 Stick Free Static Longitudinal Stability The tailplane volume of the present aircraft is follows with equation (8-26), Vh V Vh V
2 Sh lh S . c

385 Using this value at Ve = 240 km/h it

CNh

= +0.0303 (e in degrees)
2

CNh

= +1.7353 (e in radians)

Next, the position of the neutral point, stick xed is derived by means of equation (8-22). 4W Again at Ve = 240 km/h is 0 V 3 S = 0.0204 and at xcg = 0.277 c the slope of the elevator trim curve is
de dVe

= +0.0502o /km/h or +0.187o /m/s. This results in, xcg xnf ix = 0.187 0.0282 48.9 = 0.258 c

With xcg = 0.277 c follows xnf ix = 0.485 c. If the same calculation is repeated for the trim curve measured at xcg = 0.285 c, the result is xnf ix = 0.490 c. Using the data obtained thus far the moment curves Cm versus V and Cm versus CN can be derived. If in addition the relation CN versus of the aircraft is known, the moment curves Cm versus can be calculated as well. Figure 8-16a shows the calculated values of CNh Vh as a function V of .Figure 8-16b gives the Cm versus curves and gure 8-16c shows the position of the neutral point, stick xed, as a function of and CN .
2

8-2

Stick Free Static Longitudinal Stability

It was already noted in the previous section that it makes sense to study the static stability in the case of stick free. Two reasons can be given to study the static stability in this condition. The rst is the requirement for the aircraft to be remain stable if the pilot would release the stick. The second reason lies in the close relationship between the static stability, with stick free, and a certain characteristic of the way in which the elevator force required to maintain the elevator in the position for steady ight, varies with airspeed. This characteristic will be called the elevator stick (or control) force stability. As in section 8-1, the concept of static stability is based entirely on the variation of the pitching moment with angle of attack. Leaving the elevator free only changes the contribution of the horizontal tailplane, and the elevator, to the pitching moment. First we consider the behavior of the elevator in the control free situation. If the control stick or wheel is free, the control force Fe is zero. Neglecting the friction in the control mechanism this means that in the stick free situation the hinge moment is zero, He = 0, or, Chef ree = 0
Flight Dynamics

386

Longitudinal Stability and Control in Steady Flight

Figure 8-16: Some stability characteristics derived from the measured trim curves shown in gure 8-15 (Fokker F-27)

8-2 Stick Free Static Longitudinal Stability Chef ree = 0 h V ef ree

387

Chef ree = Ch h + Ch ef ree + Cht te = 0 te

Figure 8-17: The equilibrium of the free elevator

In this situation the elevator angle assumes a certain value, depending on the way the elevator is aerodynamically balanced and on the trim tab angle. This elevator angle, stick free, indicated as ef ree see gure 8-17, varies with angle of attack. If the airfoil of the tailplane is symmetric, see equation (5-63), Ch0 = 0 and then, Chef ree = Ch h + Ch ef ree + Cht te = 0 This means that ef ree becomes, ef ree = Cht Ch h t Ch Ch e (8-27)

Dierentiation with respect to h results in the variation of ef ree , with h at constant trim tab angle, de dh and the variation with , de d Using equation (5-66), dh d =1 d d results in, de d = Ch Ch 1 d d (8-29) = Ch dh Ch d (8-28) = Ch Ch

f ree

f ree

f ree

Flight Dynamics

388

Longitudinal Stability and Control in Steady Flight

8-2-1

Stick free static longitudinal stability in gliding ight

The contribution of the horizontal tailplane to the pitching moment is, see equations (5-60) and (8-5), Cmh = CNh h + CNh e Vh V
2

Sh lh S c

(8-30)

When studying the static longitudinal stability in section 8-1, the elevator control was assumed to be held xed, the elevator angle e was constant and did not change when was slightly varied. If, however, in equation (8-30) the constant e is replaced by the variable ef ree , dierentiation with respect to results in the contribution of the horizontal tailplane with free elevator to the static longitudinal stability, stick free, Cmf ree , dCmh d = Cmh = CNh dh + CNh d de d Vh V
2

f ree

f ree

f ree

Sh lh S c

(8-31)

Substituting equation (5-66), dh = d and equation (8-29), de d results in, Cmh = CNh CNh Ch Ch 1 d d Vh V
2

d d

f ree

Ch Ch

d d

f ree

Sh lh S c

(8-32)

If equation (8-32) is compared with the corresponding expression (8-7) for the contribution of the horizontal tailplane in the stick xed situation, when e = 0 is constant, Cmh = CNh d 1 d Vh V
2

f ix

Sh lh S c

the concept of the normal force gradient of the tailplane in the stick free situation arises, CNh = CNh CNh Ch Ch (8-33)

f ree

Using this gradient in equation (8-32) results in close analogy with equation (8-7), Cmh = CNh 1 d d Vh V
2

f ree

f ree

Sh lh S c

(8-34)

8-2 Stick Free Static Longitudinal Stability

389

The static stability, stick free, Cmf ree , follows by adding to equation (8-34) the contribution of the wing, fuselage and nacelles, Cmf ree = CNw xcg xw CNh f ree c 1 d d Vh V
2

Sh lh S c

(8-35)

Previously, the static stability, stick xed, was derived as, Cm = CNw xcg xw CNh c 1
f ree

d d

Vh V

Sh lh S c

The expressions (5-75), (8-35) and (8-33) for CNh

indicate that the dierence between

C

the static stability stick xed and stick free is entirely due to the term CNh Ch . In this h term CNh is positive. The hinge moment derivatives Ch and Ch were already discussed in chapter 5. They depend on the way the elevator is aerodynamically balanced. To keep the control forces at reasonably low levels, it is essential that both Ch and Ch are made small in the absolute sense. The required sign of Ch follows from the behaviour of the free elevator in steady, trimmed ight, i.e. the situation where Fe = 0. In that situation Che = 0. Suppose the elevator now obtains a small deviation de from the equilibrium position, due to some disturbance. The immediate eect is a hinge moment, dChe = Ch de if Ch > 0, the hinge moment will be positive if de is positive. Due to the hinge moment the change in elevator angle will further increase. The general conclusion is that an equilibrium position of a control surface in the controls free condition will be unstable if Ch is positive; the control surface does not return to the equilibrium position after a disturbance has occurred. For this reason it is strictly necessary that Ch is negative. The above argument needs a slight extension. Suppose a control surface has a Ch nearly equal to zero. In such a case actually two hinge moments are present, of nearly equal magnitude but of opposite sign, see the pressure distribution in gure 5-71. The resultant value of Ch is due to the dierence of these two hinge moments. A small variation in one of these moments, as may be caused by a small change in the shape of the control surface, due for instance to the tolerance in the manufacturing process, will have a large eect on Ch if compared to the intended value. For this reason Ch has to be not only negative, but should not be allowed to approach zero too closely. Common values are: 0.2 to0.3 < Ch < 0.1 (e in Radians). The sign of Ch will be discussed in detail in section 8-2-3 in relation with the elevator control force stability. Here it can be said that Ch may be either negative or positive. Too large positive values of Ch , e.g. Ch > approximately 0.1 (with h in Radians), cannot be used because they may lead to dynamic instability, stick free. In the decision on the required values of Ch and Ch the inuence which ice accretion on the stabilizer and elevator has on the aerodynamic balance of the control surface, has to be taken into account as well. In the
Flight Dynamics

390 Zr

Longitudinal Stability and Control in Steady Flight dCmf ree = dCNf ree
xcg xnf ree c

dCNw xw

dCNf ree lh dCNhf ree

Vh V 2 Sh S

acw

xcg xnf ree

ach cg n.p.f ree xh

Xr
Figure 8-18: The change in the moment dCmf ree due to a change in angle of attack

nal choice of the elevator balance the non-linear relation between the hinge moment and h and e at large angles is an additional important consideration. From equations (8-33) and (8-35) it can be seen that depending on the sign of Ch the aircraft with free elevator control will be less statically stable (Ch < 0), equally stable (Ch = 0) or more stable (Ch > 0) than it is in the control xed situation.

8-2-2

Neutral point, stick free

As for the aircraft with xed elevator control a certain center of gravity position exists also in the control free situation where Cm , here Cmf ree , is equal to zero. This center of gravity position is called the neutral point, stick free, indicated as n.p.f ree . The abscissa of this point is xnf ree . As in equation (8-11) the position of the neutral point, stick free, is obtained by letting Cmf ree = 0, see equation (8-35), CNh xnf ree xw f ree = c CN d 1 d Vh V
2

Sh lh S c

(8-36)

Just like the neutral point, stick xed, the neutral point, stick free, can be interpreted in a second way, in addition to the one just given. It is the point of action (on the m.ac) of the total change in normal force coecient dCN due to a change in angle of attack d if the elevator control is left free. In analogy with equation (8-12) it follows, Cmf ree = CNf ree xcg xnf ree c

8-2 Stick Free Static Longitudinal Stability with, CNf ree CNf ix

391

The latter expression is obtained by neglecting in equation (8-10) the dierence in normal force gradient CN due to the dierence between CNh and CNh , as the contribuf ree tion of the tailplane to CN is relatively small when compared to the contribution of the wing. The change in pitching moment dCmf ree due to a change in angle of attack d with free elevator control then is, dCmf ree = dCNf ree xcg xnf ree c (8-37)

Figure 8-18 gives an illustration of this second interpretation of the neutral point, stick free. The relative position of the two neutral points, stick xed and stick free, follows from equations (8-11) and (8-36), CNh CNh xnf ree xnf ix f ree f ix = c CN Using equation (8-33), CNh the result is, CNh Ch xnf ree xnf ix = c CN Ch
2

d d

Vh V

Sh lh S c

(8-38)

f ree

= CNh CNh

Ch Ch

d d

Vh V

Cm Ch Sh lh = S c CN Ch

d d

(8-39)

According to this latter expression the relative position of the two neutral points is determined primarily by the sign of Ch . A simple explanation of the inuence of the sign of Ch on the static stability, stick free, is possible by looking at the behaviour of the free elevator, see gure 8-19. If Ch = 0 the elevator angle ef ree does not vary with angle of attack according to equation (8-29). It is as if the controls were xed, see subgure b of gure 8-19. There is no dierence between the static stability stick xed and stick free, the two neutral points coincide. If Ch < 0 the elevator turns with the direction of the ow as the angle of attack varies, see subgure a of gure 8-19. With an increase in angle of attack the trailing edge of the elevator moves up. The eect of freeing the elevator control is an extra downward force on the tailplane. This extra force promotes a further increase in angle of attack. As a consequence the static stability stick free is less than it is for stick xed and xnf ree is smaller than xnf ix , if Ch < 0. The inuence of a positive Ch can be understood in the same way, see subgure c of gure 8-19.
Flight Dynamics

392

Longitudinal Stability and Control in Steady Flight ef ree ef ree (< 0) V (A) Ch < 0 ef ree = 0 V (B) Ch = 0 ef ree

V (C) Ch > 0

ef ree ef ree

Figure 8-19: The behaviour of the free elevator after a change of the angle of attack

8-2-3

Elevator stick force curves and elevator stick force Stability

A. The control force curve The elevator control force needed for equilibrium was derived in chapter 5 as, Fe = de 1 2 V Se ce Che dse 2 h

The following discussion is intended primarily for qualitative purposes. It is, therefore, permissible to express Che here as a linear function of h , e and te , see equation (5-63). This leads to the following expression for Fe , see also equation (5-83), Fe = de 1 2 V Se ce dse 2 h Ch h + Ch e + Cht te

For a more accurate calculation of the control force in a given ight condition Che has to be obtained from wind tunnel measurements for the correct values of h , e and te . In chapter 5 h was derived as, see equation (5-65), h = ( 0 ) 1 d d + (0 + ih )

The elevator angle e required for equilibrium about the lateral axis is, see equation (5-80),

8-2 Stick Free Static Longitudinal Stability

393

e =

1 Cme

Cm0 + Cmf ix ( 0 )

Substitution of equations (5-65) and (5-80) in equation (5-83) results after some elaboration in, Fe = where,
Ch0 =

de 1 2 V Se ce dse 2 h

Ch0 + Ch ( 0 )

(8-40)

Ch Ch Cmac CNh (0 + ih ) + Cht te f ree Cm CNh 1 d d Ch Ch Cmf ix = Cmf ree Cm Cm

(8-41)

Ch = Ch

= Furthermore, ( 0 )

xcg xnf ree Ch CN Cm c

(8-42)

W
1 2 2 V S

1 1 W 1 2 CL CN 2 V S

(8-43)

Substituting equation (8-43) in equation (8-40) results in, de Se ce dse Vh V

2 Ch0

Fe =

1 2 1 W V + Ch 2 S CN

(8-44)

According to equation (8-44) the elevator control force required to maintain steady ight consists of two parts. One is proportional to the dynamic pressure and thus varies with airspeed. The other part is independent of airspeed. The part depending on airspeed is considered rst. It can be seen from equations (8-44) and (8-41) that this part varies by changing the trim tab angle or the stabilizer angle of incidence or both. Now suppose that ih is constant. At a certain value of te , Ch0 will be equal to zero. This particular trim tab angle is called te0 . It is derived from equation (8-41) by letting,
Ch0 = 0

The result is,

Ch Cm

Cmac +

te0 =

Ch CNh

CNh

f ree

(0 + ih ) (8-45)

Cht

Flight Dynamics

394

Longitudinal Stability and Control in Steady Flight

Fe () te < te0

Fe () te < te0 te = te0

te = te0

FeV =0

te > te0

FeV =0

te > te0

1 2 2 Vmin

1 2 2 V

Vemin (B)

Ve

(A)

Figure 8-20: Schematic form of the elevator control force curve, Fe as a function of (a) dynamic pressure 1 2 2 V and (b) Fe as a function of equivalent airspeed Ve

8-2 Stick Free Static Longitudinal Stability

Using equations (8-41) and (8-45), Ch0 can now be written as, Ch0 = Cht

395

te te0

(8-46)

After some elaboration the nal expression for the elevator control force can now be obtained from equation (8-44), using equations (8-46) and (8-42). The result is, de Se ce dse Vh V
2

Fe =

W Ch xcg xnf ree 1 V 2 Cht S Cme c 2

te te0

(8-47)

The variation of the elevator control force with dynamic pressure or airspeed as expressed by equation (8-47) is shown schematically in gure 8-20. Negative elevator control forces, i.e. pull forces exerted by the pilot, are plotted upward in the gure, just like negative elevator angles. If the various aerodynamic characteristics in equation (8-47) may indeed be considered as constants, then Fe varies linearly with the dynamic pressure 1 V 2 , see subgure a of gure 2 8-20, or quadratically with V , see subgure b of gure 8-20. The measured trim curves and elevator control force curves shown in gure 8-21 show that in reality the various simplifying assumptions are not always satised. The second part of the elevator control force is independent of airspeed, it is indicated as FeV =O . According to equation (8-47) this part is determined by the position of the cg relative to the n.p.f ree or by the sign of the static stability, stick free. If the aircraft is statically stable, stick free (xcg < xnf ree ), FeV =O is negative. The variation of the part of Fe depending on airspeed is determined by te . If te = te0 , Fe does not vary with airspeed. If te > te0 the control force increases in the positive sense with increasing airspeed, see gure 8-20, independent of the cg position. It was assumed that Ch , Cme and Cht have the normal, negative sign. The airspeed at which the control force is zero, is called the trim speed, indicated as Vtr . So, Vtr = VFe =O If the elevator control is let free at this airspeed, the control position does not change.

B. Elevator stick force stability If the elevator control force curve at the trim speed satises the condition, dFe dV

>0
Fe =0

(8-48)

the aircraft is said to have elevator control force stability in that particular ight condition. Elevator control force stability is thus seen to depend only on the slope of the control force curve in the trimmed, i.e. Fe = 0, condition.

Flight Dynamics

396

Longitudinal Stability and Control in Steady Flight

The expression will now be further analyzed. Dierentiating equation (8-47) with respect to V yields the slope of the control force curve as, de dFe = Se ce dV dse Vh V
2

V Cht

te te0

(8-49)

This expression holds for both trimmed and untrimmed conditions. The elevator control force stability follows by substituting in equation (8-49) the values of te pertaining to Vtr . The trim speed Vtr is obtained from equation (8-47) by letting Fe = 0, 1 2 V Cht 2 tr te te0 = W Ch xcg xnf ree S Cme c (8-50)

Combining equations (8-49) and (8-50) results in the elevator control force stability, dFe dV de Se ce dse Vh V
2

Fe =0

= 2

W Ch xcg xnf ree 1 S Cme c Vtr

(8-51)

The direct relation between the control force stability as dened in equation (8-48) and the position of cg relative to the n.p.f ree position can be seen from equation (8-51). If the aircraft is statically stable, stick free, it is control force stable according to equation (8-51) and vice versa. Just like the control position stability, discussed in section 8-1, control force stability is important in two respects, 1. In the rst place, the control force stability provides the possibility to ascertain from measurements in ight if the aircraft is statically stable, stick free, in a certain aircraft conguration and ight condition, see equation (8-51). 2. In the second place, and perhaps most important, an aircraft that is control force stable is more pleasant and safer to y by the pilot. The latter subject will now be discussed. As a general rule it can be stated that it is highly desirable, if not imperative, that control displacements and the control forces required to generate and maintain these displacements both in manoeuvres and in steady ight conditions, have equal directions. This requirement can be expressed simply as, dFe >0 dse (8-52)

The control manipulator, whether it is a stick or a wheel, then behaves as if it were pulled back to the neutral position by springs, see gure 8-22. The eect to the pilot is that he feels through the force he has to apply on the control manipulator in which direction and to what extent he has deected the control. Generally it can be said that the pilot is far more sensitive to changes in the control forces he has to exert than to changes in the control

8-2 Stick Free Static Longitudinal Stability

397

Figure 8-21: Measured trim curves and elevator control force curves for the De Havilland D.H.98 Mosquito M II F in gliding ight (from reference [27])

Flight Dynamics

398

Longitudinal Stability and Control in Steady Flight

Figure 8-22: Schematic representation of the concept of positive feel,

dFe dse

column displacement. Due to the adapted sign convention be written as,

de dse

> 0, the requirement as in equation (8-52) can then

dFe >0 de It was argued in subsection B of section 8-1-3 that the initial and the ultimate control displacement should be in the same direction. From the above it follows that this applies equally to the changes in the exerted control force. Using the same sign convention for control displacements and control forces and considering the requirement for elevator control position stability, see equation (8-19), de >0 dV the corresponding requirement for elevator stick force stability is written as in equation (8-48), dFe dV >0
Fe =0

Inevitably, the control mechanism of the aircraft has a certain friction which was neglected so far. The magnitude of this friction, which is primarily static friction, has been expressed in section 5-2-7 in terms of an equivalent hinge moment Hef . A certain control force Fef is required to overcome this friction. If the airspeed in steady ight diers only slightly from Vtr the required control may be small as well. If |Fe | < Fef , no control force is required from the pilot. He can leave the control stick or wheel free, the required control position is maintained by the friction in the control mechanism, see gure 8-23.

8-2 Stick Free Static Longitudinal Stability Fe ()

399

Range of possible values of Vtrim due to Fef = 0

Fef Fef V

Vtrim if Fef = 0

Figure 8-23: Uncertainty in trim speed due to friction in the control mechanism

The result is that in actual ight not one single trim speed exists, and any airspeed within the range where Fe > Fef is then a trim speed. In order to keep this uncertainty of the trim speed within acceptable limits the elevator control force stability must have a certain minimum positive value and, in addition, Fef has to be suciently small. The U.S. military regulations, see reference [19], require, dFe dV corresponding with, dFe dV > 0.041 (kg/km/h)
Fe =0

> 0.5 (lbs/3 kts)

Fe =0

According to the same regulations the control force required to overcome the friction is limited to, Light aircraft control stick Fef < 1.3 kg
Flight Dynamics

400 control wheel Fef < 1.8 kg Heavy aircraft

Longitudinal Stability and Control in Steady Flight

control stick Fef < 2.3 kg control wheel Fef < 3.2 kg

This control force is to be measured on the ground, the engines not running. In ight the control force needed to overcome the friction may be considerably less due to vibrations of the aircraft. The exact denitions of the terms light and heavy aircraft can be found in reference [19].

A quick derivation of stick force stability We start with the symmetrical equations of motion, in which, for simplicity, the derivatives CZ and CZq were set to zero. CXu 2c Dc CX CZo 0 u CZu CZ 2c Dc 0 2c + 0 0 Dc 1 q c 2 Cm + Cm Dc 0 Cmq 2c KY Dc Cmu V 0 CZ 0 e = 0 Cm

c c In stationary straight ight conditions q = Dc = Dc u = Dc q = 0, so: V V CXu CX CZo u 0 CZu CZ 0 . + CZ .e = 0 Cmu Cm 0 Cm

The stick force can be written as: Fe = writing it for simplicity as

de 1 / V 2 .Se ce .(Ch .h + Ch .e + Cht .te ) dse 2 h

Fe = K.(Ch .h + Ch .e + Cht .te ) where K= de 1 / V 2 .Se ce dse 2 h

8-2 Stick Free Static Longitudinal Stability

401

We will concern ourselves with small variations Fe of Fe and corresponding small deviations from a trimmed steady state condition at V0 , 0 , 0 . After some time, a new state of equilibrium will be reached: V = Vo + V ( = u V ), =o + , =o + Vo

As we are interested in stick force, we assume that the trim tab is not changed, so, Fe = K.(Ch .h + Ch .e ) where, h = (1 so Fe = K {Ch (1 and e = d ) d d ) + Ch e } d

1 Ch d Fe (1 ) K Ch Ch d

If we now substitute e into the equations of motion for the case of steady state we get CXu CX CZo u 0 Ch d CZu CZ 0 + CZ { (1 ) + Ch d Cmu Cm 0 Cm 1 Fe = 0 K Ch or CXu CZu Cmu CX Ch Ch .(1
Ch Ch

CZ

Cm

.(1

d d ).CZ d d ).Cm

Here we introduce two new stability derivatives for the case of stick free, CZ f ree = CZ Cm f ree = Cm so, A xss + B Ch d (1 ). CZ Ch d Ch d .(1 ).Cm Ch d 1 Fe = 0 K.Ch

0 CZ Cm

CZo u 0 + 0 1 Fe = 0 K Ch

Flight Dynamics

402 with

Longitudinal Stability and Control in Steady Flight

Multiplication by the inverse of A results in:

CXu A = CZu Cmu

CX CZ f ree Cm f ree

CZo 0 u 0 , B = CZ ,xss = Cm 0 1 Fe K.Ch

xss = A1 B or
u Fe Fe Fe

and

xss =

K Ch K Ch = A1 B K Ch

F K Ch = u xss (1)
V V

Finally, since = u

, we get:

K Ch F de 1 2 = = Ch Se ce 1/2Vh V xss (1) V dse xss (1)V Applying Cramers rule, we nd: 0 CZ Cm CXu CZu Cmu CX CZf ree Cmf ree CX CZf ree Cmf ree = Now for simplicity we set CZ = 0, Cmu = 0 resulting in: xss (1) = Since CZ CN and CZu 2CL = 2 W 1/ V 2 S 2

x (1) = ss

CZo 0 0 CZo 0 0 = -

CZo CZo

CZ Cm CZu Cmu

CZf ree Cmf ree CZf ree Cmf ree

CZu Cmf ree Cmu CZf ree

CZ Cmf ree Cm CZf ree

Cm CZf ree CZu Cmf ree

8-2 Stick Free Static Longitudinal Stability we get: xss (1) =

403

Cm CZf ree CZu Cmf ree

W 1/ V 2 S , 2

With CZf ree CZf ree and CZu 2 CL = 2 Cm CZf ree xss (1) = . W 2 1/ V 2 S Cmf ree 2 We already derived:

F K.Ch de 1 1 2 = V = , Ch .Se ce 1/2Vh V xss (1) dse xss (1) V so: Cmf ree 1 F K.Ch de W 2 2 = V = Ch .Se ce 1/2Vh . = V xss (1) dse Cm 1/2V 2 S CNf ree V = xc.g. W de 2 2 Ch .Se ce 1/2Vh . 1/ V 2 S dse Cm 2 c

and since this formula applies to trim speed V = Vtr we nd: dF de Vh W Ch xc.g. xnf ree 1 = 2 Ch .Se c( )2 dV dse V S Cm c Vtr

8-2-4

Eect of center of gravity and mass unbalance on stick force stability

A. Inuence of the center of gravity position In the foregoing it was shown that the center of gravity has an inuence only on the part FeV =0 of the control force which is independent of airspeed, see equation (8-47). If the cg is shifted in X-direction this constant term in equation (8-47) varies and, at a constant trim tab angle, the entire control force curve will shift up or down, parallel to itself. If the cg is moved forward the control force curve will shift upwards (a larger pull force will be required) at constant te and ih . This agrees with the measured control force curves shown in gures 8-21 and 8-24. The slope of the control force does not change at constant te , but with the new cg position corresponds to another, higher trim speed. If the original trim speed is re-established by an additional downward deection of the trim tab, it turns out, see for instance subgure d of gure 8-24, that the elevator control force stability at constant trim speed has increased. So, strictly speaking, the elevator control force stability cannot be considered independent of the trim speed. B. Trim tab angle and the stabilizer angle of incidence

Flight Dynamics

404

Longitudinal Stability and Control in Steady Flight

Figure 8-24: Measured elevator control force curves for the North American Harvard II B in gliding ight (from reference [15])

8-2 Stick Free Static Longitudinal Stability

405

From equation (8-47) it can be concluded that for a stick free, statically stable aircraft, any trim tab angle te > te0 results in a certain airspeed where Fe = 0. The trim tab angle for a given Vtr is obtained from equation (8-50), te = te0 + W
1 2 2 Vtr S

1 Ch xcg xnf ree Cht Cme c

(8-53)

It follows from equation (8-53) that for a given cg position the pilot can choose a certain trim speed where he can y hands-o by selecting the appropriate te . If the stability margin stick free is positive, cg in front of the n.p.f ree , the trim speed is reduced by increasing te , i.e. with increasing downward tab deection. The trim wheel in the cockpit is turned backward to this end. Once a certain trim speed has been chosen the variation of the control force with airspeed in steady ight is xed, according to equation (8-51), and so is the elevator control force stability. It can be clearly seen in gure 8-24 from the measured elevator control forces that the slope of the control force curve indeed increases at a given airspeed by a downward deection of the trim tab. From the foregoing it becomes clear that for a given cg position a xed relation exists between the trim tab angle and the correspondig trim speed. The aircraft possesses elevator control force stability if according to equation (8-53), dte <0 dVtr The trim tab angle curve may be used to investigate the elevator control force stability. An example of a measured trim tab angle curve is shown in gure 8-25. As already mentioned in section 5-2-6, in some aircraft the elevator control forces are reduced to zero by adjusting the stabilizer angle of incidence. In such cases the elevator does not need to have a trim tab. It follows from equations (8-41) and (8-44) that both the horizontal stabilizer setting ih and the trim tab angle te gure in Ch0 and FeV =0 . It also follows from (8-41) and (8-44) that it is irrelevant whether the stabilizer or the trim tab is used to trim the control force to zero. If the stabilizer is used Ch0 can be written in analogy with equation (8-46) as,
Ch0 = Ch (ih ih0 )

(8-54)

where ih0 is the value of ih for which Ch0 = 0. The inuence of ih on the elevator control force stability is entirely comparable to that of te .

