Professional Documents
Culture Documents
P
C
G
A
T
E
C
H
3
8
© Career Avenues A1
IPC
Instumentation
and Process
Control
© Career Avenues A2
IPC
IPC
1. GENERAL CONCEPTS OF PROCESS CONTROL SYSTEMS
Elements of Process Control systems
2. LAPLACE TRANSFORMS
Dynamics systems
Review of Complex Variables and Complex Functions
Complex Functions
Review of Laplace Transforms
Important Properties of Laplace Transforms
Final Value Theorem, Initial Value Theorem
5. CONTROL ACTION
Introduction
Proportional Control
Derivative Control
Integral Control
Proportional Integral derivative Control
PID Controller: Theory and Explanation
© Career Avenues A3
IPC
6. STABILITY
Introduction
Poles, Zeros and system Response
Routh-Hurwitz Criterion
Generating a Basic Routh Table
Routh-Hurwitz Criterion: Special Cases
Routh-Hurwitz Criterion: additional Examples
Root Locus Techniques
The Root Locus Concept
Construction Rules
7. BODE PLOTS
Introduction
Plotting Frequency Response
Asymptotic Approximations: Bode Plots
Stability, Gain Margin and Phase margin Via Bode plots
8. ZN SETTINGS
Ziegler – Nichols Controller Settings
© Career Avenues A4
IPC
to suppress the influence of external disturbances, i.e. the effect of surroundings on the process;
to ensure the stability of a chemical process, i.e. to keep process parameters (variables) as close as
possible to their desired values;
to optimise the performance of a chemical process, i.e. to meet the requirements of safety,
satisfaction of production specification and maximising of economic objectives.
Let's consider two systems for the control of temperature and level of liquid in a stirred tank shown in
Fig. 1.1. These two control systems have all elements inherent to every control system, namely:
a process - the stirred tank;
controlled variables - temperature of the effluent liquid (or temperature of the liquid in the tank)
and level of the liquid in the tank;
a transducer, a measuring instrument, a transmitter;
a transmission line (electrical, pneumatic or hydraulic) that carries a measurement signal from a
measuring instrument (or transducer) to a controller;
a control signal from a controller to a final control element (for example, a control valve);
a controlling device or a controller;
a final control element - a control valve.
Fin , Fout , Fst - flowrates of inlet and outlet streams of the liquid, and steam, respectively;
Hs - a desired value (set point) of the liquid level in the stirred tank;
LT - a level transmitter;
© Career Avenues A5
IPC
TE - a thermocouple;
Tin , Tout , Tst - temperatures of inlet and outlet streams of the liquid, and steam,
respectively;
Controller
Desired
LT value Hs
CV
Desired
value Ts Condensate
Steam
Transmission lines
Fst, Tst
A process is an equipment or apparatus for which supply of energy, material, etc., and demand must
be balanced. We don't include in this concept control hardware. There are many types of processes:
from simple, such as our example or a tank for storage of liquid hydrocarbons, where liquid level
should be kept within an acceptable range, to complex processes such as distillation columns and
reactors.
A process variable is a physical or chemical property, quantity or other condition which can be
changed. Here are several examples of process variables: a flow rate of a feed stream into a distillation
column; a pressure of gas in a storage tank; a level in a condenser; a temperature of a liquid stream
(feed to a reactor) exiting from a furnace; pH of a solution in a stirred tank; density and viscosity of
liquid products of oil refining (these parameters are usually used as an indicator of desirable quality of
products); a resistance of an electrical circuit; a speed of rotor revolution, etc.
© Career Avenues A6
IPC
In general, process variables can be classified as shown in Fig. 1.2. Now we can give explanations of
types of process variables with the examples from Fig.1.1.
Variable
Input Output
Measured Unmeasured
Output variables denotes the effect of the process on the surroundings ( Fout , Tout , H ).
Manipulated variables can be adjusted freely by the human operator or a controller ( Fst , Fout ).
Disturbances are not the result of an adjustment by an operator or a controller( Fin , Tin , Tst ).
Examples
A manometer converts the change of pressure into the movement of an arrow along a scale of an
instrument;
A mercury-in-glass thermometer converts the change of temperature into the change of the length
of a mercury column.
A transmitter is a device that converts a process variable into a form of a signal suitable for
transmission to another location. Temperature is detected by a temperature transmitter, then it is
converted to an analog electrical (4-20 mA, 0-5 mA, 0-20 mA, 0-10 V, 0-5 V, -10 to +10 V, -5 to +5
V dc), or pneumatic (20-100 kPa), or digital signal which is proportional to the temperature under
measurement. This signal is sent to a controller. A measuring instrument or transducer must be
© Career Avenues A7
IPC
capable faithfully and accurately detect any changes that occur with the measured process
parameter.
Transmission lines are used for carrying a measurement signal from a measuring instrument or
transmitter to a controller, and from a controller to a final control element (control valve, for
example). Very often these lines are equipped with amplifiers to increase the measurement signal,
since it is very weak (for example, a thermal electromotive force of a thermocouple has the
magnitude of several dosens of millivolts).
A controller is an element that compares a current value of a controlled variable (the input variable
for a controller) with a desired value (the set point) and takes appropriate control actions to adjust
values of manipulated variables in the way to reach the desired value of process variable. In our
example (see Fig.1.1) the controller changes flowrates of steam entering the heat exchanger and
effluent liquid.
As the result of this action, the process variable (the temperature of the liquid in the outlet of the
stirred tank in Fig. 1.1) changes in such a way, that it approaches closer to the desired value (set-
point). This is a closed-loop or a feed-back control system. Feedback is an information about the status
(or the magnitude) of the controlled variable which can be compared with its desired value. A control
system without feed-back is called an open-loop control system. In other words, if we remove the
temperature transducer and level transmitter, controllers and control valves, or simply brake links
between thermocouple and level transmitter and controllers, or between controllers and control valves
(see Fig.1.1) and leave the tank not under control we will get an open-loop or unregulated system. In
this case information about controlled variables (the temperature and level of the liquid in the tank) is
not used for control of these process variables.
It is interesting to note that an unregulated system can be self-regulated. Assume that the flowrate of
the inlet liquid suddenly has increased to another value. Since Fin Fout , the liquid will be
accumulated in the tank (the level will increase). Thus increased liquid level will inevitably increase
the head of the liquid, which in its turn will increase the value of Fout . This will continue until that
moment when Fin Fout . After that, the level will be maintained on another value.
Have the feedback path. Does not have the feedback path.
Less sensitive to noise, disturbances and changes in Sensitive to noise, disturbances and changes in
the environment. the environment.
The system can compare the output response with The system cannot correct the disturbances.
the input and make a correction if there is any
difference.
© Career Avenues A8
IPC
A final control element is a device which receives the control signal from a controller and changes the
amount of matter or energy entering the process in a way to bring the controlled variable (process
variable) to its set point. As examples, we can mention control valves, relay switches providing on-off
control, variable-speed pumps, etc.
© Career Avenues A9
IPC
Laplace Transforms
DYNAMIC SYSTEMS
This course is about controlling dynamic systems. What makes dynamic systems unique and
distinctive is that dynamic systems have a memory. Their present output, y(t), does not just depend on
the present input signal value u(t), but on past inputs u() for t.
The way this “memory” is described in mathematical form is through integral/differential equations. In
this course, we will learn that even simple dynamic systems are modeled by derivative and integral
terms.
The Laplace transform allows us to transform these integral/differential equations into simpler
algebraic equations and solving them. It is a method for solving differential equations and
corresponding initial and boundary value problems.
COMPLEX FUNCTION
A complex function F(s), a function of s, has a real part and an imaginary part or F(s) = Fx + jFy, where
Fx is the real part and jFy is the imaginary part. F(s) can be represented graphically in the complex
plane as follows:
Fx2 Fy2
The magnitude of F(s) is denoted by |F| and is equal to . Its angle is obtained as
1
tan ( Fy / Fx )
. The angle is measured counterclockwise from the positive real axis.
© Career Avenues A 10
IPC
F Fx jFy
The complex conjugate of F(s) is defined as . Another form in which complex conjugate
*
is denoted is F . Note that the magnitude of F(s), denoted by |F|, can also be defined as:
F FF Fx2 Fy2
An advantage of the Laplace transform method is that it allows the use of graphical techniques for
predicting the system performance without actually solving system differential equations. Another
advantage of the Laplace transform method is that, when we solve the differential equation, both the
transient component and steady-state component of the solution can be obtained simultaneously.
Let us define
f(t) = a function of time t such that f (t) = 0 for t < 0
s = a complex variable
L = an operational symbol indicating that the quantity that it prefixes is to be transformed by the
0 f (t)e
st
Laplace integral dt .
Then the Laplace transform of f (t) is given by L[f (t)] F(s) f (t)est dt
0
Laplace transform of common functions have been tabulated [see the text book page 40] and are used
to transform the time domain functions f(t) into their Laplace form directly from tables and, vice versa.
Similarly, we obtain the following relationship for the second derivative of f(t) applies:
© Career Avenues A 11
IPC
d 2 f (t )
L[ 2
] s 2 F ( s) f (0) sf (0) th
dt and for the n derivative:
( n 2 ) ( n 1)
d n f (t ) n 1 n2
L[ ] s n
F ( s ) s f ( 0) s f ( 0) ... s f ( 0) f (0)
dt n
In most of the problems that we will consider in this course, it is assumed that the initial values of f (t)
th
and its derivatives are equal to zero, then the Laplace transform of the n derivative of f (t) is given by
s n F (s) .
Real-integration theorem
L f (t )dt F ( s)
s .
for zero initial value,
FINAL-VALUE THEOREM
The final-value theorem is extensively used throughout the course and relates the steady-state behavior
of f (t) to the behavior of sF(s) in the neighborhood of s = 0. The final-value theorem may be stated as
follows:
If f (t) and df (t)/dt are Laplace transformable, if F(s) is the Laplace transform of f (t), and if
lim t f (t )
exists, then
lim t f (t ) lim s0 sF (s)
This formula allows you to calculate the final steady state value of f(t) by using F(s) without having to
convert F(s) back into its time domain representation f(t).
Evaluating the inversion integral appears complicated. In practice, we seldom use this integral for
finding f(t). Instead, we usually decompose the F(s) into a simpler form by using partial fraction
© Career Avenues A 12
IPC
expansion and, then use a conversion table to obtain the time domain equivalent of the partial
fractions.
