Ground experiments are executed and the results are used to characterize the nonlin-
ear structural dynamics of store-loaded F-16 pylons. Results from a modal-space imple-
mentation of the restoring-force-surface method and results from homogeneity tests are
consistent with linear modal stiffness and nonlinear structural damping. Different theories
exist regarding what nonlinear mechanisms bound F-16 aeroelastic limit-cycle oscillations;
the results in this paper help legitimize nonlinear structural damping as a bounding mech-
anism.
Nomenclature
Superscript
( ˙ ) = derivative with respect to time
+ = matrix pseudoinverse
> = matrix transpose
∗
This material is declared a work of the U.S. Government and is not subject to copyright protection in the United States.
†
Carriage Mechanics Engineer, SURVICE Engineering Co., 205 West D Avenue, Suite 348.; pasquale@vt.edu, AIAA
Member.
‡
Principal Technical Advisor, Carriage Mechanics Division, 205 West D Avenue, Suite 348.; AIAA Associate Fellow.
§
Carriage Mechanics Engineer, SURVICE Engineering Co., 205 West D Avenue, Suite 348.; dustin.coleman@survice.com,
AIAA Member.
1 of 13
Background
F-16 limit cycle oscillations & nonlinear aeroelasticity
Flutter-induced limit cycle oscillations (LCO) are encountered regularly on F-16 aircraft.1 Constraints
on allowable LCO amplitude result in reduced airspeed envelopes for many external store configurations.
The nondivergent behavior of LCO is a result of intrinsically nonlinear physics, so classical linear flutter
analysis is unable to predict LCO amplitude. This shortfall of classical flutter analysis has historically
resulted in a heavy reliance on flight test for certification of new F-16 store configurations. Recent work
by Denegri et. al 2 includes a detailed review of historic and current nonlinear analysis efforts attempting
to increase predictive capabilities and reduce flight test requirements. It is understood that classical
flutter initiates F-16 LCO,3 but the nonlinear mechanisms that bound the oscillations are uncertain. This
uncertainty has resulted in only limited success in F-16 LCO prediction, and a continued reliance on flight
test for certification.
Foundational work in nonlinear aeroelasticity is documented by Woodcock4 with continued early de-
velopments documented by Haidl5 and Breitbach.6 The works include identification of significant aircraft
structural nonlinearities with detailed experimental results and comparisons to classical nonlinear analysis
techniques. Breitbach6 categorizes aircraft structural nonlinearities as concentrated or distributed. Con-
centrated nonlinearities are common in mechanical control-system devices and structural interfaces; these
nonlinearities can manifest as Coulomb friction, free play, hysteresis, and piecewise-linear stiffness. Similar
nonlinear dynamics can occur at riveted, screwed, and bolted connections, which are commonly distributed
throughout aircraft structures. The resulting damping and stiffness nonlinearities are effectively charac-
terized as distributed. Breitbach6 shows distributed nonlinear damping results for an F-104G aircraft and
a composite wing spar.
Meijer and Cunningham7 argue that nonlinear aerodynamics, specifically shock-induced trailing-edge
separation, is plausibly bounding F-16 LCO. The authors use a semi-empirical approach with nonlinear
aerodynamics to attain results that are well-correlated with flight test data. Cunningham and Holman8
modified the semi-empirical approach to include Coulomb friction, a nonlinear structural dynamic effect,
at the wing-tip launcher attachment. The computational results were sensitive to Coulomb friction. Re-
cent work by Denegri et. al 2 and Gabbard et. al 9 demonstrates computational aeroelastic analysis using
nonlinear Euler aerodynamics coupled with nonlinear structural dynamics. Structural nonlinearities are
based on assumed nonlinear modal damping functions. Results in Refs. [2, 9] resemble F-16 LCO observed
in flight tests.
This paper presents results from ground experiments that characterize the nonlinear structural dynam-
ics of F-16 pylons. The results help legitimize nonlinear structural damping as a bounding mechanism for
F-16 LCO.