C. Inuence of a spring or an unbalanced mass in the control mechanism Sometimes it is not possible to obtain satisfactory control forces and elevator control force stability in all required aircraft congurations and ight conditions by aerodynamic balancing of the elevator only. If at a given cg position and trim speed the elevator control force stability
Flight Dynamics

406

Longitudinal Stability and Control in Steady Flight

Figure 8-25: Trim curves, hinge moment coecients and required trim tab angles as functions of CL for the Siebel 204-D-1 aircraft (from reference [16])

8-2 Stick Free Static Longitudinal Stability

407

Figure 8-26: The inuence of a bobweight in the control mechanism

Figure 8-27: The inuence of a spring in the control mechanism

Flight Dynamics

408

Longitudinal Stability and Control in Steady Flight

Figure 8-28: The inuence of a spring or bobweight on the elevator control force curve

should become too low an unbalanced mass, a bobweight, see gure 8-26, or a spring, see gure 8-27, may be used in the control mechanism, see also gure 5-90. These devices increase FeV =0 by a constant force FeV =0 , see gure 8-28. A spring to be used for this purpose is chosen in principle such that the applied force remains approximately constant over the entire range of control deections. The entire control force curve will then be shifted upward, parallel to itself. A more detailed analysis shows, that now Cmf ree has become a function of airspeed. It also turns out that the n.p.f ree position has moved backwards and consequently the stability margin, stick free, has been increased. The use of a spring or an unbalanced spring in the control mechanism is thus seen to have the same eect, as far as the control force curve is concerned, as a forward shift of the center of gravity.

8-2-5

Inuence of design variables on control forces

Another look at equation (8-47), de Se ce dse Vh V

2

Fe =

W Ch xcg xnf ree 1 V 2 Cht S Cme c 2

te et0

reveals that the variables of direct interest to the pilot are: 1 V 2 , xcg , te and ih (they have 2 already been discussed in previous sections). Other variables occuring in equation (8-47) are d hl of interest in the design of the aircraft, such as dse , Se , ce , the tailvolume SSh and CNh (or c e Cme ), Ch , Cht , fcree . The following can be said on these variables. The control forces increase with the wing loading W . the control forces are in addition proportional to Se and S
xn

8-2 Stick Free Static Longitudinal Stability

409

ce . Hence, for aircraft of identical shape the control forces vary with the third power of the dimensions. The control gearing de is determined by the fact that the extreme deection of the control dse surface must correspond with the extreme displacement of the cockpit control column or stick. The range of elevator deections emax is restricted aerodynamically, amounting to about 50o or 60o (25o to 30o to either side). The permissable range of control displacements semax is restricted by the dimensions of the human pilot and sometimes by the dimensions d of the cockpit. The maximum value is about 40 cm.. The resulting value of dse turns out to e be approximately 1.25 to 1.5o /cm. (2.2 to 2.6 Radians/m). Some aircraft have a variable gear ratio depending on the aircraft conguration. The required value of Cme is determined by the maximum required elevator power. Usually either the take-o or the landing at the most forward cg position is the most critital condition see section 5-2-6. The derivative Cht determines the required size of the trim tab, taking into account the (limited) range of the trim tab angles. For a trim tab to be eective at any occurring value of h and e the deection should not be larger than 15o . The required size of the trim tab then follows from the requirement that, according to the regulations, it must be possible to reduce the control force to zero in certain aircraft congurations, including the cg position, and over certain ranges of airspeed, see section 5-2-7. The only remaining variables which the designer can use to inuence the control forces are the stability margin, stick free and Ch . Earlier was found, see equation (8-39), CNh Ch xnf ree xnf ix = c CN Ch 1 d d Vh V
2

Cme Ch Sh lh = S c CN Ch

d d

This means that at a given position of the neutral point, stick xed, Ch and Ch are the only remaining variables with which the designer can manipulate the control forces. From the foregoing it will have become clear that the general requirement is for the center of gravity to be positioned forward of the neutral points, both stick xed and stick free. The aircraft then possesses both elevator control position and force stability. If Ch of the elevator as well as Ch are negative, the n.p.f ree lies forward of the n.p.f ix position. The aft permissable cg position is limited by the n.p.f ree . It is possible to shift this n.p. aft by reducing |Ch |. The n.p.f ree will even lie aft of the n.p.f ix , if Ch > 0. This is one reason why a slight overbalance with respect to angle of attack is sometimes aimed for when balancing the elevator. For many modern aircraft it turns out to be no longer possible to obtain acceptable control forces in all required aircraft congurations and ight conditions, merely by choosing Ch and Ch . In particular this is true for aircraft having a high wing loading W , a high maximum airS speed Vmax and/or a large ratio of maximum to minimum airspeed Vmax (V/STOL-aircraft). Vmin If purely aerodynamic means do not suce additional use can be made in the rst place of a spring or an unbalanced mass or both in the control mechanism. These devices also inuence the dynamic stability, stick free, and sometimes in an unfavourable sense. In more advanced cases, such as transonic, supersonic and some V/STOL-aircraft, springs and masses cannot oer a satisfactory solution. In those cases the aerodynamic hinge moments
Flight Dynamics

410

Longitudinal Stability and Control in Steady Flight

are balanced in total or in part by, usually hydraulic, control force ampliers, i.e. control boosters, or by servo controls. When hydraulic servo controls are used it would be sucient for the pilot to apply only the very small forces needed to operate the servo valve. Experience has shown, however, that in the interest of safety of ight the pilot must have to exert forces in his control stick or wheel having the usual levels of magnitude and varying in the normal way with airspeed, control deection and trim deection. This requirement is met by the use of a separate installation in the ight control system, a so called articial feel unit. This mechanism applies a force in the control manipulator which has to be balanced by the pilots control force. These articially generated control forces are made to vary with 1 V 2 , e and te in such a way that the pilots 2 control force shows indeed the variations found in an aircraft where the control manipulator has a direct mechanical linkage with the control surface.

8-2-6

Control forces a pilot can exert

The discussion in the previous sections centered around the control forces required to y the aircraft. The following deals with forces a pilot can exert on his controls. Evidently the designer must ensure that the required control forces remain within the capabilities of even the least strongest amongst the pilot population. Appropriate regulations aim to safeguard this requirement. The maximum control force a pilot can apply depends primarily on, 1. the individual person; strength and endurance dier appreciably from one person to another 2. the type of control manipulator (control wheel, center stick, side stick, rudder pedals) 3. the time during which the force must be exerted 4. the location of the control manipulator relative to the pilot Regarding to these items, the following can be remarked. Figure 8-29 clearly shows the large dierences in the maximum possible control forces due to individual dierences in strength and endurance of the test subjects. It appears that the largest force a pilot can exert is the rudder pedal force (curve a), the side force for aileron control is the smallest (curves d and e). It is also seen that the forces on a control wheel can be larger than those on a control stick. In the measurements of gure 8-29 the control manipulators were all handled by the right hand. This is in agreement with the situation in most single seat cockpits where the pilot uses his left hand to operate the engine throttle. Figure 8-29 gives an indication of the maximum possible control forces as a function of the uninterrupted time period over which these forces have to be applied. The maximum possible control force decreases approximately logarithmically with increasing duration. The location

8-2 Stick Free Static Longitudinal Stability

411

Figure 8-29: Maximum control forces as a function of the duration (from reference [140])

of the control manipulator is not only important because it inuences the maximum possible control forces. It also inuences the pilots comfort, see references [140, 65, 111, 36, 39]. For a large amount of information on the preferable locations of controls in the cockpit. Occasionally the aircraft ight condition can experience a rather sudden change. In view of such occurrances it is necessary to know how fast a pilot can change the control deections under various levels of control force gradients. Information on this can be found in references [162, 120, 69]. It is evident that the required control forces may not surpass the possible control forces. The various Airworthiness Regulations contain requirements concerning the maximum permissable control forces under various circumstances. Table 8-1 gives the maximum forces as stipulated in the U.S. and the British Airworthiness Regulations, see references [6] and [7] respectively.

8-2-7

Airworthiness requirements for steady straight symmetric ight

The Airworthiness Regulations, see references [11, 19, 6, 20, 12, 13, 7], require that the aircraft is statically stable in the most important aircraft congurations and ight conditions. This means that the cg must be in front of the neutral point in those conditions. Sometimes
Flight Dynamics

412

Longitudinal Stability and Control in Steady Flight

the requirement is expressed in terms of an equivalent requirement regarding the control characteristics. In view of the fact that the pilot is in general more sensitive to control forces and changes in control forces than to control positions and control displacements, the requirement usually stipulates that the aircraft possesses elevator control force stability in the relevant aircraft congurations and ight conditions at all permissable center of gravity positions. As discussed in section 8-2-2 this requirement can easily be reformulated into the condition that the center of gravity must be positioned in front of the neutral point stick free. For the designer the emphasis may be slightly dierent. In section 8-2-4 it was seen that several possibilities exist to inuence the control force through non-aerodynamic means. The arsenal of available tools is certainly not exhausted with the springs and unbalanced masses discussed in section 8-2-4. It is of great practical interest that several of these means may still be applied relatively easy in a late stage of the development of the aircraft, if the need arises, without great expense or drastic changes in the design. To change the elevator control position stability or the static stability, stick xed, usually turns out to be much more dicult since it requires considerably more drastic modications in the design, see section 8-1-2. This argument leads to an important design rule. Already in an early stage of the design of an aircraft it is essentual to ensure that, in addition to other requirements to be met, the neutral point, stick xed, in the prescribed aircraft congurations and ight conditions will lie behind the rearmost envisaged center of gravity position. Experience has shown that during the development of an aircraft the center of gravity often shows the tendency to move a bit due to the successive design modications. For this reason it may be wise to provide an additional margin when choosing the position of the neutral point, stick xed, in the various aircraft congurations and ight conditions. In the next section some control characteristics will be discussed concerning turning ight. Requirements to be met by the control characteristics in such ight conditions will prove to inuence also the forward and rear limits of the permissable center of gravity positions.

8-3

Longitudinal Control in Pull-Up Manoeuvres and Steady Turns

The previous two sections dealt with the longitudinal control characteristics in steady straight ight. Control in turning and non steady ight, i.e. in manoeuvres, was not considered so far. It is by no means certain that an aircraft showing good control characteristics in steady, straight ight, also has similarly good control characteristics in manoeuvres. It is thus necessary to study these latter control characteristics and to this end express them quantitatively in appropriate terms. This way also allows to formulate quantitative requirements for desired control characteristics. Of the many possible turning manoeuvres only two of the greatest general interest are considered here: the pull-up from a dive and the steady, coordinated, i.e. slip-free, horizontal

8-3 Longitudinal Control in Pull-Up Manoeuvres and Steady Turns

413

Requirements according to

U.S. Civil (reference [7])

British Civil (reference [6])

Rudder pedals

82 kg (short intervals) 9 kg (long intervals)

82 kg (short intervals) 22.5 kg (long intervals)

Elevator control

35 kg (short intervals) 4.5 kg (long intervals)

22.5 kg (short intervals) 2.6 kg (long intervals)

Aileron control

27 kg (short intervals) 2.3 kg (long intervals)

note that a short interval is regarded as a period of time lasting a few seconds while a long interval is regarded as a period of time lasting several minutes
Table 8-1: Maximum permissable values of the required control forces, according to the U.S. and British Civil Airworthiness Regulations

Flight Dynamics

414 turn.

Longitudinal Stability and Control in Steady Flight

During these two manoeuvres the aircraft has an angular, pitching, velocity about the lateral axis. In turns the pitching velocity q about the lateral axis is always accompanied by a yawing velocity r about the top axis, but because of the symmetry of the aircraft the inuence of the asymmetric component r on the symmetric forces and moments can be neglected, see chapter 4. An important purpose of the following discussion is to make longitudinal control in these manoeuvres accessible to quantitative study. Therefore, the control characteristics have to be expressed in characteristic variables, suciently simple to be manageable for the designer. To arrive at such characteristic variables some simplifying assumptions have to be made. Suppose the pilot of a normal stable aircraft wants to pull his aircraft up from a steady, straight, symmetric dive. In this context any descending ight will be called a dive. To achieve his purpose, he moves the elevator control back to a new position, which is assumed to be constant to simplify the discussion. The eect of this schematic control movement can be described as follows. Assuming a stable aircraft, rst the pitching velocity q and the angle of attack quickly increase to new approximately constant values depending on the ultimate control displacement. This motion occurs usually within a few seconds, often via a more or less well damped oscillation during which the airspeed remains very nearly constant. The aircraft motion, usually called the short period oscillation will be discussed in more detail in chapter 10. Figure 8-30 gives an example of such a symmetric motion caused by a step elevator deection. Figure 8-31 presents the various measured components, the elevator deection and the control force for a pull-up manoeuvre of an Auster J-5B Autocar. Due to the increase in angle of attack the total lift on the aircraft becomes larger than the aircraft weight. This causes the trajectory of the cg to be curved, the aircraft is pulled up from the dive. If the pilot keeps the elevator angle constant, a second oscillation will occur. In contrast with the rst oscillation, the second one is relatively slow and usually has a very low damping. After the second oscillation has also dampened out, the pitching velocity q is zero and the airspeed as well as the angle of attack have assumed the new constant values corresponding to the new elevator angle in steady, straight ight. In actual ight, when pulling up from a dive the pilot will suppress the second, long-period, oscillation by means of suitable, small corrective elevator movements. This second oscillation, usually called the slow oscillation, or phugoid will also be discussed in more detail in chapter 10. In the following, only the rst and relatively fast part of the motion is studied. The assumption is made that the airspeed remains constant at the level of the original steady ight condition. In addition to the assumption of a constant airspeed another simplication can be made. The normal load factor, n, is equal to 1 in the original straight ight. Due to the control displacement initiating the pull-up from the dive, the value of n increases. If the short-period oscillation is suciently damped, the normal load factor remains very nearly constant during a brief time interval, see gures 8-30 and 8-31. This fact allows the normal load factor to be considered constant in the following, just like the airspeed.

8-3 Longitudinal Control in Pull-Up Manoeuvres and Steady Turns

415

Figure 8-30: Response curves due to an elevator step deection, Lockheed 1049 C Super Constellation

Flight Dynamics

416

Longitudinal Stability and Control in Steady Flight

Figure 8-31: Response curves due to an elevator step deection, Auster J-5B Autocar (from reference [23])

8-3 Longitudinal Control in Pull-Up Manoeuvres and Steady Turns

417

Due to these two assumptions the pull-up trajectory is idealized to a steady ight condition. For steady turns the two simplifying and approximating assumptions concerning the aircraft motion need not to be made. The aircraft has by denition in a steady turn not only a constant airspeed but also a constant normal load factor. In the following two notions will be derived characterizing longitudinal control in steady pull-ups and in steady, horizontal, coordinated turns.

8-3-1

Characteristics of longitudinal control in turning ight

A notion characteristic for longitudinal control in turning ight should relate two aspects of the manoeuvre: on the one side the pilots action, the control displacement and the control force, and on the other hand a characteristic element of the resultant aircraft motion. For the latter the normal load factor n, mentioned before, is used, n= N W (8-55)

In the initial steady ight condition N = W , or n = 1. Using this load factor, the stick displacement per g and the stick force per g are obtained, where g is the acceleration due to gravity. Just as with the elevator control position stability, the elevator angle is used rather than the stick displacement itself, so, stick displacement per g, stick force per g,
dFe dn de dn

The load factor in these two expressions is not only characteristic for the change in the trajectory of the aircraft, it is also closely related to the loads imposed on the aircraft. The two characteristics de and dFe express the extent to which the control position and dn dn control force have changed after the aircraft has changed at constant airspeed from a condition of steady, straight ight to another condition of steady, but turning ight or to steady turn at a load factor n. For the case of the pull-up manoeuvre it can be argued that de dFe dn and dn should have a negative sign. This argument is as follows. In order to achieve a pull-up from a dive at a certain airspeed, an increase in angle of attack is needed. This requires an increase in the angle of pitch of the aircraft. This is the initial aircraft motion, the transition from the steady dive to the steady pull-up manoeuvre. For this initial motion a tail-heavy moment (Cm > 0) is needed and because Cm < 0, an initial backward elevator control movement and an elevator deection trailing edge up (ei < 0) are required.

Flight Dynamics

418

Longitudinal Stability and Control in Steady Flight

Maximum

Minimum

Trainer and ghter aircraft

n25.4 L 1 n54.4 L 1

9.5 nL 1

Transport aircraft and bombers

n20.4 L 1

Table 8-2: Permissable values of the stick force per g in kg

It is argued in section 8-1-3 that a very general desirable control characteristic requires the nal control displacement to have the same sign as the initial displacement. During the steady pull-up manoeuvre at a constant, positive n this general characteristic requires that also the nal e is negative. This implies that the required sign is negative, de <0 dn Here it is argued that the change in control force required for a control displacement should have the same direction as the control displacement itself. As a consequence, since de > 0, ds dFe <0 dn It is of course desirable that also in a turn both the change in control position and control force are negative. In the pull-up manoeuvre as well as in a turn, a backward control movement and an extra pull force should be needed. The stick force per g is generally assigned both a maximum and a minimum permissable value. If the stick force per g is too large in the absolute sense, the pilot will become tired too soon if he has to perform many manoeuvres. If, on the other hand, the stick force per g is too low, the aircraft is too sensitive to small unintentional variations in the control force. There is the additional risk that the aircraft may become overstrained due to a sudden control force. The dynamic characteristics of the aircraft are also important here. They have a direct bearing on the desired values of de and dFe , see reference [86]. The dn dn following table 8-2 gives the maximum and minimum permissible values of the stick force per g, derived from the U.S. military requirements, see reference [19]. It will be seen that these values are related to the maximum permissable load factor n and thus to the strength of the aircraft. A general rule is, that the absolute value of the stick force per g should not be lower than 1.4 kg. If the requirements of table 8-2 are applied to ghter aircraft having a nL = 6, the result is that the stick force per g should be between 5.1 kg and 1.9 kg. For transport aircraft,

8-3 Longitudinal Control in Pull-Up Manoeuvres and Steady Turns

419

de dn

spongy

ineective

heavy and ineective too light light and eective oversensitive too heavy
dFe dn

too sensitive
Figure 8-32: Description of the character of the control feel for various ratios of the control force per g to the control displacement per g (from reference [170]).

dse dn

(cm) 2.0

1.5 poor 1.0 acceptable satisfactory

0.5

10

dFe dn

(kg)

Figure 8-33: The inuence of the stick displacement per g and the stick force per g on the pilots opinion of the control characteristics at constant airspeed (from reference [86]).

Flight Dynamics

420

Longitudinal Stability and Control in Steady Flight

usually having a nL = 2.5, the boundaries are 32.2 kg and 13.6 kg. It is not sucient to give a maximum and a minimum permissable value of the stick force per g. The relation between the stick displacement per g and the stick force per g is also important to achieve pleasant control characteristics. Figure 8-32 shows how various combinations of the stick displacement per g and the stick force per g were judged. Figure 8-33 is another example of the inuences of the two criteria on the pilots opinion on the control characteristics of a ghter aircraft at constant airspeed. Quantitative requirements for the stick displacement per g do not exist. In the following the stick displacement per g and the stick force per g are expressed as functions of the aerodynamic coecients and the mass of the aircraft for the two cases, 1. The idealized pull-up manoeuvre previously discussed, where n and V are assumed to be constant. The trajectory of the aircraft then is an arc of a circle in the vertical plane. The extent to which this idealized pull-up manoeuvre agrees with the actual motion of the aircraft may be judged from gures 8-30 and 8-31. It will be seen that n and V indeed remain very nearly constant over a brief time interval. 2. The steady, horizontal, coordinated turn.

8-3-2

The stick displacement per g

The calculation is based on the forces and moments acting on the aircraft both before and after the transition from the steady, straight ight to the idealized pull-up manoeuvre or the steady, horizontal turn. During both manoeuvres the airspeed is assumed to be constant. This implies, that the forces in X-direction are assumed to be in equilibrium and need not be further considered. In the steady initial condition is, N0 = W and, M0 = 0 or in non-dimensional form, CN0 = and, W
1 2 2 V S

8-3 Longitudinal Control in Pull-Up Manoeuvres and Steady Turns

421

Cm0 = 0 The index 0 here indicates the steady initial condition. After the transition is, using equation (8-55), CN = and, Cm = 0 (8-57) W
1 2 2 V S

(8-56)

After the transition the angle of attack has increased an amount and the aircraft has obtained an angular velocity q about the lateral axis. The elevator angle has changed by e . The variables describing the aircraft motion now are, , q or q c and e (V = constant) V

whereas in straight ight, considered so far, the variables are, , V and e ( q c = 0) V

In chapter 3 the assumption was made that V would have no eect on the aerodynamic coecients. Due to the change in angle of attack the aircraft experiences an extra normal force CN and an extra moment Cm . The moment due to the elevator deection is Cme e . The pitching velocity q causes an aerodynamic force along the Z-axis. In section 6-4 a nondimensional stability derivative CZq was introduced, CZq = CZ CN q = q c V Vc

c Using this stability derivative, the extra force along the Z-axis can be expressed as CZq q . V q c q Here V is the non-dimensional pitching velocity, see also section 6-4. A positive force CZq Vc is directed along the positive Z-axis (downward). Usually CZq is negative. The pitching velocity q gives also rise to an aerodynamic pitching moment about the lateral axis. This moment was expressed using another stability derivative, Cmq , also discussed in 6-4,

Cmq = The pitching moment due to q is Cmq

q c V .

Cm c q V

Usually Cmq is negative.

Flight Dynamics

422

Longitudinal Stability and Control in Steady Flight

From the foregoing follows for equations (8-56) and (8-57), CN = CN CZq Cm = Cm + Cmq Equation (8-59) results in, e = 1 Cme Cm + Cmq q c V q c W = 1 2 n V 2 V S q c + Cme e = 0 V (8-58)

(8-59)

The stick displacement per g results from dierentiating this latter expression with respect to n, de 1 = dn Cme
d dn c d q d + Cmq V dn dn

Cm

follows from equation (8-58), d 1 = dn CN

c CZq d q W V + 1 2 CN dn 2 V S

(8-60)

This results in, de 1 = dn Cme Cm CN W + 1 2 2 V S Cm CZq + Cmq CN

c d q V dn

(8-61)

c The relation between the non-dimensional pitching velocity q and the normal load factor n V is yet to be determined. A distinction has to be made here between the idealized pull-up manoeuvre and the steady, horizontal turn.

A. Derivation of

c d q V dn

for the pull-up manoeuvre

In the pull-up manoeuvre the bottom of the trajectory is considered, see gure 8-34. In this lowest point of the trajectory, the Z-axis is very nearly vertical and the total force along the Z-axis is N W . The centripetal acceleration along the Z-axis is V q. As a consequence, N W =mV q or, using equation (8-55), q c g c = 2 (n 1) V V

8-3 Longitudinal Control in Pull-Up Manoeuvres and Steady Turns

423

center of rotation N

W
Figure 8-34: The forces and angular velocity q in the idealized pull-up manoeuvre

This results in,

c d q g c V = 2 dn V

(8-62)

In chapter 4 a non-dimensional measure of the aircrafts mass c was introduced. This socalled relative density is dened as, c = m W = Sc gSc

Using c , equation (8-62) can be written as,

c d q 1 W V = 1 dn 2c 2 V 2 S

(8-63)

B. Derivation of

c d q V dn

for the steady, horizontal turn

In this steady ight condition the resultant force along the Z-axis, see gure 8-35, is N W cos . The centripetal acceleration along the Z-axis is again V q, so, N W cos = m V q
Flight Dynamics

424

Longitudinal Stability and Control in Steady Flight N cos N

q r W cos W rear view YB

ZB

Figure 8-35: The forces and angular velocity in a steady horizontal turn

From gure 8-35 follows also, N cos = W This relation and equation (8-55) result in, q c g c = 2 V V and after dierentiation,
c d q 1 W V = 1 dn 2c 2 V 2 S

1 n

1+

1 n2

(8-64)

c It follows from (8-64) that in steady, horizontal turns q is a non-linear function of n. If equaV tion (8-63) is substituted in equation (8-61), the resulting expression for the stick displacement per g in the pull-up manoeuvre can be written as,

de 1 = dn Cme

W
1 2 2 V S

Cm CN

1+

CZq 2c

Cmq 2c

(8-65)

Using equations (8-64) and (8-61) the resulting expression for the steady turns reads, de 1 = dn Cme W
1 2 2 V S

Cm + CN

Cmq Cm CZq + CN 2c 2c

1+

1 n2

(8-66)

8-3 Longitudinal Control in Pull-Up Manoeuvres and Steady Turns

425

A few simplications can now be made. In equation (8-65) 2c is relatively large if compared to CZq . Also, in equation (8-66) Cmq is large in the absolute sense, if compared to Cm CZq . CN This permits both equations (8-65) and (8-66) to be simplied to,

Pull-up manoeuvres

de 1 = dn Cme

W
1 2 2 V S

Cmq Cm + CN 2c

(8-67)

Steady turns

1 de = dn Cme

W
1 2 2 V S

Cmq Cm + CN 2c

1+

1 n2

(8-68)

It follows from equation (8-67) that in pull-up manoeuvres the control deection and the elevator angle are linear functions of n. In steady, horizontal turns this is, according to equation (8-68), not the case. Then de varies with n. The dierence between de pullup and de turns decreases with dn dn dn increasing n.

From equations (8-67) and (8-68) it follows also that both for the pull-up manoeuvres and 1 the turns de is proportional to the wingloading W and to 1 V 2 . Figures 8-36 and 8-37 show dn S measurements of the stick displacement per g in pull-up manoeuvres and steady turns of the Auster J-5B Autocar.
2

For both types of manoeuvres de appears to have the desired negative sign. For this aircraft dn the relation between e and n in pull-up manoeuvres is non-linear. From the measurements in gures 8-36 and 8-37 it is clear that the required control displacement for a steady turn is larger than for the pull-up manoeuvre at the same load factor n.
Flight Dynamics

426

Longitudinal Stability and Control in Steady Flight

A quick derivation of the stick displacement per g We start with the approximated equations of motion for constant speed, see chapter 10, CZ 2uc + CZq Cm Cmq .
q c V

CZ Cm

e = 0

We consider a small deviation from steady straight ight with = , q = q and n = n 1 = V = V ( ) = V (q ). As we consider a steady state pull up manoeuvre we g g g have = 0, so V q c g c n = q, so = 2 n g V V Substitution results in CZ Cm 2uc +C Zq Cmq .
g c .n V2

CZ Cm

. e = 0

or with Axss +B e = 0, in which A= This can be solved as: xss = g c .n V2

CZ Cm

2uc +C Zq Cmq

, and B =

CZ Cm

= A

Be , or

e g c .n V2 e

= A1 B

By applying Cramers rule we get: n = e or: CZ Cm CZ Cm CZ Cm CZ Cm Cm CZ V2 . CZ Cmq Cm (2uc + CZq ) g c

2uc + CZq Cmq

When simplifying this expression by settingCZ CZq 0, and replacing CZ by CN and W uc by gS we get: c Cmq de W e 1 Cm = = .( + ) 1/ V 2 S CN dn n Cm 2 2c

CZ Cmq Cm (2uc + CZq ) g e c = . 2 n CZ Cm Cm CZ V

8-3-3

The manoeuvre point, stick xed

In equations (8-67) and (8-68) the rst term between parantheses is Cm . From the discussion CN of the static longitudinal stability and the neutral point stick xed, from section 8-1-2 follows, xcg xnf ix Cm = CN c (8-69)

8-3 Longitudinal Control in Pull-Up Manoeuvres and Steady Turns

427

Figure 8-36: The incremental elevator deection e as a function of the incremental load factor n in pull-up manoeuvres, Auster J-5B Autocar (from reference [23]).

