Input
Device
Signal Conditioning
Processing Variable
Output
Device
Process
Variable
Measurement
Reading
Figure 1. Instrumentation
3. Marine Science/Biology
4. Metrology
5. Automation/Robotics
6. Automatic Process Control
7. Biomedical/Pharmaceutical
8. Cryogenics
9. Nuclear
10. Power
11. Analytical
12. Telemetry
Process Variables
1. Controlled Variables (CVs): The process variables that are controlled. These are the
variables which quantify the performance or quality of the final product, which also are called
output variables. The desired value of a controlled variable is called the set point.
2. Manipulated Variable (MVs): The process variable that can be adjusted in order to keep
the controlled variables at or near the set point. These are input variables that are adjusted
dynamically to keep the controlled variables at their set-points. Typically the manipulated
variable is the flow rates.
3. Disturbance Variables (DVs): The process variables that affect the controlled variables
but cannot be manipulated. These are called the load variables and represent input variables
that can cause the controlled variables to deviate from their respective set points.
Set-point change: implementing a change in the operating conditions. The set point signal is
changed and the manipulated variable is adjusted appropriately to achieve a new operating
conditions. This is also called servomechanism (or servo) control.
Disturbance change: the process transient behavior when a disturbance enters, also called
regulatory control or load change. A control system should be able to return each controlled
variable back to its set-point.
It is important to note, however, that control systems of interest for analysis or design
purposes include not only those manufactured by humans, but those that normally exist in
nature, and control systems with both manufactured and natural components.
The incentives of using a control system are as follows:
1. For the plant to come up with a safe operation of a process.
2. For the plant to produce the desired amounts and quality of the final products.
3. For the plant to adhere to certain environmental regulations imposed by the state.
4. For the plant processes to work within certain operational constraints.
5. For the plant to conform with certain market conditions, that is, the availability of raw
materials and the demand of final products in order to minimize operating costs and to
maximize profit.
A. Variables in a Process
1. Input variables the stimulus applied to a control system, which produces a specified
response. Their values may be adjusted, in which case, they are called manipulated variable,
otherwise, they are disturbances. Input variables can be further classified into:
a. Manipulated/adjustable variables are those values that can be adjusted freely by
the human operator or a control mechanism.
b. Disturbances are those values that are not the result of adjustment by an operator or a
control system.
2. Output variables the actual response obtained from a control system the value of which
is affected by the value of the input variable. Output variables can be further classified into:
a. Measured Output variables are those values that are known by directly measuring
them.
b. Unmeasured Output variables are those values that cannot be directly measured.
B. Design Elements of a Control System
In attempting to design a control system that meets the process requirements, the following
elements must be considered.
1. Control objectives, which are defined qualitatively and quantitatively in terms of the
output variables
2. Measurements of certain process variables, which aid in monitoring the process
performance
3. Manipulated variables to be used
4. Control configuration
a. Feedback, Feedforward, Inferential, etc.
5. Design the mode of controller
a. Proportional, Integral, Derivative or Combination
Command
Input
Reference
Selector
Reference
Input
Dynamic
Unit
Output
Motor
Powering
Device
Controller
Command
Input
Output
Load
Control
Effort
In contrast closed loop systems are dependent on the output. They are also called feedback.
They permit the output to be compared with the input to the system so that the appropriate
5
control action may be formed as function of the input and output. A block diagram of
feedback control is shown in Figure 7.
Controller
Final Control
Element
Process
Measuring
Device
Figure 7. Feedback Control.
Feedback is that property of a closed system which permits the output (or some other control
variable) to be compared with the input to the system (or an input to some other internally
situated component or subsystem) so that the appropriate control action may be formed as
some function of the output and input.
Characteristics of Feedback
1. Increased accuracy (say the ability to faithfully reproduce the output)
2. Tendency toward oscillations or instability
3. Reduced sensitivity of the ratio of output to input to variations in system parameters and
other characteristics.
4. Reduced effects of nonlinearities.
5. Reduced effects of external disturbances or noise.
6. Increased bandwidth. The bandwidth of a system is a frequency response measure of how
well the system responds to variations or frequencies in the input signal.
PROBLEMS
1. Identify the input and output for the pivoted adjustable mirror.
2. Identify a possible input and a possible output for a rotational generator for electricity.
3. Identify the organ system components, and input and output, and describe the operation of
the biological control system consisting of a human reaching for an object.
4. How is the following open loop systems calibrated?
a. automatic washing machine
b. automatic toaster
c. voltmeter
5. Explain the operation of ordinary traffic signals, which control the automobile traffic at
roadway intersection. Why is it open loop control system? How can traffic be controlled more
efficiently? Why is the system considered to e closed loop?
6. Describe in a simplified way, the components and variables of the biological control system
involved in walking in a specified direction. Why is walking a closed loop operation? Under
what conditions would the human walking apparatus become an open loop system? A
sampled data system? Assume the person has normal vision.
10
4. MATHEMATICAL MODELING
In the process industry control systems are designed based on how the systems behave under
the influence of changes in manipulated variables and external disturbances. Mathematical
models are mathematical representations of the physical and chemical phenomena that are
taking place in a chemical process.
B. Classification of Models
Mathematical models can be considered in three different classifications, depending on how
they are derived.
1. Theoretical models are developed using principles of chemistry and physics.
2. Empirical models obtained from a mathematical statistical analysis of process
operating data
3. Semi-empirical models combine 1 and 2 with one or more parameters to be evaluated
from plant data.
d1 d2
d3.......di
Manipulated variables
m1
m2
m3
mi
disturbances
Measured variables
PROCESS
y1
y2
y3
yi
Unmeasured outputs
Figure 2. Representation of the Input and Output Variables of a Process.
11
We apply the input-output model, that is, the cause and effect relationship in a process:
output f (input var iables )
yi f (m1 , m2 , m3 ,...., d1 , d 2 , d 3 ,..., d i )
(1)
Degrees of freedom of a system refer to the independent variables that must be specified in
order to define the process completely.
f V E
(2)
steam
h
F, T
Mass Balance:
d ( Ah )
Fi F
dt
(3)
Assuming constant
dh
Fi F
dt
(4)
Energy balance:
d [ AhC (T Tr )]
FiC (Ti Tr ) FC (T Tr) Q
dt
(5)
d (hT )
Q
FiTi FT
dt
C
12
(6)
Ah
dT
dh
Q
AT
FiTi FT
dt
dt
C
(7)
Ah
dT
Q
T ( Fi F ) FiTi FT
dt
C
(8)
Substituting Eq (K 4) gives
Ah
dT
Q
Fi (Ti T )
dt
C
(9)
13
(10)
( X X o ) df
( X X o )2
1!
2!
dt X o
d2 f
2
dt
( X X o )n
...
n!
Xo
dn f
n
dt
(11)
df
f ( x) f ( x0 ) ( x x0 )
dX X o
(12)
dX
df
f ( X o ) ( X X o )
dt
dX X o
(13)
Substituting in Eq (10).
This is the linearized form of equation 10. The error introduced in the approximation is the
same order of magnitude as the term
I
( X X o )2
2!
d2 f
2
dX
Xo
(14)
The linear approximation is satisfactory only when X is very close to Xo, when the value of I
is very small.
Deviation Variables (Perturbation)
Let Xs = steady state value of X
dX s
0 f (X s )
dt
Then
From
dX
dt
f ( X s ) , a non-linear equation
Linearizing around Xs
14
(15)
dX
f
f ( X s ) ( X X s )
dt
X X s
d(X Xs )
f
(X Xs)
dt
X X s
Subtracting
Let
(16)
(17)
X X s , a deviation variable
d f
dt X X s
(18)
(19)
dX 2
f2 (X1, X 2 )
dt
f
f1 ( X 1 , X 2 ) f1 ( X 1,o , X 2,o ) ( X 1 X 1,o ) 1
X 1 X1,o
(20)
( X X 1,0 ) 2 f1
f
( X 2 X 2,o ) 1
1
X 2
2!
X 2 X 2 ,o
1 X 1, o
2
( X 2 X 2,o ) 2 f
n f1
..
(
X
X
)(
X
X
)
1
1,o
2
2 ,o
2
2!
X 1X 2
X 2 X 2 ,o
f
f 2 ( X 1 , X 2 ) f 2 ( X 1,o , X 2,o ) ( X 1 X 1,o ) 2
X 1 X1,o
(21)
( X X 1,0 ) 2 f 2
f
( X 2 X 2,o ) 2 1
X 2
2!