2 of 13
Methods
Experimental Program
Four configurations of F-16 underwing weapon pylons carrying inert warheads were mounted on a
fixture and tested. A geometric model of a single test configuration is shown in Fig. 1a. Testing of the
empty fixture was also accomplished. The test articles, shown in Fig. 1b, include:
3 of 13
4 of 13
Point Description
P1 Pylon side of forward pylon-to-fixture attachment point
P2 Pylon side of aft pylon-to-fixture attachment point
P3 Weapon side of forward hook/lug
P4 Pylon/adapter side of forward hook/lug
P5 Fixture side of aft pylon-to-fixture attachment point
P6 Weapon side of forward/left sway brace contact point
P7 Weapon side of aft hook/lug
P8 Pylon/adapter side of aft hook/lug
P9 Pylon side of pylon/adapter aft connection
P10 Side of aft pylon section near countermeasures
P11 Bottom of aft pylon section near countermeasures
P12 Back of pylon
P13 Weapon nose cone
P14 Back of weapon
P15 Fixture side of forward pylon-to-fixture attachment point
P16 Sway brace side of forward/left sway brace contact point
The shaker armature is attached to the store using a modified nose cone. Both vertical and horizontal
force-input directions are used, as shown in Fig. 4. Each pylon-store configuration is tested using: 1)
random excitation at multiple force levels, 2) forward and backward linear swept-sine input at multiple
force levels, and 3) stepped-sine input at multiple force levels around several resonance peaks.
5 of 13
• Complex Exponential
curve fitting methods provide frequency domain, z domain, and time domain pole estimates, respectively.18
Stability diagrams are used for locating stable poles. Frequency domain curve fitting is used for all residue
estimates. Agreement between curve fitting techniques provides confidence in the linear parameter esti-
mates.
For a linear system fr (η, η̇) = cη̇ + kη, and the modal degrees of freedom are independent (i.e., the system
matrices are diagonal) if damping is proportional . Modal restoring-force surfaces for linear systems can be
interpreted as single degree of freedom restoring-force surfaces. Modal coupling is inevitable for nonlinear
systems, but modal restoring-force surfaces provide valuable insight to the system nonlinear stiffness and
damping properties.
Time histories of physical acceleration and physical applied force are measured using accelerometers
and force transducers. The physical time histories are transformed to modal space using the mode shape
matrix from linear parameter estimation
η̈ = Φ+ ü (2)
>
fa = Φ pa . (3)
Trapezoidal rule is used to integrate η̈ and produce time histories for η̇ and η; high-pass filters are used
to remove integration error. The scattered time-history points that make up the modal restoring-force
6 of 13
Results
Results are presented for a single configuration with a inert 2,000 lb warhead mounted on a countermeasure-
dispensing pylon.
7 of 13
MAC
4 0.612 13.7 0.00 0.01 0.00 0.94 0.00 0.01
5 0.737 18.5 0.01 0.27 0.00 0.00 0.97 0.36
6 0.694 25 0.00 0.08 0.00 0.01 0.39 0.90
% Diff. 0.0% 0.0% 0.0% 0.0% 0.5% 0.0%
% Diff. 25.7% 4.7% 11.3% 6.3% 13.0% 36.9%
(a) Linear parameter estimates comparison. (b) Pylon lateral mode – 8.9 Hz.
(c) Pylon pitch mode – 13.1 Hz. (d) Pylon yaw mode – 13.7 Hz.
Differences in mode shape and frequency estimates from runs with different force-input directions are
negligible; however, there are substantive differences in modal mass and damping estimates based on runs
with different force-input directions. For restoring-force-surface method inputs, linear parameters for pylon
pitch were estimated from vertical input runs and linear parameters for pylon lateral and pylon yaw modes
are estimated from horizontal input runs. Low-level random excitation was used for both cases.
8 of 13
Homogeneity Test
A homogeneity test is implemented by superimposing frequency response functions (FRFs) from dif-
ferent input amplitudes. Figure 7 shows FRFs for two measurements and from 10 lb RMS and 22 lb RMS
random input. The peaks of modes 2, 3, and 4 show reductions in amplitude and frequency as the input
level is increased. These results are consistent with an increase in damping at higher input levels.
9 of 13
10 of 13
(c) Pylon pitch mode stiffness force vs. displacement. (d) Pylon pitch mode damping force vs. velocity.
(e) Pylon yaw mode stiffness force vs. displacement. (f) Pylon yaw mode damping force vs. velocity.
11 of 13
Ground experiments were executed and the results were used to characterize the nonlinear structural
dynamics of store-loaded F-16 pylons. Time history results and time-frequency results clearly indicated
the presence of structural nonlinearity. Results from a modal-space implementation of the restoring-force-
surface method and results from homogeneity tests are consistent with linear modal stiffness and nonlinear
structural damping. Cubic-like modal damping was observed for the three most relevant pylon modes.