In equation (8-69) xnf ix is the x-coordinate of the neutral point, stick xed, in gliding ight. The rst contribution to de then appears to be proportional to the distance of the center of dn gravity to the neutral point. As a consequence the stick displacement per g will decrease in the absolute sense if the cg is shifted backward. This can clearly be seen in gures 8-36 and 8-37. It follows that there is a certain cg position where the stick displacement per g is zero. This is the case for the pull-up manoeuvres, if, see equation (8-67), xcg xnf ix Cmq Cmq Cm + = + =0 CN 2c c 2c (8-70)

The cg position at which the stick displacement per g is zero, is called the manoeuvre point, stick xed, abbreviated as m.p.f ix . The x-coordinate of this point is xmf ix . From equation (8-70) follows, if xcg = xmf ix , xmf ix xnf ix Cmq = c 2c (8-71)

Here Cmq has been assumed independent of the cg position, see chapter 6. Combining equations (8-69) and (8-71) results in, xcg xmf ix Cmq Cm = + c CN 2c (8-72)
Flight Dynamics

428

Longitudinal Stability and Control in Steady Flight

Figure 8-37: The incremental elevator angle e as a function of the incremental load factor in turns, Auster J-5B Autocar (from reference [23])

Next, equation (8-72) is substituted in equation (8-67). This results in an expression for the stick displacement per g as a function of cg position relative to the m.p.f ix , de 1 = dn Cme W
1 2 2 V S

xcg xmf ix c

(8-73)

But the manoeuvre point, stick xed, is not just the cg position at which the stick displacement per g vanishes. The manoeuvre point can be interpreted in yet another way, as explained in the following. The normal force N can generally be written as, see equation (8-55), N =nW Then, dCN W = 1 2 dn 2 V S Substitution of equation (8-74) in equation (8-73) results in,

(8-74)

8-3 Longitudinal Control in Pull-Up Manoeuvres and Steady Turns

429

de 1 dCN xcg xmf ix = dn Cme dn c or, Cme de = dCN xcg xmf ix c (8-75)

Here Cme de is the change in the pitching moment due to the elevator deection de and dCN is the change in the normal force coecient. According to equation (8-75) Cme de
f ix f ix must be balanced by dCN . This means that is the arm of the force dCN . c c Evidently, the second interpretation of the manoeuvre point, stick xed, is the point where the resultant change in normal force acts after the transition from a condition of steady, straight ight to a steady pull-up manoeuvre, if the elevator angle were kept constant during this transition.

xcg xm

xcg xm

Since Cmq is negative, it follows from equation (8-71) that the m.p.f ix always lies behind the n.p.f ix . With increasing c , i.e. with increasing ight altitude, the manoeuvre point moves forward. This can also be seen in gure 8-38 where the calculated position of the m.p.f ix of the Fokker F-27 Friendship is shown. Subgure a of gure 8-38 clearly shows the forward shift of m.p.f ix and also the limiting Cm position: n.p.f ix . The latter position is attained when the term 2cq becomes vanishingly small. Thus far the manoeuvre point was discussed only in relation to the pull-up manoeuvres. Both interpretations of the manoeuvre point hold equally for the steady turns. The dierence, however, is that then the position of the manoeuvre point is a function also of the load factor n. This can be seen as follows. From equations (8-68) and (8-69) follows for the steady turn, xmf ix xnf ix Cmq = c 2c and by consequence, xcg xmf ix Cmq Cm = + c CN 2c 1+ 1 n2 (8-77) 1+ 1 n2 (8-76)

The expression (8-73) is valid also for the stick displacement per g in steady turns, although the actual position of the manoeuvre point will be dierent from the position in the pull-up manoeuvre. From equation (8-76) it can be seen that xmf ix now depends not only on the ight altitude, but also on the load factor. If n = 1, xmf ixpullup xnf ix xmf ixturns xnf ix =2 c c (8-78)

With increasing n the manoeuvre point in turns shifts forward. At high values of n the manoeuvre points in the pull-up and in the turn very nearly coincide. This can be seen also
Flight Dynamics

430

Longitudinal Stability and Control in Steady Flight

Figure 8-38: Calculated positions of the manoeuvre point, stick xed, and the stick displacement per g of the Fokker F-27 Friendship

8-3 Longitudinal Control in Pull-Up Manoeuvres and Steady Turns

431

in subgure b of gure 8-38. It should be remembered that n = 2 in a steady, horizontal turn 1 corresponds to an angle of roll = 60o , (n = cos ). Apart from the stick displacement per g, the stick force per g is of particular interest. It is discussed in the following section.

8-3-4

The stick force per g

The elevator control force was written in section 5-2-7 as, see equation (5-83), Fe = de 1 2 V Se ce dse 2 h Ch h + Ch e + Cht te

The stick force per g is the derivative of the elevator control force with respect to the load factor in a pull-up manoeuvre or a steady turn. It is a measure of the change in elevator control force after the transition from a condition of steady, straight ight to a steady pullup manoeuvre or a steady turn. In equation (5-83) both h and e are functions of the load factor n. It is assumed that the trim tab angle does not vary in the pull-up manoeuvre or the turn, te is constant. The airspeed V and thus also Vh in equation (5-83) are assumed to be constant. The stick force per g is then obtained from equation (5-83), de 1 2 dFe = V Se ce dn dse 2 h Ch dh de + Ch dn dn (8-79)

The stick displacement per g, the factor de on the right hand side of equation (8-79), has dn already been determined in section 8-3-2. In order to nd dFe the derivative dh needs to be dn dn determined. The angle of attack h of the horizontal tailplane varies with n in the steady manoeuvres c because both and q, q , vary with n and, V h = f , q c V

The partial variation of h with alone was already discussed in section 5-2-4. For steady, straight ight it was found that, see equation (5-65), h = ( 0 ) 1 d d + (0 + ih )

The partial variation of h with q alone will now be derived. In the idealized pull-up manoeuvre and in the steady turn the symmetric motion of the aircraft relative to the air can be considered as a pure rotation, with angular velocity q, about a center of rotation situated on the Z-axis above the center of gravity, see gure 8-39. This position of the center of rotation is such that the local angle of attack at the center of gravity does not change due to the rotation. When in chapter 4 the stability derivatives with respect to pitching velocity were discussed, it was shown that the principal eect of such a q-motion is a variation of the local, geometric
Flight Dynamics

432

Longitudinal Stability and Control in Steady Flight q

center of rotation =
xxcg R

xxcg q c c V

R=

V q

Zr

xcg x Xr

Figure 8-39: The variation of the angle of attack in an arbitrary point of the aircraft caused by a pure q-motion

angle of attack, proportional to the angular velocity q and the distance in x-direction to the center of gravity, see gure 8-39, = x xcg q x xcg c = R c V (8-80)

At the horizontal tailplane the change in geometric angle of attack due to a q-motion is, h = xh xcg q lh q c c c V c V (8-81)

In steady, turning ight the total h is found by adding equation (5-65) and equation (8-81), h = ( 0 ) 1 d d + (0 + ih ) + lh q c c V (8-82)

From equation (8-82) follows the derivative with respect to n, dh = dn The derivatives (8-64).
d dn

d d

q c d lh d V + dn c dn

(8-83)

and

c d q V dn

were aready obtained previously, see equations (8-60), (8-63) and

All variables needed to determine and (8-64) results in,

dh dn

are now known. Substituting equations (8-60), (8-63)

8-3 Longitudinal Control in Pull-Up Manoeuvres and Steady Turns Pull-up manoeuvres dh W = 1 2 dn V S 2 Steady turns dh W = 1 2 dn 2 V S 1 d d lh 1 1 + CN c 2c 1+ 1 n2 1 d d lh 1 1 + CN c 2c

433

(8-84)

(8-85)

All factors in (8-79) are now known. Substituting and using the expressions already derived in sections 5-2-6 and 8-2-1, i.e. equation (5-76), Vh V
2

Cme = CNh equation (8-35),

Sh lh S c

Cmf ree = CNw and equation (8-33),

xcg xw CNh f ree c

d d

Vh V

Sh lh S c

CNh

f ree

= CNh CNh

Ch Ch

the stick force per g is obtained after some reformulation, Pull-up manoeuvres dFe de W = dn dse S Vh V
2

Se ce

Ch Cme

Cmf ree CN

1 2c

Cmq Cme

Ch lh Ch c

(8-86)

Steady turns dFe de W = dn dse S Vh V

2

Se ce

Ch Cme 1+

Cmf ree CN 1 n2

1 2c

Cmq Cme

Ch lh Ch c

(8-87)

The rst term between the brackets of the expressions (8-86) and (8-87) is due to the static stability, stick free. The second term between the brackets in equation (8-86) can be simplied
Flight Dynamics

434

Longitudinal Stability and Control in Steady Flight

h considerably. From equation (8-81) It can be seen that lc is a measure of the change in angle of attack of the horizontal tailplane caused by the q-motion, i.e.,

dh lh q = c c dV

(8-88)

The variation of the elevator angle with the angle of attack of the horizontal tailplane in the stick free situation was expressed in equation (8-27), de dh = Ch Ch

f ree

Combining equations (8-88) and (8-27) results in, de c d q V

C

=
f ree

de dh

f ree

Ch lh dh q = c Ch c dV

h Evidently, the term Cme Ch lc in equations (8-86) and (8-87) represents the moment h due to the q-movement, caused by freeing the elevator for,

dCm c d q V

= Cmq = Cme

de c d q V

f ree

= Cme

Ch lh Ch c

(8-89)

If a new derivative, Cmq stick free, is introduced it can be written as, Cmqf ree = Cmqf ix + Cmq or, with equation (8-89), Cmqf ree = Cmq Cme Ch lh Ch c (8-90)

With equations (8-90), (8-86) and (8-87) this can be written as, Pull-up manoeuvres dFe de W = dn dse S Vh V
2

Se ce

Ch Cme

Cmf ree CN

Cmqf ree 2c

(8-91)

Steady turns dFe de W = dn dse S Vh V

2

Se ce

Ch Cme

Cmf ree CN

Cmqf ree 2c

1+

1 n2

(8-92)

8-3 Longitudinal Control in Pull-Up Manoeuvres and Steady Turns

435

From equations (8-91) and (8-92) the inuence of the cg position on the stick force per g can be understood. If the cg moves aft, Cmf ree decreases in the absolute sense and as a consequence the stick force per g decreases in the absolute sense.

From the above expressions it will be seen that contrary to the stick displacement per g the stick force per g is independent of airspeed if the aerodynamic derivatives in equations (8-91) and (8-92) are invariant with airspeed. The variation of the stick force per g with the cg position is also evident from the measurements shown in gures 8-40 and 8-41. The stick force per g in steady turns is indeed, in the absolute sense, larger than in pull-up manoeuvres at the same load factor n. The relatively large scatter in the data points, visible mainly in gure 8-40, is thought to be caused by friction in the control system.

A quick derivation of stick force per g Starting from the approximated equations of motion for constant speed, see chapter 10, CZ 2uc + CZq Cm Cmq .
q c V

CZ Cm

.e = 0
V g

we substitute = , q = q and n = n 1 = V = V ( ) = g g in a steady state pull up: V q c g c n = q, so = 2 n g V V which results in: CZ Cm 2uc + CZq Cmq . c q V + CZ Cm

(q ), knowing that

.e = 0

The corresponding stick force is : Fe = , with K= and h = (1 So e = de 1 2 . /2Vh Se ce {Ch .h + Ch .e } = K.{Ch .h + Ch .e } dse de 1 2 . /2Vh Se ce dse d h q c ). + . d c V

Fe 1 Ch d Ch h q c (1 ). . K Ch Ch d Ch c V
Flight Dynamics

436

Longitudinal Stability and Control in Steady Flight

Substituting this in the equations of motion results in: CZ Cm 2uc + CZq Cmq CZ Cm .{ . c q V +

CZ Cm

Ch d Fe 1 Ch h q c (1 ). . } = K Ch Ch d Ch c V C C d h CZ Ch (1 d ) 2uc + CZq CZ Ch c h h . + c Ch Ch h d q Cm Ch (1 d ) Cmq Cm Ch c V

CZ Cm

Fe 1 =0 K Ch

Now we introduce: d Ch (1 )=CZ f ree Ch d Ch h CZq CZ = CZq f ree Ch c Ch d Cm Cm (1 )=Cm f ree Ch d Ch h Cmq Cm = Cmq f ree Ch c CZ CZ resulting in: CZ f ree 2uc + CZq f ree Cm f ree Cmq f ree or: A.xss + B. Of which the solution is: xss = A1 B
c We may solve for q as: V

c q V

CZ Cm

F 1 =0 K Ch

F 1 =0 K Ch c q V

F 1 with xss = K Ch

q c = V

CZ f ree CZ Cm f ree Cm CZ f ree 2uc + CZq f ree Cm f ree Cmq f ree

F 1 K Ch

F 1 CZ f ree Cmq f ree Cm f ree (2uc + CZq f ree ) K Ch .

Cm .CZ f ree CZ .Cm f ree

8-3 Longitudinal Control in Pull-Up Manoeuvres and Steady Turns

437

With good approximation we may set CZq = CZ = 0, so that CZf ree = CZ and CZqf ree = CZq = 0 so q c Fe 1 Cm .CZ = . V CZ .Cmqf ree 2c Cmf ree K Ch
c and using q = V g c V 2 n

and CZ = CN : Cmf ree CN + Cmqf ree 2c CN Cmf ree CN + = Cmqf ree 2c Cmqf ree 2c

Fe g Ch c = K.2c 2 . n V Cm or:

c de 1 2.W g Ch 2 . . /2Vh Se ce . dse gS V 2 Cm c

Cmf ree

Ch dFe de W Vh 2 = .( ) Se ce . dn dse S V Cm

8-3-5

The manoeuvre point, stick free

In section 8-3-3 it was shown that at a certain cg position corresponding with the manoeuvre point, stick xed, the stick displacement per g becomes zero. For the stick force per g also a certain cg position exists where dFe is zero. This point is called the manoeuvre point, stick dn free. It is also indicated as m.p.f ree , the abscissa of this point is xmf ree . From the discussion in section 8-2-2 follows, Cmf ree CN If xcg = xmf ree then
dFe dn

xcg xnf ree c

(8-93)

= 0, and, Cmf ree CN Cmqf ree 2c

=0

For the pull-up manoeuvre the above can be written as, see equations (8-91) and (8-93), Cmqf ree xmf ree xnf ree = c 2c

(8-94)

Since Cmqf ree is again negative, the manoeuvre point, stick free, lies behind the neutral point, stick free in gliding ight. According to equations (8-93) and (8-94) is for the pull-up manoeuvres, Cmf ree Cmqf ree xcg xmf ree = + c CN 2c

(8-95)
Flight Dynamics

438

Longitudinal Stability and Control in Steady Flight

Figure 8-40: The incremental control force Fe as a function of the incremental load factor n in pull-up manoeuvres, Auster J-5B Autocar (from reference [23])

Figure 8-41: The incremental control force Fe as a function of the incremental load factor n in turns, Auster J-5B Autocar (from reference [23])

8-3 Longitudinal Control in Pull-Up Manoeuvres and Steady Turns For the stick force per g follows then with equations (8-91) and (8-95), dFe de W = dn dse S Vh V
2

439

Se ce

Ch xcg xmf ree Cme c

(8-96)

For the steady turns it can be derived in the same way, Cmqf ree xmf ree xnf ree = c 2c and consequently, Cmqf ree Cmf ree xcg xmf ree = + c CN 2c 1+ 1 n2 (8-98) 1+ 1 n2 (8-97)

This means that equation (8-96) is also true for the stick force per g in steady turns, although the position of the manoeuvre point, stick free, diers for the two types of manoeuvres, according to equations (8-94) and (8-97). Just as in the case of the manoeuvre point, stick xed, a second interpretation can be given of the manoeuvre point, stick free. To this end, the transition from a condition of steady, straight ight to a manoeuvre is separated in two phases. These will be considered separately one after another. In the rst phase, the aircraft is supposed to change with free elevator, from steady, straight ight to a turning manoeuvre. As far as the forces along the Z-axis are concerned this is a steady manoeuvre: equilibrium exists along the Z-axis. The change in normal force dCNf ree due to this transition acts in a certain point, the abscissa of which is provisionally indicated as xdCNf ree . The accompanying change in the pitching moment due to this transition then is, dCmf ree = dCN xcg xdCNf ree c

The second and subsequent phase takes place at constant , q and V . In this phase equilibrium of the pitching moment is obtained. The moment dCmf ree is balanced by applying an appropriate elevator deection de and the control force dFe . The required elevator angle de is, de = xcg xdCNf ree dCmf ree 1 = dCNf ree Cme Cme c (8-99)

The required elevator control force dFe follows from the change in the hinge moment occuring in the second phase. Note that in the rst phase the elevator was free: Fe = 0. Thus, dFe = de 1 2 V Se ce dChe dse 2 h (8-100)
Flight Dynamics

440

Longitudinal Stability and Control in Steady Flight

Figure 8-42: Calculated positions of the manoeuvre point, stick free, and the stick force per g of the Fokker F-27 Friendship

8-3 Longitudinal Control in Pull-Up Manoeuvres and Steady Turns

441

altitude

Ch < 0

Ch = 0

Ch > 0

xmf ree h1

xmf ree = xmf ix

xmf ree

xnf ree
dFe () dn

xnf ree = xnf ix

xnf ree

x c

Ch < 0

Ch = 0

Ch > 0

0 h1 h1

0 h1

xmf ree = xmf ix

xmf ree = xmf ix

x c

Figure 8-43: Inuence of altitude and magnitude and sign of Ch on the position of the manoeuvre point, stick free, and the stick force per g

Flight Dynamics

442

Longitudinal Stability and Control in Steady Flight

Since h is constant in the second phase ( and q remain constant), dChe is, dChe = Ch de (8-101)

Substituting de according to equation (8-99) in equation (8-101) and subsequently in equation (8-100) results in, xcg xdCNf ree Ch de 1 2 Vh Se ce dCNf ree dse 2 Cme c

dFe =

(8-102)

Next, the change in elevator control force dFe is considered, following from the expression (8-96) for the stick force per g. With equation (8-74), where dCN = dCNf ree , from equation (8-96) follows for both types of manoeuvres, dFe = xcg xmf ree de 1 2 Ch Vh Se ce dCNf ree dse 2 Cme c (8-103)

If the expressions (8-102) and (8-103) are now compared, it will be clear that the manoeuvre point, stick free, in equation (8-103) corresponds with the point of action of the change in normal force dCNf ree in equation (8-102), caused by the transition with free elevator from steady, straight ight to a manoeuvre. This explains the second interpretation of the manoeuvre point, stick free. Using equations (8-94) and (8-97) the position of the manoeuvre point was calculated for the Fokker F-27. The results are shown in gure 8-42. The Ch of the elevator of this aircraft is approximately zero. As a result, there is no dierence between the situations: control free and control xed. It can be seen from the gure that the manoeuvre point moves forward with increasing ight altitude, approaching in the limit the neutral point. With increasing load factor n the manoeuvre point in steady turns approaches the manoeuvre point in pull-up manoeuvres. Figure 8-43 indicates schematically how the position of the manoeuvre point, stick free, and dFe dn change with ight altitude and with Ch . Figure 8-44 shows the inuence of Ch on the range of permissible center of gravity positions. The discussion in section 8-3-1 referred to the fact that the stick force per g has an upper and a lower limit of permissable values. Other requirements remain also in force, such as those relating to the elevator control position and force stability. As a consequence it is sometimes not an easy matter to satisfy all requirements on the stability and control characteristics in all aircraft congurations and ight conditions stipulated in the regulations.

8-3-6

Non-aerodynamic means to inuence the stick force Per g

Sometimes springs or unbalanced masses are installed in the control mechanism, to inuence the elevator control force stability or the stick force per g or both.

8-3 Longitudinal Control in Pull-Up Manoeuvres and Steady Turns

443

Figure 8-44: Increasing the range of permissible cg positions by decreasing |Ch |

In section 8-2-4 the inuence was discussed of a spring or an unbalanced mass on the elevator control force stability and the position of the neutral point, stick free. The spring was assumed to exert a constant hinge moment. Such a spring has no inuence on the stick force per g, but an unbalanced mass has. Suppose, a mass has been installed in the control mechanism, see gure 8-45, such that in straight and level ight a static hinge moment Hew is generated. In a manoeuvre at a load factor n an extra hinge moment (n 1) Hew occurs. This hinge moment must be balanced by an extra control force Fe , Fe = de n Hew dse

The inuence of the mass, sometimes called a bobweight, on the stick force per g is, dFe de = H ew dn dse

(8-104)

In this way it is seen that a modication of the stick force per g can be obtained which is independent of cg position and ight altitude. In some cases this may bring dFe within the dn required limits. By choosing the right combination of springs and bobweights it is possible in principle to vary both the elevator control force stability and the stick force per g independent of one another using non-aerodynamic means. If, however, the hinge moment generated by springs or bobweights or both becomes too large, the dynamic stability, stick free, may be inuenced
Flight Dynamics

444

Longitudinal Stability and Control in Steady Flight

Figure 8-45: The incremental hinge moment due to a bobweight in the control mechanism in a pull-up manoeuvre at a load factor n

unfavorably. In addition, the possibility of the occurrence of utter has to be investigated. It was shown in this chapter that the stick displacement per g and the stick force per g must have a negative sign, whereas the magnitude of the stick force per g is additionally limited to lie within a certain range. These requirements may restrict both the forward and the rear limits of permissible cg positions. This remark is supplementary to what was noted in section 8-2-7. There the rear limit of cg positions was said to be determined by the requirement that the aircraft must possess elevator control position and elevator control force stability. The forward limit of permissible cg positions was there shown to be dictated by the available elevator power in take-o or landing.

Chapter 9 Lateral Stability and Control in Steady Flight

9-1

Introduction

Using the equilibrium equations for asymmetric ight, the lateral control characteristics in some important steady, asymmetric ight conditions will be discussed in the following. To simplify the discussion only horizontal ight will be considered. In a condition of steady ight, in which the roll angle is constant by denition, the rolling velocity about the XB -axis is zero. This is the case when the XB -axis, along the velocity vector of the c.g., lies in the horizontal plane, see equation (3-158) on page 116 From the above follows that in steady, asymmetric horizontal ight the rolling velocity is always zero. The steady rolling ight to be considered in section 9-6, is an approximated, quasi-steady ight condition giving in a simple manner insight in some important requirements on roll control for aircraft.

9-2

Equations of Equilibrium

In the most general case of a horizontal, steady, asymmetric condition of ight, the aircraft rb sideslips and yaws at constant and 2V . The angle of roll of the aircraft diers from zero but the rolling velocity is zero. To maintain this ight condition, generally the ailerons as well as the rudder are deected. Not only the aerodynamic force Y along the YB -axis and the aerodynamic moments L and N about the XB - and ZB -axis respectively act on the aircraft, but also the component of the weight along the YB -axis. Since the pitch angle , measured in the stability reference frame, is zero in horizontal ight, the component of the weight along the YB -axis (or YS -axis) is,
Flight Dynamics

446

Lateral Stability and Control in Steady Flight

r W cos W

W sin Y

ZB

YB

Figure 9-1: The forces along the YB -axis of an aircraft in steady, horizontal, asymmetric ight

W sin The combined forces along the YB -axis cause a centripetal acceleration in the YB -direction: V r, see gure 9-1. The resultant equation for the forces along the YB -axis, both in the aircraft and the stability reference frame, then is, W sin + Y = m V r whereas the equilibrium of the moments about the XB - and ZB -axes is expressed by, L=0 (9-2) (9-1)

N =0

(9-3)

The equations are made non-dimensional by dividing equation (9-1) by 1 V 2 S, and both 2 equations (9-2) and (9-3) by 1 V 2 Sb. This results for small values of in, 2 CL 4b
rb 2V

+ CY

= 0 (9-4)

C = 0 Cn = 0 where,

9-2 Equations of Equilibrium

447

CL = and

W
1 2 2 V S

b =

m Sb

is the non-dimensional mass parameter for the asymmetric motions, see chapter 4. If the aerodynamic forces and moments in equation (9-4) are now expressed in the contrirb butions arising from , 2V , a and r , according to equations (7-2) and (7-3), the resulting asymmetric equilibrium equations for horizontal steady asymmetric ight become, CL + CY + (CYr 4b ) C Cn +Cr +Cnr
rb 2V rb 2V rb 2V

+ CYa a + CYr r = 0 + Ca a + Cr r = 0 + Cna a + Cnr r = 0 (9-5)

Based on the discussions in chapter 7, the control derivatives CYa and Cr will be neglected in the following. In a rst approximation used to study the various types of steady ight, CYr , Cna and CYr will also be dropped. In the thus simplied form the equilibrium equations read as, CL + CY 4b C + Cr rb =0 2V (9-6)

rb + Ca a = 0 2V rb + Cnr r = 0 2V

(9-7)

Cn + Cnr Or, using matrix notation,

(9-8)

CL CY 0 C 0

4b Cr Cnr

0 Ca 0

0 0 Cnr

Cn

rb 2V = 0 a r
Flight Dynamics

448

Lateral Stability and Control in Steady Flight

In section 9-3 to 9-5 the lateral control characteristics of the aircraft in various steady, asymmetric ight conditions are studied, using the above equations. Because of the simplications introduced, the considerations are primarily qualitative in nature, especially for larger deviations from symmetric ight.

9-3

Steady Horizontal Turns

In the three equations (9-6), (9-7) and (9-8) ve variables occur. This implies, that steady rb turns at a given airspeed and yaw-rate 2V can be own in principle at innitely many combinations of the remaining four variables r , a , and . Only if one of these four variables has rb been xed, the remaining three can be expressed as functions of 2V . In the following steady turns are studied in which each of the four variables mentioned are assumed separately to be equal to zero.

9-3-1

Turns Using the Ailerons Only, r = 0

Using equations (9-7) and (9-8) and the condition r = 0, it follows that the variation of the aileron angles with rate of yaw can be expressed as, da 1 C Cnr Cr Cn = rb Ca Cn d 2V Here Ca < 0 and Cn > 0. To initiate a turn to the right using the ailerons only, a negative aileron deection must be given to obtain a positive rolling velocity and a positive angle of roll. Using the same arguments as in chapter 8, it is desirable that in the ultimate steady ight condition the da aileron control remains deected in the direction of the initial control deection, so d rb < 0. According to equation (9-9) the necessary condition for this is, C Cnr Cr Cn > 0
2V

(9-9)

(9-10)

This latter condition corresponds to the condition for spiral stability which is discussed in chapter 11. If r = 0 in the expressions (9-6) and (9-8), the variation of angle of roll with rate of yaw can be derived as, 4b + CY d = rb CL d 2V
Cnr Cn

>0

(for

CL > 0)

(9-11)

9-3 Steady Horizontal Turns a [o ]

2

449

-12

-8

-4

rb 2V

103

-2

[o ]
1

-12

-8

-4

rb 2V

103

-1

[o ]
60

40

20

0 -12 -8 -4 0 4 8

rb 2V

103

-20

-40

-60

Figure 9-2: Steady turns using ailerons only, North American Harvard II B, gliding ight, CL = 0.31, xc.g. = 0.304 c, V = 78 m/sec (from reference [14])
Flight Dynamics

450 and also, d Cn = r >0 rb Cn d 2V

Lateral Stability and Control in Steady Flight

Cnr < 0,

Cn > 0

(9-12)

In a turn to the right, with r = 0, the aircraft has a roll angle to the right ( > 0) and the rb sideslip is towards the inside of the turn ( > 0). At practical values of and 2V the angle of sideslip usually remains limited to a few degrees. Figure 9-2 shows the results of some measurements made in steady ight, using the ailerons only.

9-3-2

Turns Using the Rudder Only, a = 0

Using equations (9-7) and (9-8) and the condition a = 0, the variation of the rudder angle with rate of yaw is obtained as, dr 1 C Cnr Cr Cn = rb Cnr C d 2V (9-13)

Here Cnr < 0 and C < 0. It then follows that if the condition equation (9-10) for spiral stability is satised, in a turn to the right the rudder is also deected to the right (r < 0). In analogy with equations (9-11) and (9-12) the following expressions for the variations of angle of roll and angle of sideslip hold, 4b + CY d = rb CL d 2V and, d C = r >0 rb C d 2V Cr > 0, C < 0 (9-15)
Cr C

>0

(for

CL > 0)

(9-14)

In a turn to the right, using the rudder only, the aircraft assumes again a positive angle of roll (to the right) and a positive, usually very small, angle of sideslip. Figure 9-3 shows the results of some measurements made in steady ight using the rudder only. As in gure 9-2, the control surface deection and the angle of sideslip remain very small, even at angles of roll of 40o . As a consequence, the inuence of the limited accuracy of the measurements and possible non-linearities is large enough to obscure a clear relation between the slopes of some measured curves and the expressions derived in this section, such as equation (9-15).