X 2 X 2 ,o
1 X 1o
2
( X 2 X 2,o ) 2 f 2
n f2
....
(
X
X
)(
X
X
)
1
1,o
2
2 ,o
2
2!
X 1X 2
X 2 X 2 ,o
f
f1 ( X 1 , X 2 ) f1 ( X 1,o , X 2,o ) ( X 1 X 1,o ) 1
X 1 X1,o
f
( X 2 X 2,o ) 1
X 2 X 2 ,o
15
(22)
f
f 2 ( X 1 , X 2 ) f 2 ( X 1,o , X 2,o ) ( X 1 X 1,o ) 2
X 1 X1,o
(23)
f
( X 2 X 2,o ) 2
X 2 X 2 ,o
Substituting to the original differential equation
f
f
dX 1
f1 ( X 1 X 1,o ) ( X 1 X 1,o ) 1 ( X 2 X 2,o ) 1
dt
X 1 X
X 2 X
1, o
(24)
2 ,o
f
f
dX 2
f 2 ( X 2 X 2,o ) ( X 1 X 1,o ) 2 ( X 2 X 2,o ) 2
dt
X 1 X
X 2 X
1, o
(25)
2,o
Expressing the linearized equation in terms of deviation variables, let the steady state X1,s and
X2,s as the point of linearization, that is
X 1,o X 1,s and X 2,o X 2,s
At steady state:
f1 ( X 1,s , X 2,s ) 0
f 2 ( X 1,s , X 2,s ) 0
Subtracting,
d ( X 1 X 1, s )
dt
d ( X 2 X 2, s )
dt
f
f
( X 1 X 1, s ) 1 ( X 2 X 2, s ) 1
X 1 s
X 2 s
(26)
f
f
( X 1 X 1, s ) 2 ( X 2 X 2,s ) 2
X 1 s
X 2 s
(27)
f
f
d1
1 1 2 1
dt
1 s
2 s
(28)
f
f
d 2
1 2 2 2
dt
1 s
2 s
(29)
d1
a111 a12 2
dt
(30)
d 2
a211 a22 2
dt
(31)
16
where:
f
a11 1
1 s
f
a12 1
2 s
f
a 21 2
1 s
f
a 22 2
2 s
f1 f1
1 s 2
f 2 f 2
1 s 2
(32)
an
dny
d n1 y
dy
a
...a1
a0 y K p x(t )
n 1
n
n 1
dt
dt
dt
(33)
where y is the output variable, x is the input variable, Kp is the process gain and an, an-1, etc are
the constants. Taking the Laplace Transform with initial condition y (0) =0, yields
(34)
Kp
y ( s)
G( s)
n
x ( s) (a n s a n1 s n 1 ...a1 s a0 )
(35)
Simplifying yields,
y ( s)
is called the transfer function of the system. The transfer function is the ratio of
x ( s)
the Laplace Transform of the output variable y (s) to the Laplace transform of the input
variable x (s)
G( s)
Transfer function
LaplaceTransform of output y ( s)
LaplaceTransform of input x ( s)
(36)
17
From the previous section, we were able to get the transfer function of the system as shown in
equation 35, which can be simplified as
G( s)
y ( s)
x ( s)
(37)
The system dynamics in s domain can be conveniently shown using block diagram
representation below.
x (s)
y (s)
G(s)
Figure 3. Block Diagram Representation of the System.
The arrow entering the block represent the input and arrow leaving the block represent the
output. The transform of the output is conveniently obtained by multiplying the input with the
transfer function inside the block. The Laplace Transform of the output can be obtained from
y ( s ) G( s ) x ( s )
For a double input single output system with dynamics represented by
(38)
G1 ( s)
x2 ( s)
y (s)
G2 ( s )
A double input, double output system can be represented by the equations shown below.
y1 ( s) G1 ( s) x1 ( s) G2 y 2 ( s)
y 2 ( s) G3 ( s) x1 ( s) G4 ( s) x2 ( s)
(39)
The dynamics of the system can be represented by the block diagram shown in Figure 5.
If it is required to determine the transfer function relating one input to a particular output, it
can be obtained by letting the other inputs equal to zero.
y1 ( s )
, let x1 ( s) 0 , so that y1 (s) y 2 (s)G2 (s) .
x2 (s)
The second output y 2 can be obtained from y 2 (s) x2 (s)G4 (s) . Substituting this equation to
the equation for y1 yields y1 (s) x2 (s)G4 (s)G2 (s) .
18
y1 ( s)
G2 ( s)G4 ( s) .
x2 ( s)
G1 ( s)
x1 ( s)
y1 ( s)
G2 ( s )
G3 ( s)
y 2 ( s)
x2 ( s)
G4 ( s )
Response
Output variable
Process
Y(s)
G(s)
G( s)
Y ( s)
X ( s)
(40)
2. It completely describes the dynamic characteristics of the system. From the Transfer
Function, the response of the system to a forcing function can be obtained.
Y (s) G(s).X (s)
19
(41)
Y (t ) L1 Y (s)
(42)
a1 X 1 ( s )
a 2 X 2 (( s )
Y(s)
G(s)
20
(43)
(44)
G( s)
y ( s)
f ( s)
y (s)
(45)
where
f (s) = Laplace Transform of the input function
y (s) = Laplace Transform of the output function
G(s) = Ratio of the output to the input.
The output of the system is determined by the nature of the transfer function.
The response of the system can be determined by taking the inverse LaplaceTransform of the
function:
(46)
y ( s ) G( s) f ( s )
dy
a0 y bf (t )
dt
(47)
dy
y K p f (t )
dt
(48)
a1
b
and K p
. is known as the time constant and Kp is the steady state gain
a0
a0
or static gain.
where
Getting the Laplace Transform of the first order differential equation yields
G( s)
kp
Y ( s)
X ( s) s 1
(49)
This is the Transfer Function of the First Order system. It is defined only for linear timeinvariant system. For non-linear systems, the mathematical model must first be linearized
around a steady state. When defining the transfer function, all initial conditions must be set to
zero.
The time constant is related to the speed of response of the process once the process starts to
responds to a forcing function. The slower the process the larger is the value of .
21
fluid film
Hg
glass wall
Let x = temperature of surrounding fluid
y = temperature of mercury
thermometer reading
Assumptions:
1. Resistance of glass and mercury are negligible.
2. No temperature gradient in the mercury.
3. Heat capacity of glass is negligible.
4. No expansion or contraction of glass walls.
5. For t 0, the thermometer is at steady state; at t 0, a change in the surrounding
temperature x(t) occurs.
Under transient (unsteady state) condition
Heat Balance:
Input Output = Accumulation
hA( x y ) mC p
(50)
dy
dt
(51)
xs y
(52)
At steady state, t 0
hA( xs y s ) 0 ;
hA( x x s ) ( y y s ) mC p
22
d ( y ys )
dt
(53)
Let x x s X ;
y ys Y
where
dY
Y X
dt
(54)
mC p
X ( s) s 1
(55)
(56)
Let
A
qo
h
R
q(t ) qo (t )
qo (t )
d ( A h )
dt
h
R
h Adh
q(t ) R dt
(57)
(58)
(59)
At steady state
qs hs 0
23
(60)
q(t ) qs
A
R
dt
Let q(t ) q s Q ;
(61)
h hs H
Q
H
dH
A
R
dt
(62)
Q( s) ARs 1
Q( s )
(63)
(64)
Let AR
G( s)
R
s 1
(65)
A
x(t)
0
Zero time is the time at which the sudden change in magnitude A occurs.
b. Ramp Input. This corresponds to inputs that drift, that is, they change relatively slowly
upward or downward for some period with a roughly constant slope.
x(t) = 0
x(t) = at
t< 0
t 0
slope = a
x(t)
t
c. Impulse. Impulse functions are not encountered in usual plant operations, nor are they easy
to generate physically by the engineer. To obtain an impulse input it is necessary to inject a
finite amount of energy or material into a process in an infinitesimal length of time.
x(t) = A(t)
1/b
24
(66)
(t ) lim
b 0
u (t ) u (t b)
b
(67)
tw
t<0
(68)
0 t 0
The step input function is given by x(t )
(69)
A t 0
where A is the magnitude of step chance. Taking the Laplace transform, the input in the s
domain is
A
(70)
x( s )
s
The response of the system to the first order system is given by the equation
y ( s)
A KP
s s 1
(71)
y(t ) AK p (1 e t / )
25
(72)
1
0.9
0.8
0.7
0.632
0.6
y(t)/AKp
0.5
0.4
0.3
0.2
0.1
0
3
t/
Figure 6. Response of the First Order System to Step Change in the Input.