If a particular transform F(s) cannot be found in a table, then we expand it into partial fractions and
write F(s) in terms of simple functions of s for which the inverse Laplace transforms are already
known.
For problems in control systems analysis, F(s), the Laplace transform of f (t), frequently occurs in the
form F(s) =B(s)/ A(s) where A(s) and B(s) are polynomials in s. In the expansion of F(s) = B(s)/A(s)
into a partial-fraction form, it is important that the highest power of s in A(s) be greater than the
highest power of s in B(s). If such is not the case, the numerator B(s) must be divided by the
denominator A(s) in order to produce a polynomial in s plus a remainder (a ratio of polynomials in s
whose numerator is of lower degree than the denominator).
If F(s) is broken up into components, F(s) F1 (s) ... Fn (s) then its inverse Laplace transform is
obtained as:
L1 F (s) L1 F1 (s) ... L1 Fn ( s) f1 (t ) ... f n (t )
The advantage of the partial-fraction expansion approach is that the individual terms of F(s), resulting
from the expansion into partial-fraction form, are very simple functions of s; consequently, it is not
necessary to refer to a Laplace transform table if we memorize several simple Laplace transform pairs.
It should be noted, however, that in applying the partial-fraction expansion technique in the search for
the inverse Laplace transform of F(s) = B(s)/A(s) the roots of the denominator polynomial A(s) must
be obtained in advance. That is, this method does not apply until the denominator polynomial has
been factored.
If F(s) involves distinct poles only, then it can be expanded into a sum of simple partial fractions as
follows:
B( s ) a a
F ( s) 1 ... n
A( s) s p1 s pn
where a k (k = 1, 2, . . . , n) are constants. The coefficient a k is called the residue at the pole at
s pk
. The value of a k can be found by multiplying both sides of the above Equation by (s pk )
and letting s pk which gives:
© Career Avenues A 13
IPC
B( s )
ak s pk
A( s) s pk
a
L1 k ak e pk t
s pk
pt p t
then f(t) is obtained as f (t ) a1e ... an e for all t 0.
1 n
since:
If can be shown that the order of the integration and limit operation on the left side of this equation can
be interchanged if the conditions of the theorem hold. Doing this gives
df
0 dt lim sf s f 0
dt s/0
Example 4.1 Find the final value of the function x(t) for which the Laplace transform is
1
n(s) =
s s3 3s 2 3s 1
Direct application of the final-value theorem yields
1
lim n t s/0 lim 1
t / s 3s 3s 1
3 2
As a check, note that this transform was inverted in Example 3.6 to give
r t
2
x(t) = 1 e 1 1 .
2
Which approaches unity as t approaches infinity. Note that since the denominator of sx(s) can be
factored to (s + 1)3, the conditions of the theorem are satisfied; that is, (s + 1) 3 ≠ 0 unless s = - 1.
© Career Avenues A 14
IPC
The conditions on this theorem are not so stringent as those for the previous one because for functions
of interest to us the order of integration and limiting process need not be interchanged to establish the
result.
Example 4.3: Find the initial value x(0) of the function that has the following transform
s 4 6sh 2 9s 8
x(s) =
s s 2 s3 2s 2 s 2
The function sx(s)is written in the form
s 4 6s 2 9s 8
sx(s) =
s 4 5s 2 4
s 2 9s 12
sx(s) = 1
s 4 5s 2 4
x(0) = 1
GATE 2006
The Laplace transform of the input function, x(t), given in the figure below is given by
2
1 e2s 2
1 e2s
1 1
A) B)
2s 2s
C) 2 1 e2s D) 2 1 e2s
1 1
s s
Correct Answer: 1
Solution:
Ans. (1)
Given function,
© Career Avenues A 15
IPC
1
X(t) = t, t<2
2
=1 t>2
L x t est x t dt
0
1 2
= t est dt 1.est dt
02 2
2
1 st t 1 e st
∴ X(s) = e
2 s S 2 0 s 2
1 1 2s 2 1 e2s e
e
s zero
=
2 S2 s S 2 s
1 1 1 1
= 2 e2s 2 e2s e2s
2s s 2s s
∴ X(s) = 2 1 e2s
1
2s
GATE 2009
1
The inverse Laplace transform of
2s 3s 1is
2
Correct Answer: 1
Solution:
Ans: 1
1 1 2 1
2s 3s 1 2s 1 s 1 2x 1 s 1
2
1 2 1 t /2 t
L 2 L L e e
2x 3x 1 2x 1 s 1
GATE 2004
t
For the time domain function f(t) = t, the Laplace transform of f t dt is given by
0
3 3 3
A) 1/(2s ) B) 2/s C) 1/s D) 2/s2
27 Ans.
f(t) = t
0
t
t t2
L f t dt L t dt L
0 2
= L t2
1 1 2! 1
3
2 2 s 2 1 s
Correct answer (C)
© Career Avenues A 16
IPC
GATE 1999
A control system has the following transfer function,
s 1 s 1
F(S) =
s s 2 s 4
The initial value of the corresponding time function is
A) 1 B) 1/8 C) 7/8 D) – 1
Ans.
From the initial value theorem
Lt Lt
Ft sF s
t 0 s
So
Lt s 1 s 1 Lt s 1s 1
s
t 0 s s 2 s 4 s s s 2 s 4
= 1/8
© Career Avenues A 17
IPC
TRANSFER FUNCTIONS
A mathematical model of a dynamic system is defined as a set of equations that represents the
dynamics of the system accurately or, at least, fairly well. Note that a mathematical model is not
unique to a given system. A system may be represented in many different ways and, therefore, may
have many mathematical models, depending on one's perspective. The degree of accuracy and
complexity of models is very much dependent on the use and application that they are intended for.
The dynamics of many systems, whether they are mechanical, electrical, thermal, economic,
biological, and so on, may be described in terms of differential equations. Such differential equations
may be obtained by using physical laws governing a particular system, for example, Newton's laws for
mechanical systems and Kirchhoff's laws for electrical systems. We must always keep in mind that
deriving a reasonable mathematical model is the most important part of the entire analysis.
MATHEMATICAL MODELS
Mathematical models may assume many different forms. We will consider be considering the
transient-response or frequency-response analysis of single-input-single-output, linear, time-invariant
systems. Here, a transfer function representation is more convenient than any other. Once a
mathematical model of a system is obtained, various analytical and computer tools can be used for
analysis and synthesis purposes.
In general, in solving a new problem, we find it desirable first to build a simplified model so that we
can get a general feeling for the solution. A more complete mathematical model may then be built and
used for a more complete analysis.
LINEAR SYSTEMS
A system is called linear if the principle of superposition applies. The principle of superposition states
that the response produced by the simultaneous application of two different forcing functions is the
sum of the two individual responses. Hence, for the linear system, the response to several inputs can
© Career Avenues A 18
IPC
be calculated by treating one input at a time and adding the results. It is this principle that allows one
to build up complicated solutions to the linear differential equation from simple solutions.
NON-LINEAR SYSTEMS
A system is nonlinear if the principle of super-position does not apply. Thus, for a nonlinear system
the response to two inputs cannot be calculated by treating one input at a time and adding the results.
Although many physical relationships are often represented by linear equations, in most cases actual
relationships are not quite linear. In fact, a careful study of physical systems reveals that even so-
called "linear systems" are really linear only in limited operating ranges. In practice, many
electromechanical systems, hydraulic systems, pneumatic systems, and so on, involve nonlinear
relationships among the variables. For example, the output of a component may saturate for large
input signals. There may be a dead space that affects small signals. (The dead space of a component is
a small range of input variations to which the component is insensitive). Square-law nonlinearity may
occur in some components. For instance, dampers used in physical systems may be linear for low-
velocity operations but may become nonlinear at high velocities, and the damping force may become
proportional to the square of the operating velocity. Examples of characteristic curves for these
nonlinearities are shown in your text.
TRANSFER FUNCTIONS
In control theory, functions called transfer functions are commonly used to characterize the input-
output relationships of components or systems that can be described by linear, time-invariant,
differential equation.
The transfer function of a linear, time-invariant, differential equation system is defined as the ratio of
the Laplace transform of the output (response function) to the Laplace transform of the input (driving
function) under the assumption that all initial conditions are zero. Consider the linear time-invariant
system defined by the following differential equation:
(n) ( n 1) ( m) ( m1)
ao y a1 y ... an1 y an y bo x b1 x ... bm1 x bm x
where y is the output of the system and x is the input. The transfer function of this system is obtained
by taking the Laplace transforms of both sides of the above equation, under the assumption that all
initial conditions are zero.
© Career Avenues A 19
IPC
By using the concept of transfer function, it is possible to represent system dynamics by algebraic
equations in s. If the highest power of s in the denominator of the transfer function is equal to n, the
system is called an nth-order system.
Forcing Functions
Step function. Mathematically, the step function of magnitude A can be expressed as
X(t) = Au(t)
where u(t) is the unit-step function defined in Chap.2. A graphical representation is shown in fig.5.3.
© Career Avenues A 20
IPC
The transform of this function is X(s) = A/s, A step function can be approximated very closely
in practice. For example, a step change in flow rate can be obtained by the sudden opening of a value.
Step response
Y(t) = 0 t<0
Y(t) = A(1 – e-t/τ) t>0
The value of Y(t) reaches 63.2 percent of its ultimate value when the time elapsed is equal to one time
constant τ. When the time elapsed is 2τ, 3τ and 4τ, the percent response is 86.5, 95, and 98,
respectively. From these facts, one can consider the response essentially completed in three to four
time constants.
The true impulse function, obtained by letting b l 0 in Fig. 5.4, has Laplace transform A. It is used
more frequently as a mathematical aid than as an actual input to a physical system. For some systems
it is difficult even to approximate an impulse forcing function. For this reason the representation of
Fig. 5.4 is valuable, since this form can usually be approximated physically by application and
removal of a step function. If the time duration b is sufficiently
Where A is the amplitude and ω is the radian frequency. The radian frequency ω is related to the
frequency f is cycles per unit time by ω = 2πf. Figure 5.5 shows the graphical representation of this
function. The transform is X(s) = Aω/(s2 + ω2). This forcing function forms the basis of an important
branch of control theory known as frequency response. Historically, a large segment of the
development of control theory was based on frequency-response methods, which will be presented in
Chaps. 16 and 17. Physically, it is more difficult to obtain a sinusoidal forcing function in most
process variables than to obtain a step function.