9-3-3

Coordinated Turns, = 0

When ying a steady turn, the pilots aim is in principle to keep the angle of sideslip zero. Then the drag is at a minmum. In additiion, the coordinated turn is the most comfortable turn for passengers.

9-3 Steady Horizontal Turns r [o ]

1

451

-12

-8

-4

rb 2V

103

-1

[o ]
1

-12

-8

-4

rb 2V

103

-1

[o ]
60

40

20

0 -12 -8 -4 0 4 8

rb 2V

103

-20

-40

-60

Figure 9-3: Steady turns using rudder only, North American Harvard II B, gliding ight, CL = 0.31, xc.g. = 0.304 c, V = 78 m/sec (from reference [14])
Flight Dynamics

452 If = 0, it follows from equation (9-6) that,

Lateral Stability and Control in Steady Flight

CL = 4b

rb 2V

(9-16)

This means that in the coordinated turn the component along the YB -axis of the centripetal acceleration is caused only by the lateral component of the aircraft weight. This is true, also for all objects in the aircraft. They do not experience a side force during the coordinated turn. As a consequence the pilot can also feel if he ies a well coordinated turn. The variation of the roll angle with rate of yaw follows from equation (9-16), d 4b = >0 rb CL d 2V (for CL > 0) (9-17)

The required control surface deections follow from equations (9-7) and (9-8) with = 0, da C = r >0 rb Ca d 2V and, Cn dr = r <0 rb Cnr d 2V Cnr < 0, Cnr < 0 (9-19) Cr > 0, Ca < 0 (9-18)

In a coordinated turn to the right the aircraft has an angle of roll to the right. The rudder is deected in the desirable negative direction (to the right). But the aileron is deected to the left, in the direction opposite to that of the initial control deection. As a matter of fact, this is less desirable, however it is inevitable in normal aircraft. Figure 9-4 shows the results of measurements made in steady, coordinated turns. The aileron angle turns out to be very nearly constant. Quantitative calculations show very simply that the control surface deections in each of the three types of steady turns just described remain usually small, see also gures 9-2, 9-3 and 9-4. It follows then that these ight conditions are not at all critical for the sizing of the control surfaces. In cruising ight and usually also in the approach the dierences between the coordinated turn ( = 0) and a turn using the ailerons only (r = 0) is so small as to make it attractive to the pilot to perform the lateral control using the roll control only. If the initiation of a turn is done not too quickly, the adverse yaw of a conventional aircraft will be small enough to render the use of the rudder unnecessary. Obviously, if the rudder need not be employed, the control of the aircraft is much easier, thereby improving the accuracy of the control of the aircraft.

9-4 Steady, Straight, Sideslipping Flight

453

9-3-4

Flat Turns

From equation (9-6) follows, if = 0, d 1 = 4b < 0 rb CY d 2V Using equation (9-20) it is simple to derive that, da >0 rb d 2V dr <0 rb d 2V CY < 0 (9-20)

In a at turn to the right, the rudder angle is negative, just as in a turn to the right using the rudder only. In a at turn the aircraft must perform a sideslip to the outside of the turn ( < 0 if r > 0) in order to obtain the lateral force required for the centripetal acceleration. Due to the combined sideslipping and yawing motions, a positive rolling moment (to the right) is generated. This requires a positive aileron deection (to the left) for equilibrium about the XB -axis. Figure 9-5 shows the results of some measurements made in steady, at turns.

9-4

Steady, Straight, Sideslipping Flight

In general the pilot will try to avoid sideslipping in ight. The intentional use of steady, straight, sideslipping ight is limited to landings with cross wind and to the approach in apless aircraft in order to control the ight path angle. A further discussion of steady, straight sideslipping ight is justied, however, because the aircraft may enter into a sideslip even without the pilots intention. Using the lateral control characteristics in steady, straight, sideslips, it is simple to verify if the eective dihedral, C , and the static dimensional stability, Cn , have the correct sign. These two stability derivatives have a strong inuence on the lateral control characteristics in general. They are also two important parameters in obtaining good lateral stability characteristics. The equilibrium equations for steady, straight sideslip follow directly from the expressions rb (9-6), (9-7) and (9-8), by letting 2V = 0. Neglecting small contributions, the equilibrium equations then become,

CL CY 0 C 0

0 Ca 0

0 0 Cnr

Cn

=0 a r

(9-21)

Flight Dynamics

454 a [o ]
2

Lateral Stability and Control in Steady Flight

-12

-8

-4

rb 2V

103

-2

r [o ]
1

-12

-8

-4

rb 2V

103

-1

[o ]
60

40

20

0 -12 -8 -4 0 4 8

rb 2V

103

-20

-40

-60

Figure 9-4: Steady coordinated turns, North American Harvard II B, gliding ight, CL = 0.31, xc.g. = 0.304 c, V = 78 m/sec (from reference [14])

9-4 Steady, Straight, Sideslipping Flight The variation of the angle of roll with follows from the rst row of equation (9-21). CY d = >0 d CL CY < 0, CL > 0

455

(9-22)

Apparently, a positive angle of sideslip is associated with a positive angle of roll. A given angle of sideslip creates a large lateral force CY which has to be balanced in steady ight by a component of the weight along the YB -axis. If the absolute value of CY is large, the equilibrium at a given angle of side-slip necessitates a large roll angle, as can be seen from equation (9-22). A suddenly occurring angle of sideslip, caused for instance by a gust, gives also rise to a force along the YB -axis and hence to a sideward acceleration. The pilot senses this acceleration as a sideward force exerted upon him by his seat and seat belts. If, again CY is large in the absolute sense, this force will be relatively large, making it easier for the pilot to detect the presence of the sideslipping motion. These arguments lead to the conclusion that a large negative value of CY may be considered as desirable. To initiate a straight sideslip with a positive (velocity vector pointing to the right of Xb axis) the aircraft has to be given a positive rolling moment (to the right) and a negative yawing moment (to the left). This requires an initial aileron deection to the right (negative) and an initial rudder deection to the left (positive). It is then desirable that in the ultimate steady ight condition the aileron control is also deected to the right and the rudder control to the left or, da <0 d and dr >0 d

The required surface deections follow from equation (9-21), C da = d Ca Cn dr = d Cnr (9-23)

(9-24)

As both Ca and Cnr are negative, the required directions of the control deections in steady, straight sideslips are obained, if, C < 0 and Cn > 0

It will be shown in chapter 11 that the latter conditions have to be satised also to obtain lateral dynamic stability. Measurement of the control deections in steady, straight sideslips oers a simple method to determine if these conditions are satised. Figure 9-6 shows some
Flight Dynamics

456 a [o ]
8

Lateral Stability and Control in Steady Flight

-3

-2

-1

rb 2V

103

-4

-8

r [o ]
8

-3

-2

-1

rb 2V

103

-4

-8

[o ]
8

-3

-2

-1

rb 2V

103

-4

-8

Figure 9-5: Steady at turns, North American Harvard II B, gliding ight, CL = 0.31, xc.g. = 0.304 c, V = 78 m/sec (from reference [14])

9-4 Steady, Straight, Sideslipping Flight a [o ]

8

457

-12

-8

-4

[o ]

-4

-8

r [o ]
8

-12

-8

-4

[o ]

-4

-8

[o ]
8

-12

-8

-4

[o ]

-4

-8

Figure 9-6: Steady, straight, sideslipping ight, North American Harvard II B, gliding ight, CL = 0.31, xc.g. = 0.304 c, V = 78 m/sec (from reference [14])

Flight Dynamics

458

Lateral Stability and Control in Steady Flight

results of measurements, made in steady, straight sideslips. As a nal remark it can be said that the center of gravity position has an inuence on the magnitude Cn , in analogy with the inuence on Cm . As a consequence, the c.g. position also has some inuence on the lateral stability and control. The importance of this inuence is, however, generally much less than is the case for the longitudinal stability and control characteristics.

9-5

Steady Straight Flight with One or More Engines Inoperative

For multi-engined aircraft the requirement exists that the aircraft can continue ight if one engine becomes inoperative. This imposes requirements on the controllability and usually in particular on the available rudder power. The most important consequences of the loss of the engine thrust from an engine not mounted in the plane of symmetry are a yawing moment due to the asymmetric thrust distribution and, in the case of a propeller-driven aircraft, a rolling moment caused by the resultant asymmetric lift distribution. The yawing moment can be written in non-dimensional form as, Cne = k Tp ye 1 2 2 V Sb (9-25)

Here Tp is the dierence in thrust from the operative engine and the drag of the inoperative engine; ye is the distance from the line of action of the thrust or drag to the plane of symmetry of the aircraft. Tp may be considerably larger than the thrust of the operative engine if the inoperative engine is windmilling in the airstream. Tn the case of a propeller-driven aircraft, Tp is minimal if the inoperative propeller is feathered, whereby the rotation is stopped. The factor k in equation (9-25), which for propeller-driven aircraft may assume values of 1.5 to 2.0 in some instances, expresses the fact that the yawing moment actually occurring may be larger than would result from the changes in thrust only. Figure 9-7 shows some results of measurements of Cne conrming this fact. The explanation of this eect lies in the increased lift of the part of the wing submerged in the slipstream. This increase in lift creates an extra downwash behind the wing which is now asymmetric. The downwash behind the operative propeller is larger than the downwash behind the inoperative propeller. In analogy with the airow behind a high or low mounted wing in sideslipping ight, the dierences in downwash create a circulation around the rear fuselage, thereby causing a cross-ow at the vertical tailplane. It is not dicult to verify that this cross-ow or sidewash causes an increase in the disturbing yawing moment. The experimental evidence in gure 9-7 shows that the direction of rotation of the propeller and the vertical position of the wing relative to the fuselage have an important inuence on the magnitude of Cne . Reference [57] discusses this matter in detail. In general it can

9-5 Steady Straight Flight with One or More Engines Inoperative

459

Figure 9-7: The magnitude of Cne due to an engine failure, as a function of the location and the sense of rotation of the propeller and the vertical position of the wing, measured on a model of a twin-engined propeller-driven aircraft (from reference [105]).

Flight Dynamics

460

Lateral Stability and Control in Steady Flight

be concluded, that if the propellers turn clockwise, the loss of the left engine causes the largest disturbing yawing moment Cne . This engine is then called the critical engine. The rolling moment Ce due to an inoperative engine is caused mainly by the dierence in lift on both sides of the wing, behind the operative and the in-operative engine. Some results of measurements on Ce are shown in gure 9-8. If the right engine cuts out, both a positive Cne and a positive Ce arise. Due to Cne the aircraft assumes a positive rate of yaw leading to a negative . The rolling moment Ce causes a positive rate of roll and thus a positive roll angle . The rolling motion is amplied rb by the contributions C (> 0) and Cr 2V (> 0) due to the sideslipping and yawing motions. As a consequence, the roll angle may increase very rapidly. The motion is counteracted by the yawing moments generated by p, r and (Cnp , Cnr , Cn , and by the application of both a positive rudder deection and a positive aileron deection. A new condition of steady, straight ight may now be established, described by the following equilibrium equations in which a few minor contributions have been neglected,

CL CY C Cn

0 Ca 0

CYr 0 Cnr

0 = C e a Cne r

(9-26)

In ight with asymmetrically distributed engine thrust very large rudder deections may be required. As a consequence the lateral force caused by r cannot be omitted in equation (9-26). As the four variables , , a and r occur in the equilibrium equations (9-26) there is not just one, single steady ight condition possible, but a whole range of conditions. In practice two of the many possible conditions have special signicance: steady ight at a roll angle equal to zero and the condition at which the angle of sideslip is equal to zero. If = 0 the aircraft drag is at a minimum, resulting in maximum ight performance. If the right engine is inoperative, a positive rudder deection (to the left) is required to balance Cne . In the ight condition at = 0, a positive angle of sideslip is necessary according to the top row of equation (9-26), to obtain an aerodynamic lateral force to balance the lateral force CYr r , see gure 9-9. The aircraft performs a sideslipping motion in the direction of the inoperative engine. In the ight condition with = 0 the pilot must apply a negative roll angle, assuming again that the right engine is stopped, such that the component of the weight CL balances the lateral force CYr r . The wing with the operative engine has to be kept low, see gure 9-10.

9-5 Steady Straight Flight with One or More Engines Inoperative

461

Figure 9-8: The magnitude of Cne due to an engine failure, as a function of the location and the sense of rotation of the propeller and the vertical position of the wing, measured on a model of a twin-engined propeller-driven aircraft (from reference [105]).

Flight Dynamics

462

Lateral Stability and Control in Steady Flight

Figure 9-9: Steady, straight, single-engined ight at = 0

In this situation the rudder deection required for equilibrium is smaller than in the ight condition with = 0, where not only Cne but also the moment Cn acting in the same direction has to be balanced, see the bottom row in equation (9-26). The ight condition at = 0 is, therefore, more critical for the determination of the required rudder power than the ight condition at = 0. Contrary to the moments Cnr r and Cn , the disturbing yawing moment Cne depends strongly on airspeed. Taking the case of the propeller-driven aircraft, the engine power is constant at constant throttle setting. To a rst approximation this means, Tp V = constant From equation (9-25) then follows, that Cne in that case is inversely proportional to the cube of airspeed, V 3 . There will be an airspeed below which it is not possible to balance the yawing moment due to an engine failure, Cne , see gure 9-11. This airspeed is called the minimum control speed, Vm.c. . When determining Vm.c. in ight, the regulations stipulate that || may not be larger than 5o . High engine power and low airspeeds, and consequently large values of Cne at engine failure, occur during take-o and in climb immediately after take-o. In the take-o conguration at high CL -values, the extra interference yawing moment, expressed by the factor k in equation (9-25), is also large. An additional factor is the fact that during take-o the aircraft is not allowed to loose height or to obtain large angles of roll.

9-6 Steady Rolling Flight

463

Figure 9-10: Steady, straight, single-engined ight at = 0

The importance of the minimum control speed as an element in the choice of the take-o procedure to be followed, depends strongly on the type and location of the engine. Jetpropelled aircraft, having the engines mounted on the rear fuselage, often have a minimum control speed lower than the minimum airspeed for sustained ight. For such aircraft the minimum control speed loses its signicance. Handling the aircraft on the ground with failed engine may then be an important consideration. For twin-engined, propeller-driven aircraft, however, the mininimum control speed in the take-o conguration is of extreme importance. Figure 9-12 shows some results of measurements on the Fokker F-27 with the right propeller feathered.

9-6

Steady Rolling Flight

One of the measures for the manoeuvrability of an aircraft is the rate of roll obtained by giving a certain aileron deection. The time required to establish a steady turn, is determined largely by this rate of roll. This is the reason why requirements exist for the attainable rates of roll. In many cases this xes the required aileron power. If the ailerons are deected and then maintained in their deected position, the aircraft will experience at rst an angular acceleration about the XB -axis. When the growing damping pb moment Cp 2V has become equal to the rolling moment Ca a , the rate of roll will remain
Flight Dynamics

464 Cn

Lateral Stability and Control in Steady Flight

Cne

Cne + Cn (if = 0)

Cnr r

max

V (Vm.c. )=0 (Vm.c. )=0

Figure 9-11: Minimum control speed

constant, if the simplifying assumption is made that during this rolling motion the aircraft will neither sideslip nor yaw. The resulting ight condition is that of steady rolling ight. It is possible to demonstrate, that this steady condition is usually established relatively quickly. This makes it sensible to assess the manoeuvrability of an aircraft by means of the attainable rate of roll.

Under the inuence of for instance the component of the weight CL and the yawing moments due to a and p, the aircraft will not only roll but also sideslip and yaw. This means, that continuous rudder control would be required to make the aircraft perform a rolling ight at constant rate of roll without any sideslippiag and yawing motions.

A real exactly steady rolling ight is thus not possible. The steady rolling ight to be discussed below is, therefore, a quasi-steady ight condition in straight ight (r = 0) at constant rudder angle (r = 0), at small angles of roll and to the neglect of the angle of sideslip actually occurring ( = 0). For this idealized ight condition the equation for the equilibrium of the aerodynamic moments about the XB -axis is,

Cp

pb + Ca a = 0 2V

(9-27)

9-6 Steady Rolling Flight a [o ]

40

465

30

20

10

-12

-8

-4

[o ]

r [o ]
40

30

20

10

-12

-8

-4

[o ]

[o ]
2

-12

-8

-4

[o ]

-2

-4

-6

Figure 9-12: Steady, straight, sideslipping ight with the right propeller feathered, Fokker F-27, h = 1850 m, CL = 1.92, xc.g. = 0.28 c, V = 46.3 m/sec (from reference [51])
Flight Dynamics

466 or,

Lateral Stability and Control in Steady Flight

C pb = a a 2V Cp

(9-28)

where both Ca and Cp are negative. A positive aileron deection is thus seen to cause a negative rate of roll. In the discussion of Ca and Cp it was shown that these two derivatives are to a rst approximation independent of CL or the airspeed. According to equation (9-28) the rate of roll at a given aileron angle is thus proportional to airspeed, but the non-dimensional rate of roll is independent of V . From early investigations, see reference [60], it appeared that the pilots assessment of the manoeuvrability of the then considered aircraft, showed pb a better correlation with 2V than with pmax . On this basis the requirements on max manoeuvrability were expressed using the non-dimensional rate of roll. As an example, older U.S. military regulations, see reference [19], required for a ghter aircraft, pb 2V and for the other categories of aircraft, pb 2V > 0.07
max

> 0.09
max

Since the time when the requirements were formulated the range of possible airspeeds has pb increased so much, that the non-dimensional rate of roll, 2V , as the only criterion for the manoeuvrability of an aircraft about the XB -axis, has lost some of its signicance. For V/STOL aircraft capable to operate at very low airspeeds, the rate of roll p resulting pb from the above requirements on 2V would be too low at low values of V , to ensure good manoeuvrability. For such aircraft the additional requirement is sometimes used, that in the pb approach 2V , which can be seen as the maximum vertical speed of the wing tip due to rolling, must be at least be 10 ft/sec. On the other hand, the rolling velocity at supersonic airspeed of a ghter aircraft having pb a small wing span according to the requirement on 2V , would be unreasonably large. At those high rates of roll, diculties with the dynamic stability might occur, as at high rolling velocities the symmetric and the asymmetric aircraft motions at high rates of roll may become quite violently unstable. In later U.S. military Regulations, see reference [13], the requirement is given that ghter aircraft in the combat conguration must be capable to attain an angle of roll of = 50o within one second. For other aircraft congurations and ight conditions, as well as for other categories of aircraft, requirements on non-dimensional rate of roll are maintained.
max

9-6 Steady Rolling Flight p

467

a3

a2 a1

Reversal Speed V

Vmin
Figure 9-13: The limitation in the maximum attainable roll-rate due to elastic wing deformation

It was discussed in section 7-7 that due the elastic deformation of the wing, Ca will decrease in absolute value at high airspeed, or rather at high dynamic pressures. Due to this eect, the rate of roll no longer increases linearly with airspeed at a constant aileron deection, see gure 9-13. At high airspeeds the rate of roll will even decrease. The reversal speed is reached when Ca = 0 and as a consequence also p = 0. The elastic wing twist may be reduced by locating the ailerons not at the wing tips, but more inboard. The sketch In gure 9-14 shows that this may considerably reduce the total twist at the wing tip, thereby reducing the loss in rolling moment generated by aileron deection. If the ailerons are located inboard, the outer part of the wing, generally possessing less torsional stiness, will experience no further increase in wing twist. In cases were no satisfactory roll controll can be obtained in this way, it may be necessary to use spoilers for roll control, as discussed in section 7-7. The maximum achievable rate of roll p may be limited at high airspeeds by the maximum possible roll control force, rather than by the maximum aileron deection. If no hydraulic servos are applied and if no dierential aileron deections are used, the roll control force can be written according to equation (9-40), if in addition ta = 0, as, Fa = da 1 2 1 V Sa ca Ch a + Ch a dsa 2 2 (9-29)

It will be seen, that the required control force strongly increases with airspeed at a given
Flight Dynamics

468

Lateral Stability and Control in Steady Flight

(due to a )

Figure 9-14: Variation of the wing twist angle along the wing span due to aileron deection for two locations of the ailerons

9-6 Steady Rolling Flight

469

geometric
pb 2V

ef f ective ym ba

average = 0

V
b 2

Figure 9-15: The average value of along the aileron span in steady, rolling ight

aileron deection. In equation (9-29) a is the average eective change of angle of attack due to rolling over the span of the aileron. This average eective value can be equated to a geometric change in angle of attack at a distance ym from the plane of symmetry, see gure 9-15. The expression is, pb 2ym 2V b

a =

(9-30)

For a given wing, ym can be determined using reference [155]. The aileron deection a required for a steady, rolling ight at = 0 follows from equation (9-28), pb Cp 2V Ca

a =

(9-31)

Substitution of equation (9-30) and (9-31) in (9-29) results in the following expression for the roll control force in a steady, rolling ight at = 0, as a function of rate of roll p and airspeed V, da 1 2 2ym 1 Cp V Sa ca Ch Ch dsa 2 b 2 Ca pb 2V

Fa =

(9-32)
Flight Dynamics

470 p

Lateral Stability and Control in Steady Flight

a = constant |Fa | increasing

amax a = constant elastic wing a = constant Fa = constant Famax Fa = constant Fa = constant V |a | increasing

Figure 9-16: Roll-rate p as a function of airspeed V for various values of a and Fa

or, Fa = da 1 b 2ym Ch Cp V p Sa ca Ch dsa 2 2 b 2 Ca (9-33)

If for a given aircraft the aerodynamic derivatives Ch , Ch , Cp and Ca may be assumed to remain constant, at a given ight altitude (i.e. value of ) the roll control force is, Fa = constant p V (9-34)

In a diagram showing rate of roll as a function of V , curves for a constant control force will be hyperbolas, see gure 9-16. Due to wing twist, Ca will decrease in absolute value with increasing airspeed, as discussed in section 7-7. For a real, elastic wing, the rate of roll achieved at a given control force will be smaller than for a rigid wing, just like the rate of roll resulting from a given aileron deection, see gure 9-16.

9-7
9-7-1

Control Forces and Hinge Moments for Lateral Control

Roll Control

Using the same argument as in chapter 5 for the elevator control, the following expression applies to roll control in analogy with equation (5-77), Fa dsa + Har dar + Hal dal = 0 (9-35)

9-7 Control Forces and Hinge Moments for Lateral Control or, Fa = dar dal Har + Hal dsa dsa

471

(9-36)

The hinge moments Har and Hal can be written in the usual way as, Har = Char 1 2 V Sar car 2 1 2 V Sal cal 2

Hal = Chal

If Sar = Sal = Sa (the area of one single aileron) and, car = cal = ca the aileron control force in equation (9-36) becomes, Fa = dar dal Char + Chal dsa dsa 1 2 V Sa ca 2 (9-37)

If no dierential deection of the ailerons is used, this expression can be futher developed in a straightforward manner. Using equation (7-32), ar = al = or, dar da 1 da = l = dsa dsa 2 dsa Substitution in equation (9-37), Fa = where, Char = Ch0r + Ch r + Ch ar + Cht tar and, Chal = Ch0l + Ch l + Ch al + Cht tal In asymmetric ight is r = l . Suppose now,
Flight Dynamics

a 2

1 da 1 2 V Sa ca 2 dsa 2

Char Chal

(9-38)

472

Lateral Stability and Control in Steady Flight

r = + a then, l = a The result is, r l = 2 a In addition, is, ar al = a tar tal = ta Ch0r Ch0l = 0 and thus, Char Chal = Ch 2 a + Ch a + Cht ta Substitution of equation (9-39) in equation (9-38) nally results in, Fa = da 1 2 V Sa ca dsa 2 Ch a + Ch a t + Cht a 2 2 (9-40) (9-39)

The expression between parentheses is Ch of a single aileron.

9-7-2

Rudder Control

The applicable expression for the control force on the rudder pedals is entirely in analogy with equation (5-77) for the elevator control, Fr = dr 1 2 V Sr cr dsr 2 v Ch v + Ch r + Cht tr (9-41)

Calculation methods for the hinge moment derivatives in equation (9-40) and (9-41) have been mentioned in section 5-2-3.

Part III

Dynamic Stability Analysis

Flight Dynamics

Chapter 10 Analysis of Symmetric Equations of Motion

Once the equations of motion of an aircraft have been derived and the required aerodynamic derivatives are known, it is possible to calculate the aircraft motions caused by an arbitrary control deection or external disturbance. In the most general case, if the equations of motion to be used are non-linear, only a numerical integration of the equations is possible. Many aircraft motions, however, may be considered as deviations from a condition of steady, straight, symmetric ight, small enough to permit linearization of the equations about that reference condition. In chapter 4 the linearization has been given. Thereafter it is possible to use analytical methods, such as the Laplace-transform, to solve the linearized equations. Even then, the computer remains a nearly indispensable tool to carry out the calculations. If the motions of an aircraft about a given equilibrium condition are studied, it is, quite independent of the type of disturbance to be considered, of primary interest to know if the equilibrium is stable. The following discusses how this question is answered. Stability is a characteristic of any equilibrium condition of any arbitrary system. The question of the stability of an equilibrium condition relates to the behaviour of the system after it has been subjected to a small deviation from the equilibrium due to some disturbance, once the disturbance has ceased to act. If, after some time, the system returns to the original equilibrium condition, that condition is stable. If the system does not return to the original condition but deviates increasingly with time, the equilibrium was unstable. Finally, the equilibrium condition is neutrally stable, if the deviation caused by the disturbance neither disappears nor keeps increasing with time. If the motions of the system under study about the initial reference condition can be described by linear equations of motion, the stability of the equilibrium condition is independent of the
Flight Dynamics

476

Analysis of Symmetric Equations of Motion

type and magnitude of the disturbance and of the magnitude of the initial deviation from the equilibrium condition. The condition of linearity is, also in the case of aircraft, not always satised. In the following, stability is a characteristic of an arbitrary steady ight condition. For the sake of simplicity the stability of an aircraft is often referred to, whereas the stability of a certain equilibrium condition of that aircraft is actually meant. This is not quite correct, since any particular aircraft may be in a stable and sometimes also in an unstable equilibrium condition. The behaviour of the aircraft once it has acquired a small deviation from an equilibrium condition is described by the linearized deviation equations, (4-37) and (4-40). For simplicitys sake, only the longitudinal stability is discussed in the following sections. The stability criteria to be derived and the methods used to solve the equations of motion are equally applicable to the asymmetric motions, see chapter 11. To begin with, the linearized deviation equations for the symmetric motions, where CXq = 0, are repeated here, see equation (4-38), CXu 2c Dc CZu 0 Cmu CX CZ0 0 CZq + 2c 1
2 Cmq 2c KY Dc

CZ + (CZ 2c ) Dc CX0 0 Cm + Cm Dc CXe Dc 0

=
q c V

CZ e = 0

Cme

This is a group of four simultaneous, constant coecient, linear dierential equations of rst order. The independent variable in the equations is, in the physical sense, the disturbance variable, i.e. the elevator deection. The dependent variables are the four components of the motion. According to the foregoing, the stability of the equilibrium condition is apparent from the aircrafts motion, once it has acquired a deviation from the equilibrium situation due to a disturbance and the disturbance has ceased to act. In the equations of motion this means that the time history of the dependent variable is studied, assuming the independent variables to be all zero and assuming furthermore, that the dependent variables have been given non-zero initial conditions. As the equations are linear, the magnitude of the initial

10-1 Solution of the equations of motion

477

conditions, i.e. the magnitude of the assumed disturbances and the resulting deviations, have no inuence on the stability of the equilibrium, as noted before. In the more general case, where the motions are described by non-linear equations, the magnitude of the disturbances may inuence stability. Since stability is then independent of the input, or the disturbance, the equations to be solved for the disturbed symmetric aircraft motions can now be written in the following homogeneous form,
CXu 2c Dc CZu 0 Cmu CX CZ + (CZ 2c ) Dc 0 Cm + Cm Dc CZ0 CX0 Dc 0 0 CZq + 2c 1
2 Cmq 2c KY Dc

=0
q c V

(10-1) As the elevator angle has been assumed to remain constant, the stability to be determined from these equations is the so-called stability, stick xed. The assumption is thus made that the pilot holds the control manipulator and thereby the elevator xed, such that the elevator angle does indeed remain constant. A highly realistic alternative is the situation where the pilot has trimmed the control force to zero and takes his hands o the control manipulator. In that case the control mechanism is free during the disturbed motion and the elevator may vary. The actual time history of the elevator angle during the disturbed motion then follows from one or more additional equations of motion describing the behaviour of the control mechanism in the stick free condition. The stability thus to be determined, is the so-called stability, stick free. The following general discussions of the stability of an equilibrium condition are valid both for the stability, stick xed and the stability, stick free.