1
0.9
0.8
Increasing Kp
0.7
0.6
y(t)/AKp
0.5
0.4
0.3
0.2
0.1
0
Time (sec)
Figure 7. Effect of Varying Kp on the Response of First Order system to Step Change.
26
1
0.9
Increasing
0.8
0.7
0.6
y(t)/AKp
0.5
0.4
0.3
0.2
0.1
0
Time (sec)
on the Response of the First Order System to Step Change in the Input.
of
the
system
is
Therefore, the time constant is the time needed for the response to reach 63.2% of the
maximum value of the response of the system. The time constant determines the speed of the
response of the system. The smaller the value of , the faster is the response; the bigger the
value of the slower is the response of the system. The effect of the time constant is shown
in Figure 8.
3. Impulse Response of the First Order System
Impulse function is given by the equation,
0
x(t ) A
0
t0
t 0
(73)
t 0
y(t ) AK p e t /
The plot of the response to various magnitude of impulse change is shown in Figure 9.
27
(74)
0 t0
x(t )
A sin wt t 0
(75)
Y(t)/Kp
Increasing A
2.5
2
1.5
1
0.5
0
0
Time (sec)
Figure 9. Effect of Varying the Magnitude of Impulse Change to the Impulse Response of a
First Order System.
KP
A
2
s s 1
2
(76)
(77)
Taking the inverse Laplace Transform of the system, the response of the first order system to
sinusoidal input in time domain is given by
Ae t /
A
A
y(t ) 2 2
2 2
cos t 2 2
sin t
1 1
1
(78)
This equation can be written in another form using the trigonometric identity
p cos A q sin A r sin( A )
where r
p 2 q 2 and tan
p
.
q
28
(79)
(80)
The effect of different values of amplitude A is shown in Figure 10 and the effect of different
values of frequency is shown in Figure 11.
As the time approaches infinity, the first term of the right side of the output equation (80)
vanishes and leaves only the ultimate periodic steady state solution
A
(81)
y(t ) t y(t ) s 2 2
sin(t )
1
Comparing the input function and the ultimate periodic steady state response of the system
shows that
a) The output is a sine wave with the same frequency as the input signal.
1
b) The ratio of the output amplitude to the input amplitude is 2 2
. This is always
1
smaller than 1, thus the output signal is attenuated. The comparison between input and
output waves is shown in Figure 12.
c) The output lags behind the input signal by an angle . If 0 then there is a phase
lag. If 0 then there is phase lead. Lag always occurs in a first order system since
the value of is always negative.
2.5
Increasing A
1.5
1
0.5
y(t)/Kp
0
-0.5
-1
-1.5
-2
0
10
12
Time (sec)
Figure 10. Effect of Varying the Amplitude of the Sinusoidal Input Function to the Response of a First
Order System.
29
0.8
Decreasing
0.6
0.5pi
0.4
pi
y(t)/AKp
0.2
1.5pi
0
-0.2
-0.4
-0.6
0
10
Time (sec)
Figure 11. Effect of Varying the Radian Frequency of the Sinusoidal Input to the Response of a First
Order System.
1.5
input wave
1
phase lag
0.5
output wave
y(t)/AKp
-0.5
-1
-1.5
10
Time (sec)
Figure 12. Comparison of the Input Sinusoidal Curve to the Output Sinusoidal Curve of the First Order
System.
30
q1
qo
h1 , A1
R1
h2 , A2
q2
R2
In the figure qs are flow rates, hs are the heights of the tanks, As are the cross sectional
areas of the tanks and Rs are the valve resistances.
Applying overall mass balance of Tank 1 yields,
dm1 d V1 d A1h1
qo q1
dt
dt
dt
(82)
1 dm dV1 dA1h1
qo q1
dt
dt
dt
(83)
If the tank height is taken as the state variable and tank cross sectional area is constant, the
following state equation can be obtained
A1dh1
qo q1
dt
(84)
For linear valves, the outlet flow is related to the tank height, so for tank 1, the outlet flow rate
is
h
(85)
q1 1
R1
The state equation will become
A1dh1
h
qo 1
dt
R1
(86)
This can be further simplified to the standard form of the first order system state equation
A1R1
dh1
h1 R1qo
dt
(87)
Comparing equation K-87 with the standard equation for the first order system, it can be seen
that the time constant is 1 A1 R1 and the process gain K p1 R1 , the state variable (output
variable) is the height in tank l, h1 and the input variable is q0 , the inlet flow rate to tank 1.
The state equation can now be written as
31
At steady state,
dh1
h1 K p1qo
dt
(88)
dh1
0 , h1 h1s and q0 q0s . Applying the steady state condition yields
dt
h1s K p1qos
(89)
Subtracting the steady state equation from the unsteady steady state equation yields
dh1
h1 h1s K p1 (qo qos )
dt
(90)
dh1 dh1
. These are called
dt
dt
deviation variables. The advantage of using the deviation variables is that at t 0 , the
deviation is zero. The state equation in terms of the deviation variables is
If we let h1 h1 h1s and qo qo q0 s , it can be seen that
d h1
h1 K p1qo
dt
(91)
(92)
.
Simplifying the equation yields]
G1 ( s)
K p1
h1 ( s)
q o ( s ) 1s 1
(93)
This the transfer function relating the inlet flow rate to tank 1 to the height of liquid level in
the tank.
Applying overall mass balance to Tank 2 yields
dm2 d V2 d A2 h2
q1 q2
dt
dt
dt
(94)
dh2 h1 h2
dt
R1 R2
(95)
A2 R2
dh2
R
h2 2 h1
dt
R1
(96)
Comparing equation 96 with the standard form, it can be seen that the time constant is
R
2 A2 R2 and the process gain is K p 2 2 . The standard equation becomes
R1
32
dh2
h2 K p 2 h1
dt
(97)
Applying the deviation variables and getting the Laplace transforms yields
dh2
h2 K p 2 h1
dt
(98)
and
G2 ( s )
K p2
h2 ( s)
h1 ( s) 2 s 1
(99)
This is the transfer function relating the output of tank 2 (h2) to the input of tank 2 (h1).
The height of tank 2 can be taken using h2 ( s) G2 ( s)h1 ( s)
K p2
2s 1
h1 ( s) .
K p2
h2 ( s) G2 ( s)h1 ( s) G2 ( s)G1 ( s) q o ( s)
K p1
2 s 1 1s 1
q o ( s)
(100)
.
The block diagram representation of the system is shown in Figure 14.
q o ( s)
G1 ( s)
h1 ( s)
K p1
G2 ( s)
1s 1
h2 ( s)
K p2
(101)
2s 1
K p1 K p 2
h2 ( s)
G1 ( s)G2 ( s)
1s 1 2 s 1
q o ( s)
(102)
It can be seen that for two non-interacting first order systems in series, the resulting overall
transfer function is second order. In general, for a system of n first order systems is series
shown in Figure 15, the overall transfer function is nth order as shown in the equation below.
Yn ( s)
G1 ( s)G2 ( s)...Gn1 ( s)Gn ( s)
X o ( s)
X o (s)
G1 ( s)
Y1 ( s)
G2 ( s )
Y2 ( s)
Yn2 ( s)
Gn1 ( s)
Yn1 ( s)
(103)
Gn (s)
Yn (s)
Second order systems are systems whose dynamics are governed by second order differential
equations. Although most of the systems especially those who are very complex have
dynamics that are higher than second order, the systems of order higher than two can be
transformed into a combination of first and second order systems. The overall response of the
higher order systems to changes in the input is the sum of all the first and second order
responses. Thus, it is very important to understand the dynamic behavior of the second order.
This program automatically generates the response of the second order system to step change
in the input. Second order characteristic parameters can be varied to see the effect of these
parameters to the response of the system. This can be used as the starting guide for process
control software development.
d2y
dy
2
y K p x(t )
2
dt
dt
2
(104)
Taking the Laplace transform of the differential equation, the transfer function of the second
order system is
G(s)
K
y (s)
2 2 P
x( s) s 2 s 1
(105)
0
The step input function is given by x(t )
A
t0
,
t0
where A is the magnitude of step change. Taking the Laplace Transform, the input in the s
domain is
A
x( s )
(106)
s
34
KP
A
s 2 s 1 s
2
(107)
Getting the inverse Laplace Transform of equation 107 yields the response in the time
domain.