© Career Avenues A 21
IPC
Sinusoidal response
To investigate the response of a first-order system to a sinusoidal forcing function, the example of the
mercury thermometer will be considered again. Consider a thermometer to be in equilibrium with a
temperature bath at temperature xs. At some time t = 0, the bath temperature begins to vary according
to the relationship
x = xs + A sin ωt t>0 (5.17)
Where x = temperature of bath
xs = temperature of bath before sinusoidal is applied
A = amplitude of variation in temperature
ω = radian frequency, rad/time
From Eq. (5.25), the amplitude of the response and the phase angle are calculated; thus
A 2
0.8960F
1
2 2
4 1
= –tan-12 = – 63.5’
or
Phase lag = 63.50
© Career Avenues A 22
IPC
To obtain the lag in terms of time rather than angle, we proceed as follows: A frequency of
10/π cycles/min means that a complete cycle (peak’ to peak) occurs in (10/π)-1 min. Since one cycle is
equivalent to 3600 and the lag is 63.50, the time corresponding to this lag is
63.5
time for 1 cycle
360
or
63.5
Lag = = 0.0555 min
360 10
In general, the lag in units of time is given by
Lag =
360f
when is expressed in degree
The transfer function of the first order system is 1/(τs+1) where τ is time constant.
Step response
Y(t) = 0 t<0
Y(t) = A(1 – e-t/τ) t≥0
The value of Y(t) reaches 63.2 percent of its ultimate value when the time elapsed is equal to one time
constant τ. When the time elapsed is 2τ, 3τ, and 4τ, the percent response is 86.5, 95, and 98,
respectively. From these facts, one can consider the response essentially completed in three to four
time constants.
Impulse response
Sinusoidal response
To investigate the response of a first-order system to a sinusoidal forcing function, the example of the
mercury thermometer will be considered again. Consider a thermometer to be in equilibrium with a
temperature bath at temperature xs. At some time t = 0, the bath temperature begins to vary according
to the relationship
x = xs + A sin ωt t>0 (5.17)
© Career Avenues A 23
IPC
tan 1
Figure 7.1
© Career Avenues A 24
IPC
Several non-interacting systems are connected in series then the overall transfer function is product of
individual transfer functions.
A second-order system exhibits a wide range of responses that must be analyzed and described.
Varying a first-order system’s parameter simply changes the speed of the response; changes in the
parameters of a second-order system can change the form of the response.
In this section we define the following natural responses and find their characteristics:
Natural response
Two exponentials with time constants equal to the reciprocal of the pole locations, or
cn(t) = K1 e -1t + K2 e-2t
© Career Avenues A 25
IPC
Natural response
Damped sinusoid with an exponential envelope whose time constant is equal to the reciprocal of the
pole’s real part. The radian frequency of the sinusoid, the damped frequency of oscillation, is equal to
the imaginary part of the poles, or
cn(t) = A e -dt cos d t -
Undamped responses
Natural response
Undamped sinusoid with radian frequency equal to the imaginary part of the poles, or
cn(t) = A cos (1 t - )
Natural response
One term is an exponential whose time constant is equal to the reciprocal of the pole location. Another
term is the product of time, t , and an exponential with time constant equal to the reciprocal of the pole
location, or
cn(t) = K1 e-1t + K2 te-1t
The step responses for the four cases of damping discussed in this section are superimposed in Figure
5.10.
© Career Avenues A 26
IPC
Notice that the critically damped case is the division between the overdamped cases and the
underdamped cases and is the fastest response without overshoot.
© Career Avenues A 27
IPC
1. Overshoot. Overshoot is a measure of how much the response exceeds the ultimate value
following a step change and is expressed as the ratio A/B in Fig.8.3
The overshoot for a unit step is related to by the expression
Overshoot = exp (8.24)
1 2
This relation is plotted in Fig. 8.4. The overshoot increases for decreasing .
2. Decay ratio. The decay ratio is defined as the ratio of the sizes of successive peaks and in
given by CIA in Fig.8.3. The decay ratio is related to by the expression
2
Decay ratio = exp = (overshoot)2 (8.25)
1 2
Which is plotted in Fig. 8.4. Notice that larger means greater damping, hence greater decay.
3. Rise time. This is the time required for the response to first reach is ultimate value and is
labeled tr in Fig. 8.3. The reader can verify from Fig. 8.2 that t r increase with increasing .
4. Response time. This is the time required for the response to come within ± 5 percent of its
ultimate value and remain there. The response time is indicated in Fig. 8.3. The limits ± 5 percent are
arbitrary, and other limits have been used in other texts for defining a response time.
5. Period of oscillation. From Eq. (8.17), the radian frequency (radians/time) is the coefficient of
t in the sine term; thus,
1 2
ω, radian frequency = (8.26)
Since the radian frequency ω is related to the cyclical frequency f by ω = 2πf, it follows that
1 1
2
1
f= (8.27)
T 2
Where T is the period of oscillation (time/cycle). In terms of Fig. 8.3, T is the time elapsed
between peaks. It is also the time elapsed between alternate crossings of the line Y = 1.
© Career Avenues A 28
IPC
6. Natural period of oscillation. If the damping is eliminated [C = 0 in Eq. (8.1). or = 0], the
system oscillates continuously without attenuation in amplitude. Under these “natural” or undamped
conditions, the radian frequency is 1/τ, as shown by Eq. (8.26) when = 0. This frequency is referred
to as the natural frequency n .
1
n (8.28)
The corresponding natural cyclical frequency fn = and period Tn are related by the expression
1 1
fn = (8.29)
Tn 2
Thus, τ has the significance of the undamped period.
From Eqs. (8.27) and (8.29), the natural frequency is related to the actual frequency by the
expression
f
1 2
fn
In summary, it is evident that is a measure of the degree of damping, or the oscillatory character,
and τ is a measure of the period, or speed, of the response of a second-order system.
1. The roots of the characteristic equation of an under damped second order system are
A) Real, negative and equal, B) Real, negative and unequal
C) Real, positive and unequal, D) Complex conjugates
Ans: D
GATE 1999
Ans: (D)
GATE 2005
The unit step response of a first order system with time constant τ and steady state gain Kp is
given by
-t/τ -t/τ -2t/τ
K p e t /
A) Kp(1 – e ) B) Kp(1 + e ) C) Kp(1 – e ) D)
Ans: (A)
© Career Avenues A 29
IPC
GATE 2008
s 2
A unit step input is given to a process that is represented by the transfer function . The
s 5
initial value (t = 0+) of the response of the process to the step input is
A) 0 B) 2/5 C) 1 D) ∞
y S S 2
G(S) =
x S S 5
X(t) = u(t)
1
X(S) =
S
S 2
Y(S) =
S S 5
S 2
Y(S) =
S S 5 S S 5
1 2
Y(D) =
S 5 S S 5
Y(t) = e-5t + 2(1 – e-5t)
Y(t) = 2 – e-5t
Y(t) = 2 – e-5 × 0 =2–1 =1
GATE 2006
A process is perturbed by a sinusoidal input, u(t) = A sin ωt. The resulting process output is Y(s)
KA
= if y(0), the differential equation representing the process is
s 1 s2 2
dy t dy t
A) y (t) = Ku(t) B) + y(t) = KAu(t)
dt dt
dy t dy t
C) +y(t) = Ku(t) D) y t = KAu(t)
dt dt
© Career Avenues A 30
IPC
GATE 2008
The unit inpulse response of a first order process is given by 2e -0.5t. The gain and time constant
of the process are, respectively
A) 4 and 2 B) 2 and 2 C) 2 and 0.5 D) 1 and 0.5
GATE 2008
A tank of volume 0.25m3 and height 1 m has water flowing in at 0.05m3/min. The outlet flow
rate is governed by the relation.
Fout = 0.1 h
Where he is the height of the water in tank in m and Fout is the outlet flow rate in m3/min.
The inlet flow rate changes suddenly from its nominal value of 0.05m3/min to 0.15m3/min and
remains there. The time (in minutes) at which the tank will begin to overflow is given by
A) 0.28 B) 1.01 C) 1.73 D) ∞
Answer 3:
Solution:
(V = A × h)
Massin – massout = Acumulation
dh
Fi – F0 = A
dt
© Career Avenues A 31
IPC
h dh
Fi = =A (1)
R dt
h dh
Fis = s = A s (2)
R dt
1 d h hs
(Fi – Fis) – (h – hs) = A
R dt
H dH
F– =A
R dt
H S
F(S) – = ASH(S)
R
1
H(S) AS = F(S)
R
H S R
F S S 1
τ = AR.
1
Fout = 0.1h R = = 10
0.1
V = 0.25 m3, h = 1
∴ A = 0.25 m2
∴ τ = AR = 0.25 × 10 = 2.5
F(S) = (0.15 – 0.05) u(t) = 0.1 u(t)
0.1
F(S) =
S
10 0.1 1
H(S) =
2.5S 1 S S 2.5S 1
t
H(t) = 1 e
2.5
At overflow h(t) = 1 m
H(t) = h(t) – hs
F 0.05
hs = = 0.5m
0.1 0.1
t
∴ 1 – 0.5 = 1 – e 2.5
t
0.5 = 1 – e 2.5
t = 1.73 mm
GATE 2009
For a tank of cross-sectional area 100 cm2 and inlet flow rate (Q1 in cm3/s), the outlet flow rate
(Q0 in cm3/s) is related to the liquid (H in cm) as Q0 = 3 H (see figure below).
© Career Avenues A 32
IPC
H s
Then the transfer function (overbar indicates deviation variables) of the process around
Q1 s
the steady-state point, Qi,s = 18 cm3/s and Hs = 36 cm, is
1 2 3 4
A) B) C) D)
100s 1 200s 1 300s 1 400s 1
Correct Answer: 4
Solution
dH
Qi – Q0 = A
dt
dH
Qi – 3 H = A
dt
Qi – 3 H = A dH
dt
f H = f HS +f HS H- HS
H Hs
d H H Hs
dH
Hs
1
= 36 H Hs
2 Hs
1
=6+ (H – Hs)
12
H
=6+ –3
12
H
=3+
12
H dH
Qi – 3 3 A
12 dt
H dH
Qi – 9 A
4 dt
Hs dHs
Qi(s) – 9 A
4 dt
H s
Qi(s) – = 100H(s)
4
© Career Avenues A 33
IPC
H s 4
Qi s 400s 1
GATE 2005
An example of an open-loop second order under damped system is
A) liquid level in a tank
B) U-tube manometer
C) thermocouple in a thermo-well
D) two non-interacting first order system in series.