10-1

Solution of the equations of motion

The analytical solution of the linear dierential equations considered here is based on the transformation of these equations into algbraic equations. This can be done in various ways. The solution of the above homogeneous equations has the following general form, see for instance reference [104]. For the symmetric motions, x = Ax ec sc (10-2)
Flight Dynamics

478

Analysis of Symmetric Equations of Motion

where x represents any of the components of the motion, and the variables sc is the time, made non-dimensional, see table 4-1, where, sc = V t c

The coecient Ax in equation (10-2) is determined partly by the initial conditions given to the equations of motion. The variable c in (10-2) to be discussed in detail later, can be either real or complex. From the solution (10-2) it can be seen at once if the equilibrium condition under study is stable. For stability it is both necessary and sucient that x goes to zero with increasing time. The requirement for stability then is, lim x = 0

sc

For nite values of the coecient Ax it is the variable c that determines entirely and only, if the stability requirement is met. It is, therefore, of primary interest to see how this c is to be found. To this end it is only necessary to substitute the solution (10-2) of the equations of motion back in these equations. The value or values of c turning these equations into equalities are the ones looked for. Before performing the substitution, it may be realized that in equation (10-1), Dc x = c d Ax ec V dt
V c

= c x

(10-3)
2 Dc x = 2 x c

In the following, the equations for the symmetric motions are further analyzed. Equations (10-2) and (10-3) are substituted in equation (10-1),
CXu 2c c CZu 0 Cmu CX CZ + (CZ 2c ) c 0 Cm + Cm c CZ0 CX0 c 0 Cmq 0 CZq + 2c Au

A A 1 2 Aq 2c KY c

s e c c =0

(10-4) or, in a more compact notation,

10-1 Solution of the equations of motion

479

[] A ec sc = 0 where [] is the matrix and A is the vector with elements Au , A , A and Aq . By the substitution the equations of motion change from dierential equations into simultaneous, linear, homogeneous algebraic equations. From these equations the non-zero common factor ec sc can be omitted without inuencing the values of c to be obtained, [] A = 0 These homogeneous equations are always satised by the so-called trivial solution, A=0 which is a short notation for, Au = A = A = Aq = 0 In equation (10-2) this means, u=== q c =0 V (10-5)

This is the original equilibrium ight condition, which is evidently a particular case of the disturbed motion. This trivial solution is not further considered here. The homogeneous, algebraic equations represented by equation (10-5) are in general inconsistent. A simple example may serve to illustrate this fact. Suppose, the following two homogeneous, algebraic equations have been given, a11 x1 +a12 x2 = 0 a21 x1 +a22 x2 = 0 Or, a11 a12 a21 a22 x1 x2

=0

From the rst equation follows the solution, x2 = a11 x1 a12

Flight Dynamics

480 and from the second equation, x2 = These two solutions are consistent if, or,

Analysis of Symmetric Equations of Motion

a21 x1 a22

a11 a21 = a12 a22

a11 a22 a12 a21 = 0 Or, expressed in a dierent way, a11 a12 =0 a21 a22 If the coecient determinant of the homogeneous, algebraic equations is zero, the equations are no longer inconsistent. But then they have become dependent. In the case of aircraft motions considered here, the equations (10-5) and (10-4) have a solution only for a few particular values of c . These are the values of c resolving the inconsistency of the equations by making the equations dependent. These values of c called the eigenvalues of the dierential equations, are found according to the foregoing by equating the determinant of the square matrix [] to zero. The determinant of [] is called the characteristic determinant. The eigenvalues are then found from, [] = 0 This characteristic equation can be written more explicitly as, CXu 2c c CZu 0 Cmu CX CZ0 0 CZq + 2c =0 1
2 Cmq 2c KY c

CZ + (CZ 2c ) c CX0 0 Cm + Cm c c 0

For a further study of stability via the eigenvalues it is necessary to expand the above characteristic equation. The result may be expressed as, A 4 + B 3 + C 2 + D c + E = 0 c c c (10-6)

10-1 Solution of the equations of motion

481

The four roots of this characteristic equation are the four eigenvalues c to be found. According to equation (10-2) they determine if the original equilibrium condition is stable. The coecients A to E in equation (10-6) are given with the more detailed discussion of dynamic longitudinal stability in section 10-3. The resulting eigenvalues may nally be substituted in (10-5). For each c follow an arbitrary combination of three out of the four equations the three ratios, real or complex, of the four components of the motion. They provide further insight into the characteristics of the disturbed motion, but this subject is not further pursued here. In general the roots of the characteristic equation are all dierent. After substitution in equation (10-2) they give the aircraft motion after a disturbance from the equilibrium situation. Assuming four dierent eigenvalues, the solution for the symmetric motions is, u ec1 sc (10-7)

= [A]
q c V

s e c2 c ec3 sc ec4 sc

where the matrix [A] can be written more explicitely as, Au1 Au2 A2 A2 Aq2 Au3 A3 A3 Aq3 Au4 A4 A4 Aq4 (10-8)

A1 [A] = A 1 Aq1

Combination of equation (10-7) and (10-8) results in the solution of the homogeneous dierential equations, u
q c V

= Au1 ec1 = A1 ec1 = A1 ec1 = Aq1 ec1

sc sc sc sc

+Au2 ec2 +A2 ec2 +A2 ec2 +Aq2 ec2

sc sc sc sc

+Au3 ec3 +A3 ec3 +A3 ec3 +Aq3 ec3

sc sc sc sc

+Au4 ec4 +A4 ec4 +A4 ec4 +Aq4 ec4

sc sc sc sc

Four out of the sixteen constants in [A] are xed by four initial conditions which must be given if the disturbed motion is to be determined completely. The remaining twelve constants then follow from twelve real or complex ratios of the components of the motion, mentioned
Flight Dynamics

482

Analysis of Symmetric Equations of Motion

earlier. From this brief discussion it may be seen, that the constants in [A] have no inuence whatsoever on the stability or instability of the equilibrium conditions. The various possible types of eigenvalues c and the corresponding motions, the so-called eigenmotions are now studied, under the assumption that all eigenvalues are dierent. Real eigenvalues The part of the total aircraft response, i.e. the eigenmotion, corresponding to a real eigenvalue c is described by an aperiodic exponential function, see gure 10-1. The equilibrium can only be stable, if all real eigenvalues are negative. Only then is, lim Ax ec sc = 0

sc

If c = 0, then, Ax ec sc = Ax The eigenmotion corresponding to this particular value is a constant for any of the components of the motion. For c > 0, all components of the motion tend to go to innity with increasing time, because for any value of Ax , lim Ax ec sc =

sc

The speed at which for a given negative, real c the corresponding eigenmotion converges to zero, is commonly expressed by two dierent characteristic times, The time to damp to half the amplitude, T 1 . This is the time interval in which the 2 exponential function decreases to half its original value. This means, x(t + T 1 ) =
2

1 x(t) 2

or,
c V c t+T 1

e From this follows,

1 c V e c 2

T1 =
2

ln 1 c 0.693 c 2 = c V c V

10-1 Solution of the equations of motion

483

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

(a) Convergent motion, < 0

2

6 t [sec]

10

1.8

1.6

1.4

1.2

0.8

0.6

0.4

0.2

(b) Constant deviation, = 0

150

6 t [sec]

10

100

x
50 0 0 1

(c) Divergent motion, > 0

6 t [sec]

10

Figure 10-1: Aperiodic motions corresponding to a real eigenvalue

Flight Dynamics

484

Analysis of Symmetric Equations of Motion The time constant . This is the time interval in which the exponent decreases by 1 and the exponential function itself decreases by the factor 1 . Then, e x(t + ) = or,
ec c (t+ ) = V

1 x(t) e

1 c V e c e

or c V = 1 c 1 c c V

= The relations between T 1 and are,

2

= 1.443 T 1

T 1 = 0.693
2

Using the time constant, the eigenmotion corresponding to the real c can be written as,
Ax ec c t = Ax e V t

If c is positive and an aperiodic divergence results, T 1 and may still be used to indicate the 2 level of divergence. In that case both T 1 and are negative. In those situations use is often 2 made of the time to double amplitude T2 . It is the time interval in which the exponential function increases to twice its original value, T2 = T 1
2

10-1 Solution of the equations of motion Complex Eigenvalues In the case of a complex c , the eigenvalue is written as, c = c + j c j= 1

485

Substituting of such an eigenvalue in equation (10-2) or (10-3) would lead to complex expressions for the components of the motion. But complex eigenvalues occur always in complex conjugate pairs, i.e., c1,2 = c j c The corresponding constants Ax1 and Ax2 in equation (10-2) are also complex conjugate, as the resultant motion can only be real. This means, Ax1,2 = Rx j Qx The eigenmotion belonging to the two complex conjugate roots of the characteristic equation, can now be written as, Ax1 ec1 sc + Ax2 ec2 sc = (Rx + j Qx ) e(c +j c )sc + (Rx j Qx ) e(c j c )sc =
2 Rx + Q2 ec sc cos c sc + arctan x

Qx Rx

(10-9)

This part of the total response of the components of the motion is evidently a damped oscillation, i.e. c is negative. The period P of the oscillation follows from the imaginary part c of the two eigenvalues. If the argument of the harmonic function has increased by 2, a time P has elapsed, c or, P = 2 c c V V P = 2 c

According to equation (10-9), the amplitude of the oscillation is,

2 2 Rx + Q2 ec sc x Flight Dynamics

486

Analysis of Symmetric Equations of Motion

0.5

x
0 0.5 0

(a) Convergent motion, < 0

1.5

6 t [sec]

10

0.5

0.5

1.5

(b) Constant amplitude, = 0

150

6 t [sec]

10

100

50

50

100

150

(c) Divergent motion, > 0

6 t [sec]

10

Figure 10-2: Periodic motion corresponding to a pair of complex, conjugate eigenvalues 1,2 = j

10-1 Solution of the equations of motion

487

Apparently, only the real part c of the eigenvalues determines if the amplitude converges to zero with increasing time, see gure 10-2. For stability it is necessary, that the real parts of all complex eigenvalues are negative. Only then is,
2 lim 2 Rx + Q2 ec sc = 0 x

sc

Various measures of the damping of an oscillation are in use. The time to damp to half amplitude, T 1 . This characteristic has the same meaning as 2 for the aperiodic motions. T 1 is now expressed by,
2

T1 =
2

0.693 c c V

If an oscillation has negative damping, i.e. it diverges, it means that c > 0 and T 1 is 2 negative. Usually the time to double amplitude, T2 of the oscillation is then given, T2 = T 1
2

The number of periods, C 1 , in which the amplitude decreases to half its original value,
2

C1 =
2

T1
2

c 0.693 c = 0.110 2 c c

In analogy with T2 is, C2 = C 1

2

The logarithmic decrement . This is the natural logarithm of the ratio of the oscillations amplitude in two successive maxima,
ec c (t+P ) V

= ln or,

c V c

= c

V P c

= 2

c 0.693 = c C1
2

The damping ratio . The complex eigenvalues c1,2 are often written as follows, c1,2 = 0 j0 1 2 c V (10-10)

Here is the damping ratio and 0 is the undamped natural frequency (in rad/sec). If = 0, the eigenvalues are,
Flight Dynamics

488

Analysis of Symmetric Equations of Motion

c1,2 = j0

c = c V

The real part of c1,2 is then zero, the oscillation is accordingly undamped and the angular frequency of the oscillation is 0 . The relation between the period and the eigenfrequency n of the damped oscillation is, P = With P =
2 c

2 n

c V

it follows from equation (10-10) that, n = 0 1 2

The relations between 0 , , c and c can be derived to be, 0 =

2 2 c + c

V c

2 2 c + c

if 0 < < 1 a damped oscillation occurs, because then c1 and c2 are complex and the real c part 0 V is negative. With increasing the damping increases as well. The relations between C 1 , and are,
2

= 2 and,

1 2

C 1 = 0.110
2

1 2

If = 1, a transition occurs from a periodic to an aperiodic motion. The motion is then called critically damped. The characteristic equation in that particular case has two equal roots. If > 1, the matIoa is aperiodic, as both c1 and c2 are then real. Finally, if < 0, the real part of the eigenvalues is positive, see equation (10-10). The considered equilibrium condition is then unstable. The motions diverge, in a periodic manner if 1 < < 0 and aperiodic if < 1. At the end of this review of the stability of an equilibrium condition for the case where all eigenvalues are dierent, the ndings can be summarized as follows.

10-2 Stability criteria

489

The equilibrium is stable if all real eigenvalues and the real parts of the complex eigenvalues are negative. The disturbed motion then converges back to the original equilibrium condition. If at least one real or complex eigenvalue has a positive real part, then the equilibrium condition is unstable. In this case the disturbed motion diverges increasingly from the steady equilibrium ight condition. A transitional situation exists, where no eigenvalue has a positive real opart, but a real eigenvalue or the real part of a pair of complex eigenvalues is just equal to zero. In such a situation, a deviation of constant magnitude or amplitude occurs. The equilibrium condition is then called neutrally stable. Often, the eigenvalues, i. e. the roots of the characteristic equation, are depicted as points in the complex plane. Real eigenvalues are situated on the real axis at a distance of c of the origin and complex eigenvalues have as coordinates (c j c ). The above stability criterion implies that all eigenvalues must be situated to the left of the imaginary axis. As soon as at least one eigenvalue is placed to the right of the imaginary axes, the equilibrium condition is unstable. If one or more eigenvalues are placed on the imaginary axis, whether or not in the origin of the system axis, and none lies in the right half of the plane, the equilibrium condition is neutrally stable. The solution of the equations of motion for the case where two or more roots of the characteristic equation are equal, is not further discussed here. This topic is further discussed for instance in reference [104]. It appears then that what was said previously about distinct eigenvalues is generally true also for multiple eigenvalues. Dierences exist only, if multiple eigenvalues lie on the imaginary axis and none to the right of that axis. The equilibrium is then either neutrally stable or unstable, depending on the multiplicity of the relevant eigenvalues and on the rank of the characteristic determinant. In order to determine the stability of the equilibrium condition according to the foregoing rules, it is always necessary to construct the characteristic determinant and to determine from it the values and types of eigenvalues. In contrast with the foregoing, the following section discusses a method to determine the stability without an explicit calculation of the eigenvalues.

10-2

Stability criteria

Often a need exists to determine the stability of an equilibrium condition without resorting at once to the solution of the characteristic equation. Sometimes it is necessary to study the inuence on the stability of systematic changes in the system under consideration. In such cases, fruitful use may be made of various so called stability criteria. In this chapter only the Routh-Hurwitz criteria will be discussed. A more general stability criterion is that of Lyapunov of which the Routh-Hurwitz criteria are a particular case, see reference [104].

Flight Dynamics

490 The Routh-Hurwitz Stability Criteria

Analysis of Symmetric Equations of Motion

In the foregoing it was shown, that the criterion for stability is that all real eigenvalues and all real parts of the complex eigenvalues are negative. Routh, see reference [137] and Hurwitz, see reference [77], derived criteria which the coecients of an algebraic equation have to satisfy for all real roots and the real parts of the complex roots to be negative. It will be sucient here to present only the stability criteria as they apply to a quartic characteristic equation. The quartic is, A 4 + B 3 + C 2 + D c + E = 0 c c c (10-11)

The coecients A, B, C, D and E are all real. Without loss in generality of the following discussions it is always assumed that, A>0 The stability criteria then are, A > 0, B > 0, C > 0, D > 0, E > 0 and, BCD AD2 B 2 E > 0 The latter expression, R = BCD AD2 B 2 E is often called Rouths discriminant. If A < 0, the signs of B, C, D, E and R must also be negative to obtain negative real parts of the eigenvalues.

10-3

The complete solution of the equations of motion

As indicated already in section 10-1, an idea of the aircraft motions after a disturbance can be obtained by determining the roots of the characteristic equation (10-6). The general character of the symmetric motions is discussed in the following. To this end, a numerical example is discussed, along the lines presented in section 10-1.

10-3 The complete solution of the equations of motion According to equation (10-6) the characteristic equation for the symmetric motions is, A 4 + B 3 + C 2 + D c + E = 0 c c c

491

The coecients A to E are obtained by expanding the characteristic determinant, see section 10-1. The coecients A to E are,
2 A = 42 KY (CZ 2c ) c

B = Cm 2c CZq + 2c Cmq 2c (CZ 2c )

2 2c KY {CXu (CZ 2c ) 2c CZ }

C = Cm 2c CZq + 2c Cm 2c CX0 + CXu CZq + 2c

2 Cmq {CXu (CZ 2c ) 2c CZ } + 2c KY (CX CZu CZ CXu )

(10-12)

D = Cmu CX CZq + 2c CZ0 (CZ 2c ) Cm 2c CX0 + CXu CZq + 2c +

Cm (CX0 CXu CZ0 CZu ) + Cmq (CXu CZ CZu CX )

E = Cmu (CX0 CX + CZ0 CZ ) + Cm (CX0 CXu + CZ0 CZu ) If the various aerodynamic and inertial parameters are known, the quantitative values of the coecients A to E can be calculated. The eigenvalues c are next obtained as the roots of the quartic characteristic equation (10-6). To nd the roots of the characteristic equation, use is made of the program package MATLAB. In many practical situations a computer is available to obtain the eigenvalues c . Very often the starting point for the calculations is then the characteristic determinant rather than the characteristic equation (10-6). The required program, e.g. MATLAB, is commonly
Flight Dynamics

492

Analysis of Symmetric Equations of Motion

V S
2 KY

= = =

59.9 m/sec 24.2 m2 0.980

m lh xcg

= = =

4547.8 kg 5.5 m 0.30 c

c c

= =

2.022 m 102.7

CX0 CXu CX CX CXq CX

= = = = = =

0 0.2199 0.4653 0 0 0

CZ0 CZu CZ CZ CZq CZ

= = = = = =

1.1360 2.2720 5.1600 1.4300 3.8600 0.6238 Cmu Cm Cm Cmq Cm = = = = = 0 0.4300 3.7000 7.0400 1.5530

Table 10-1: Symmetric stability and control derivatives for the Cessna Ce500 Citation

available as a standard routine. It is then merely necessary to provide the numerical values of the stability derivatives and the inertial parameters. As an example the symmetric disturbed motions of a Cessna Ce500 Citation are considered. The required data are given in table 10-1. From these data follow the coecients A to E, according to equation (10-12) and from these parameters the characteristic equation. From the coecients A to E the stability can directly be assessed. The eigenvalues are actually determined by MATLAB using the roots.m routine for root nding of polynomials. The resulting two pairs of complex conjugate eigenvalues are, c1,2 = 2.9107 104 j 6.6006 103 c3,4 = 3.9161 102 j 3.7971 102 or with = c
V c,

1,2 = 8.6226 103 j 1.9544 101 3,4 = 1.1601 100 j 1.1240 100

10-3 The complete solution of the equations of motion

1.5

493

3
1

0.5

1 j
0

2
0.5

4
1.5 1.5

0.5

0.5

Figure 10-3: The location of the eigenvalues for the symmetric motions

Apparently, the disturbed motion is the sum of two periodic motions. The one corresponding to the eigenvalues c1,2 is a relatively slow, lightly damped oscillation, the other mode, corresponding to eigenvalues c3,4 has a much shorter period and is highly damped. Two such oscillations occur in nearly all situations with most categories of conventional aircraft. They are indicated as the long period oscillation, or phugoid, and the short period oscillation respectively. The various characteristics of these motions derived in section 10-1 are next calculated from the eigenvalues c . The results are, Phugoid mode the period P = 32.1391 seconds c1,2 = 2.9107 104 j 6.6006 103
2

the time to damp to half the amplitude T 1 81 seconds the undamped natural frequency 0 = 0.1957 rad/sec. the damping ratio = 0.0441 Short period mode the period P = 5.5900 seconds c3,4 = 3.9161 102 j 3.7971 102

Flight Dynamics

494

Analysis of Symmetric Equations of Motion the time to damp to half the amplitude T 1 0.6 seconds
2

the undamped natural frequency 0 = 1.6153 rad/sec. the damping ratio = 0.7182

Figure 10-3 shows the position of the eigenvalues in the complex plane. Here the dimensional eigenvalues are used, = c V c V c V c

= c

= c

Figures 10-4 and 10-5 show the time responses of the aircraft after some small disturbance. This calculation was made using the program package MATLAB. The actual disturbance used was a negative step elevator deection (e = 0.005 Rad.). The parameters used in the simulation, as well as the ight condition, are given in table 10-1. In gure 10-4 the simulation of the phugoid mode is clearly seen, the period is approximately 32 seconds. The short period oscillation is best seen in gure 10-5, see the simulation of the pitch-rate q at the larger time scale. The high damping of this motion is also apparent.

10-4
.

Approximate solutions

Figures 10-4 and 10-5 show some characteristic dierences between the phugoid mode and the short period oscillation. The airspeed varies hardly during the short period oscillation. Due to the phugoid oscillation, airspeed and pitch angle vary in particular. The angular acceleration about the lateral axis, q is almost zero during the long period oscillation. Also during this oscillation the variations of the angle of attack are relatively small if compared with the variations in angle of pitch. Since = + , the changes in ight path angle will be nearly equal to those of . Based on these characteristic dierences, some approximating calculation methods for the short period and the long period oscillations can be given. The approximate solutions are useful to quickly and easily get some idea of the various characteristics of the two eigenmotions. Of course, the results are very handy for a quick check on the order of magnitude of results of complete computer calculations. The accuracy of the approximating calculations depends of course on the inacceptability which depends on the type of aircraft. The following gives some calculation methods that generally produce satisfactory results for conventional aircraft. Even among the approximating

10-4 Approximate solutions

495

60

59.5

59

58.5

V [m/sec]

58

57.5

57

56.5

56

55.5

55

50

t [sec]

100

150

0.05

0.04

0.03

0.02

[Rad]

0.01

0.01

0.02

0.03

0.04

0.05

50

t [sec]

100

150

Figure 10-4: Response curves for a step elevator deection (e = 0.005 [Rad]) for the Cessna Ce500 Citation, phugoid response

Flight Dynamics

496

Analysis of Symmetric Equations of Motion

0.1

0.08

0.06

0.04

[Rad]

0.02

0.02

0.04

0.06

0.08

0.1

50

t [sec]

100

150

0.02

0.015

0.01

q [Rad/sec]

0.005

0.005

0.01

0.015

0.02

50

t [sec]

100

150

Figure 10-4: (Continued) Response curves for a step elevator deection (e = 0.005 [Rad]) for the Cessna Ce500 Citation, phugoid response

10-4 Approximate solutions

497

60

59.5

59

58.5

V [m/sec]

58

57.5

57

56.5

56

55.5

55

t [sec]

10

0.05

0.04

0.03

0.02

[Rad]

0.01

0.01

0.02

0.03

0.04

0.05

t [sec]

10

Figure 10-5: Response curves for a step elevator deection (e = 0.005 [Rad]) for the Cessna Ce500 Citation, magnication of gure 10-4, short period response

Flight Dynamics

498

Analysis of Symmetric Equations of Motion

0.1

0.08

0.06

0.04

[Rad]

0.02

0.02

0.04

0.06

0.08

0.1

t [sec]

10

0.02

0.015

0.01

q [Rad/sec]

0.005

0.005

0.01

0.015

0.02

t [sec]

10

Figure 10-5: (Continued) Response curves for a step elevator deection (e = 0.005 [Rad]) for the Cessna Ce500 Citation, magnication of gure 10-4, short period response

10-4 Approximate solutions

499

calculations, some are more rened than others. Very often an apparently more rened calculation does not produce more accurate results, if compared with the complete calculation. A more detailed motivation for the various approximations can be found using the so-called eigenvectors of the eigenmotions. They are not discussed here. 1. Short period oscillation (a) V = constant The approximating calculation of the short period oscillation is based on the assumption that during this motion the airspeed remains constant. This implies that in the equations of motion u = 0 and that the rst column in equation (10-1) disappears. This also implies that the forces in the XB -direction must remain in equilibrium during the entire motion. As a consequence, the XB -equation can be entirely omitted. As an additional simplication, the initial steady ight condition is assumed to be level, so 0 = 0 and CX0 = 0. This latter assumption causes the angle of pitch to disappear from the Z- and M equations, allowing the kinematic c relation Dc + q = 0 to be omitted as well. In this way equations (10-1) are V thus reduced to the following set of simpler equations, CZ + (CZ 2c ) Dc Cm + Cm Dc CZq + 2c Cmq
2 2c KY Dc

q c V

The characteristic equation is obtained, just as in the discussion given in section 10-1, by equating the characteristic determinant to zero, CZ + (CZ 2c ) c Cm + Cm c Expanding the determinant results in, A2 + Bc + C = 0 c where,
2 A = 2c KY (2c CZ ) 2 B = 2c KY CZ 2c + CZq Cm (2c CZ ) Cmq

=0

CZq + 2c =0
2 Cmq 2c KY c

C = CZ Cmq 2c + CZq Cm The roots of the characteristic equation are the two eigenvalues c1,2 , c1,2 = c j c = B j 4AC B 2 2A
Flight Dynamics

500

Analysis of Symmetric Equations of Motion A further simplication is possible by omitting the derivatives CZ and CZq . They occur in the equations of motion in combination with the mass parameter 2c compared to which they are negligible. The coecients A, B and C are then reduced to,
2 A = 42 KY c

2 B = 2c KY CZ + Cm + Cmq

C = CZ Cmq 2c Cm (b) V = constant, rotations in pitch only ( is constant) A further simplifying assumption is, that during the short period oscillation the trajectory of the aircraft c.g. is a straight line. If the initial steady ight condition is level ight the c.g. moves along a horizontal straight line. This means that the forces in the ZB -direction remain in balance during the entire motion. As a consequence the Z-equation can be omitted. The only remaining motion is the rotation in pitch, i.e. only the M -equation remains. Because = = 0, can be replaced by and Dc by Dc in the M -equation. c 2 With Dc q = Dc the M -equation is then written as, V
2 2 Cm + Cm Dc + Cmq Dc 2c KY Dc = 0

The characteristic equation has the coecients,

2 A = 2c KY

B = Cm + Cmq C = Cm 2. Phugoid oscillation (a) q = 0, = 0 In this rst, coarsest approximation of the long period oscillation the assumption is made that the angle of attack remains constant during this motion. As a consequence = 0 and also = 0. This makes the -column to disappear in equation (10-1) and one of the equations has become superuous. Because q = 0, the M equation is omitted. The fact that = 0, may be seen as a consequence of a relatively high negative value of Cm . In such a case is also very nearly Cm = 0. If, nally, CZq is neglected relative to 2c and again CX0 = 0, the equations (10-1) now become,

10-4 Approximate solutions

501

CXu 2c Dc CZu 0

CZ0 0 Dc

The characteristic equation is reduced to a quadratic, the coecients are, A = 42 c B = 2c CXu C = CZu CZ0 The result permits a very simple approximation of the period. The undamped natural frequency 0 and the damping ratio follow from the eigenvalues c1,2 = c j c expressed in the coecients A, B and C. 0 = V c C V = A c CZu CZ0 42 c (sec1 )

2c = 0 q c 1 V

B CXu = = 2 CZu CZ0 2 AC For CXu , CZ0 and CZu the following approximations are introduced as discussed in sections 6-1, 6-2, CXu = 2 CD CZ0 = CL CZu = 2 CL For 0 and then follows, V 0 = c = 2 2CL g 2 = 42 V c 2CD
2 2 2CL

(sec1 )

1 CD 2 2 CL

According to this highly simplied calculation the damping of the long period oscillation is determined by the drag to lift ratio CD . This ratio usually is in the CL order of 0.1. The damping of the long period oscillation is, therefore, nearly always very low. In that case the period can be written as,
Flight Dynamics

502

Analysis of Symmetric Equations of Motion

P =

2 0 1 2

2 0

(sec)

With the above result for 0 , the period of the long period oscillation becomes, 2 P = V = 0.453 V g 2 with V [m/sec]. According to this aprroximating calculation, the period in seconds of the long period oscillation is roughly 0.5 times the airspeed in [m/sec]. It should be remarked that for powered ight, the inuence of Cmu (neglected here) can be of considerable inuence on the phugoid motion. (b) q = 0, = 0 A second, slightly more rened approximation of the long period oscillation goes less far than the previous one. Variations in are now permitted, however, they are assumed to occur so slowly that remains negligible. In the equations (10-1) c the contributions due to Dc q and Dc disappear, but none of the equations can V be omitted. If CZq and CX0 are again neglected, the result is, CXu 2c Dc CX CZu 0 Cmu CZ 0 Cm CZ0 0 Dc 0 0 2c 1 Cmq u (sec)

The coecients of the characteristic equation, following from an expansion of the characteristic determinant, are, A = 2c CZ Cmq 2c Cm B = 2c (CXu Cm Cmu CX ) + Cmq (CZu CX CXu CZ ) C = CZ0 (Cmu CZ CZu Cm )

=0
q c V

Chapter 11 Analysis of Asymmetric Equations of Motion

The asymmetric equations of motion will be treated along the same line as in chapter 10. Furthermore, the characteristic modes for the asymmetric aircraft motions will be discussed in this chapter, including several approximations for these modes. All simulations will be performed using a model of the Cessna Ce500 Citation.