There are three cases to be considered in taking the inverse Laplace transform of the second
order systems. The form of function depends upon the value of the damping factor. The
damping factor depends upon the roots of the characteristics polynomial of the second order
function
2 s 2 2 s 1 0
(108)
2 1
p2
2 1
(109)
t
y (t ) AK p 1 e t / (cosh 2 1
sinh 2 1
2 1
(110)
t
y (t ) AK p 1 1 e t /
(111)
1
y (t ) AK p 1
e t / sin t
1 2
1 2
1 2
tan 1
The response of the second order system to step change input is shown in Figure 16.
35
(112)
(113)
(114)
1.6
1.4
1.2
y(t)/AKp
0.8
0.6
0.4
0.2
0
t/
10
Figure 16. Effect of Varying the Damping Factor to the Response of a Second
Order System to Step Change in the Input.
A
exp
B
1 2
(115)
b. Decay Ratio. Decay ratio is the ratio of the sizes of the successive peaks. This computed
using the formula
C
exp 2
A
A
1 2
B
(116)
36
e. Period of oscillations (T). The period of oscillations is the time between two peaks. The
radian frequency (radians/time) is the coefficient of t in the sine term. It is related to and
by the formula,
1 2
(117)
1
1
T 2 2
1 2
(118)
f. Natural period of oscillations. If the damping factor approaches zero, the system oscillates
continuously without attenuation. The frequency for sustained oscillation is called the natural
1
frequency. This frequency is given by the formula n and the corresponding cyclical
frequency is f n
expression
1
1
. The natural frequency is related to the actual frequency by the
Tn 2
f
1 2 .
fn
1.6
Period, T
1.4
Response time
limit
A
1.2
y(t)/AKp
0.8
0.6
T
B
0.4
0.2
Response time
Rise time, Tr
0
10
Time (sec)
37
12
14
16
18
20
0.3
0.2
0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
The characteristics of the step response of an underdamped system are shown in Figure K18.
2 1
t
e t / sinh
2
(120)
(121)
(122)
The response of the second order system to a unit impulse input function is shown in Fig 19.
38
0.2
0.8
0.4
0.6
0.6
0.8
0.4
y(t)/AKp
1.0
0.2
1.2
1.4
-0.2
-0.4
0
5
6
7
8
9
10
t/
Figure 19. Effect of the Damping Factor to the Response of a Second Order System to Impulse Change in the Input
Function.
Another way of deriving the response of the system to impulse change is by obtaining it from
the response of the system to the step changes. It can be shown that the transform of the
output to impulse response is related to the transform of the output to the step response by the
formula
y ( s) impulse sy (s) step
(123)
Using a theorem of Laplace transforms yields the relationship between the time domain
response of the system to the impulse change in the input and the time domain response of the
system to the step change in the input as shown in the equation
y(t ) impulse
d
y(t ) step
dt
(124)
Thus, the response of the system to impulse change is obtained from the first derivative of the
response of the system to step changes.
To summarize the nature of the response of the system to impulse change, it can be seen in
Figure 16 that the general behavior is the same as the step response curves except that the
impulse response always returns to zero while the step change response settled to a new
steady state value.
Sinusoidal Response of Second Order System
The sinusoidal input function is x(t ) A sin t and the Laplace tranform of the function is
A
x ( s) 2
. Applying this transform of the input to the equation y (s) G(s) x (s) , yields
s 2
the transform of the response of the second order system to sinusoidal change,
39
y ( s)
KP
A
2
2 2
s s 2 s 1
(125)
In order to inverse the function, it is necesssary to factor the two terms of the denominator.
y ( s)
AK p / 2
KP
( s i )(s i ) ( s s1 )(s s 2 )
(126)
Here s1 and s2 are the roots of the modified denominator of the second order transfer function
1
s 2 2 s 2 =0. For the case of underdamped system, the roots of the denominator are a
pair of pure imaginary roots ( i , i contirbuted by the sinusoiodal input function and a
pair of complex conjugate roots of the main transfer function
2
2
i 1 , i 1
(127)
1 2
1 2
y (t ) C1 cos t C 2 sin t e t / C3 cos
t C 4 sin
(128)
(129)
Upon evaluating C1 and C2 and simplifying using trigonometric identitites, the response of the
second order system to a sinusoidal input function is ultimately sinusoidal and has the same
frequency as the input sinusoidal curve. The ultimate response of the system is given by
y (t )
A
[1 ( ) 2 ] 2 (2 ) 2
sin(t )
2
tan
1 ( ) 2
(130)
Figure 17 compares the response of the system to the input forcing function at differrent input
frequencies. If we compare the ultimate response of the system to the sinusoidal input
function, it can be seen that:
a) The ratio of the the output amplitude to the input amplitude is
1
[1 ( ) 2 ]2 (2 ) 2
(131)
This amplitude ratio may be less than or greater than 1 depending on the values of and .
It can be seen in Figure 20 that for sinusoidal input with frequency 0.25 , the amplitude
ratio is greater than 1 while for the input with frequency 0.75 , the amplitude ratio is
less than 1.
40
b) The output lags the input by phase angle . It can be shown later that approaches
180o assymtotically as the frequency approaches infinity. On the other hand, the phase lag
of the first order system can never exceed 90o.
Most systems are not truly second order but they exhibit similar step responses to second
order system, thus such systems are often approximated as second order in mathematical
analysis. Therefore, second order system is very important in control theory.
0.5
1 .25
2 .75
1.5
0.5
y(t)/AKp
-0.5
-1
Output 1
Input 1
-1.5
0
Input 2 Output 2
10
15
Time (sec)
Figure 20. Comparison of the Response of a Second Order System to Sinusoidal Input
Function at Various Radian Frequencies.
8. CLOSED-LOOP SYSTEMS
8.1. Introduction to Feedback Control Systems
This section covers the analysis and design of feedback controls systems. This closed loop
system represents the control schemes most commonly encountered in a chemical plant. The
emphasis of this section is in understanding the effects, which various feedback controllers
have on the response of controlled processes. This will also help the designer decides what is
the most appropriate controller selection.
the open loop systems. The feedback of a closed loop system permits the output to be
compared with the input to the system so that appropriate action may be formed as some
function of the output and input.
Consider the generalized process shown in Figure 21. This system has an output y; a potential
disturbance d and a variable can be a potential manipulated variable.
disturbance, d
manipulated
variable, m
Process
output variable, y
In this system the disturbance changes in an unpredictable manner and the control objective is
to keep the value of the output variable at desired level, which is the set point. A feedback
control system is installed and the generalized feedback control system is shown in Figure 22.
Controller mechanism
y sp
c
Controller
ym
Final
Control
Element
Process
Measuring
device
System.
d) Controller: this includes the function of the comparator. This is the unit that decides
the control action to be taken in order to reduce the error and make the output value
attain the setpoint value.
e) Final control element: this is usually a control valve or a variable speed-metering
pump. This device receives the control signal and physically adjusts the value of the
manipulated variable.
C. Types of Controllers
There are three basic types of feedback controllers: 1) proportional, 2) proportional-integral,
and 3) proportional-integral-derivative.
Proportional controller (P controller)
The control signal or the actuating output is proportional to the magnitude of the error:
c(t ) K c e(t ) cs
(132)
where c is the control signal, Kc = the proportional gain of the controller and cs = controller
bias signal. This is the signal from the controller if there is no deviation from the desired
output. This is to maintain the steady state operation of the system. The proportional
controller is characterized by the value of its proportional gain Kc or equivalently by this
proportional band PB. Proportional band PB = 100/Kc. The proportional band characterizes
the range over which the error must change in order to drive the actuating signal of the
controller over its full range.
The control signal at steady state, assuming the steady state error is zero, is given by
c() c s
Subtracting the steady state control signal to the actual actuating signal yield
c(t ) c(t ) c() c(t ) cs K c e(t )
(133)
(134)
For convenient let us drop the sign to denote the deviation variable.
Taking the Laplace Transform yields
c ( s) K c e ( s)
(135)
Gc ( s )
c ( s)
Kc
e ( s)
(136)
1 t
(137)
c(t ) K c e(t ) e(t )dt c s
0
where I is the integral time constant or reset time. This has usually a unit of minutes.