Ans. B
GATE 2004
Match first order system giving in Group I with the appropriate time constant in Group II
Group I Group II
P. Thermometer 1. (mCp) / (hA)
Q. Mixing 2. q/V
3. V/q
r. (hA)/(mCp)
A) P-4, Q-2 B) P-4, Q-3 C) P-1, Q-2 D) P-1, Q-3
Ans. (D)
GATE 2004
The experimental response of the controlled variable y(t) for a step change of magnitude P in the
manipulated variable x(t) is shown below,
P
Q e R s
Q
Q e d s
P
Q P s 1
A) B)
d s 1
© Career Avenues A 34
IPC
P P
Q e d s Q e Q s
Q R s 1
C) A)
d s 1
Ans. (C)
kp e sD
Q e s
P
D
T.F. =
s 1 Q s 1
P
(C) Correct Answer
GATE 2003
Water is entering a storage tank at a temperature TO and flow rate QO and leaving at a flow rate
Q and temperature T. There are negligible heat losses in the tank. L the area of cross section of
the tank is AC. The model that describes the dynamic variation of water temperature in the tank
with time is given as
dT dT
A) Q0(T0 – T) = Ach B) Q0T0 – QT = Ach
dt dt
dT d Th
C) Q(T0 – T) = Ach D) Q0(T0 – T) = Ach
dt dt
Answer:
Mass balance
d
(AchPw) = Q0Pw – QPw
dt
dh
or Ac = Q0 – Q
dt
Energy balance (Ref temp = O0C
d
Ac h Pw CpT Q0 Pw Cp T0 QPw CpT
dt
© Career Avenues A 35
IPC
d
or Ac (h T) = Q0T0 – QT
dt
dh dT
or AcT Ac h = Q0T0 – QT
dT dt
dT
N1 T(Q0 – Q) + Ach = Q0T0 – QT (using mass balance equation)\ N
dt
dT
N1 Ach Q0 T0 Q T TQ0 Q T
dt
dT
Ac h Q0 T0 T
dt
GATE 2007
Match the transfer functions with the responses to a unit step shown in the figure,
Diagram
2.5 4s 1 2e10s
A) B)
4s 2 4s 1 10s 1
5 0.1
C) D)
20s 1 s
4s 3
E)
2s 1
Y s
D) G(s) =
X s
for unit step input
G s 0.1
∴ Y(s) = 2
s s
∴ Y(t) = – 0.1t
Observing the figure, curve ‘d’ fits the above equation.
5
C) y(s) =
20s 1 s
5 1
=
20 1
s s 20
1 A B
y(s) =
4s 1
s
20
( by partial fraction)
Taking inverse Laplace transform
© Career Avenues A 36
IPC
1
y(t) = 1 e
1/20 t
80
Observing figure, curve ‘a’ fits above equation.
GATE 2000
2s 1
The unit step response of the transfer function reaches its final steady state
3s 1 4s 1
asymptotically after
A) a montonic increase
B) a monotonic decrease
C) initially increasing and then decreasing
D) initially decreasing and then increasing
Answer is (A)
GATE 2000
1
The unit step response of the transfer function
s 2s 3
2
Answer is (B)
GATE 1999
Answer is B
GATE 1998
4
The second order system with the transfer function has a damping ratio of
s 2s 1
2
Ans.
By comparing s2 + 2s + 1 with 2 s2 2s 1 we get
2 = 1 2 = 2
So =1
=1
© Career Avenues A 37
IPC
GATE 1997
1
The transfer function of a process is . If a step change is introduced into the system,
16s 8s 4
2
GATE 1997
The transfer function for a first-order process with time delay is
e s
D
e s
D
A) B)
s 1 s 1
1 D s
C) D)
s 1 D s 1 s 1
Answer is (B)
GATE 1996
The open loop transfer function of a control system is KR/(1 + s ), This represents:
A) a first order system B) dead time system
C) a first order time lag D) a second order system
Ans. (A)
GATE 1995
The response of two tanks of same size and resistance in series is –
A) under damped B) critical damped
C) over damped D) none of the above
Ans. (C)
GATE 1995
The transfer function of a pure dead time system with dead time D is –
1
A) B) D s 1 C) e rDs D) e Xs
D s 1
© Career Avenues A 38
IPC
GATE 1995
The time constant of a first order process with resistance R and capacitance C is –
A) R + C B) R – C C) R C D) 1/RC
Ans. C
GATE 1995
Identify an unbounded input from four inputs whose functions are given below:
A) 1 B) 1/S C) 1/S2 D) 1/(S2 + 1)
Ans. (C)
GATE 1991
A certain thermocouple has a specific time constant of 2s. If the process temperature changes
abruptly from 8000C to 9000C, the temperature reading in an indicator attached to the
thermocouple after 6s will be approximately
A) 8600C B) 9000C C) 8900C D) 8950C
Ans:
1st order system
Y(t) = X(t) [1 – e-t/q] diagram
at t = 6s 8000C → 9000C
-6/2
Y(t) = 100 [1 – e ] x(t) = 900 – 800
0
Y(t) = 95 C = 1000C
∴ Temperature read g thermocouple = 800 + 950C
= 8950C
GATE 1992
When a bare thermocouple is covered by a protective sheath, the response becomes
A) Faster and Oscillatory B) Faster and non-oscillatory
C) Slower and Oscillatory D) Slower and non-oscillatory
Ans. (A)
GATE 2007
The dynamic model for a mixing tank open to atmosphere at its top as shown below is to be
written. The objective of mixing is to cool the hot water stream entering the tank at a flow rate q 2
and feed temperature of Ts with a cold water feed stream entering the tank at a flow rate q1 and
feed temperature of T0. A water stream is drawn from the tank bottom at a flow rate of q 4 by a
pump and the level in the tank is proposed to be controlled by drawing another water stream at a
flow rate q3. Neglect evaporation and other losses from the tank.
© Career Avenues A 39
IPC
dV dT
1. = q1 + q2 + q3 V = q1 T0 + q2Ts – q3T
dt dt
dV d VT
2. = q1 – q4 = q1 Ts – q4T
dt dt
dV d VT
3. = q1 + q2 + q4 = q1 Ts + q2Ts – q4T
dt dt
dV dT
4. = q1 + q2 – q3 – q4 V = q1 (T0 – T) + q2 (Ts – T)
dt dt
A) i – e, ii – c, iii – a, iv – d, v – b B) i – b, ii – a, iii – c, iv – e, v – d
C) i – a, ii – b, iii – c, iv – d, v – e D) i – e, ii – a, iii – c, iv – b, v – d
Correct Answer: 4
Solution
Answer (D)
dT
V = Q1(T0 – T) + Q2(TS – T)
dt
© Career Avenues A 40
IPC
SUMMARY
Transfer function of a simple first order system (Mercury thermometer):
Y s 1
X s s 1
Impulse response:
Y(t) = e-t/τ
A
Yt s sin t
2 2 1
Y s 1
X s s 1
Transfer function of RC circuit:
Ec s 1
V s s 1
Time constant for different first order systems:
Thermometer:
mC
hA
Liquid level process:
= AR
Mixing process:
V
=
q
© Career Avenues A 41
IPC
RC circuit:
= RC
H2 s 1 R2
Q s 1s 1 2 s 1
Transfer function of first order interacting tanks in series:
H2 s R2
Q s 12 s 1 2 A1 R 2 s 1
2
Y s 1
X s s 2s 1
2 2
Overshoot is a measure of how much the response exceeds the ultimate value following a step change:
Overshoot = exp / 1 2
Decay ratio is the ratio of sizes of succesive peaks of response curve.
Decay ratio = exp 2 / 1 2 = (overshoot)
2
© Career Avenues A 42
IPC
INTRODUCTION
We have been working with individual subsystems represented by a block with its input and output.
More complicated systems are represented by the interconnection of many subsystems. Since the
response of a single transfer function can be calculated, we want to represent multiple subsystems as a
single transfer function.
Then, we can apply the analytical techniques of the previous chapters and obtain transient response
information about the entire system.
BLOCK DIAGRAMS
In the previous section the concept of a transfer function for a linear time-invariant system was
presented. By definition the transfer function is the ratio of the Laplace transform of the output
variable to the Laplace transform of the input variable. Let X(s) be the (Laplace transform of the)
input variable, Y(s) be the output variable, and G(s) be the transfer function. One method of
graphically denoting the relationship Y(s) = X(s)U(s) is through a block diagram, as shown in the
following figure. The block represents a transfer function corresponding to a system’s mathematical
model and the arrows represent signals (e.g. electrical voltage from a position sensor).
X(s) Y(s)
G(s)
Many systems are composed of multiple subsystems. When multiple subsystems are interconnected, a
few more elements such as summing junction and pickoff points must be added to the block. All
component parts of a block diagram for a linear, time-invariant system are shown in Figure 3.1.
Figure 3.1
© Career Avenues A 43
IPC
We will now examine some common topologies for interconnecting subsystems and derive the single
transfer function representation for each of them. These common topologies will form the basis for
reducing more complicated systems to a single block.
CASCADE FORM
Figure 4.2(a) shows an example of cascaded subsystems and Figure 4.2(b) shows the equivalent single
transfer function.
Figure 2.2
PARALLEL FORM
Figure 4.3(a) shows an example of parallel subsystems. The equivalent transfer function, Ge(s) appears
in Figure 4.3(b).
Figure 3.3
© Career Avenues A 44
IPC
FEEDBACK FORM
The third topology is the feedback form as shown in Figure 4.4(a) and Figure 4.4(b) shows a
simplified model. We obtain the equivalent, or closed-loop, transfer function shown in Figure 4.4(c).
Figure 4.6 shows equivalent block diagrams formed when transfer functions are moved left or right
past a pickoff point.
© Career Avenues A 45
IPC
Figure 3.6
Block diagram
algebra for pickoff
points—
equivalent forms
for moving a block
© Career Avenues A 46
IPC
Ex 3.1 Reduce the block diagram shown in Figure 3.7 to a single transfer function.