11-1

Solution of the equations of motion

The equations for the disturbed asymmetric motions are obtained in the homogeneous form from equation (4-41), by omitting the terms due to the control surface deections,
CY + CY 2b Db 0 C Cn + Cn Db CL
1 2 Db

CYp 1
2 Cp 4b KX Db

CYr 4b

0 0

Cr

Cnp + 4b KXZ Db

2 Cnr 4b KZ Db

0 =0 pb + 4b KXZ Db 2V
rb 2V

(11-1) The character of the asymmetric motions is discussed below, using a quantitative example. Some characteristic parameters were derived for the symmetric eigenmotions in section 10-1. The same expressions also apply to the asymmetric motions, if the mean aerodynamic chord c is replaced by the wing span b. The eigenvalues resulting from the characteristic equation are now indicated as b .

Flight Dynamics

504

Analysis of Asymmetric Equations of Motion

The expressions for the time to damp to half an amplitude, and for the period, now read as follows, Aperiodic motion (b real), T1 =
2

ln 1 b 0.693 b 2 = b V b V

Periodic motion corresponding to two complex, conjungate eigenvalues b = b j b P = 2 b b V 0.693 b b V j= 1

T1 =
2

C1 =
2

T1
2

P
V

0.693 b b = 0.110 2 b b = b V b 0.693 P = 2 = b b C1

2

= ln

eb b (t+P ) e
b V b t

In chapter 9 it was stated that CY and Cn are usually neglected, so it will be assumed here that, CY = Cn = 0 Expanding the characteristic determinant of equation (11-1),
CY + CY 2b b 0 C Cn + Cn b CL 1 b 2 0 0 CYp 1
2 Cp 4b KX b

(11-2)

CYr 4b 0 Cr + 4b KXZ b
2 Cnr 4b KZ b

=0

Cnp + 4b KXZ b

(11-3) with the assumptions of equation (11-2), results in a quartic characteristic equation, A 4 + B 3 + C 2 + D b + E = 0 b b b

11-1 Solution of the equations of motion

505

The coecients A to E for the asymmetric motions are derived from the characteristic determinant, see (11-3),
2 2 2 A = 163 KX KZ KXZ b

2 2 2 2 2 B = 42 2 CY KX KZ KXZ + Cnr KX + Cp KZ + Cr + Cnp KXZ b

C = 2b

2 2 CY Cnr CYr Cn KX + CY Cp C CYp KZ +

CY Cnp Cn CYp + CY Cr C CYr

2 4b Cn KX + 4b C KXZ +

KXZ +

1 Cp Cnr Cnp Cr 2

2 D = 4b CL C KZ + Cn KXZ + 2b C Cnp Cn Cp +

1 1 CY Cr Cnp Cnr Cp + CYp C Cnr Cn Cr + 2 2 1 CY Cp Cn Cnp C 2 r

E = CL C Cnr Cn Cr If the stability derivatives and inertial parameters are known for a certain aircraft conguration and ight condition, the coecients A to E can be calculated. Using one of the common numerical methods, the eigenvalues may be obtained. As an example the asymmetric disturbed motions of a Cessna Ce500 Citation are studied. The required data are given in table 11-1. From these data follow the coecients A to E for the characteristic equation, A 4 + B 3 + C 2 + D b + E = 0 b b b which can be obtained from equation(11-3). The roots of the characteristic polynomial are obtained by the program package MATLAB using the routine roots.m.

Flight Dynamics

506

Analysis of Asymmetric Equations of Motion

V S b CL

= = = =

59.9 m/sec 24.2 m2 13.36 m 1.1360

b
2 KX 2 KZ

= = = =

15.5 0.012 0.037 0.002

KXZ

CY CYp CYr CYa CYr

= = = = =

0.9896 0.0870 0.4300 0 0.3037

C Cp Cr Ca Cr

= = = = =

0.0772 0.3444 0.2800 0.2349 0.0286

Cn Cnp Cnr Cna Cnr

= = = = =

0.1638 0.0108 0.1930 0.0286 0.1261

Table 11-1: Asymmetric stability and control derivatives for the Cessna Ce500 Citation

For a stable aircraft it follows that all eigenvalues b must have negative real parts and for the disturbed motion to be positively damped. An explicit determination of the four eigenvalues, see also gure 11-1, for the Cessna Ce500 Citation example become, b1 = 0.3291 b2 = 0.0108 b3,4 = 0.0313 j 0.3314 The disturbed asymmetric motions consist apparently of two aperiodic eigenmotions and one periodic motion. One of the two aperiodic motions appears to be well damped, the other one is only very lightly damped or may be even slightly undamped (which is the case for most aircraft. The characteristics of the eigenmotions are as follows, First aperiodic mode b1 = 0.3291 the time to damp to half the amplitude T 1 0.223 seconds
2

Second aperiodic mode

11-2 General character of the asymmetric motions

4

507

3
3

4 3.5

2.5

1.5

0.5

Figure 11-1: The location of the eigenvalues = b V for the asymmetric motions b

b2 = 0.0108

the time to damp to half the amplitude T 1 6.782 seconds

2

Periodic mode the period P = 2.004 seconds the time to damp to half the amplitude T 1 2.340 seconds
2

b3,4 = 0.0313 j 0.3314

the undamped natural frequency 0 = 3.1495 rad/sec. the damping ratio = 0.0940 Figures 11-2 and 11-3 show the calculated responses for the aircraft, simulated by the program package MATLAB. In this particular case the disturbances are pulse-shaped rudder and aileron deections.

11-2

General character of the asymmetric motions

From the time responses in gures 11-2 and 11-3 three eigenmotions may be distinguished, as was evident also from the eigenvalues.
Flight Dynamics

508

Analysis of Asymmetric Equations of Motion

0.03

0.02

0.01

[Rad]

0.01

0.02

0.03 0

t [sec]

10

15

0.01

0.01

[Rad]

0.02

0.03

0.04

0.05

0.06

0.07 0

t [sec]

10

15

Figure 11-2: Response curves for a pulse-shaped rudder deection (r = +0.025 [Rad] during 1 second) for the Cessna Ce500 Citation

11-2 General character of the asymmetric motions

509

0.1 0.08 0.06

p [Rad/sec]

0.04 0.02 0 0.02 0.04 0.06 0.08 0.1 0

t [sec]

10

15

0.08

0.06

r [Rad/sec]

0.04

0.02

0.02

0.04

0.06 0

t [sec]

10

15

Figure 11-2: (Continued) Response curves for a pulse-shaped rudder deection (r = +0.025 [Rad] during 1 second) for the Cessna Ce500 Citation

Flight Dynamics

510
0.015 0.01 0.005 0

Analysis of Asymmetric Equations of Motion

[Rad]

0.005 0.01 0.015 0.02 0.025 0.03 0.035 0

t [sec]

10

15

Figure 11-2: (Continued) Response curves for a pulse-shaped rudder deection (r = +0.025 [Rad] during 1 second) for the Cessna Ce500 Citation

A highly damped aperiodic motion The corresponding eigenvalue is b11 , real and strongly negative. The motion is apparent mostly in the rate of roll. The damping of this motion is due primarily to the damping rolling moment generated by the wing when rotating about the longitudinal axis, i.e. pb the moment Cp 2V . Aperiodic mode in which the aircraft sideslips, yaws and rolls The motion is characterized by the eigenvalue b2 real, lightly positively or negatively damped. Depending on the sign of b2 this eigenmotion converges or diverges. If the motion diverges (b2 > 0), the aircraft after a disturbance enters into a sideslipping turn with ever increasing angle of roll and decreasing radius. During this motion the aircraft sideslips towards the inside of the turn, see gure 11-4. The aircraft is than called spirally unstable, as the aircraft describes a descending spiral. The eigenmotion is called the spiral motion. If it is damped, the aircraft has spiral stability. Periodic mode in which the aircraft sideslips, yaws and rolls This eigenmotion is determined by the eigenvalues b3,4 = b j b . Figure 11-5 shows the general character of the motion. Rolling and yawing velocity are presented for the case where the motion is damped b < 0. This motion is generally called Dutch roll. The name seems to be inspired by the resemblance to the motion of a skater on the ice.

11-2 General character of the asymmetric motions

511

x 10

[Rad]

8 0

t [sec]

10

15

0.05

0.1

[Rad]

0.15

0.2

0.25

0.3 0

t [sec]

10

15

Figure 11-3: Response curves for a pulse-shaped aileron deection (a = +0.025 [Rad] during 1 second) for the Cessna Ce500 Citation

Flight Dynamics

512

Analysis of Asymmetric Equations of Motion

0.05

p [Rad/sec]

0.05

0.1

0.15

0.2

0.25

0.3 0

t [sec]

10

15

0.005

0.01

r [Rad/sec]

0.015

0.02

0.025

0.03

0.035

0.04 0

t [sec]

10

15

Figure 11-3: (Continued) Response curves for a pulse-shaped aileron deection (a = +0.025 [Rad] during 1 second) for the Cessna Ce500 Citation

11-3 Routh-Hurwitz stability criteria for the asymmetric motions

0 0.02 0.04 0.06

513

[Rad]

0.08 0.1 0.12 0.14 0.16 0.18 0

t [sec]

10

15

Figure 11-3: (Continued) Response curves for a pulse-shaped aileron deection (a = +0.025 [Rad] during 1 second) for the Cessna Ce500 Citation

11-3

Routh-Hurwitz stability criteria for the asymmetric motions

The criteria for stability were discussed already in section 11-1. It can be proved, that in the transition from stability to instability the conditions E > 0 and R > 0 are the critical stability criteria. If only E becomes negative, one real eigenvalue changes sign. This means that one of the aperiodic motions changes from convergence to divergence. If R becomes more negative, the real part of the two complex, conjugate eigenvalues changes signs implying periodic divergence. According to the foregoing, the coecient E determines the character of the aperiodic spiral motion and from the discriminant R follows the character of the (periodic) Dutch roll motion. In the following the stability criteria are used for a further study of the spiral motion and the Dutch roll oscillation of the aircraft. To prepare for this discussion, the factors determining the signs of the coecients A to E are investigated. As a rst approximation, the relatively small derivative CYp is equated to zero and CYr is neglected relative to 4b . For aircraft having a straight wing without a large dihedral these assumptions are acceptable. Also, the coecient KXZ , the product of inertia, is omitted. 2 2 KXZ is relatively small, relative to KX and KY . The remaining stability derivatives determining the magnitude and the signs of the coecients A to E have, according to chapter 9, the following signs for conventional aircraft in normal ight,
Flight Dynamics

514

Analysis of Asymmetric Equations of Motion

Figure 11-4: The spiral motion of an aircraft

11-3 Routh-Hurwitz stability criteria for the asymmetric motions

515

Figure 11-5: Characteristics of the Dutch Roll motion

Flight Dynamics

516
0.1 0.08 0.06

Analysis of Asymmetric Equations of Motion

rollrate yawrate

p, r [Rad/sec]

0.04 0.02 0 0.02 0.04 0.06 0.08 0.1 0

t [sec]

10

15

Figure 11-6: Roll- and yaw-rate characteristics of the Dutch Roll motion

CY < 0

Cp < 0

Cnp < 0

Cr > 0

Cnr < 0

The stability derivatives C and Cn are practically the only ones the designer can manipulate to inuence the lateral stability. As discussed in chapter 9, pleasant control characteristics require that, Cn > 0 C < 0

Neglecting CYp , CYr and KXZ and using the known signs of the other derivatives and the inertial parameters the signs of the coecients A to E can be established as follows,
2 2 A = 163 KX KZ > 0 b

2 2 2 2 B = 42 2 CY KX KZ + Cnr KX + Cp KZ > 0 b 2 2 2 C = 2b CY Cnr KX + CY Cp KZ + 4b Cn KX +

1 Cp Cnr Cnp Cr 2

If Cn > 0 (the aircraft is then statically directionally stable), then C > 0. 1 2 D = 4b CL C KZ + 2b C Cnp Cn Cp + CY Cr Cnp Cnr Cp 2

11-4 Spiral and Dutch roll mode, the lateral stability diagram If CL > 0, C < 0 and Cn > 0, then also D > 0. E = CL C Cnr Cn Cr At positive values of CL , the positive sign of E is determined by the condition, C Cnr Cn Cr > 0

517

It appears that the above requirement for dynamic stability is identical to the one derived in chapter 9, see section 9-3. The latter was based on the criterion for good control characteristics in steady turns using the ailerons or the rudder only.

11-4

Spiral and Dutch roll mode, the lateral stability diagram

In the previous section it was stated that positive damping of the spiral motion required that E > 0 and for damping of the Dutch roll motion, R > 0. The magnitudes of E and R are inuenced to a large extent by the eective dihedral C and the static directional stability Cn . The derivative C can be varied by changing the geometric dihedral of the wing; Cn is modied by varying the surface Sv of the vertical tailplane and the taillength lv . But it should be remembered that changing Sv and/or lv inuences also some other stability derivatives, such as Cnr and CY . The eect of changes in Cn and C on E and R is graphically depicted in the so-called lateral stability diagram, along the axes, C and Cn are plotted, see gure 11-7. For a particular aircraft, the boundaries E = 0 and R = 0 are plotted for a given aircraft conguration and ight condition in the lateral stability diagram. The remaining stability derivatives are assumed to remain constant, or to be adjusted to the external shape of the aircraft corresponding to each combination of C and Cn . The two curves E = 0 and R = 0 indicate the boundaries of areas in the diagram for spiral stability and a damped Dutch roll oscillation. The diagram indicates for any combination of C and Cn the state of the dynamic lateral stability for the aircraft conguration and ight condition under study. Only the most common case, where C < 0 and Cn > 0 is further discussed. Using the lateral stability diagram, the spiral motion is considered rst. The criterion for spiral stability is, C Cnr Cn Cr > 0 It follows, that spiral stability can be increased in two ways, e.g. by decreasing Cn or by increasing C . Because other requirements on the control characteristic have to be met as well, a decrease in Cn may be hardly acceptable. A value of C too large negative may not
Flight Dynamics

518

Analysis of Asymmetric Equations of Motion

Cn

E=0 E<0

Spiral instability E > 0, R > 0

Spiral stability and convergent Dutch Roll R=0

R<0 Divergent Dutch Roll

C
Figure 11-7: Lateral stability diagram

11-5 Approximate solutions be acceptable either, as will be discussed below.

519

The eect of these conicting requirements is, that it may be dicult to choose the external form of the aircraft such that it is spirally stable in all required aircraft congurations and ight conditions. The divergence from a spirally unstable equilibrium ight condition is often relatively slow. In order to avoid even less desirable characteristics, a certain slight degree of spiral instability may be unavoidable. The curve for R = 0, determining stability of the m Dutch roll motion, appears to depend largely on the relative density b = Sb of the aircraft. The curve for R = 0 moves upward in the lateral stability diagram with increasing b . This means that for given aerodynamic characteristics (stability derivatives) the damping of the Dutch roll motion decreases with increasing b . In general, the Dutch roll motion of aircraft having a low wing loading, ying at relatively low altitudes are relatively well damped. A characteristic value of the damping for such aircraft is 0.15 or C 1 0.7.
2

For the Dutch roll motion it does not suce to determine only the sign of R, because good ying qualities imply a certain minimal damping of the Dutch roll motion. This minimal required damping has been made a function of the period of the motion, see reference [13]. This may be explained as follows. If the period is short, the reactions of the pilot will be too slow to stabilize the motion by suitable control deections. In such a case a relatively high damping or a low value of T 1 , or C 1 , is required. If the period is longer, the 2 2 damping of the motion may be lower, because in such a situation the pilot has less diculty to improve the damping of the aircrafts motions via his control deections. However, the requirement relating the period and the minimal acceptable time to damp to half amplitude has turned out to provide insucient guarantee for good ying qualities. More recent requirements, see e.g. reference [13], stipulate that the damping of the Dutch roll has a certain minimal value depending on the roll to yaw ratio |p| occurring during the |r| Dutch roll motion. The interpretation of this ratio
|p| |r|

is shown in gure 11-6.

11-5

Approximate solutions

1. Heavily damped aperiodic rolling motion The aperiodic rolling moment may be approximated by assuming that the aircraft can only roll about the longitudinal axis. This is permissable because this motion often disappears before the other eigenmotions of the aircraft have really started. If only the rolling motion is considered, in the equations (11-1) the angle of sideslip , and the rb non-dimensional yaw-rate 2V disappear and the equations for the lateral force and the yawing moment can be omitted. In the remaining equation for the rolling moments the pb angle of roll does not occur, so the kinematic relation 1 Db + 2V = 0 is no longer 2 needed. The rolling moment equation then reads,
2 Cp 4 b KX Db

pb =0 2V
Flight Dynamics

520

Analysis of Asymmetric Equations of Motion It is easy to see, that this expression gives the real eigenvalue, b1 = Cp 2 4 b KX

2. Dutch roll motion (a) =

pb 2V

=0

For a large category of conventionial aircraft a certain approximation of the Dutch roll motion results if the rolling component is discarded in the Dutch roll motion. pb pb If and 2V are set at zero, the - and 2V -columns in the equations of motion disappear and the rolling moment equation can be omitted since the rolling moments have to remain in balance. Only the aerodynamic force Y - and the moment N -equations remain. As usual, CY and Cn are neglected, whereas CYr is insignif icant relative to 4 b , CY 2b Db 4b Cn Cnr
2 4b KZ Db

rb 2V

The coecients of the quadratic characteristic equation become,

2 A = 8 2 KZ b

=0

2 B = 2b Cnr + 2KZ CY

C = 4b Cn + CY Cnr From these coecients 0 , , P and T 1 can easily be obtained.

2

(b) =

pb 2V

= 0, yawing rotation only

A more restrictive assumption is that the trajectory of the aircrafts c.g. is a straight line during the oscillation. This implies that the course angle is constant, where = + = 0, or = . It implies as well that the Y -equation is superuous. The remaining yawing moment equation is further simplied by the rb fact that 2V = 1 Db and = . The result is, 2 1 2 2 Cn + Cnr Db 2b KZ Db = 0 2 The coecients A, B and C of the characteristic equation are,
2 A = 2b KZ

11-5 Approximate solutions

521

1 B = Cnr 2 C = Cn The expressions for 0 and are obtained very simply as, V b C V = A b Cn 2 2b KZ

0 =

B Cnr = = 2 2 AC 4 2b KZ Cn 3. Aperiodic spiral motion This is usually a very slow eigenmotion, in which the aircraft sideslips, yaws and rolls. It is, therefore, permissible to approximate the motion by assuming all linear and angular accelerations to be negligible. In the equations of motion this means, Db = Db pb rb = Db =0 2V 2V

In addition to CYr , also CYp is neglected. CY CL 1 Db 2 0 0 0 1 Cp Cnp 4b 0 Cr Cnr

0 C

Cn

pb = 0 2V
rb 2V

Expanding the characteristic equation results in the eigenvalue, 2CL C Cnr Cn Cr

b4 =

Cp CY Cnr + 4b Cn Cnp CY Cr + 4b C

It is easy to verify that the dominator is negative if all stability derivatives have their normal signs. For convergence of this motion it is necessary that b4 < 0, which means, CL C Cnr Cn Cr > 0 This corresponds to the requirement E > 0, derived in section 11-3.
Flight Dynamics

522

Analysis of Asymmetric Equations of Motion

4. Dutch roll motion and the aperiodic rolling motion If the assumption is made that the c.g. moves along a straight line, but in addition the rolling motion is permitted, only the Y -equation disappears. The following equations pb in and 2V result,
2 2 C + 1 Cr Db + 2b KXZ Db Cp 4b KX Db 2

Cn +

1 2 Cnr Db

2 2 2b KZ Db

Cnp + 4b KXZ Db

pb 2V

The characteristic equation is a cubic, A 3 + B 2 + C b + D = 0 b b The coecients are then,

2 2 2 A = 42 KX KZ KXZ b

=0

B = b

2 2 Cr + Cnp KXZ + Cnr KX + Cp KZ

2 C = 2b C KXZ + Cn KX +

1 Cp Cnr Cnp Cr 4

D=

1 C Cnp Cn Cp 2

Evidently no explicit analytical solution to this cubic can be given. By using the approximation for the eigenvalue of the motion in roll, the characteristic equation may be reduced to a quadratic equation.

Bibliography

[1] J.H. Abbot and A.E. von Doenho. Theory of wing sections. Dover Publications, New York, 1960, (see also NACA Rep. 824). [2] H.J. Allan. Estimation of the forces and moments acting on inclined bodies of revolution at low mach numbers. Technical report, NACA R.M. A9 126, 1949. [3] H.J. Allen and E.W. Perkins. A study of eects of viscosity on ow over slender inclined bodies of revolution. Technical report, NACA Report 1048, 1951. [4] M.B. Ames and R.E. Sears. Determination of control surface characteristics from naca plain-ap and tab data. Technical report, NACA Rep. 721, 1941. [5] R.F. Anderson. Determination of the characteristics of tapered wings. Technical report, NACA Rep. 572, 1936. [6] Anonymous. British civil airworthiness requirements section d, aeroplanes. Technical report, Air Registration Board, London. [7] Anonymous. Civil air regulations (c.a.r.) 4b, airplane airworthiness-transport categories. Technical report, Civil Aeronautics Board, Washington D.C. [8] Anonymous. Data sheets, aerodynamics. Technical report, R. Ae. Soc. London. [9] Anonymous. Data sheets, aerodynamics, vols. i, ii and iii. Technical report, Royal Aeronautical Society. [10] Anonymous. Data sheets aerodynamics, vols, i, ii and iii. Technical report, Royal Aeron. Society, London. [11] Anonymous. Federal aviation regulations, subchapter c, aircraft, part 25. Technical report, F.A.A. Washington. [12] Anonymous. Icao, airworthiness of aircraft, annex 8, part iii. Technical report, Chapters 2 and 3.
Flight Dynamics

524

Bibliography

[13] Anonymous. Specication for ying qualities of piloted aeroplanes. Technical report, Bureau of Aeronautics, SR-119B, 1948. [14] Anonymous. Verslag der vliegproeven met het vliegtuig ft 404, type north-american harvard ii b, deel d: Statische stuurstanddwarsstabiliteit. Technical report, N.L.L. Rapport V. 1422 d, 1950 (in Dutch). [15] Anonymous. Samenvatting der resultaten van het experimentele onderzoek der vliegeigenschappen van het vliegtuig ft 404, type north american harvard ii b. Technical report, N.L.L. Rapport V 1599, 1951 (in Dutch). [16] Anonymous. Verslag van vliegproeven ter bepaling van langsstabiliteit en besturingseigenschappen in glijvlucht met beide schroeven in vaanstand van het vliegtuig ph-nll type siebel 204-d-1. Technical report, N.L.L. Rapport V 1698, 1953 (in Dutch). [17] Anonymous. F-27 staartvlakmodel 1:9. eigenschappen van het horizontale staartvlak. Technical report, N.L.L. Rapport A 1389, 1955 (in Dutch). [18] Anonymous. Verzameling van grootheden betreende stabiliteit en besturing van moderne vliegtuigen. Technical report, N.L.L. T.M. V. 1864, 1960 (in Dutch). [19] Anonymous. Military specication, ying qualities of piloted airplanes. Technical report, MIL-F-8785B (A.S.G.), 1969. [20] Anonymouys. Ministry of aviation. design requirements for aircraft for the royal air force and royal navy. Technical report, Aviation publication 970, Part 6, 1948. [21] W.R. Bates. Collection and analysis of windtunnel data on the characteristics of isolated tail surfaces with and without end plates. Technical report, NACA T.N. 1291, 1947. [22] W.R. Bates. Static stability of fuselages having a relatively at cross section. Technical report, NACA T.N. 3429, 1955. [23] H. Binkhorst. Metingen in de vlucht van de stuurverplaatsing per g en de stuurkracht per g in afvangmanoeuvres en in zuivere bochten van het vliegtuig auster j-5b autocar, ph-neh. Technical report, V.T.H.-Rapport 95, 1960 (in Dutch). [24] H. Binkhorst. Windtunnelmetingen aan een model van de fokker f-27 friendship. Technical report, VTH M47, 1961 (in Dutch). [25] J. Bird. Some theoretical low-speed span loading characteristics of swept wings in roll and sideslip. Technical report, NACA Report 969, 1950. [26] P.L. Bisgood and D.J. Lyons. Preliminary report of the ight measurement of aeroelastic distortions in relation to its eect on the control and longitudinal stability of a mosquito aircraft. Technical report, R. and M. 2371, 1949. [27] P.L. Bisgood and D.J. Lyons. Preliminary report on the ight-maesurement of aeroelastic distorsion in relation to its eects on the control and longitudinal stability of a mosquito aircraft. Technical report, R. and M. 2371, 1949.