Converting the control signal to its deviation variable form yields
43
1
c(t ) c(t ) c s K c e(t )
I
e(t )dt
(138)
Taking Laplace Transform and dropping the for the deviation variable yields
1
c ( s) K c e ( s)
e ( s)
Is
(139)
Gc ( s)
c ( s)
1
K c 1
e ( s)
s
I
(140)
The reset time I of the PI controller is an adjustable parameter and it is sometimes referred
as time per repeat. Some manufacturers calibrate their controller in terms of the reciprocal of
the reset time 1 . This is known as the reset rate.
Kc
e(t )dt
e I
(141)
This means that the control action has repeated the response of the proportional action after
this period. This repetition takes place every I minute. This is the reason why this term is
called reset time.
Reset time is the time needed to repeat the initial proportional action change in its output.
The integral action causes the controller output c(t ) to change as long as an error exists in the
process output. Such controllers can eliminate even smaller errors.
Proportional-Integral-Derivative controller (PID controller)
This is also known as proportional-plus-reset-plus-rate controller.
Its actuating signal output is given by the equation
1
c(t ) K c e(t )
I
e(t )dt
0
de(t )
cs
dt
(142)
1
c(t ) c(t ) c s ) K c e(t )
I
e(t )dt D
de(t )
dt
Taking the Laplace Transform and dropping the for the deviation variables yields
44
(143)
1
c ( s) K c e ( s)
e ( s) D se ( s)
Is
(144)
Gc ( s)
c ( s)
1
K c 1
D s
e ( s)
Is
(145)
The derivative control action is sometimes referred as anticipatory control because the
presence of the derivative term in the PID controller enables the controller to anticipate what
the error will be in the immediate future.
The major disadvantage of the derivative control action is that: 1) for a response with constant
non zero error it gives no control action since de dt 0 , and 2) for a noisy response with
almost zero error, it can compute large derivatives giving large control action although it is
not necessary.
d (s)
Gd (s)
y sp (s)
(s)
Gc (s)
y m (s)
c(s)
G f (s)
m(s)
G p (s)
y(s)
Gm (s)
System.
From Figure 23, we can get the corresponding output of the block from its input and transfer
function. In particular we can get:
For the process we can get:
(146)
(147)
(148)
c(s) GC (s)(s)
(149)
m( s) G f ( s)c( s)
(150)
where GP , Gd , Gm , GC , and G f are the transfer functions between the corresponding inputs
and outputs.
Algebraic manipulations of the equation will yield
(151)
(152)
(153)
Putting the last expression into y(s) GP (s)m(s) Gd (s)d (s) yields
(154)
y( s)
GP ( s)G f ( s)GC ( s)
1 GP ( s)G f ( s)GC ( s)Gm ( s)
y SP ( s)
Gd ( s)
d ( s)
1 GP ( s)G f ( s)GC ( s)Gm ( s)
(155)
This is the closed loop response of the process. The first terms show the effect on the output
of a change in the setpoint while the second terms constitutes the effect on the output of a
change in the disturbances (load).
From this equation we get
GP G f GC
1 GP G f GC Gm
G
GSP
1 GGm
the transfer function relating the output variable to the set point and
46
(156)
Gd
Gd
Gload
1 GP G f GC Gm 1 GGm
(157)
(158)
(159)
(160)
Putting the corresponding transfer function to this formula yields the transfer function relating
the output to the setpoint
GP G f GC
y ( s)
G
GSP
y sp ( s) 1 GP G f GC Gm 1 GGm
(161)
We can also get the relationship between the output and the disturbance using the same
process.
The forward path transfer function for disturbance is Gd (s) , and the feedback loop path
transfer function is the same as above. Therefore the transfer function relating the output to
the disturbance using the formula presented above is
Gd
Gd
y ( s)
Gload
d ( s) 1 GP G f GC Gm 1 GGm
(162)
Gd (s)
(s)
Gc (s)
y m (s)
c(s)
G f (s)
m(s)
G p (s)
Gm (s)
47
y(s)
y s p (s)
GSP (s)
y (s)
(163)
For every feedback control system we can distinguish two types of control system problems:
the servo problem and the regulator problem. This will be discussed below:
Servo Problem
In servo problem the disturbance does not change while the setpoint changes. The feedback
controller acts in such a way that the output y will follow the changing setpoint ysp as close as
possible. For this problem the relationship between the setpoint and the output is given by
(164)
Regulator Problem
In this problem the set point remains the same while the load changes. The feedback
controller suppresses the error caused by changing disturbances and forcing the value of the
output back to the desired setpoint level. In this case the relationship between the disturbance
and the output is
y (s) Gload (s)d (s)
(165)
Gload (s) is the overall regulator transfer function.
48
dy
y K pm Kd d
dt
(166)
Kd
KP
m( s)
d ( s)
P s 1
P s 1
(167)
d(s)
Gd ( s )
ysp(s)
G p (s)
Gc ( s) K c
Kd
ps 1
Kp
y(s)
ps 1
For the uncontrolled system we got the original time constant p , and original static gains Kp
for the manipulated variable and Kd for the disturbance.
The close loop response of the system is
y ( s)
Gc G p
1 Gc G p
y s p ( s)
Gd
d ( s)
1 Gc G p
(168)
y( s)
K P Kc
Kd
y SP ( s)
d ( s)
p s 1 K p Kc
p s 1 K p Kc
(169)
Rearranging yields
where
'P
K 'P
K 'd
y ( s) '
y ( s) '
d ( s)
P s 1 SP
P s 1
(170)
1 K P KC
K P KC
K 'P
1 K P KC
Kd
K 'd
1 K P KC
The parameters Kp and Kd are called closed loop static gains.
Let us consider a unit step change in the set point and examine the resulting closed loop
responses. For a unit step change in the set point with constant load (servo problem), we get
the expression for the output
y( s)
49
K 'P 1
'P s 1 s
(171)
y(t ) K ' p (1 e
t ' p
(172)
As time approaches infinity, the new steady state value approaches y() K ' p . For a unit
step change in the setpoint, the desired steady state value for the output is y() y sp () 1.
We noticed that there is an offset, which is equal to
offset
1 K p Kc 1 K p K c
(173)
K 'd 1
'Ps 1 s
(174)
y(t ) K ' d (1 e
The ultimate value of the response is
t ' p
(175)
y() K ' d .
Applying the equation for the offset, with setpoint for regulator problem equals 0, we get
Offset = 0 K ' d 0
Kd
Kd
1 K p Kc 1 K p K c
(176)
It can be seen that the offset will approach zero when K c approaches infinity.
The effect of increasing Kc on the closed loop response of the system is shown in Figure 27.
The effects of the proportional controller on the response of a first order system are:
a) The system remains first order with respect to load and set point changes.
b) The time constant has reduced ( ' p p ) indicating that the closed loop response is
faster than the uncontrolled system.
c) The static gains decreased.
d) There is an offset for both regulator and servo problems.
e) Although generally the offset approaches zero as the controller gain is increased,
extremely large values of the gain cannot be used for proportional control especially
for higher order system due to problems of instability.
50
1
0.9
0.8
Offset
0.7
Increasing Kc
0.6
y(t) 0.5
0.4
0.3
0.2
0.1
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Time (sec)
Figure 27. Effect of Increasing the Controller Gain Kc on the Closed Loop Response
of the First Order System.
y( s)
K 'P
G( s)
2 2
'
y SP ( s) ( ) s 2 s 1
(177)
The simplified block diagram for the closed loop system is shown in Figure 28.
ysp(s)
Gc ( s) K c
G p ( s)
Kp
y(s)
p 2 s 2 2 s 1
Gcl
Gc G p
y ( s)
y sp ( s) 1 Gc G p
(178)
y( s)
where
'
K 'P
y SP ( s)
( ' ) 2 s 2 2 ' ' s 1
1 K P KC
51
(179)
'
K 'P
1 K P KC
K P KC
1 K P Kc
Let us consider a unit step change in the set point, then the transform of the output becomes
y( s)
K 'P
1
' 2 2
'
( ) s 2 ' s 1 s
(180)
K p Kc
(181)
1 K p Kc
Applying the equation for offset with new set pint equals 1 for servo problem with unit step
change in the setpoint yields
offset = 1
K p Kc
(182)
K p Kc
Kp 1
Kc = 41
1.4
1
p 1
1.2
y(t)
0.8
0.6
Increasing Kc
0.4
Kc = 1
0.2
Time (sec)
52
Kc
Gc
Is
G p ( s)
Kp
y(s)
ps 1
Figure 30. Simplified Closed Loop Block Diagram for the First
Order System Controlled by Integral Controller System.