Different theories exist regarding what nonlinear mechanisms bound F-16 aeroelastic limit-cycle oscilla-
tions, but the results in this paper help legitimize nonlinear structural damping as a bounding mechanism.
Assumed nonlinear damping profiles used in Ref. [2] approximated similar nonlinear damping behavior for
a full aircraft, and computational results in Ref. [2] resembled LCO observed in F-16 fight tests. Future
ground experiments on additional substructures or a full F-16 aircraft could result in better understanding
of the aircraft dynamics and could supply data for measured estimates of nonlinear sturctural damping.
A least-squares polynomial fit of the restoring force surface in modal space could supply modal damping
models for aeroelastic analysis.
Acknowledgments
This material is based on research sponsored by the U.S. Air Force Research Laboratory (AFRL) with
funding provided by the U.S. Air Force Office of Scientific Research (AFOSR) under agreement number
16RWCOR359.
The authors would like to thank our collaborators at AFSEO, AFRL, and the 780th Test Squadron.
We would especially like to thank Crystal Pasiliao and Daniel Reasor for leading the AFRL program;
Michael Kendra for leading the AFOSR program; Vin Sharma and Lynn Neergaard for for helping with
the test planning and execution; Mike Prestarri for helping with test-fixture modeling; Gaëtan Kerschen,
Thibaut Detroux, Jean-Philippe Noël, Jay Northington, and James Dubben for sharing their insights on
F-16 testing; Phillip Douglas for helping with geometric modeling, Glen Beck and Mike Symons for test
management; Ed Santiago for designing the test fixture modifications; and Jim LaBerge, Adam Sanders,
and Warren Gould for assembling, instrumenting, and executing the tests.
References
1
Denegri Jr., C. M., “Limit Cycle Oscillation Flight Test Results of a Fighter with External Stores,” Journal of Aircraft,
Vol. 37, No. 5, September 2000, pp. 761–769, doi: 10.2514/6.2000-1394.
2
Denegri Jr, C. M., Sharma, V. K., and Northington, J. S., “F-16 Limit-Cycle Oscillation Analysis Using Nonlinear
Damping,” Journal of Aircraft, 2015, pp. 1–8, doi: 10.2514/1.C033315.
3
Denegri, C. M., Dubben, J. A., and Maxwell, D. L., “In-Flight Wing Deformation Characteristics During Limit Cycle
Oscillations,” Journal of Aircraft, Vol. 42, No. 2, 2005, pp. 500–508, doi: 10.2514/1.1345.
4
Woodcock, D. L., “Structural Non-Linearities,” North Atlantic Treaty Organization Advisory Group for Aerospace
Research and Development (AGARD) Manual on Aeroelasticity Part 1. Chapter 6 , 1960, pp. 1–55.
5
Haidl, G., “Non-Linear Effects in Aircraft Ground and Flight Vibration Tests,” Tech. Rep. AGARD-R-652, North
Atlantic Treaty Organization Advisory Group for Aerospace Research and Development, 7 Rue Ancelle 92200 Neuilly Sur
Seine, France, 1976.
6
Breitbach, E., “Effects of Structural Non-Linearities on Aircraft Vibration and Flutter,” Tech. Rep. AGARD-R-665,
North Atlantic Treaty Organization Advisory Group for Aerospace Research and Development, 7 Rue Ancelle 92200 Neuilly
Sur Seine, France, 1978.
7
Meijer, J. and Cunningham Jr., A. M., “A Semi-Empirical Unsteady Nonlinear Aerodynamic Model to Predict Tran-
sonic LCO Characteristics of Fighter Aircraft,” 36th Structures, Structural Dynamics and Materials Conference, Structures,
Structural Dynamics, and Materials and Co-located Conferences, 1995, doi: 10.2514/6.1995-1340.
8
Cunningham Jr., A. M. and Holman, R. J., “Time Domain Aeroelastic Solutions: A Critical Need for Future Analytical
Methods Developments,” NATO Rept. RTA-MP-AVT-154, Brussels, 2008.
9
Gabbard, M. D., Lindsley, N. J., and Kunz, D. L., “Modeling the Effects of Underwing Missile Canards on F-16 Limit
Cycle Oscillations,” Journal of Aircraft, Vol. 53, No. 1, 2015, pp. 202–216, doi: 10.2514/1.C033301.
12 of 13
13 of 13