Figure 3.7
Answer:
© Career Avenues A 47
IPC
Ex 4.2 Reduce the block diagram shown in Figure below to a single transfer function.
Answer
© Career Avenues A 48
IPC
Do-it-yourself: Find the equivalent transfer function, T(s) = C(s)/R(s) for the system shown in
Figure 3.9.
Figure 3.9
Answer:
s3 1
T ( s)
s 4 s 3 22 1
S2
1 S 3 1
2 1
S2
1 S 1
1 S2 S S2
2
S
2 1 S 1
3
S
1
2 1 1 S2
1 S S S2
1
1 S2
S 5
© Career Avenues A 49
IPC
S3 1 S S 3 1
2
S2 S S3 1 S 4 S 3 1
S3 1 S S 3 1
1 2 3 1
S2 S S 1 S S 3 S 2 1
(b)
S3 S4 S3 S S3
4 3
S S 5 S S 3 S 2 1 S 3 1 S 4 S 3 S S 3 1
S 3
1
S 4 2S 3 S 1
R(s) C(s)
G(s)
R(s) C(s)
G(s)
Moving a
starting point 1
B(s) G(s)
B(s)
R(s) + C(s)
Closed loop G(s) R(s) G( s) C(s)
-
system 1 H ( s)G ( s)
H(s)
© Career Avenues A 50
IPC
Exercise: Consider the transfer function of the system shown in Fig. 3-9a. The final transfer function
is shown in Fig. 3-9d. Note that the first reduction involves a parallel combination; the second
involves a cascade combination.
First reduction
G3(s)
R(s) + +
G1(s) G2(s) C(s)
+ G4(s)
- +
-
H1(s)
H2(s)
R(s) + C(s)
G1(s) G2(s)+G3(s) G4(s)
- +
-
H 1 (s)
G4 (s)
H2(s)
(b)
[G 2 ( s ) G3 ( s )]G 4 ( s )
R(s) + G5 ( s ) C(s)
G1(s) 1 H 1 ( s )[G 2 ( s ) G3 ( s )]
-
H2(s)
(c)
G1(s)G5 (s)
R(s) C(s)
1 G1(s)G5 (s)H2 (s)
(d)
Fig.3-9 Obtaining transfer function by block diagram reduction
© Career Avenues A 51
IPC
GATE 1996
In a feed-back control system G and H denote open loop and closed loop transfer functions
respectively. The output-input relationship is:
A) G / (1 + H) B) H / (1 + G) C) G / H D) H / G
Answer: (B)
GATE 2006
The block diagram of a closed loop control system is shown in the Fig. below. Y is the
controlled variable, D is disturbance, Ysp is the set point, G1, G2 and G3 are transfer functions,
and Ke is the proportional controller.
GATE 2006
1
Let G1(s) = 1 and G2(s) = G3(s) = A step change of magnitude M is made in the
s 1
© Career Avenues A 52
IPC
set point. The steady state offset of the closed loop response Y is
M M K c 1 M
A) B) C) A) zero
1 2K c 1 2K c 1 Kc
Answer is (A)
Solution:
The offset is defined as the difference in the magnitude of the set point and the controlled
variable at the steady state condition of the system.
Offset = R(∞) – C(∞)
where, R(∞) is the ultimate value of point,
C(∞) is ultimate value of controlled variable
Y s G3 1 / s 1
Here,
D s 1 G1G 3 G 2 K c 1 1
1 1 Kc
s 1 s 1
Y s 1
D s s 1 2K c
Y s 1 1
or
D s 1 2K c 1
1 2K s 1
c
1
Defining, '
1 2K c
Y s 1 1
We have
D s 1 2K c 's 1
for step change of Magnitude, M,
D(t) = M or D(s) = M/s
1 M 1
Y s '
1 2K c s s 1
© Career Avenues A 53
IPC
GATE 2008
Which one of the following transfer functions corresponds to an inverse response process with a
positive gain?
1 2 2 5
A) B)
2s 1 3s 1 s 1 s 10
3 0.5s 1 5 3
C) D)
2s 1 s 1 s 1 2s 1
Correct Answer: 3
GATE 2009
The block diagram for a control system is shown below:
For a unit step change in the set point, R(s), the steady state offset in the output Y(s) is
A) 0.2 B) 0.3 C) 0.4 D) 0.5
Solution:
Answer (A)
© Career Avenues A 54
IPC
Gate 2010
A block diagram for a control system is shown below:
The steady state gain of the closed loop system, between output Y(s) and set point R(s), is
A) 5/9 B) 4/9 C) 1/3 D) 2/9
Answer:
2 1 0.2s 2
. 2
Y s 1 0.02s s 11s 1
0.2R s 2 1 0.23 2
1 0.2
1 0.02s s 11s 1
2
4 1 0.2s
=
1 0.02s s2 11s 1 0.8 1 0.2s
Y s 4 0.2 1 0.2s
or
R s 1 0.02s s 2 11s 1 0.8 1.0.2s
For steady state gain of the closed loop
put s = 0
Y s 0.8 1 0.2 0
R s 1 0.2 0 0 11 0 1 0.8 1 0.2 0
2
0.8 0.8 4
=
1 1 0.8 1.8 9
GATE 2004
For the block diagram shown below
© Career Avenues A 55
IPC
Answer:
Characteristic equation
1 + G1G2G3 = 0
1 K ped s Km
1 + Kc 1 . . 0
s 1 s 1
1p p m
K c K p K m 1 1s
or 1 + 0
1s 1 ps 1 ns 1
or 1s 1 ps 1 ms 1 KcKpKm 1 1s 0
Correct answer is (C)
GATE 2001
The block diagram an integrating level process is given below. For unit step change in the set
point h set = 1 with d = 0, the offset exhibited by the system is
Kc 1 2K c
A) B) C) 0 D)
1 Kc 1 Kc 1 Kc
Answer:
1
2 Kc
Y s 2p Kc
G(s) =
X s 1 2 K 1 s Kc
c
2p
1
x s (unit step change)
p
1 Kc A B
Y(s) = .
p s Kc p s Kc
Kc = A(p + Kc) + Bp = AKc + p(A + B)
∴ A = 1, A + B = 0, B = – 1
1 1
Y(p) =
p p Kc
∴ Y(t) = 1 – e-kct
© Career Avenues A 56
IPC
offset y t t y t t 0
=1–1=0
© Career Avenues A 57
IPC
SUMMARY
The control in which set point is kept constant and the load variable changes is called regulatory
problem.
The set point is varied in case of servo problem where the load variable is maintained constant.
Diagram = TR T
' '
© Career Avenues A 58
IPC
CONTROL ACTION
INTRODUCTION
In many cases the transient and steady state response may not satisfy the system requirements. We
may wish to modify the transient or steady state characteristics, or both, by introducing a controller in
the control system. We shall show that an appropriate controller does indeed alter the response of the
system although it also brings about additional changes that may be undesirable. It is not always
possible to optimize everything and the particular price that is to be paid must be weighed against the
advantages to be gained.
A general feedback control system is shown in Fig.6-3. The output signal M (s) of the controller is the
actuating signal that is employed for making necessary corrections so that the output corresponds to
the input in some manner. The relationship of the actuating signal M (s) to the error signal is directly
dependent upon Gc (s) ,
M ( s)
Gc ( s )
E ( s)
C (s)
Gc ( s)G L ( s)
E (s)
C ( s) Gc ( s)G L ( s)
R( s) 1 Gc ( s)G L ( s) H ( S )
We note that response c(t ) will depend upon the nature of Gc (s) . Since we can select Gc (s) , we have
some control upon C (s) for a given input. There are, in general, three basic types of controls that we
may select,
Proportional control
Derivative control
Integral control
We may, of course, select any combination of the three types. In the remaining section the effect of
different controllers on system response is investigated.
© Career Avenues A 59
IPC
PROPORTIONAL CONTROL
With this control the actuating signal is proportional to the error signal,
M (s) K p [ R(S ) B(s)]
K p E (s)
(6-12)
A feedback control system with proportional control is shown in Fig.6-4. In such a system a
compromise is often necessary in selecting a proper gain so that the steady state error and maximum
overshoot are within acceptable limits. Practically, however, a compromise cannot always be reached
since an optimum value of K may satisfy the steady state error but may cause excessive overshoot or
even instability. This problem can be overcome if we employ proportional control in conjunction with
some other type of control.
R(s) KP G(s) C(s)
+ -
H(s)
The proportional term makes a change to the output that is proportional to the current error value. The
proportional response can be adjusted by multiplying the error by a constant Kp, called the
proportional gain.
Pout K P e t
Where
Pout: Proportional term of output
Kp: Proportional gain, a tuning parameter
e: Error = SP − PV
t: Time or instantaneous time (the present)
© Career Avenues A 60
IPC
Plot of PV vs. time, for three values of Kp (Ki and Kd held constant)
A high proportional gain results in a large change in the output for a given change in the error. If the
proportional gain is too high, the system can become unstable (See the section on loop tuning). In
contrast, a small gain results in a small output response to a large input error, and a less responsive (or
sensitive) controller. If the proportional gain is too low, the control action may be too small when
responding to system disturbances.
In the absence of disturbances, pure proportional control will not settle at its target value, but will
retain a steady state error that is a function of the proportional gain and the process gain. Despite the
steady-state offset, both tuning theory and industrial practice indicate that it is the proportional term
that should contribute the bulk of the output change.
DERIVATIVE CONTROL
The actuating error signal for derivative control action consists of proportional error signal added with
derivative of the error signal. Thus the actuating signal is given as,
de(t )
ea (t ) e(t ) Td
dt
© Career Avenues A 61
IPC
(1 sTd )2n
T(s)
s2 (2n n2 Td )s n2
Or,
n Td
2
Thus, the effective damping is increased using the derivative control, resulting in decreasing
overshoot. The natural frequency is not changed. Now, the steady state error due to unit ramp input is,
1 (1 sTd )n2
ess lim ;G(s)
s0 sG(s) s(s 2n )
2
ess ,
n
Thus the steady state error is not affected by derivative control action. It is also seen that a zero has
been added at
s 1 / Td which will result in a small reduction in rise time, tr .
INTEGRAL CONTROL
The actuating error signal consists of proportional error signal added with integral of control action.