GP G f GC
y( s)
y SP ( s)
(183)
1 GP G f GC Gm
To simplify the system, both measuring device and final control element transfer functions are
taken as unity. Putting the expression for the transfer functions into the output equations
yields
K P
1
K c
s
I
y ( s) P
y SP ( s )
(184)
K P
1
K C
s
I
P
1
y SP ( s)
s 2 s 1
2
(185)
where
I P
K P KC
I
1
2 P K P KC
It can be seen that for first order uncontrolled process, the response of the closed loop
becomes second order. Since the order is increased the response will be slower compared to
the original system. Thus, integral action alone tends to make the response of the control
system slower.
Let us consider a unit step change in the set point for the integral controlled system at hand.
The equation for the output becomes
y ( s)
1
1
s 2 s 1 s
2
53
(186)
(187)
Kp 1
Kc = 5
p 1
I 1
y
(
t 0.5
)
Kc = 4
Kc = 3
Kc = 2
Kc = 1
10
12
Time (sec)
Figure 31. Effect of Change the Controller Gain Kc on the Closed Loop Response of the
First Order System Controlled by Integral Controller.
1.4
1.2
Kp 1
0.8
Kc 1
y(t)
p 1
0.6
Increasing
I 1
0.4
0.2
10
12
14
16
18
20
Time (sec)
54
G p ( s)
Gc K c d s
y(s)
Kp
ps 1
Figure K 33. Simplified Closed Loop Block Diagram for the First Order System
Controlled by Derivative Controller System.
The closed loop response of a first order system with derivative action is
KP
K C D s
P s 1
y( s)
y SP ( s)
KP
1
K C D s
P s 1
(188)
Simplifying yields
y( s)
K P K C D s
y ( s)
P K P K C D s 1 SP
(189)
From these equations, we can observe the following effects of the derivative action on the
closed loop response of a first order system:
a) The order of the response remains the same.
b) The effective time constant of the closed lop response is larger than the original
uncontrolled system. This implies that the derivative control action slows down the
response of the process. Increasing the value of Kc further make the response slower.
Second order system
The block diagram of a second order system controlled by derivative control action only is
shown in Figure 34. For simplicity, both the valve and measuring element transfer function
were taken as unity.
ysp(s)
Gc K c d s
G p ( s)
Kp
y(s)
p 2 s 2 2 1
Figure 34. Simplified Closed Loop Block Diagram for the Second Order System Controlled
by Derivative Controller System.
The closed loop response of a second order system with derivative action is
y( s)
K P K C D s
y SP ( s)
s 2 K P K C D s 1
2 2
(190)
From this equation, the following effects of the derivative action on the closed loop response
of a second order system can be observed:
55
a) The natural period of the closed loop response is the same as the period of the
uncontrolled response.
b) The effective damping factor is larger than the original damping factor. This implies
that the closed loop response is more damped. Increasing the Kc further increases the
damping factor. This increased damping factor produces a more robust response.
c) The derivative control adds more robustness and stability to the system. This
advantage is offset by a slower closed loop response.
56
with a 0 .
The graph of the function is shown in Figure 35. If a 0 this function is unbounded and the
graph of the input is shown in Figure 36.
an
dny
d n1 y
dy
a
... a1
ao y KF (t )
n 1
n
n 1
dt
dt
dt
(191)
The solution of this differential equation depends upon the nature of the characteristic
equation, which can be obtained from the left hand side of the equation, and the nature of the
forcing function F (t ) .
Bounded Input
Unounded Input
1.5
3
2.5
2
1.5
F(t)
F(t)
1
0.5
0
0.5
-0.5
-1
-1.5
10
15
20
25
10
t
t
Consider a linear system
defined by a differential equation
Figure 35
Figure
36 the characteristics
For bounded input, the stability of the response of the system depends
upon
solution y c . This characteristic solution depends upon the roots of the characteristic equation
For the equation with n finite and distinct roots, the characteristics solution
yc c1e r1t c2 e r2t ..cn e rn1t .
(192)
(193)
(194)
If the set of roots contains a least one pair of complex roots r a bi the characteristic
equation will have the form
yc e at (c1 sin bt c2 cos bt ) y3 y 4 .. y n
(195)
where y n are linearly independent functions representing the remaining roots. These
functions are either of the form of equation (193) or (194) depending upon the nature of the
remaining roots.
The response of the system depends upon signs of the roots. If at least one root is positive, the
term represented by this root will be infinity as time approaches infinity. This root will cause
unbounded response even for bounded input, thus the system is unstable. For a pair of
complex roots, the sinusoidal response will have a decreasing amplitude if the sign of the real
part is negative otherwise the amplitude will be increasing.
57
In general, the system will be stable if all the roots of the characteristic equation lie on the left
hand side of the imaginary plane. The region on the imagery plane where the roots are stable
is shown in Figure 37. If at least one root lies on the right hand side of the imaginary plane,
then the system is unstable. Examples of stable roots r 1 2i , r 3 these roots lie on the
left hand side of Figure 38. Examples of unstable roots are r 1 i and r 4 , they lie on the
right hand side of the imaginary plane as shown in Figure 38.
Imaginary Axis
Stable roots
Imaginary Axis
Real Axis
2
1
0
-4
-2
-1
-2
-3
Real Axis
Figure 37
Figure 38
(196 )
y ( s)
K
n
F ( s) a n s a n1 s n1 ..a1 s ao
(197)
an s n an1 s n1 ..a1 s ao 0
(198)
Equation 198 is of the same form to the characteristic equation shown in Equation (192), thus
the denominator of the transfer function of the system can be taken as the characteristic
equation of the system.
Stability of the closed loop response
Consider the standard block diagram of a feedback control system
Gd (s)
y sp (s )
(s )
Gc (s)
c(s )
y m (s )
G f (s)
m(s )
G p (s)
Gm (s)
y (s )
1
2s 1
1
(199) Gm
s 1
Gc K c
G p Gv
(198 ) Gv
1
2s 1
(200)
(201)
It can be shown that the closed-loop system produces unstable responses if Kc is too large.
The closed loop transfer function for set-point changes is
Gservo
Gc G f G p
y ( s)
d ( s ) 1 Gc G f G p G m
(202)
1
substituting with y sp s , step four
s
K c ( s 1)
1
y s .
3
s 10s 17 s 2 2s 1 K c
(203)
If Kc is specified, y t can be determined from the inverse Laplace transform with the roots
of the polynomial by partial fractions. With this procedure, the response may be plotted for
various Kc as shown in Figure 40.
Step Response
3
2.5
Kc = 15
Kc = 6
1.5
y(t)
1
0.5
0
Kc = 2
-0.5
-1
-1.5
10
12
14
16
18
20
time
Figure 40
Gregulator
Gd
y ( s)
d ( s ) 1 Gc G f G p G m
(204)
Take note that the denominator of the servo and regulator transfer functions are the same.
This function is called the characteristic equation of the system
1 Gc G f G p Gm 0
(205)
The roots of this characteristics equation determine the stability of the system. As discussed
above, of the roots of this equation determines the stability of the system. If all roots of this
equation lie on the left hand side of the imaginary plane, then the system is stable. Otherwise,
if at least one root lies on the right hand side of the imaginary plane then the system is
unstable,
59
Gc G f G p
1 GOL
y sp
Gd
d
1 GOS
(206)
y sp
1 GOL
and for disturbance changes
Gd
y
d 1 GOL
(207)
(208)
Simplifying the set point change equation and factoring into poles ( Pi ) and zeroes ( Z i )
s z1 s z 2 ....s z m
y
K'
s pi s p2 ....s pm
y sp
(209)
'
1 GOL 0
(210)
1
S
K ' s z1 s z 2 ....s z m
.
s s pi s p 2 ....s p m
(211)
A0
An
A1
A2
.......
s s p1 s p 2
s pn
Taking the inverse Laplace transform
y s
(212)
(213)
If one of the poles is a positive real number, Pn 0, then C (t ) is unbounded thus the close
loop system is unstable.
If Pk is a complex number, Pk ak jbk with a k 0, the system is unstable.
General stability Criterion A feedback control system is stable if and only if all roots of the
characteristic equation are negative or have negative real parts. Otherwise the system is
unstable.
60
an s n an1 s n1 .... a1 s a0 0
(214)
with a n 0.