Bibliography

525

[28] P.J. Bobbitt. Tables for the rapid estimation of downwash and sidewash behind wings performing various motions at supersonic speeds. Technical report, NASA Memo 2-2059 K, 1959. [29] P.J. Bobbitt and D.A. Malvestuto Jr. Estimation of forces and moments due to rolling for several slender-tail congurations at supersonic speeds. Technical report, NACA T.N. 2955, 1953. [30] J.L. Boier. The Dynamics Of Flight, The Equations. John Wiley & Sons, 1998. [31] K.W. Booth. Eect of horizontal-tail chord on the calculated subsonic span loads and stability derivatives of isolated unswept tail assemblies in sideslip and steady roll. Technical report, NASA Memo 4-1-59 L, 1959. [32] C. Brady. Boeing 737. http://www.b737.org.uk/, 1999. [33] J.P. Campbell and A. Goodman. A semi-empirical method for estimating the rolling moment due to yawing of airplanes. Technical report, NACA T.N. 1984, 1949. [34] J.P. Campbell and M.O. McKinney. ummary of methods for calculating dynamic lateral stability and response and for estimating lateral stability derivatives. Technical report, NACA Report 1098 and T.N. 2409, 1952, 1951. [35] NASA Dryden Flight Research Center. X-29 at high angle of attack with smoke generators. http://www.dfrc.nasa.gov/gallery/photo/index.html, 1991. [36] D.C. Cheatman. A study of the characteristics of human pilot control response to simulated lateral motions. Technical report, NACA Rep. 1197, 1954. [37] K.L. Coin. Equations and charts for the rapid estimation of hinge moment and eectiveness parameters for trailing edge controls having leading and trailing edges swept ahead of the mach lines. Technical report, NACA Rep. 1041, 1959. [38] W.L. Cowley and H. Glauert. The eect of the lag of the downwash on the longitudinal stability of an aeroplane on the rotary derivative. Technical report, R. and M. 718, 1921. [39] A. Damon, H.W. Stoudt, and R.A. McFarland. The human body in equipment design. Havard University Press., Cambridge (Mass.), 1967. [40] P. de Lange, J.H. Pelleboer, and J. Zaaijer. De coordinaten van het zwaartepunt en de grootte van het gewicht en massatraagheidsmoment van het vliegtuig dhc-2 beaver, ph-vth, ten behoeve van in oktober/november 1968 uitgevoerde metingen in niet-stationare symmetrische vlucht. Technical report, V.T.H. Memorandum M-165, 1970 (in Dutch). [41] P. de Lange, J.H. Pelleboer, and J. Zaaijer. De coordinaten van het zwaartepunt en de grootte van het gewicht en massatraagheidsmoment van het vliegtuig dhc-2 beaver, ph-vth, ten behoeve van in oktober/november 1968 uitgevoerde metingen in niet-stationare symmetrische vlucht. Technical report, V.T.H. Memorandum M-165, 1970 (in Dutch).
Flight Dynamics

526

Bibliography

[42] J. de Young. Spanwise loading for wings and control surfaces of low aspect ratio. Technical report, NACA T.N. 2011, 1950. [43] J. de Young. Spanwise loading for wings and control surfaces of low aspect ratio. Technical report, NACA T.N. 2011, 1950. [44] J. de Young. Theoretical antisymmetric span loading for wings of arbitrary plan at subsonic speeds. Technical report, NACA Report 1056, 1951. [45] J. de Young. Theoretical symmetric span loading due to ap deection for wings of arbitrary plan form at subsonic speeds. Technical report, NACA Rep. 1071, 1952. [46] J. de Young. Theoretical symmetric span loading due to ap deection for wings of arbitrary plan form at subsonic speeds. Technical report, NACA Report 1071, 1952. [47] J. de Young. Theoretical symmetric span loading due to ap deection for wings of arbitrary plan forms at subsonic speeds. Technical report, NACA Rep. 1071, 1952. [48] J. de Young and W.H. Barling. Prediction of downwash behind swept-wing airplanes at subsonic speed. Technical report, NACA T.N. 3346, 1955. [49] J. de Young and C.W. Harper. Theoretical symmetric span loading at subsonic speeds for wings having arbitrary plan form. Technical report, NACA Rep. 921, 1948. [50] G.L. Decker. rediction of downwash at various angles of attack for arbitrary tail locations. Aero. eng. Rev., 15(7):22 to 27, 61, 1956. [51] F.E. Douwes Dekker and W.J. Molier. Vliegeigenschappen van het fokker f-27 vliegtuig. Technical report, N.L.L. Rapport V. 1902, 1962 (in Dutch). [52] F.D. Diederich. Charts and tables for use in calculations of downwash of wings of arbitrary plan form. Technical report, NACA T.N. 2353, 1951. [53] W.J. Duncan. The principles of the control and stability of aircraft. University Press, Cambridge, 1952. [54] V.M. Falkner and D. Lehrians. Calculated loadings due to incidence of a number of straight and swept-back wings. Technical report, R. and M. 2596, 1948. [55] L.R. Fischer and H.S. Fletcher. Eect of lag of sidewash on the vertical-tail contribution to oscillatory damping in yaw of airplane models. Technical report, NACA T.N. 3356, 1955. [56] L.R. Fisher. Approximate corrections for the eects of compressibility on the subsonic stability derivatives of swept wings. Technical report, NACA T.N. 1854, 1949. [57] R.W. Franks. The application of a simplied lifting-surface theory to the prediction of the rolling eectiveness of plain spoiler ailerons at subsonic speeds. Technical report, NACA R.M. A54 H 26a, 1954. [58] G.C. Furlong and J.G. McHugh. A summary and analysis of the low-speed longitudinal characteristics of swept wings at high reynolds numbers. Technical report, NACA Rep. 1339, 1957.

Bibliography

527

[59] K. Gersten. Berechnung der aerodynamischen beiwerte von trag u geln endlicher spannweite in bodenn he. Abhandlungen der Braunschw. Wissensch. Gesell sch., a XII:95 to 115, 1960 (in German). [60] R.R. Gilruth and W.N. Turner. Lateral control required for satisfactory ying qualities based on ight tests of numerous airplanes. Technical report, NACA Report 715, 1941. [61] H.J. Goett and J.P. Reeder. Eects of elevator nose shape, gap balance and tabs on the aerodynamic characteristics of a horizontal tail surface. Technical report, NACA Rep. 675, 1939. [62] A. Goodman and G.H. Adair. Estimation of the damping in roll of wings through the normal ight range of lift coecient. Technical report, NACA T.N. 1924, 1949. [63] A. Goodman and J.D. Brewer. Investigation at low-speeds of the eect of aspect ratio and sweep on static and yawing stability derivatives of untapered wings. Technical report, NACA T.N. 1669, 1948. [64] A. Goodman and L.E. Fisher. Investigations at low speeds of the eect of aspect ratio and sweep on rolling stability derivatives of untapered wings. Technical report, NACA Report 968, 1950. [65] M.N. Gough and A.P. Beard. Limitations of the pilot in applying forces to airplane controls. Technical report, NACA T.N. 550, 1936. [66] X. Hafer. Untersuchungen zur aerodynamik der u gel-rumpf-anordnungen. Technical report, W.G.L. Jb., 1957, pages 181 to 208 (See also thesis of the same name, Fakult a t f u r Machinenwesen, Braunschweig) (in German). [67] S.H. Harman and L. Jereys. Theoretical lift and damping-in-roll of thin wings with arbitrary sweep and taper at supersonic speeds. supersonic leading and trailing edges. Technical report, NACA T.N. 2114, 1950. [68] K.G. Hecks. The high speed shape. pitch-up and palliatives adopted on swept-wing aircraft. Flight International, 85:13 to 18, 1964. [69] H. Hertel. Ermittlung der gr ssten aufbringbaren steuer kr fte und erreichbaren o a geschwindigkeiten der steuerbet tigung. Zeitschrift f u r Flugtechnik und Motora luftschiahrt, 21 Jrg., page 36 to 45, 1930 (in German). [70] D.E. Hoak and J.W. Carlson. USAF Stability and Control Handbook. Douglas Aircraft Company Inc., 1961. [71] D.E. Hoak and J.W. Carlson. USAF, Stability and Control Handbook. Douglas Aircraft Co. Inc, 2 edition, 1968. [72] D.E. Hoak and J.W. Carlson. USAF Stability and Control Handbook, 2nd ed. Douglas Aircraft Co., Inc., 1968. [73] H.P. Hoggard and J.R. Hagerman. Downwash and wake behind untapered wings of various aspect ratios and angles of sweep. Technical report, NACA T.N. 1703, 1948.
Flight Dynamics

528

Bibliography

[74] V. Holmboe. Charts for the position of the aerodynamic centre at low speeds and small angles of attack for a large family of tapered wings. Technical report, Saab T.N. 27, 1954. [75] E.J. Hopkins. Lift, pitching moment and span load characteristics of wings at low speed as aected by variations of sweep and aspect ratio. Technical report, NACA T.N. 2284, 1951. [76] E.J. Hopkins. A semi-empirical method for calculating the pitching moment of bodies of revolution at low mach numbers. Technical report, NACA R.M. C14, 1951. [77] A. Hurwitz. Ueber die bedingungen, unter welchen eine gleichung nur wurzeln mit negativen re llen teilen besitzt. Mathematische Werke, Band II, Zahlentheorie und e Geometrie:533 to 545, 1933. [78] W. Jacobs. The inuence of the induced sidewind on the eciency of the vertical tail. a simplied method for calculation. Technical report, F.A.A. Report 35, 1950. [79] W. Jacobs. Lift and moment changes due to the fuselage for a yawed aeroplane with unswept and swept wings. Technical report, F.A.A. Report 34, 1950. [80] W. Jacobs. The induced sidewind behind swept wings at subsonic velocities. Technical report, F.A.A. Report 41, 1951. [81] W. Jacobs. Theoretical and experimental investigations of interference eects of delta wing-vertical tail combinations with yaw. Technical report, F.A.A. Report 49, 1953. [82] W. Jacobs and E. Truckenbrodt. Der induzierte seitenwind von ugzeugen. Ing. Archiv, 21:1 to 22, 1953 (in German). [83] A.L. Jones and A. Alksne. A summary of lateral stability derivatives calculated for wing plan forms in supersonic ow. Technical report, NACA Report 1052, 1951. [84] F.S. Malvestuto Jr. and K. Margolis. Theoretical stability derivatives of thin sweptback wings tapered to a point with swept back or sweptforward trailing edges for a limited range of supersonic speeds. Technical report, NACA Report 971, 1950. [85] J.B. Dods Jr. and B.E. Tinling. Summary of results of a windtunnel investigation of nine related horizontal tails. Technical report, NACA T.N. 3497, 1955. [86] R.P. Harper Jr. Flight evaluations of various longitudinal handling qualities in a variable stability jet ghter. Technical report, Cornell Aero. Lab. WADC T.R. 55-299, 1955. [87] T.J. King Jr. and T.B. Pasteur Jr. Wind-tunnel investigation of the aerodynamic characteristics in pitch of wing-fuselage combinations at high subsonic speeds. taperratio series. Technical report, NACA T.N. 3867, 1956. [88] W.E. Cotter Jr. Summary and analysis of data on damping in yaw and pitch for a number of airplane models. Technical report, NACA T.N. 1980, 1946. [89] W.H. Michael Jr. Analysis of the eects of wing interference on the tail contributions to the rolling derivatives. Technical report, NACA Report 1086, 1952.

Bibliography

529

[90] W.H. Michael Jr. Investigation of mutual interference eects of several vertical tailfuselage congurations in sideslip. Technical report, NACA T.N. 3135, 1954. [91] S. Katzo and W. Mutterperl. The end-plate eect of a horizontal-tail surface on a vertical-tail surface. Technical report, NACA T.N. 797, 1941. [92] S. Katzo and H.H. Sweburg. Ground eect on downwash angles and wake location. Technical report, NACA Rep. 738, 1943. [93] M. Kayton and W.R. Fried. Avionics Navigation Systems. John Wiley & Sons, Inc., 2 edition, 1997. [94] V. Klein. Estimation of aircraft aerodynamic parameters from ight data. Aerospace Sci., 26:1 to 77, 1989. [95] C.C. Kranenburg, P. de Lange, and J.A. Mulder. Determination of the rigid body inertial characteristics of the de havilland dhc-2 beaver experimental aircraft from high accuracy measurements of free oscillations. Technical report, V.T.H. Report, LR337, Delft University of Technology, Department of Aerospace, Delft, 1982. [96] C.C. Kranenburg, P. de Lange, and J.A. Mulder. Determination of the rigid body inertial characteristics of the de havilland dhc-2 beaver experimental aircraft from high accuracy measurements of free oscillations. Technical report, V.T.H. Report, LR337, Delft University of Technology, Department of Aerospace Engineering, Delft, 1982. [97] R.E. Kuhn and J.W. Wiggings. Wind-tunnel investigations of the aerodynamic characteristics in pitch of wing-fuseage combinations at high subsonic speeds. aspect-ratio series. Technical report, NACA R.M. L52 A29, 1952. [98] R.A. Lagerstrom and M.E. Graham. Linearized theory of supersonic control surfaces. Aero. Sci., 16:31 to 34, 1949. [99] R.A. Lagerstrom and M.E. Graham. Linearized theory of supersonic control surfaces. Aero. Sci., 16:31 to 34, 1949. [100] J.G. Lowry. Data on spoiler-type ailerons. Technical report, NACA R.M. L53 124a, 1953. [101] D. Lyons and P.L. Bisgood. An analysis of the lift slope of aerofoils of small aspect ratio, including ns, with design charts for aerofoils and control surfaces. Technical report, R. and M. 2308, 1945. [102] D.J. Lyons and P.L. Bisgood. An analysis of the lift slope of aerofoils of small aspect ratio including ns with design charts for aerofoils and control surfaces. Technical report, R. and M. 2038,, 1945. [103] D.L. Lyons and P.L. Bisgood. An analysis of the lift slope of aerofoils of small aspect ratio including ns, with design charts for aerofoils and control surfaces. Technical report, R. and M. 2308, 1945. [104] J.G. Malkin. Theorie der stabilit t einer bewegung. Technical report, R. Oldenburg, a M u nchen, 1959 (in German).
Flight Dynamics

530

Bibliography

[105] J. Man e. Windtunnel investigation of the inuence of the aircraft conguration on the e yawing and rolling moment of a twin-engined, propeller-driven aircraft with one engine inoperative. Technical report, N.L.L. Rapport A 1508 B (see also N.L.L. Rapport A 1508 A), 1963. [106] K. Margolis and P.J. Bobbitt. Theoretical calculations of the pressures, forces and moments at supersonic speeds due to various lateral motions acting on thin isolated vertical-tails. Technical report, NACA Report 1268, 1956. [107] K. Margolis and M.H. Elliot. Theoretical calculations of the pressures, forces and moments due to various lateral motions acting on tapered sweptback vertical tails with supersonic leading and trailing edges. Technical report, NASA T.N. D-383, 1960. [108] K. Margolis, W.L. Sherman, and M.E. Hannah. Theoretical calculation of the pressure distribution, span loading, and rolling moment due to sideslip at supersonic speeds for thin sweptback tapered wings with supersonic trailing edges and wing tips parallel to the axis of wing symmetry. Technical report, NACA T.N. 2898, 1953. [109] J.C. Martin and F.S. Mavestuto Jr. Theoretical forces and moments due to sideslip of a number of vertical tail congurations at supersonic speeds. Technical report, NACA T.N. 2412, 1951. [110] A.P. Martina. Method for calculating the rolling and yawing moments due to rolling for unswept wings with or without aps or ailerons by use of non-linear section lift data. Technical report, NACA T.N. 2937, 1953. [111] W.H. McAvoy. Maximum forces applied by pilots to wheeltype controls. Technical report, NACA T.N. 628, 1937. [112] S.W. McCuskey. An Introduction To Advanced Dynamics. Addison-Wesley Publ. Co. Inc., 1958. [113] Silent Thunder Models. http://www.silentthundermodels.com/, October 2005. [114] H. Multhopp. Zur aerodynamik des ugzeugrumpfes. Luftfahrtforschung, 18:52 to 66 (in German), 1941. [115] M. Munk. The aerodynamic forces on airship hulls. Technical report, NACA Rep. 184, NACA Rep. 184,. [116] H.E. Murray. Wind-tunnel investigation of end-plate eects of horizontal tails on a vertical tail compared with available theory. Technical report, NACA T.N. 1050, 1946. [117] R.H. Neely and R.F. Griner. Summary and analysis of horizontal-tail contribution to longitudinal stability of swept wing airplanes at low speeds. Technical report, NACA Rep. R-49, 1959. [118] W.H. Nelson and W.J. Krumm. The transonic characteristics of 38 cambered rectangular wings of varying aspect-ratio and thickness as determined by the transonic-bump technique. Technical report, NACA T.N. 3502, 1955.

Bibliography

531

[119] W.H. Nelson and J.B. McDevitt. The transonic characteristics of 22 rectangular symmetrical wing models of varying aspect-ratio and thickness. Technical report, NACA T.N. 3501, 1955. [120] J.L. Overbeeke. Measurements of the control force-time relations for sudden movements of the controls. Technical report, N.L.L.-T.N. V 1816, 1958. [121] H.A. Pearson and R.F. Anderson. Calculation of the aerodynamic characteristics of tapered wings with partial-span aps. Technical report, NACA Rep. 665, 1939. [122] H.A. Pearson and R.T. Jones. Theoretical stability and control characteristics of wings with various amounts of taper and twist. Technical report, NACA Report 635, 1938. [123] C.D. Perkins and R.E. Hage. Airplane performance, stability and control. John Wiley and Sons Inc., New York, Chapman and Hall Ltd. London, 1949. [124] W.J.G. Pinsker. A semi-empirical method for estimating the rotary rolling moment derivatives of swept and slender wings. Technical report, RAE T.N. Aero. 2641, 1959. [125] E.C. Polhamus. A simple method of estimating the subsonic lift and damping in roll of swept-back wings. Technical report, NACA T.N. 1862, 1949. [126] E.C. Polhamus. Some factors aecting the variation of pitching moment with sideslip of aircraft congurations. Technical report, NACA T.N. 4016, 1958. [127] E.C. Polhamus and W.C. Sleeman Jr. The rolling moment due to sideslip of swept wings at subsonic and transonic speeds. Technical report, NACA T.N. D-209, 1960. [128] E.C. Polhamus and K.P. Spreemann. Subsonic wind-tunnel investigation of the eects of fuselage afterbody on directional stability of wing-fuselage combinations at high angles of attack. Technical report, NACA T.N. 3896, 1956. [129] P.E. Purser. An approximation to the eect of geometric dihedral on the rolling moment due to sideslip for wings at transonic and superonic speeds. Technical report, NACA R.M. L52 B01, 1952. [130] M.J. Queijo. Theoretical span load distributions and rolling moments for sideslipping wings of arbitrary plan form in incompressible ow. Technical report, NACA Report 1296, 1956. [131] M.J. Queijo and D.R. Riley. Calculated subsonic span loads and resulting derivatives of unswept and 45o sweptback tail surfaces in sideslip and in steady roll. Technical report, NACA T.N. 3245, 1945. [132] E.T. Raymond and C.C. Chenoweth. Aircraft Flight Control Actuation System Design. Society of Autonotive Engineers, Inc., 400 Commonwealth Drive, Warrendale, PA 15096-0001, USA, 1993. [133] H.S. Ribner. Formulas for propellers in yaw and charts of the side-force derivative. Technical report, NACA Report 819, 1945. [134] H.S. Ribner. Propellers in yaw. Technical report, NACA Report 820, 1945.
Flight Dynamics

532

Bibliography

[135] D.R. Riley. Eect of horizontal-tail span and vertical location on the aerodynamic characteristics of an unswept tail assembly in sideslip. Technical report, NACA Report 1171 and T.N. 2907, 1954, 1953. [136] Rosebut. Wright yer 1. http://www.earlyaviator.com/, December 1903. [137] E.J. Routh. Dynamics of a system of Rigid Bodies. MacMillan and Co., London, 1905. [138] M.D. White R.R. Gilruth. Analysis and prediction of longitudinal stability of aircraft. Technical report, NACA Rep. 711, 1941. [139] A.H. Sacks. Aerodynamic forces, moments, and stability derivatives for slender bodies of general cross section. Technical report, NACA T.N. 3283, 1954. [140] J.C. Scheer and A. Marx. Metingen ter bepaling van krachten die, afhankelijk van de tijdsduur, door een vlieger op de stuurorganen kunnen worden uitgeoefend. Technical report, N.L.L. Rapport V 1255, 1942 (in Dutch). [141] H. Schlichting. Calculation of the inuence of a body on the position of the aerodynamic centre of aircraft with swept-back wings. Technical report, R. and M. 2582, 1952. [142] H. Schlichting and W. Frenz. Uber den einuss von uegel und rumpf auf das seitenleitwerk. Jahrbuch der Deutschen Luftfahrtforshung, page 300 to 314, 1941 (in German). [143] H. Schlichting and E. Truckenbrodt. Aerodynamik des ugzeuges, bd 1 und 2. Technical report, Springer-Verlag, Berlin, 1959, (in German). [144] W.L. Sherman and K. Margolis. Theoretical calculations of the eects of nite sideslip at supersonic speeds on the span loading and rolling moment for families of thin sweptback tapered wings at an angle of attack. Technical report, NACA T.N. 3046, 1952. [145] A. Silverstein and S. Katzo. Design charts for predicting downwash angles and wake characteristics behind plain and apped wings. Technical report, NACA Rep. 648, 1939. [146] A. Silverstein, S. Katzo, and W.K. Bullivant. Downwash and wake behind plain and apped airfoils. Technical report, NACA Rep. 651, 1939. [147] C.L. Spigt and A. de Gelder. Windtunnelmetingen aan een model van de f-27 friendship uitgevoerd in de lage-snelheids tunnel van de onderafdeling der vliegtuigbouwkunde. Technical report, V.T.H. Memorandum M-42, 1959 (in Dutch). [148] C.L. Spigt and A. de Gelder. Windtunnelmetingen aan een model van de romp van de fokker f-27 friendship. Technical report, VTH M41, 1959 (in Dutch). [149] C. Spoon. Windtunnel measurements on a model of the wing of the fokker f-27 friendship. Technical report, V.T.H. M 46, 1958. [150] C. Spoon. Windtunnelmetingen aan een model van de vleugel van de fokker f-27 friendship. Technical report, VTH M46, 1960 (in Dutch). [151] J.R. Spreiter and A.H. Sacks. The rolling-up of the trailing vortex sheet and its eect on the downwash behind wings. Aero. Sci., 18(1):21 to 32, 72, 1951.

Bibliography

533

[152] A. Stanbrook. The lift-curve and aerodynamic centre position of wings at supersonic and subsonic speeds. Technical report, R.A.E. T.N. Aero 2328, 1954. [153] J.L. Synge and B.A. Grith. Principles Of Mechanics. McGraw-Hill Book Co., 1959. [154] H.H.B.M. Thomas. State of the art on estimation of derivatives. Technical report, AGARD Rep. 339 and M.o.A. Current Paper 664, 1960. [155] T.A. Toll. Summary of lateral control research. Technical report, NACA Rep. 868, 1947. [156] T.A. Toll and M.J. Queijo. Approximate relations and charts for low-speed stability derivatives of swept wings. Technical report, NACA T.N. 1581, 1948. [157] E. Torenbeek. Synthesis of subsonic airplane design. Delft University Press,, Delft, 1976. [158] H. Trienes and E. Truckenbrodt. Systematische abwindmessungen an pfeil u geln. Ing. Archiv, XX:26 to 36, 1652 (in German). [159] E. Truckenbrodt. Experimentelle und theoretische untersuchungen an symmetrisch angestr mten pfeil- und delta u geln. Technical report, Z.f. Flw. Bd. 2, 1953, o pages 185 to 201 (in German). [160] E. Truckenbrodt. Beitrage zur erweiterden traglinientheorie. Technical report, Z.f.Flw. Bd. 1, 1953, pages 31 to 37 (in German). [161] H.J. van der Maas. Stuurstandslijnen van vliegtuigen; de bepaling ervan door middel van vliegproeven en hare betekenis voor de beoordeling der stabiliteit. Technical report, Verslagen en verhandelingen van de Rijks studiedienst voor Luchtvaart, Deel V, Amsterdam, 1929 (in Dutch). [162] T. van Oosterom. Measurements on the relation between magnitude and duration and on the rate of application of the control forces achieved by pilots in simulated manoeuvres. Technical report, AGARD Report 241, 1959. [163] C.H. v.d. Linden and A. Blauw. Experimental determination of the three moments of inertia and the product of inertia in the plane of symmetry of the n.l.r. laboratory aircraft siebel 204-d-1. Technical report, N.L.R. V. 1912, 1963 (in Dutch). [164] C.H. v.d. Linden and A. Blauw. Experimental determination of the three moments of inertia and the product of inertia in the plane of symmetry of the n.l.r. laboratory aircraft siebel 204-d-1. Technical report, N.L.R., V. 1912, 1963 (in Dutch). [165] J. Weissinger. The lift distribution of swept-back wings. Technical report, NACA T.M. 1120, 1947. [166] E.W. Weisstein. Rotation matrix. From MathWorldA Wolfram Web Resource. http://mathworld.wolfram.com/RotationMatrix.html, 1999. [167] N.L. Wene. Measurement of aircraft moments of inertia. Technical report, AGARD Rep. 248, 1959.
Flight Dynamics

534

Bibliography

[168] N.L. Wener. Measurement of aircraft moments of inertia. Technical report, AGARD Rep. 248, 1959. [169] J.W. Wiggings and R.E. Kuhn. Wind-tunnel investigation of the aerodynamic characteristics in pitch of wing-fuselage combinations at high subsonic speeds. sweep series. Technical report, NACA R.M. L52 JWD 18, 1952. [170] J.C. Wimpenny. Stability and control in aircraft design. J.R.A.S., 58:329 to 360, 1954. [171] R. Wurzbach. Das geschwindigkeitsfeld hinter einer auftrieb erzeugenden trag che a von endlicher spannweite. Z.f.Flw. Bd. 5,, 12:360 to 365, 1957 (in German). [172] A.H. Yates. Notes on the mean aerodynamic chord and the mean aerodynamic centre of a wing. Technical report, J. R. Ae. S. Vol. 56, 1952, pages 461 to 474.