The actuating signal for integral control action is given by,
ea (t) e(t) Ki e(t)dt
Ki
In Laplace domain, Ea (s) E(s) E(s)
s
© Career Avenues A 62
IPC
(s K i )2n
T(s)
s3 2n s2 2n s K i 2n
It is seen that the system becomes a third order system due to inclusion of the integral control.
(s K i )n2
Here, G(s)
s 2 (s 2n )
1
Now, the steady state error due to unit ramp input is, ess lim 0 ;
s0 sG(s)
The block diagram of a second order system incorporating PID control is shown in figure below.
It has been estimated that 90% of the industrial controllers are of the PI type. PI and the PID
controllers will remain exist for some time and digital controllers will be widely used in future due to
their great flexibility.
Hence:
MV(t) = Pout + Iout + Dout
where Pout, Iout, and Dout are the contributions to the output from the PID controller from each of the
three terms, as defined below.
© Career Avenues A 63
IPC
PROPORTIONAL TERM
The proportional term makes a change to the output that is proportional to the current error value. The
proportional response can be adjusted by multiplying the error by a constant Kp, called the
proportional gain.
Where
Pout: Proportional term of output
Kp: Proportional gain, a tuning parameter
e: Error = SP − PV
t: Time or instantaneous time (the present)
Plot of PV vs. time, for three values of Kp (Ki and Kd held constant)
A high proportional gain results in a large change in the output for a given change in the error. If the
proportional gain is too high, the system can become unstable (See the section on loop tuning). In
contrast, a small gain results in a small output response to a large input error, and a less responsive (or
sensitive) controller. If the proportional gain is too low, the control action may be too small when
responding to system disturbances.
In the absence of disturbances, pure proportional control will not settle at its target value, but will
retain a steady state error that is a function of the proportional gain and the process gain. Despite the
steady-state offset, both tuning theory and industrial practice indicate that it is the proportional term
that should contribute the bulk of the output change.
© Career Avenues A 64
IPC
INTEGRAL TERM
The contribution from the integral term is proportional to both the magnitude of the error and the
duration of the error. Summing the instantaneous error over time (integrating the error) gives the
accumulated offset that should have been corrected previously. The accumulated error is then
multiplied by the integral gain and added to the controller output. The magnitude of the contribution of
the integral term to the overall control action is determined by the integral gain, Ki.
The integral term is given by:
Iout = Ki 0 e d
t
Plot of PV vs. time, for three values of Ki (Kp and Kd held constant), where
Iout: Integral term of output
Ki: Integral gain, a tuning parameter
e: Error = SP − PV
τ: Time in the past contributing to the integral response
The integral term (when added to the proportional term) accelerates the movement of the process
towards set-point and eliminates the residual steady-state error that occurs with a proportional only
controller. However, since the integral term is responding to accumulated errors from the past, it can
cause the present value to overshoot the set-point value (cross over the set-point and then create a
deviation in the other direction). For further notes regarding integral gain tuning and controller
stability, see the section on loop tuning.
DERIVATIVE TERM
The rate of change of the process error is calculated by determining the slope of the error over time
(i.e. its first derivative with respect to time) and multiplying this rate of change by the derivative gain
© Career Avenues A 65
IPC
Kd. The magnitude of the contribution of the derivative term to the overall control action is termed the
derivative gain, Kd.
Plot of PV vs. time, for three values of Kd (Kp and Ki held constant)
Where
Dout: Derivative term of output
Kd: Derivative gain, a tuning parameter
e: Error = SP − PV
t: Time or instantaneous time (the present)
The derivative term slows the rate of change of the controller output and this effect is most noticeable
close to the controller set point. Hence, derivative control is used to reduce the magnitude of the
overshoot produced by the integral component and improve the combined controller-process stability.
However, differentiation of a signal amplifies noise and thus this term in the controller is highly
sensitive to noise in the error term, and can cause a process to become unstable if the noise and the
derivative gain are sufficiently large
© Career Avenues A 66
IPC
MCQS
1. A controller action in which, there is a continuous linear relation between the value of the
controlled variable and the rate of change of controlled output signal is called
A) proportional action B) integral action
C) derivative action D) none of these
2. A controller action in which there is a constant relation between the value of the
controlled variable and the value of the output signal of the controller is called.
A) proportional action B) derivative action
C) integral action D) none of these
3. The time difference by which the output of a PD controller leads the input when the input
changes linearly with time is called.
A) Error ratio B) derivative time
C) proportional sensitivity D) gain
2e s
Gp(s) = (units of time is minutes)
s 1
2
is found to be 0.6 rad/min. Assume that the measurement and valve transfer functions are
unity. The time constant, τ (in minutes) is
A) 1.14 B) 1.92 C) 3.223 D) 5.39
5. If the response of a control system is to free of offset and oscillation, most useful
controller would be
A) Proportional B) Proportional – derivative
C) Proportional – integral D) Proportional – integral – derivative
© Career Avenues A 67
IPC
GATE 1997
Answer: (B)
GATE 1997
Answer: (A)
GATE 1996
The transfer function for an ideal proportional plus rest controller (Reset time T) is:
1
A) K c 1 B) K c 1 s
s
Kc Kc
C) D)
1 s 1
s
Answer: (A)
© Career Avenues A 68
IPC
GATE 1995
GATE 1995
The offset introduced by proportional controller with gain KC in response of first order system
can be reduced by
A) reducing value of Kc B) introducing integral control
C) introducing derivative control D) none of the above
Answer: (B)
GATE 2001
A PID controller output p(t), in time domain, is given by P(t) = 30 + 5e(t) + 1.25
de t
0 e t dt 15
t
.
dt
Where, e(t) is the error at time t. The transfer function of the process to be controlled is G p(s) =
10
.
200s 1
The measurement of the controlled variable is instantaneous and accurate.
C) D)
4s 3s
t de t
p(t) = 30 + 5e(t) + 1.25 e t dt 15
0 dt
p(s) = 30
p (t) = p(t) – p(s)
t de t
= 5 e(t) + 1.25 e t dt 15
0 dt
© Career Avenues A 69
IPC
p s 1.25
Gc = 5 15s
e s s
5
=5+ + 15s
4s
1
= 5 1 3s
4s
5 12s 4s 1
2
Gc =
4s
Answer:
Assume Gm (Measuring device) = 1
Characteristic equation
1 + GpGc = 0
5 12s 2 4s 1 10
1+ . 0
4s 200s 1
4s(200 s + 1) + 50 (12 s2 + 4 s + 1) = 0
800 s2 + 4 s + 600 s2 + 200 s + 0.50 = 0
1400 s2 + 204 s + 50 = 0
700 s2 + 102 s + 25 = 0
GATE 2006
A 2-input, 2-output process can be described in the Laplace transform domain as given below
Where U1 and U2 are the inputs and Y1 and Y2 are the outputs. The gains of the transfer
functions Y1(s)/U2(s) and Y2(s)/U2(s), respectively, are
A) K2 and K3 B) K1 and K3 + K2K4
C) K2 and K3 + K1K4 D) K2 and K3 + K2 K4
© Career Avenues A 70
IPC
GATE 2003
Suppose that the gain, time constant, and dead time of a process with the following transfer
function
Gc(s) = 10 exp(– 0.1s) / (0.5s + 1)
are known with a possible error of + 20% of their values. The largest permissible gain Kc of a
proportional controller needs to be calculated by taking the value of process gain, time constant
and dead time as
A) 8, 0.6, 0.08 B) 12, 0.6, 0.12 C) 8, 0.6, 0.12 D) 12, 0.4, 0.8
e0.1s
Gc(s) = 10
0.5s 1
error = ± 20%
20
Process gain = 10 1 = 10 × 1.2 = 12
100
Three constant = 0.5 × 1.2 = 0.6
Dead time = 0.1 × 1.2 = 0.12
ANSWER IS (B)
© Career Avenues A 71
IPC
SUMMARY
Transfer function of a proportional controller:
P s
Kc
(s)
On-off control: This is a special typer of proportional controller where the control valve is either fully
opened or fully closed. Proportional controller is some time referred as reset controller and the integral
time as reset time.
P s 1
K c 1 Ds
(s) 1s
© Career Avenues A 72
IPC
STABILITY
INTRODUCTION
The next requirement that will be discussed in designing a control system is stability. The total
response of a system is the sum of the forced and natural responses or:
c(t ) c forced (t ) cnatural(t )
Using these concepts, we present the following definitions of stability, instability and marginal
stability:
A linear, time-invariant system is stable if natural response approaches zero as time approaches
infinity.
A linear, time-invariant system is unstable if natural response grows without bound as time
approaches infinity.
A linear, time-invariant system is marginal stable if natural response neither decays nor grows,
but remains constant or oscillates as time approaches infinity. Thus the definition of stability
implies that only the forced response remains constant as the natural response approaches zero.
The stable systems have closed-loop transfer functions with poles only in the left half plane.The
unstable systems have closed-loop transfer functions with at least one pole in the right half-plane
and/or poles multiplicity greater than one on the imaginary axis.The marginally stable systems have
closed-loop transfer functions with only imaginary axis poles of multiplicity 1 and poles in the left
half-plane.
© Career Avenues A 73
IPC
ROUTH-HURWITZ CRITERION
In this section, we learn a method that yields stability information without the need to solve for the
closed-loop system poles.
Using this method (Routh-Hurwitz criterion), we know how many closed-loop system poles are in the
left half-plane, in the right half-plane and on the jw-axis. The method requires two steps:
Generate a data table called Routh-table
Interpret the Routh table to tell how many closed-loop system poles are in the left half-plane, in the
right half-plane and on the jw-axis.
Figure 7.1
We first create the initial layout for Routh table shown in Table 7.1
Table 7.1
4
S a4 a2 a0
3
S a3 a1 0
S2
S1
S0
© Career Avenues A 74
IPC
S4 a4 a2 a0
S3 a3 a1 0
S2 a4 a2 a4 a0 a4 0
a3 a1 a3 0 a3 0
b1 b2 0
a3 a3 a3
S1 a3 a1 a3 0 a3 0
b1 b2 b1 0 b1 0
c1 0 0
b1 b1 b1
S0 b1 b 2 b1 0 b1 0
c1 0 c1 0 c1 0
d1 0 0
c1 c1 c1
Ex 7.1
Make the Routh table for the system shown in Figure 7.2(a).