A necessary (but not sufficient) conditions for stability is that an, an-1,..ao must be positive.
If any coefficient is negative or zero, the system is unstable.
If all coefficients are positive, construct the Routh Array to determine if the system is stable.
Row
1
2
3
4
.
.
n+1
where:
an1an2 an an3
,
an1
b a an1b2
c1 1 n3
,
b1
b1
an
an-1
b1
c1
.
.
z1
an-2 a-4 ..
an-3 an-5 .
b2 b3 ..
c2 ..
an1an2 an an3
an1
b a an1b3
c 2 1 n 3
b1
b2
Routh-Hurwitz Stability Criterion A necessary and sufficient condition for all roots of the
characteristic equation to have negative real parts is that all the elements in the leftmost
column of the Routh Array are positive.
If any of the elements of the first column is negative, then there is at least one root to the right
of the imaginary axis. The number of sign changes in the elements of the first column is equal
to the number of roots in the right side of the imaginary axis.
Characteristics of Routh Array
1) There are n+1 rows in the Routh Array.
2) The elements in the n-1 th row determine the value of the gain that will give a root that
will lie on the imaginary axis. This will be discussed in the next section.
3) The Routh stability criterion is valid only if the characteristic equation is algebraic and
all the coefficients are real, if any one of the coefficients of the characteristic equation
is a complex number or if the equation contains exponential function of s, such as the
case of a system with time delays, this criterion cannot be applied.
4) Routh Criterion is generally used only to determine the absolute stability of the
61
k
0 or (s 1)(s 2)(s 3) k 0
( s 1)( s 2)( s 3)
or
(215)
(216)
s 3 6s 2 11s (k 6) 0
For every k value, there is a corresponding set of 3 roots of the characteristic polynomial. If
we vary the value of k from zero to infinity, we can obtain 3 loci of roots when plotted on the
imaginary plane. In the example above, if we vary the value of k from 0 to 60 increment of 5,
we can obtain the roots of the characteristic equation for every k value as shown in Table 1.
Plotting all these points in the imaginary plane yields the root locus in Figure 41
Table 1
K
Root 1
Root 2
Root 3
-3
-2
-1
-3.9042
-1.0479 - 1.3112i
-1.0479 + 1.3112i
10
-4.3089
-0.8455 - 1.7316i
-0.8455 + 1.7316i
15
-4.6012
-0.6994 - 2.0186i
-0.6994 + 2.0186i
20
-4.8371
-0.5814 - 2.2443i
-0.5814 + 2.2443i
25
-5.038
-0.4810 - 2.4335i
-0.4810 + 2.4335i
30
-5.2145
-0.3928 - 2.5980i
-0.3928 + 2.5980i
35
-5.3729
-0.3135 - 2.7445i
-0.3135 + 2.7445i
40
-5.5174
-0.2413 - 2.8773i
-0.2413 + 2.8773i
45
-5.6506
-0.1747 - 2.9992i
-0.1747 + 2.9992i
50
-5.774
-0.1128 - 3.1121i
-0.1128 + 3.1121i
55
-5.8906
-0.0547 - 3.2175i
-0.0547 + 3.2175i
60
-6
-0.0000 - 3.3166i
-0.0000 + 3.3166i
62
Imaginary axis
3
2
1
st
Locus of 1 root
0
-7
-5
-3
-1
-1
-2
-3
-4
Real axis
Figure 41
N
D
(217)
The overall open loop transfer function can be written in terms of the numerator part N(s) and
denominator part D(s).
The characteristic equation is
1 GOL 0 or k
N
1
D
(218)
D
k
(219)
Since k approaches infinity, the right hand portion will vanish to zero and the characteristic
polynomial will become. N 0 . This means that at k equals infinity, the roots of the
characteristic equation are the roots of the numerator of the overall open loop transfer
function. The roots of the numeration of the open loop transfer function are called the zeroes
of the system. We can deduce that the root locus ends at the open loop zeroes when k
approaches infinity.
In terms of poles and zeroes the characteristic equation of the set can be written as
63
( s z1 )(s z 2 ).......(s z m )
1 .
( s p1 )(s p 2 )......(s p n )
(220)
The left hand size of the equation is generally complex. This equation can be written in terms
of magnitude and phase angle of the individual term.
s z1 s z 2 ........ s z m
s p1 s p 2 ........ s p n
(221)
where i is any integer or zero. The two equations above are called the magnitude and angle
criterion.
Suppose a system has an open loop transfer function
k ( s z1 )
N
D ( s p1 )(s p 2 )
(222)
Assume the s c is point on the imaginary plane lying on the root locus, this point is part of the
root locus if it follows the magnitude and angle criteria.
The point s c and the zeroes and poles of the system are plotted in Figure 42.
j
sc
sc z1
o 1
sc p1
sc p2 2
p1
z1
p2
Figure 42. Use of the angle criterion to establish the root locus.
1 1 2 (2i 1)
(223)
is followed for some value of i. The value of K corresponding to this point can be solved
using the magnitude criterion,
K sc z1
sc p1 s c p 2
(224)
sc p1 sc p 2
sc z1
(225)
( s z1 )(s z 2 ).......(s z m )
1
( s p1 )(s p 2 )......(s p n )
(226)
the following rules will be used to plot the root locus of the system.
Rule 1 The number of loci or branches is equal to the number of open loop poles, n.
Rule 2 The root loci begin at open loop poles and terminate at open loop zeros. The
termination of ( n m ) of the loci will occur at the zeros at infinity along the asymptotes. For
qth order pole q loci emerge from it and for qth order zero, q loci terminate on it.
Rule 3 Locus on real axis. The real axis is part of the root locus when the sum of the number
of poles and zeroes to the right of a point on the real axis is odd. A qth order pole (or zero)
must be counted q times in applying this rule.
Rule 4 Asymptotes. There are n m loci that approach asymptotically n m straight lines,
radiating from the center of gravity of the poles and zeroes of the open loop transfer function.
The center of asymptotes is given by
n
p z
j 1
i 1
(227)
nm
The asymptotic lines make angles of (2k 1) /( n m) with the real axis, with k = 0,1,2..nm-1.
Rule 5 Breakaway point. The point at which two root loci, emerging from adjacent poles on
the real axis, intersect and then leave the real axis. This point is determined by the solution of
the equation
m
n
1
1
i 1 s z i
j 1 s p i
(228)
If a pair of complex poles pi ai jbi add to the right side of the equation
2( s ai )
( s ai ) 2 bi
(229)
Rule 6 Angle of Departure or Approach. There are q loci emerging from each qth order open
loop pole at angles determined by
m
n
1
2k 1 p a z i p a p j
q
i 1
j 1
j a
k 0, 1, 2,.......q 1
(230)
n
m
1
2k 1 z b p j z b z i
v
i 1
i 1
i b
k 0, 1, 2,.......v 1
(231)
Deciphering symbols
ISA S5.1 defines four graphical elements-discrete instruments, shared control/display,
computer function, and programmable logic controller-and groups them into three location
categories (primary location, auxiliary location, and field mounted).
Discrete instruments are indicated by circular elements. Shared control/display elements are
circles surrounded by a square. Computer functions are indicted by a hexagon and
programmable logic controller (PLC) functions are shown as a triangle inside a square.
Adding a single horizontal bar across any of the four graphical elements indicates the function
resides in the primary location category. A double line indicates an auxiliary location, and no
line places the device or function in the field. Devices located behind a panel-board in some
other inaccessible location are shown with a dashed horizontal line
66
Letter and number combinations appear inside each graphical element and letter combinations
are defined by the ISA standard. Numbers are user assigned and schemes vary with some
companies use of sequential numbering, others tie the instrument number to the process line
number, and still others adopt unique and sometimes unusual numbering systems.
The first letter defines the measured or initiating variables such as Analysis (A), Flow (F),
Temperature (T), etc. with succeeding letters defining readout, passive, or output functions
such as Indicator (I), Record (R), Transmit (T), and so forth.
Example shows the story
Referring to the Example P&ID diagram, FT 101 represents a field-mounted flow transmitter
connected via electrical signals (dotted line) to flow indicating controller FIC 101 located in a
shared control/display device. A square root extraction of the input signal is applied as part of
FIC 101's functionality. The output of FIC 101 is an electrical signal to TY 101 located in an
inaccessible or behind-the-panel-board location. The output signal from TY 101 is a
pneumatic signal (line with double forward slash marks) making TY 101 an I/P (current to
pneumatic transducer). TT 101 and TIC 101 are similar to FT 101 and FIC 101 but are
measuring, indicating, and controlling temperature. TIC 101's output is connected via an
internal software or data link (line with bubbles) to the setpoint (SP) of FIC 101 to form a
cascade control strategy.