Appendix A List of Symbols

Roman Symbols
a A A b bf B c cf ca,e,r,t cr ct cm c cw,h,v ca,e,r,t cd CFD CD c.g. Speed of sound Center of the Earth, origin of FI Wing aspect ratio Wing span Local fuselage width Moment of momentum, angular momentum Local wing chord Wing chord ap Chord of aileron, elevator, rudder or tab, behind hinge line Chord length in the plane of symmetry Chord length at wing tip Mean wing chord (Eng. notation: c) Mean aerodynamic chord (English notation: c) Mean aerodynamic chord of wing, horizontal tailplane, or vertical tailplane Mean aerodynamic chord of control surfaces and tabs, behind hinge line Drag coecient two-dimensional ow Computational Fluid Dynamics Drag coecient three-dimensional ow Center of gravity = D
1 2 2 V S

b2 S

S b

d
1 2 2 V c

Flight Dynamics

536

List of Symbols He 1 2 2 Vh Se ce = Ch a,h,v Ch = a,e,r Ch = ta,e,r = 1 2 2 V c L

1 2 2 V Sb

Che

Hinge moment coecient of the elevator. The indices a, r or t are used to indicate the hinge moment coecients of the aileron, the rudder or a tab respectively

Ch Ch Cht c C CL ca cb c CL Ce Cp Cr C Ca Cr cm Cm cma.c. Cma.c. Coecient of the aerodynamic moment about the Y axis, two-dimensional ow Coecient of the aerodynamic moment about the Y axis, three-dimensional ow Coecient of the moment about the aerodynamic center, two-dimensional ow Coecient of the moment about the aerodynamic center, three-dimensional ow Coecient of the additional lift distribution Coecient of the basic lift distribution Derivative of the lift-coecient with respect to angle of attack, two-dimensional ow Derivative of the lift-coecient with respect to angle of attack, three-dimensional ow Coecient of the rolling moment due to one inoperative engine

= =

L
1 2 2 V S

= =

dc d

dCL d

pb 2V C = rb 2V C = C = a C = r m = 1 2 2 2 V c

M
1 2 c 2 V S

= cm0

537 Cmh Cmq Cmu Cmw Cm0 Cmf ix Cmf ree Cm Cm 2 Cm cn Cn CN Cne Cnp Cnr Cn Cna Cnr CNh CNh0 CNh CNh CNh
t

Contribution of the horizontal tailplane to Cm = = contribution of the wing with fuselage and nacelles to Cm Cm at CL = e = 0 Cm at e = constant Cm at Fe = 0 Cm c V Cm = 2 Cm = e n = 1 2 2 V c = = = N
1 2 2 V Sb

Cm c q V

1
1 2 V

M S u c

Elevator eciency Normal force coecient, two-dimensional ow Yawing moment coecient Normal force coecient, three-dimensional ow Yawing moment coecient due to one inoperative engine

N
1 2 2 V S

Static directional stability

Normal force coecient of the horizontal tailplane CNh at h = e = te = 0

pb 2V Cn = rb 2V Cn = Cn = a Cn = r Nh = 1 2 2 Vh Sh

Cn

CNh h CNh = e CNh = te =

Flight Dynamics

538

List of Symbols Np 1 2 2 V Sp Nw 1 2 2 V S r
1 2 2 V c

CNp CNw cr CR ct CT CTh CTw CX CXq CXu CX0 CX CXe CY CYp CYr CY CYa CYr

Normal force coecient on the propeller Coecient of the normal force on the wing with fuselage and nacelles Coecient of the resultant aerodynamic force, twodimensional ow Coecient of the resultant aerodynamic force, threedimensional ow Coecient of the tangential force, two-dimensional ow Coecient of the tangential force, three-dimensional ow Coecient of the tangential force of the horizontal tailplane Coecient of the tangential force on the wing with fuselage and nacelles

= =

= = = = = = =

R
1 2 2 V S

t
1 2 2 V c

T
1 2 2 V S

Th 1 2 2 Vh Sh Tw 1 V 2 S 2 X
1 2 2 V S

= = CX in steady ight 1
1 2 V

CX c q V X S u

CX CX = e Y = 1 2 2 V S = = = = = = CY
pb 2V CY rb 2V CY CY a CY r

539 Z
1 2 2 V S

CZ CZq CZu CZ0 CZ CZ CZe C1

2

= = CZ in steady ight 1
1 2 V

CZ c q V Z S u

CZ CZ = Vc CZ = e T1 = 2 P = Characteristic diameter of a jet engine Drag, two-dimensional ow Diameter of a propeller Drag, three-dimensional ow dierential operator 1 = CD V 2 S 2 d = dt d2 = 2 dt d b d = = dsb V dt = d c d = dsc V dt 1 = cd V 2 c 2

d d D D D D2 Db Dc e e Fa,e,r G h He

Non-dimensional dierential operator, asymmetric motions Non-dimensional dierential operator, symmetric motions Base of the natural logarithms Distance of the center of pressure of a prole behind the leading edge Aileron, elevator or rudder control force exerted by the pilot Vehicle (aircraft) center of mass Altitude Hinge moment about the elevator hinge line, subscripts a, e, r indicate the aileron, elevator or rudder respectively Aerodynamic hinge moment Hinge moment due to friction in the control mechanism Hinge moment due to a spring in the control mechanism Hinge moment due to static unbalance of the control mechanism

1 2 = Che Vh Se ce 2

Ha,e,ra Ha,e,rf Ha,e,rs Ha,e,rw

Flight Dynamics

540

List of Symbols

i ih,p Ix Ix0 Iy Iy0 Iz Iz0 j Jxy Jxz Jyz k k kx ky kz kxz KX KY KZ KXZ lf lh lv lw LTI

Unit vector along the X-axis Angle of incidence of the horizontal tail or the propeller axis, relative to the Xm -axis Moment of inertia about the X-axis Principle moment of inertia Moment of inertia about the Y-axis Principle moment of inertia Moment of inertia about the Z-axis Principle moment of inertia Unit vector along the Y -axis Product of inertia Product of inertia Product of inertia Factor related to the variation of thrust with speed Unit vector along the Z-axis Radius of gyration about the X-axis Radius of gyration about the Y-axis Radius of gyration about the Z-axis = = = Ix m Iy m = = = xydm xzdm yzdm = x2 + y 2 dm = x2 + z 2 dm = y 2 + z 2 dm

Non-dimensional radius of gyration about the X-axis Non-dimensional radius of gyration about the Y -axis Non-dimensional radius of gyration about the Z-axis Non-dimensional product of inertia Lift, two-dimensional ow Fuselage length Tail length, horizontal tail Tail length, vertical tail Length of the fuselage ahead of the wing Linear Time Invariant

Iz m Jxz = m kx = b ky = b kz = b kxz = 2 b 1 2 = c V c 2 [= xh xw ] [= xv xw ]

541 1 = C V 2 Sb 2 1 = CL V 2 S 2

L L Lg Lt m m mac m.p. M M M,M n n.p. N N Nh Np Nw O p p P0 P q q r r, r r R R Re sa,e,r

Rolling moment about the X-axis Lift, three-dimensional ow Longitude Latitude Aerodynamic moment about dimensional ow Mass Mean aerodynamic chord Manoeuvre point the Y -axis, two-

1 = cm V 2 c2 2 [= c] 1 = Cm V 2 Sc 2 a = V 1 = cn V 2 c 2 1 = CN V 2 S 2 1 2 = Cn V Sb 2 1 2 = CNh Vh Sh 2 1 = CNp V 2 Sp 2 1 = CNw V 2 S 2

Moment about the Y -axis, three-dimensional ow Mach number Total moment Normal force, two-dimensional ow Neutral point Normal force, three-dimensional ow Yawing moment about the Z-axis Normal force on the horizontal tailplane Normal force on the propeller Normal force on the wing with fuselage and nacelles Origin reference frames FE and FO Static pressure Angular velocity about the X-axis Static pressure in undisturbed ow Period of an oscillation Dynamic pressure Angular velocity about the Y -axis Angular velocity about the Z-axis Vector indicating a position Resultant aerodynamic force, two-dimensional ow Resultant aerodynamic force, three-dimensional ow Rouths discriminant Reynolds number Pilots control deection; aileron, elevator or rudder control respectively

1 = V 2 2

1 = cr V 2 c 2 1 2 = CR V S 2

Flight Dynamics

542

List of Symbols V t b V t c

sb sc S Sa Se,r,t Sh,v Sp t t T Tc Tc Tc Th Tw T1
2

Non-dimensional parameter of time, asymmetric motions Non-dimensional parameter of time, symmetric motions Wing area Area of one aileron, behind hinge line Area of elevator, rudder or tab respectively, behind hinge line Area of horizontal or vertical tailplane Area of the propeller disk Time Tangential force, two-dimensional ow Tangential force, three-dimensional ow Thrust coecient of a propeller Thrust coecient of a jet engine Thrust coecient Tangential force on the horizontal tailplane Tangential force on the wing with fuselage and nacelles Time to damp to half amplitude Time to double amplitude

= =

2 D 4

1 = ct V 2 c 2 1 2 = CT V S 2 Tp = V 2 D2 Tp = 1 2 2 2 V d = Tp 1 2 2 V S

T2 u u u, u v V, V v, v V Va Ve Vm.c. Vtrim Vk w w, w W

= T 1 2 du = V0

Component of V along the X-axis Change in the component of V along the X-axis Component of V along the Y -axis Velocity vector Change in the component of V along the Y -axis Magnitude of the airspeed vector V Aerodynamic velocity vector, velocity relative to a particle in the undisturbed air equivalent airspeed Minimum control speed Trimmed airspeed, V at which Fe = 0 Kinematic velocity vector, velocity relative to the normal Earth-xed reference frame Component of V along the Z-axis Change in the component of V along the Z-axis Aircraft weight

543 x xa.c. xc.g. xd xh xmf ree xmf ixed xnf ree xnf ixed xp xv xw x0 X y Y z zh zp zv zw z0 Z x-coordinate, abscissa Abscissa of the a.c. of the wing Abscissa of the c.g. Abscissa of the center of pressure Abscissa of the a.c. of the horizontal tailplane Abscissa of the manoeuver point, stick free Abscissa of the manoeuver point, stick xed Abscissa of the neutral point, stick free Abscissa of the neutral point, stick xed Abscissa of the center of the propeller disk Abscissa of the a.c. of the vertical tailplane Abscissa of the a.c. of the wing with fuselage and nacelles Abscissa of the leading edge of the mac Component of the total aerodynamic force along the Xaxis y-coordinate, ordinate Component of the total aerodynamic force along the Y axis, lateral force z-coordinate z-coordinate of the a.c. of the horizontal tailplane z-coordinate of the center of the propeller disk z-coordinate of the a.c. of the vertical tailplane z-coordinate of the a.c. of the wing with fuselage and nacelles z-coordinate of the leading edge of the m.a.c. component of the total aerodynamic force along the Zaxis 1 = CX V 2 S 2 1 = CY V 2 S 2

1 = CZ V 2 S 2

Flight Dynamics

544

List of Symbols

Greek Symbols
0 0 h,v,w 0 e a ar , a e f r s t 0 b c b c angle of attack angle of attack of the Xa -axis at CL = 0 angle of attack in steady ight angle of attack of the horizontal or vertical tailplane or the wing angle of sideslip ight path angle, angle between V , relative to the earth, and the horizontal plane ight path angle in steady ight dihedral, angle between the Yr -axis and the projection of the 1 -chord line on the Yr OZr -plane 4 eective dihedral logarithmic decrement Total aileron deection deection of the right or left aileron elevator deection ap deection rudder deection spoiler deection deection of a trim tab wing twist or washout angle downwash angle, usually at the horizontal tailplane angle between the Xr -axis and the principal inertial Xaxis damping ratio of an oscillation propeller eciency imaginary part of a complex eigenvalue angle between the principal inertial X-axis and the Xs axis of the stability reference frame angle of pitch, angle between the Xr -axis and the horizontal plane angle of pitch in steady ight Taper ratio eigenvalue non-dimensional eigenvalue, asymmetric motions non-dimensional eigenvalue, symmetric motions angle of sweep, angle between the Yr -axis and the projection of the 1 -chord line on the Xr OYr -plane 4 Relative density, asymmetric motions Relative density, symmetric motions

[= ar a ]

ct cr

m Sb m = S c =

545 1 2 0 Kinematic viscosity of the air Real part of a complex eigenvalue Air density Sidewash angle, usually at the vertical tailplane Time constant Trailing edge angle of a prole angle of roll, angle of the Y -axis and the intersection of the Y OZ-plane and the horizontal plane Bank angle, angle between the Y -axis and the horizontal plane Track angle Angle between the direction of C R and the Z-axis Angle between the neutral line and the Z-axis Yaw angle Angular velocity Circular frequency Undamped natural frequency Total angular velocity about the center of gravity Angular velocity vector

Flight Dynamics

546

List of Symbols

Subscripts
a a a a a.c. b b c.g. e e E f f f f ix f ree h i i I k k n O p r r r s s t t u v w w x y z 0 Aerodynamic reference frame Aerodynamic Aerodynamic (velocity) Aileron Aerodynamic center Balance tab Body-xed reference frame Center of gravity Elevator Engine Normal Earth-xed reference frame Flap Fuselage Friction e = constant Fe = 0, (Che = 0) Horizontal tailplane Initial Interference Inertial reference frame Kinematic reference frame Kinematic (velocity) Nacelle vehicle carried normal reference frame Relating to the propulsive system Vehicle reference frame Rudder Root of the wing Spring in the control mechanism Stability reference frame tip trim tab ultimate, nal vertical tailplane static unbalance wing or wing with fuselage and nacelles along the X-axis along the Y -axis along the Z-axis initial, steady ight condition

Appendix B Aircraft Parameters

B-1

Introduction

In this appendix a list of stability and control derivatives are given for several aircraft. Also 2 2 2 the inertial parameters, i.e. b , c , KX , KY , KZ , KXZ , etcetera, for these aircraft are provided.

B-2

Linearized Equations of Motion

In this section the linear equations of motion are given in which the inertia parameters and stability and control derivatives have to be substituted for simulation. CXu 2c Dc CZu 0 Cmu CX CZ0 CXq CZq + 2c 1
2 Cmq 2c KY Dc

CZ + (CZ 2c ) Dc CX0 0 Cm + Cm Dc CXe Dc 0

=
q c V

CZ e = 0 and as in equation (4-41),

Cme

Flight Dynamics

548

Aircraft Parameters

CY + CY 2b Db 0 C Cn + Cn Db

CL 1 Db 2 0 0

CYp 1
2 Cp 4b KX Db

Cnp + 4b KXZ Db CYa CYr 0 Cr Cnr

0 = pb Cr + 4b KXZ Db 2V
2 Cnr 4b KZ Db rb 2V

CYr 4b

0 = C a

Cna

a r

B-3

LTI-System Representation

The LTI-system, or state-space, representation of the equations of motions are given below. The symmetric and asymmetric equations of motion in state-space form are, see equations (4-46) and (4-50), Symmetric equations of motion u xu x x 0 u xe

z u z z zq = V 0 0 0 c q c mu m m mq V Asymmetric equations of motion y y yp 2V b lp np yr

ze + 0
q c V

me

0 0 = pb 2V l 0 rb n 0 2V

0 0 pb + lr 2V la nr
rb 2V

yr 0 lr

na nr

a r

B-3 LTI-System Representation

549

We refer to tables 4-9 and 4-10 for the denition of the parameters used in the above-mentioned state-space representations of the equations of motion.

In the following tables inertial parameters and stability and control derivatives for several aircraft are given.

Flight Dynamics

550

Aircraft Parameters

V S 2 KY

= = =

59.9 m/sec 24.2 m2 0.980

m lh xcg

= = =

4547.8 kg 5.5 m 0.30 c

c c

= =

2.022 m 102.7

CX0 CXu CX CX CXq CXe

= = = = = =

0 0.2199 0.4653 0 0 0

CZ0 CZu CZ CZ CZq CZe

= = = = = =

1.1360 2.2720 5.1600 1.4300 3.8600 0.6238 1.1360 0.037

Cmu Cm Cm Cmq Cme

= = = = =

0 0.4300 3.7000 7.0400 1.5530 15.5 0.002

b 2 KX

= =

13.36 m 0.012

CL 2 KZ

= =

b KXZ

= =

CY CYp CYr CYa CYr

= = = = =

0.9896 0.0870 0.4300 0 0.3037

C Cp Cr Ca Cr

= = = = =

0.0772 0.3444 0.2800 0.2349 0.0286

Cn Cnp Cnr Cna Cnr

= = = = =

0.1638 0.0108 0.1930 0.0286 0.1261

Table B-1: Symmetric and asymmetric stability and control derivatives for the Cessna Ce500 Citation, Cruise

B-3 LTI-System Representation

551

V S 2 KY

= = =

124.5 m/sec 70.0 m2 2.72

m xcg

= =

16200 kg 0.32 c

c c

= =

2.58 m 137.5

CX0 CXu CX CX CXq CXe

= = = = = =

0 0.09 0.15 0 0 0

CZ0 CZu CZ CZ CZq CZe

= = = = = =

0.45 0.90 5.90 1.59 7.36 0.44 0.45 0.0342

Cmu Cm Cm Cmq Cme

= = = = =

0 0.80 6.50 16.50 1.80 12.22 0.0

b 2 KX

= =

29.00 m 0.0127

CL 2 KZ

= =

b KXZ

= =

CY CYp CYr CYa CYr

= = = = =

0.90 0.23 0.48 0 0.29

C Cp Cr Ca Cr

= = = = =

0.09 0.60 0.23 0.086 0.029

Cn Cnp Cnr Cna Cnr

= = = = =

0.11 0.02 0.14 0.0 0.086

Table B-2: Symmetric and asymmetric stability and control derivatives for the Fokker F-27 Friendship, Cruise

V S 2 KY

= = =

66.75 m/sec 16.17 m2 0.6814

1199.8 kg

c c

= =

1.494 m 47.05

CX0 CXu CX CX CXq CXe

= = = = = =

0 0.093 0.18 0 0 0

CZ0 CZu CZ CZ CZq CZe

= = = = = =

0.310 0.620 4.631 0.850 1.95 0.430

Cmu Cm Cm Cmq Cme

= = = = =

0 0.890 2.600 6.200 1.28

Table B-3: Symmetric stability and control derivatives for the Cessna Ce-172 Skyhawk, Cruise

Flight Dynamics

552

Aircraft Parameters

V S 2 KY

= = =

51.82 m/sec 21.37 m2 0.8979

m xcg

= =

5897 kg 0.32 c

c c

= =

2.134 m 105.56

CX0 CXu CX CX CXq CXe

= = = = = =

0 0 0.580 0 0 0

CZ0 CZu CZ CZ CZq CZe

= = = = = =

1.640 3.72 5.5296 0.80 2.050 0.400

Cmu Cm Cm Cmq Cme

= = = = =

0.004 0.660 2.50 6.75 0.980

Table B-4: Symmetric stability and control derivatives for the Learjet I, Approach

V S 2 KY

= = =

103.63 m/sec 26.01 m2 1.646

3175 kg

c c

= =

1.981 m 58.35

CX0 CXu CX CX CXq CXe

= = = = = =

0 0.06 0.06 0 0 0

CZ0 CZu CZ CZ CZq CZe

= = = = = =

0.191 0.402 5.510 1.250 4.05 0.600

Cmu Cm Cm Cmq Cme

= = = = =

0 1.890 4.550 17.00 2.00

Table B-5: Symmetric stability and control derivatives for the Beechcraft M99, Cruise

B-3 LTI-System Representation

553

V S 2 KY

= = =

67.36 m/sec 510.97 m2 2.3345

m lh xcg

= = =

255830 kg 31.09 m 0.25 c

c c

= =

8.321 m 49.12

CX0 CXu CX CX CXq CXe

= = = = = =

0 0 0.630 0 0 0

CZ0 CZu CZ CZ CZq CZe

= = = = = =

1.760 3.3 5.933 3.350 2.825 0.360

Cmu Cm Cm Cmq Cme

= = = = =

0.071 1.450 1.650 10.70 1.400

Table B-6: Symmetric stability and control derivatives for the Boeing 747-100, Approach

V S 2 KY

= = =

129.1 m/sec 510.97 m2 2.488

m lh xcg

= = =

254240 kg 31.09 m 0.32 c

c c

= = =

8.321 m 56.51 1.058 kg m3

CX0 CXu CX CX CXq CXe

= = = = = =

0 0.0478 0.687 0 0 0

CZ0 CZu CZ CZ CZq CZe

= = = = = =

0.477 0.954 4.487 6.62 4.27 0.353

Cmu Cm Cm Cmq Cme

= = = = =

0.0252 0.554 3.39 19.45 1.42

Table B-7: Symmetric stability and control derivatives for the Boeing 747-100, Holding, aps up

Flight Dynamics

554

Aircraft Parameters

V S 2 KY

= = =

73.0 m/sec 510.97 m2 2.488

m lh xcg

= = =

254240 kg 31.09 m 0.32 c

c c

= = =

8.321 m 48.81 1.125 kg m3

CX0 CXu CX CX CXq CXe

= = = = = =

0 0.42 1.59 0 0 0

CZ0 CZu CZ CZ CZq CZe

= = = = = =

1.49 2.98 5.293 6.70 6.66 0.353

Cmu Cm Cm Cmq Cme

= = = = =

0.185 1.05 3.45 21.98 1.42

Table B-8: Symmetric stability and control derivatives for the Boeing 747-100, Approach, aps 33o

V S 2 KY

= = =

66.142 m/sec 510.97 m2 3.77

m lh xcg

= = =

249650 kg 31.09 m 0.25 c

c c

= = =

8.321 m 48.40 1.125 kg m3

CX0 CXu CX CX CXq CXe

= = = = = =

0.0944 0.90 1.2542 0 0 0

CZ0 CZu CZ CZ CZq CZe

= = = = = =

1.80 3.60 5.344 2.00 2.84 0.355

Cmu Cm Cm Cmq Cme

= = = = =

0 1.536 1.70 10.75 1.409

Table B-9: Symmetric stability and control derivatives for the Boeing 747-100, Landing

B-3 LTI-System Representation

555

V S b m

= = = =

145 m/sec 153.5 m2 37.49 m 59020 kg

b 2 KX 2 KZ KXZ

= = = =

17.219 0.0283 0.0471 0

6900 m

CY CYp CYr CYa CYr

= = = = =

0.5960 0 0.3690 0 0.2150

C Cp Cr Ca Cr

= = = = =

0.1374 0.5200 0.1440 0.0975 0

Cn Cnp Cnr Cna Cnr

= = = = =

0.1173 0.0210 0.1800 0.0052 0.1030

Table B-10: Asymmetric stability and control derivatives for the Lockheed L1049C Super Constellation, Cruise

V S b m

= = = =

65.4 m/sec 153.5 m2 37.49 m 51200 kg

b 2 KX 2 KZ KXZ

= = = =

7.26 0.0326 0.0543 0

0m

CY CYp CYr CYa CYr

= = = = =

0.5610 0 0.1945 0 0.2150

C Cp Cr Ca Cr

= = = = =

0.1374 0.5200 0.2750 0.0975 0

Cn Cnp Cnr Cna Cnr

= = = = =

0.1173 0.0626 0.1800 0.0052 0.1030

Table B-11: Asymmetric stability and control derivatives for the Lockheed L1049C Super Constellation, Approach

Flight Dynamics

556

Aircraft Parameters

V S b m

= = = =

77.1 m/sec 470 m2 24.4 m 100000 kg

b 2 KX 2 KZ KXZ

= = = =

7.185 0.052 0.249 0.066 0.1660 0.1410 0.2500 0.1010 0

0m

CY CYp CYr CYa CYr

= = = = =

0.3640 0 0 0 0.1290

C Cp Cr Ca Cr

= = = = =

Cn Cnp Cnr Cna Cnr

= = = = =

0.1360 0.1430 0.2100 0 0.0790

Table B-12: Asymmetric stability and control derivatives for the BAC-Aerospatiale Concorde, Approach

V S b m

= = = =

660 m/sec 18.6 m2 6.7 m 11800 kg

b 2 KX 2 KZ KXZ

= = = =

1072 0.0134 0.247 0

20000 m

CY CYp CYr CYa CYr

= = = = =

1.4200 0 0 0.0735 0.3200

C Cp Cr Ca Cr

= = = = =

0.0100 0.3150 0 0.0500 0.0100

Cn Cnp Cnr Cna Cnr

= = = = =

0.4000 0 1.4500 0.0490 0.2000

Table B-13: Asymmetric stability and control derivatives for the North-American X-15 experimental aircraft, Cruise (unspecied)

B-3 LTI-System Representation

557

V S b m

= = = =

40.2 m/sec 23.23 m2 14.63 m 2315 kg

b 2 KX 2 KZ KXZ

= = = =

5.56 0.0845 0.0192 0

0m

CY CYp CYr CYa CYr

= = = = =

0.6000 0 0.1600 0 0.2420

C Cp Cr Ca Cr

= = = = =

0.0560 0.5500 0.1200 0.1120 0

Cn Cnp Cnr Cna Cnr

= = = = =

0.0248 0.0600 0.0530 0.0015 0.0970

Table B-14: Asymmetric stability and control derivatives for the De Havilland Canada DHC-2 Beaver, Approach

Flight Dynamics

558

Aircraft Parameters

Appendix C MATLAB Files

C-1

Introduction

In this appendix a list of MATLAB les, as used in this book, is given. The presented MATLAB les are only to be used regarding the LTI, or state-space, form, i.e., x=Ax+B u y =C x+D u All les are current with MATLAB version 5.2.

C-2

Root Finding

MATLAB routine roots.m returns the roots of a polynomial, EIG = roots(POLYNOMIAL) with EIG the eigenvalues, or roots in complex form, EIG = A + j B, and POLYNOMIAL the coecients of a polynomial in descending order, for instance, POLYNOMIAL = [an an1 an2 a2 a1 a0 ] for the polynomial, an xn + an1 xn1 + an2 xn2 + + a2 x2 + a1 x + a0 or,
Flight Dynamics

560 ROOTS

MATLAB Files Find polynomial roots. ROOTS(C) computes the roots of the polynomial whose coefficients are the elements of the vector C. If C has N+1 components, the polynomial is C(1)*X^N + ... + C(N)*X + C(N+1). See also POLY, RESIDUE, FZERO.

C-3

Eigenvalue Computation

Regarding stability analysis of LTI systems, the following program is used, EIG Eigenvalues and eigenvectors. E = EIG(X) is a vector containing the eigenvalues of a square matrix X. [V,D] = EIG(X) produces a diagonal matrix D of eigenvalues and a full matrix V whose columns are the corresponding eigenvectors so that X*V = V*D. [V,D] = EIG(X,nobalance) performs the computation with balancing disabled, which sometimes gives more accurate results for certain problems with unusual scaling. E = EIG(A,B) is a vector containing the generalized eigenvalues of square matrices A and B. [V,D] = EIG(A,B) produces a diagonal matrix D of generalized eigenvalues and a full matrix V whose columns are the corresponding eigenvectors so that A*V = B*V*D. See also CONDEIG, EIGS. Overloaded methods help sym/eig.m help lti/eig.m

C-4

Simulation

For the simulation of LTI systems, the following programs are used, Step-response STEP Step response of continuous-time linear systems.

C-4 Simulation

561

Y = STEP(A,B,C,D,iu,T) calculates the response of the system: . x = Ax + Bu y = Cx + Du to a step applied to the iuth input. Vector T must be a regularly spaced time vector that specifies the time axis for the step response. STEP returns a matrix Y with as many columns as there are outputs y, and with LENGTH(T) rows. [Y,X] = STEP(A,B,C,D,iu,T) also returns the state time history. Y = STEP(NUM,DEN,T) calculates the step response from the transfer function description G(s) = NUM(s)/DEN(s) where NUM and DEN contain the polynomial coefficients in descending powers.

Time-response to arbitrary inputs LSIM Simulation of continuous-time linear systems to arbitrary inputs. LSIM(A,B,C,D,U,T) calculates and plots the time response of the system: . x = Ax + Bu y = Cx + Du to input time history U. Matrix U must have as many columns as there are inputs, U. Each row of U corresponds to a new time point, and U must have LENGTH(T) rows. Y=LSIM(A,B,C,D,U,T) returns, without plotting, a matrix Y with as many columns as there are outputs y, and with LENGTH(T) rows. [Y,X] = LSIM(A,B,C,D,U,T) also returns the state time history. LSIM(A,B,C,D,U,T,X0) can be used if initial conditions exist. LSIM(NUM,DEN,U,T) plots the time response from the transfer function description G(s) = NUM(s)/DEN(s) where NUM and DEN contain the polynomial coefficients in descending powers. LSIM now assumes a straight line approximation between the discrete values of u. This is achieved via a first order hold approximation. Left hand arguments suppress plotting. J.N. Little 4-21-85 Copyright (c) 1985-89 by the MathWorks, Inc.
Flight Dynamics

562 Revised A.C.W.Grace 8-27-89 Synopsis: lsim(a,b,c,d,u,t); lsim(num,den,u,t); lsim(a,b,c,d,u,t,x0); lsim(num,den,u,t,x0); [y,x]=lsim(a,b,c,d,u,t,.) [y,x]=lsim(num,den,u,t,.)

MATLAB Files

Index

Angle of attack Aerodynamic, 30 Kinematic, 30 Angular velocity vector, xxv Aspect ratio, xxvii Cruising Flight, xxxi Dihedral, xxvii Ecliptic, 22 Equator, 22 Greenwich meridian, 24 Landing, xxxi Latitude Geocentric, 26 Geodesic, 26 Gravitation, 26 Mean aerodynamic chord, xxvi Mean chord, geometric chord, xxvii Polar motion, 22 Powered Approach, xxxi Reference frames, xxiv, 21 Aerodynamic, xxiv, 30 Body-xed, xxiv, 27 Conventional celestial, 23 Conventional terrestrial, 24 Earth-centered inertial, 22 Earth-centered, Earth-xed, 24 Inertial, xxiv, 22 International celestial, 22

International terrestrial, 24 Kinematic, xxiv, 30 Normal Earth-xed, xxiv, 25 Principle axes, xxiv, 28 Stability, xxiv, 28 Vehicle, xxiv, 32 Vehicle carried normal Earth, xxiv, 26 Zero-lift body axes, xxiv, 28 Sideslip angle Aerodynamic, 30 Kinematic, 30 Slipping ight, xxxi Steady ight, xxxi Straight ight, xxxi Symmetric ight, xxxi Taper ratio, xxvii Time derivative, xxiv Transformation matrix, xxiv Velocity Aerodynamic, 30 Kinematic, 30 Vernal equinox, 22 Washout, wing-twist, xxvii Wing airfoil, xxx Wing area, xxvi Wing sweep, xxvii Wingspan, xxvi

Flight Dynamics