Figure 7.2
© Career Avenues A 75
IPC
Solution
First, we have to find the equivalent closed-loop system because we want to test the denominator of
this function. Using feedback formula, we obtain the equivalent system in Figure 7.2(b).Then we
create the initial layout Routh table and Table 6.3 is the completed Routh table.
Table 7.3
3
S 1 31 0
2
S 10 1 1030 103 0
S1 1 31 1 0 1 0
1 103 1 0 1 0
72 0 0
1 1 1
S0 1 103 1 0 1 0
72 0 72 0 72 0
103 0 0
72 72 72
Simply stated, the Routh-Hurwitz criterion declares that the number of roots of the polynomial that are
in the right half-plane is equal to the number of sign changes in the first column.
If the closed-loop transfer function has all poles in the left half of the s-plane, the system is stable.
Thus, a system is stable if there are no sign changes in the first column of the Routh table.
For example, Table 7.3 has two sign changes in the first column. The first sign change occurs from 1
in the s2 row to –72 in the s1 row. The second occurs from –72 in the s1 row to 103 in the s0 row. Thus,
the system of Figure 6.4 is unstable since two poles exist in the right half-plane.
Exercise
Make a Routh table and tell how many roots of the following polynomial are in the right half-plane
and in the left half-plane.
P(s) 3s 7 9s 6 6s 5 4s 4 7s 3 8s 2 2s 6
Answer: Four in the right half-plane (rhp); three in the left half-plane (lhp).
© Career Avenues A 76
IPC
If the first element of a row is zero, division by zero would be required to form the next row. To avoid
this phenomenon, an epsilon,
, is assigned to replace the zero in the first column. The value is
then allowed to approach zero from either the positive or negative side, after which the signs of the
entries in the first column can be determined.
Ex 7.2
Table 7.4
5
S 1 3 5
S4 2 6 3
7
S3 0 0
2
6 7
S2 3 0
Table 7.5 shows the first column of Table 7.4, along with the resulting signs for a choice of
positive and negative.
© Career Avenues A 77
IPC
Table 7.5
We can see, if we choose positive or negative, Table 7.5 will show two sign changes for both
conditions. Hence, the system is unstable and has two poles in the right half-plane.
Sometimes while making a Routh table, we find that an entire row consists of zeros because there is an
even polynomial that is a factor of the original polynomial.
Example 7.4
Solution:
© Career Avenues A 78
IPC
Table 7.6
S5 1 6 8
S4 7 1 42 6 56 8
S3 0 4 1 0 12 3 0 0 0
S2 3 8 0
1
S1 0 0
3
S0 8 0 0
At the second row, we multiply through by 1/7 for convenience. At third row, we can see that the
entire row consists of zeros, so we use the following procedure:
Return to the row above the row of zeros, form the auxiliary polynomial using the entries in that row
as coefficients. The polynomial will start with the power of s in the label column and continue by
skipping every other power of s. Thus the polynomial formed for this example is P(s) s 6s 8
4 2
Finally, we use the coefficients of this differentiate equation to replace the row of zeros.
There are no right-half-plane poles. For the case of zeros row, some of the roots could be on the jw-
axis. If we do not have a row of zeros, we cannot possibly have jw roots.
Example 7.5
Tell how many poles are in the right half-plane, the left half-plane, and on the jw-axis.
Solution
© Career Avenues A 79
IPC
Table 7.7
S8 1 12 39 48 20
S7 1 22 59 38 0
S6 10 1 20 2 10 1 20 2 0
S5 20 1 60 3 40 2 0 0
S4 1 3 2 0 0
S3 0 4 2 0 6 3 0 0 0 0 0
3
S2 3 2 4 0 0 0
2
1
S1 0 0 0 0
3
S0 4 0 0 0 0
Now we need to demonstrate the method’s application to a number of analysis and design problems.
Find the number of poles in the left half-plane, the right half-plane, and on the jw-axis for the system
of Figure 7.3.
Figure 7.3
© Career Avenues A 80
IPC
Solution
Table 7.9
4
S 1 11 200
S3 6 1 6 1
S2 10 1 200 20
S1 – 19
S0 20
The system is unstable, since it has two rhp and 2 lhp poles. The system cannot have jw poles since a
row of zeros did not appear in the Routh table.
Find the number of poles in the lhp, the rhp and on the jw-axis for the system in Figure 7.4.
Figure 7.4
Solution
© Career Avenues A 81
IPC
Table 7.10
S5 2 2 2
S4 3 3 1
4
S3 0
3
3 4
S2 1
12 16 3 2
S1
9 12
S0 1
Permitting to be small, positive quantity, we find that the first term of the s 2 row is negative. Thus,
there are two sign changes, and the system is unstable, with two poles in the rhp. The remaining poles
are in the lhp.
Find the number of poles in the lhp, rhp and on the jw-axis for the system of Figure 7.5. Draw the
conclusions about the stability of the closed-loop system.
Figure 7.5
Solution
Then, form the Routh table using the procedure of the entire row is zero (Table 7.11).
© Career Avenues A 82
IPC
Table 7.11
S8 1 10 48 128 128
S7 3 1 24 8 96 32 192 64
S6 2 1 16 8 64 32 128 64
S5 0 6 3 0 32 16 0 64 32 0 0 0
8 64
S4 1 8 64 24
3 3
S3 8 1 40 5
S2 3 1 24 8
S1 3
0
S 8
The summary of pole locations is shown Table 7.12. The system is unstable because of the rhp.
Table 7.12
Even Rest Total
(sixth-order) (second-order) (eighth-order)
2 rhp 0 rhp 2 rhp
2 lhp 2 lhp 4 lhp
2 jω 0 jω 2 jω
Note: rhp = right half-plane; lhp = left half-plane
An easy way to find the roots of a system when one parameter of the system is varied is to use root
locus technique, introduced by Evans. It is a graphical method of plotting the locus of roots in the s-
plane as a given system parameter is varied over a complete range of values. The roots corresponding
to a particular value of parameter or the value of the parameter for the desired root location can be
obtained from the root locus.
© Career Avenues A 83
IPC
The value of s which satisfies the above equation represents the roots or closed-loop poles of the
system. As s is a complex quantity, we can split the above equation into magnitude and phase part.
G(s)H(s) (2q 1)180 , where, q 0,1, 2, (1)
G(s)H(s) 1
and, (2)
Equation (1) is called angle condition and equation (2) is called magnitude condition. A plot of the
points of the complex plane satisfying the angle condition alone is the root locus. The gain of a system
corresponding to a point on the root locus is calculated using the magnitude condition.
Example
The above equation represents a circle with center at (-3, 0) and radius 3 .
© Career Avenues A 84
IPC
We can see that the analytical method using angle condition to determine root locus is very tedious.
Later we shall discuss how we can draw root locus graphically.
m
K s zi
n
i 1
1.
s p
p 1
i
Now,
Evans developed a graphical technique to draw the root locus of a system approximately. The
construction rule is given on the next pages.
© Career Avenues A 85
IPC
CONSTRUCTION RULES
Rule 2: Locate the open-loop poles in the s-plane. The root locus originates from an open-loop pole
with K 0 and terminates either on an open-loop zero or on infinity with K . The number of
branches terminating on infinity equals the number of open-loop poles minus zeros.
n m
(s p ) K (s z ) 0.
j 1
i
i 1
i
The characteristic equation may be written as,
When K 0 , s p j which are the open-loop poles. Thus root locus branches start at open-loop
poles.
1 n m
K j 1
( s pi )
i 1
( s zi ) 0.
The characteristic equation may also be written as,
When K , s zi which are open-loop zeros. Thus m root locus branches terminate on the open-
loop zeros.
j
If m n , the open-loop transfer function has (n - m) zeros at infinity. Putting s e in the
m
( s zi )
. This implies, K . Therefore, (n -
i 1 1 1 1
magnitude condition, we get ( n m )
n
K e K
( s pi )
j 1
m) branches of the root locus terminate on infinity.
Rule 3: Determine the root loci on the real axis. A point on the real axis lies on the root locus if the
number of open-loop poles plus zeros on the real axis to the right of this point is odd. This will be
evident from figure below and the angle condition,
m n
( s zi ) (s p j ) (2q 1)180 .
i 1 j 1
© Career Avenues A 86
IPC
s s s
Here, 0 is the test point on the real axis. The angle is zero to the left of 0 and 180 to the right of 0
for every pole and zero. The overall effect due to complex conjugate poles is zero. Thus the angle
(m n )180 (2q 1)180 .
condition on the real axis may be written as, right right Hence the rule.
Rule 4: Determine the angle of asymptotes of the (n - m) root loci branches which tend to infinity.
The angles of asymptotes are given by
(2q 1)180
A ; q=0,1,2,..., (n - m -1).
nm
Let us consider a point 0 on the root locus at infinity. It will essentially make same angle, say , to
s
each pole and zero. Using angle condition,
(2q 1)180
;
[G(s) H (s)] m n (2q 1)180 . Or, nm
Rule 5: Determine the centroid of the asymptotes. The asymptotes cross the real axis at a point
known as centroid on the real axis, determined by the following relationships:
A
real parts of poles - real parts of zeroes
n-m .
© Career Avenues A 87
IPC
K K
G ( s) H ( s) s nm
nm n m n
p j zi s nm1
m
s p j zi
j 1 i 1 s j 1 i 1
nm
n m
[ p j zi ]
j 1 i 1
Above equation will cut the real axis at = nm . This will be the equation of the centroid.
Rule 6: Find the breakaway and break-in points. The point at which multiple roots of
characteristic equation occur is called a breakaway point. The breakaway points of the root locus
dK
0
are the solutions of ds . r branches of root locus break away at an angle of 180 / r .
[The two root loci branches meet at breakaway point for some value of K as we vary it.]
Let the characteristic equation 1 G(s) H (s) 0 has a multiple root at s = - b of multiplicity r. Then,
1 G( s) H (s) ( s b)r A1 (s).
d
G( s) H ( s) 0.
At s b , the right-hand side of the equation is zero for r 2 . Therefore, at s b ds
A( s)
K
Again, from equation (01), B( s) . Differentiating K with respect to s we get,
dK A( s) B( s) A( s) B( s)
ds [ B( s)]2 ………………………………………………………..(04)
© Career Avenues A 88
IPC
dK
0
At the breakaway point ds which is evident from equation (03).