Often P&ID's include a cover page where common and typical terms, symbols, numbering
systems, etc., are defined. On the example, Typical YIC would likely appear on the cover
page and the simplified form of YIC would appear throughout the P&IDs.
Typical YIC indicates an on/off valve is controlled by a solenoid valve and is fitted with limit
switches to indicate open (ZSH) and closed (ZSL) positions. All inputs and outputs are wired
to a PLC that's accessible to the operator (diamond in a square with a solid horizontal line).
The letter 'Y' indicates an event, state, or presence. The letter 'I' depicts indication is provided,
and the letter 'C' means control takes place in this device.
Adherence to ISA's S5.1 Instrumentation Symbols and Identification standard ensures a
consistent, system independent means of communicating instrumentation, control, and
automation intent is developed for everyone to understand.
General instrument or function symbols
Primary location
accessible to
operator
Field mounted
Auxiliary location
accessible to
operator
Discrete
instruments
Shared display,
shared control
Computer
function
Programmible
logic control
1. Symbol size may vary according to the user's needs and the type of document.
2. Abbreviations of the user's choice may be used when necessary to specify location.
67
3. Inaccessible (behind the panel) devices may be depicted using the same symbol but with a
dashed horizontal bar.
Source: Control Engineering with data from ISA S5.1 standard
Identification letters
First letter
Measured or
initiating variable
Succeeding letters
Modifier
Readout or
passive function
A Analysis
Alarm
B Burner, combustion
User's choice
C User's choice
D User's choice
User's choice
Sensor (primary
element)
Ration
(fraction)
Glass, viewing
device
H Hand
High
Current (electrical)
Indication
J Power
Scan
Time rate of
change
L Level
M User's choice
Control station
Light
Middle,
intermediate
User's choice
O User's choice
Orifice, restriction
P Pressure, vacuum
Point (test
connection)
Safety
Switch
Transmit
U Multivariable
Multifunction
Vibration,
mechanical analysis
Multifunction
Well
X axis
Unclassified
Unclassified
Event, state, or
presence
Y axis
Relay, compute,
convert
Z Position, dimension
Z axis
Driver, actuator
Multifunction
Valve, damper,
louver
W Weight, force
X Unclassified
User's choice
Record
T Temperature
User's choice
Integrate,
totalizer
R Radiation
S Speed, frequency
Low
Momentary
N User's choice
Q Quantity
User's choice
Differential
G User's choice
Modifier
Control
E Voltage
F Flow rate
Output function
68
Unclassified
AE
AT
AI
AIC
AC
AFC
AR
Conductivity
CE
CT
CI
CIC
CC
CFC
CR
Density
DE
DT
DI
DIC
DC
DFC
DR
Voltage
EE
ET
EI
EIC
EC
EFC
ER
Flow
FE
FT
FI
FIC
FC
FFC
FR
Dimension
GE
GT
GI
GIC
GC
GFC
GR
Hand
HE
HT
HI
HIC
HC
HFC
HR
Current
IE
IT
II
IIC
IC
IFC
IR
Time
KE
KT
KI
KIC
KC
KFC
KR
Level
LE
LT
LI
LIC
LC
LFC
LR
Humidity
ME
MT
MI
MIC
MC
MFC
MR
Power
NE
NT
NI
NIC
NC
NFC
NR
Pressure
PE
PT
PI
PIC
PC
PFC
PR
Delta
Pressure
dP
dPE
dPT
dPI
dPIC
dPC
dPFC
dPR
Quantity
QE
QT
OI
OIC
QC
QFC
QR
Radioactivit
y
RE
RT
RI
RIC
RC
RFC
RR
Speed
SE
ST
SI
SIC
SC
SFC
SR
Temperature
TE
TT
TI
TIC
TC
TFC
TR
Delta
Temperature
dT
dTE
dTT
dTI
dTIC
dTC
dTFC
dTR
69
Viscosity
VE
VT
VI
VIC
VC
VFC
VR
Weight
WE
WT
WI
WIC
WC
WFC
WR
Vibration
YE
YT
YI
YIC
YC
YFC
YR
Position
ZE
ZT
ZI
ZIC
ZC
ZFC
ZR
Process
Measurement
HS
HV
IQ
XV
Analysis
AHS
AHV
AQ
AIQ
AXV
AY
AV
Conductivity
CHS
CHV
CQ
CIQ
CXV
CY
CV
Density
DHS
DHV
DQ
DIQ
DXV
DY
DV
Voltage
EHS
EHV
EQ
EIQ
EXV
EY
EV
Flow
FHS
FHV
FQ
FIQ
FXV
FY
FV
Dimension
GHS
GHV
GQ
GIQ
GXV
GY
GV
Hand
HHS
HHV
HQ
HIQ
HXV
HY
HV
Current
IHS
IHV
IQ
IIQ
IXV
IY
IV
Time
KHS
KHV
KQ
KIQ
KXV
KY
KV
Level
LHS
LHV
LQ
LIQ
LXV
LY
LV
Humidity
MHS
MHV
MQ
MIQ
MXV
MY
MV
Power
NHS
NHV
NQ
NIQ
NXV
NY
NV
Pressure
PHS
PHV
PQ
PIQ
PXV
PY
PV
Delta
Pressure
dP
dPHS dPHV
dPQ
dPIQ
dPXV
dPY
dPV
Quantity
QHS
QHV
QIQ
QXV
QY
QV
Radioactivity
RHS
RHV
RQ
RIQ
RXV
RY
RV
Speed
SHS
SHV
SQ
SIQ
SXV
SY
SV
Temperature
THS
THV
TQ
TIQ
TXV
TY
TV
Delta
Temperature
dT
dTHS dTHV
dTQ
dTIQ
dTXV
dTY
dTV
Viscosity
VHS
VHV
VQ
VIQ
VXV
VY
VV
Weight
WHS WHV
WQ
WIQ
WXV
WY
WV
Vibration
YHS
YHV
YQ
YIQ
YXV
YY
YV
Position
ZHS
ZHV
ZQ
ZIQ
ZXV
ZY
ZV
Process
Measurement
Element
Ratio
Switch Switch Alarm
Type Calculation Low
High
Low
Alarm
Low
Low
Alarm
High
Alarm
High
High
Code
FY
SL
SH
AL
ALL
AH
AHH
Analysis
AFY
ASL
ASH
AAL
AALL
AAH
AAHH
Conductivity
CFY
CSL
CSH
CAL
CALL
CAH
CAHH
Density
DFY
DSL
DSH
DAL
DALL
DAH
DAHH
70
Voltage
EFY
ESL
ESH
EAL
EALL
EAH
EAHH
Flow
FFY
FSL
FSH
FAL
FALL
FAH
FAHH
Dimension
GFY
GSL
GSH
GAL
GALL
GAH
GAHH
Hand
HFY
HSL
HSH
HAL
HALL
HAH
HAHH
Current
IFY
ISL
ISH
IAL
IALL
IAH
IAHH
Time
KFY
KSL
KSH
KAL
KALL
KAH
KAHH
Level
LFY
LSL
LSH
LAL
LALL
LAH
LAHH
Humidity
MFY
MSL
MSH
MAL
MALL
MAH
MAHH
Power
NFY
NSL
NSH
NAL
NALL
NAH
NAHH
Pressure
PFY
PSL
PSH
PAL
PALL
PAH
PAHH
Delta Pressure
dP
dPFY
dPSL
dPSH
dPAL
Quantity
QFY
QSL
QSH
QAL
QALL
QAH
QAHH
Radioactivity
RFY
RSL
RSH
RAL
RALL
RAH
RAHH
Speed
SFY
SSL
SSH
SAL
SALL
SAH
SAHH
Temperature
TFY
TSL
TSH
TAL
TALL
TAH
TAHH
Delta
Temperature
dT
dTFY
dTSL
dTSH
Viscosity
VFY
VSL
VSH
VAL
VALL
VAH
VAHH
Weight
WFY
WSL
WSH
WAL
WALL
WAH
WAHH
Vibration
YFY
YSL
YSH
YAL
YALL
YAH
YAHH
Position
ZFY
ZSL
ZSH
ZAL
ZALL
ZAH
ZAHH
71