Rule 7: Determine the angle of departure from an open-loop pole. The angle of departure from an
(2q 1)180 ; q 0,1, 2,
open-loop pole is given by p where is the net angle
contribution, at this pole, of all other open-loop poles and zeros.
Similarly, the angle of arrival at an open-loop zero is given by, z (2q 1)180 ; q 0,1, 2, .
The angle of departure from a real open-loop pole or the angle of arrival at a real open-loop zero is
always 0 or 180.
Rule 8: Determine the points where root loci cross the imaginary axis. This can be determined by
use of the Routh criterion.
Rule 9: Sketch the root locus taking a series of test points in the broad neighborhood of the origin.
Rule 10: Determine closed-loop poles. A particular point on the root locus branch is a closed loop
pole if the value of K at that point satisfies the magnitude condition. Conversely, the magnitude
condition enables us to determine the value of the gain K at any specific root location on the locus.
Example 01
Consider the system as shown in figure below with the open-loop transfer function:
K s 1 s 2
G(s) H(s) =
s s 3 s 4
© Career Avenues A 89
IPC
From the rules described before, the following information concerning the root locus plot is easily
obtained.
Example 02
K
1 0.
Consider the system with characteristic equation, s( s 1)( s 2) (centroid, angle of asymptote
and real axis breakaway point)
© Career Avenues A 90
IPC
(2q 1)180
60 ,180 ,300
3. Angle of asymptote, (3 0) .
2 1
A 1
4. Centroid, 3 .
dK
(3s 2 6s 2) 0.
5. Breakaway point, ds
6. The intersection with imaginary axis may be found using Routh criteria.
Example 03
(Complex breakaway point)
K
G( s) H (s) .
Consider the open-loop transfer function s( s 4)( s 2 4s 20)
© Career Avenues A 91
IPC
(2q 1)180
3. Angle of asymptote, 45 ,135 , 225 ,315 .
(4 0)
4 4
4. Centroid A 2 .
40
K s ( s 4)( s 2 4s 20)
dK
(4s 3 24s 2 72s 80) 0
ds
5. s 2, 2 j 2.45.
There is one breakaway point on the real axis and two complex conjugate breakaway points shown
left.
180
6. The root locus branches leave the breakaway point at an angle of 90 as shown in
2
figure above.
7. The angle of departure from open-loop pole is, p 180 (45 135 90 ) 90 .
Example 04
© Career Avenues A 92
IPC
1. The angle of departure at complex pole is, 180 (45 90) 135 .
s 2 2s 2
K
s2
dK ( s 2)(2s 2) ( s 2 2s 2)
0
ds ( s 2) 2
Which gives, s 4s 2 0, or, s 3.14, 0.59. The first one is the actual breakaway
2
point.
Example 05
Application of Routh criteria to the above equation gives the following Routh array.
© Career Avenues A 93
IPC
1 36 K
s4
8 80
s3
1 10
s3
26 K
s2
260 K
s1 0
26
s0
K
For stability, K 0; (260 K ) / 26 0. The critical value of K 260 , makes the s row of the
1
Routh array zero. For this value of K, the auxiliary equation is, 26s K 0 .For K = 260, s j 10
2
Solution
47. Given the characteristic equation below, select the number of roots which will be located to the
right of the imaginary axis s4 + 5s3 – s2 – 17s + 12 = 0
A) One B) Two C) Three D) Zero
Solution
f(x) = s4 + 5s3 - s2 - 17s + 12 = 0
According to Descarte's rule of signs, maximum number of +ve roots = number of sign
changes is f(x)
In f(x), number of equation changes = 2
s4 + 5s3 – s2 – 17s + 12 = 0
(s4 – s3) + (6s3 – 6s2) + (5s2 – 5s) + (– 12s + 12) = 0
Or s3(s – 1) + 6s2(s – 1) + 5s(s – 1) – 12(s – 1) = 0
Or (s – 1) (s3 + 6s2 + 5s – 12) = 0
Or (s – 1) [(s3 – s2) + (7s2 – 7s) + (12s – 12)] = 0
© Career Avenues A 94
IPC
GATE 1995-1(R)
The root locus method, a pole of a transfer function G(s) is the value of a for which G(s)
approaches:
a] – 1 b] 0 c] 1 d] ∞
Answer is [d]
GATE 2009-42
The characteristic equation of a closed loop system using a proportional controller with gain K c
is 12s3 + 19s2 + 8s + 1 + Kc = 0. At the onset of intability, the value of Kc is
a] 35/3 b] 10 c] 25/3 d] 20/3
Answer is [a]
One of instability
19 8 12 1 K c
0
19
19 × 8 = 12(1 + Kc)
© Career Avenues A 95
IPC
19 8 38
1 + Kc =
12 3
35
Kc =
3
GATE 2010-43
GATE 2010-43
s2
Kc2
s4 Kc2 s 2
G1(s) =
1 Kc2
s2 s 4 Kc2 s 2
s 4
Kc2 s 2 Kc2 s 2
G1(s) =
s 4 sKc2 2Kc2 s 1 Kc 2 2 Kc 2 2
Y s Kc1G1 s
R s 1 Kc1G1 s
Kc2 s 2
Kc1
s 1 Kc2 2 Kc 2 2
=
Kc1Kc 2 s 2
1
s 1 Kc 2 2 Kc 2 2
Y s Kc1Kc2 s 2
R s s 1 Kc2 2 Kc2 2 Kc1Kc 2 s 2
Characteristic equation
© Career Avenues A 96
IPC
Routh Asray:
1 1.5Kc2 0
3Kc 4 0
2
For stability 3Kc2 – 4 > 0
4
Kc2 > = 1.33
3
1 1.5 K c2 0
K c2 1
1.5
K c2 0.66
& K c2 1.33
3
Nowif we put K c int o determinants
4
2.175 0
the
not stable vevalue
1.75 0
similarly for K c2 1
2.5 0
1 not stable , vevalue
0
similarly for K c2 5
4
2.875 0
0.25 not stable, vevalue
0
but if K c 3
2
3.25 0
0.5 0 stable , all vevalue
Ans d ) K c2 3 is correct
2
GATE 2007-63
The first two rows of Routh’s tabulation of a third order equation are
s3 22
s2 44
© Career Avenues A 97
IPC
Solution
22
n 1
th
44
n
th
0
Elements of nth row (i.e. 3rd row) and 1st column is given by
4 2 4 2
0
4
According to theorem of Routh’s test, if the one pair of roots is on the imaginary axis,
equidistant from origin and all other roots are in the left half plane, all the elements of the n th row
will vanish and none of the elements of preceding row will vanish.
The location of pair of imaging roots can be found by solving the equation,
C s2 + D = 0
where C and D are the elements of the array in the (n – 1)th row as read from left to right
∴ Here we get 2 s2 + 2 = 0
∴s=±j
∴ The equation has two roots on the j – axis at s = j & s = – j.
Answer is [b]
GATE 1999-2.17
The system is stable only when the roots have negative real parts
GATE 2010-5
Match the location of the poles/zeros in the s-plane, listed in Group I, with the system response
characteristics in Group II.
Group I Group II
© Career Avenues A 98
IPC
GATE 2002-1.14
Answer is [a]
© Career Avenues A 99
IPC
SUMMARY
A control system is said to be stable only when all the roots of the characteristic equation have
negative real parts.
If the characteristic equation of the system is a0sn + a1sn-1 + a2sn-2 + a3sn-4 +……….+ an = 0, then the
system unstable even if one of the coefficients of the above equation is inerently negative and no need
to perform stabality check on the system. If the all the coefficients are positive then the system can be
checked for stability by using Routh stablity criterion.
According to the Routh stability criterion the system is stable only if all the element of first column of
the Routh array be positive and non zero.
If some of the elements in the first column are negative, the number of roots with a positive real parts
is equal to the numbe of sign changes in the first column.
If one pair of roots is on the imaginery axis, then all the elements of the n th row vanishes and none of
the elements of the preceding rows will vanishes.
The Routh criterion is a simple algebraic test for detecting roots of a polynomial lying in the right half
of the complex plane.
Root locus is a graphical method to find the actual values of the roots of the characteristic equation.
However this method also suffers the limitation of inapplicability for the systems with transportation
lag.
BODE PLOTS
INTRODUCTION
Frequency response methods were discovered by Nyquist and Bode in the 1930s. Frequency response
yields a new vantage point from which to view feedback control system. This technique has distinct
advantage in the following situations:
When modeling transfer function from physical data.
When finding the stability of nonlinear systems.
In settling ambiguities when sketching a root locus.
Ex 1:
Demonstrates how to obtain analytical expression for frequency response and make a plot of the
result.
Find the analytical expression for the magnitude freq response and the phase freq response for a
system G(s) = 1/(s+2). Also, plot both the separate magnitude and phase diagrams.
Solution
Substitute s = j in the system and obtaining G(j ) = 1/(j + 2)
G(j ) = 1/(j + 2) = (2 – j )/( 2 + 4)
Figure 1
Then, the magnitude frequency response is the product of the magnitude freq response of each
term, or
K ( s z1 ) s z 2 ... s z k
G( j ) m s j
s ( s p1 ) ( s p 2 ) ... ( s p n )
Thus, if we know the magnitude response of each pole and zero term, we can find the total
magnitude response. The process can be simplified by working with the logarithm of the
magnitude since the zero terms’ magnitude responses would be added and the pole terms’
magnitude responses subtracted, rather than, respectively, multiplied or divided, to yield the
logarithm of the total magnitude response. Converting the magnitude response into dB, we
obtain
Thus, if we knew the response of each term, the algebraic sum would yield the total response in
dB. Further, if we could make an approximation of each term that would consist only of straight
lines, graphic addition of terms would be greatly simplified.
s = jω, we have
G( j ) j a a j 1
a
a a
where a . Notice from the middle term that the high-frequency approximation is equal to
the low frequency approximation when a , and increases for a .
If we plot dB, 20log M , against log ω, equation
20log M 20log a 20log 20log becomes a straight line: y 20x
a
This function has a low-frequency asymptote of 20 log(1/a), which is found by letting the
frequency, s, approach zero. The Bode plot is constant until the break frequency, a rad/s, is
reached. The plot is then approximated by the high-frequency asymptote found by letting s
approach . Thus, at high frequencies
1
G j
1 1 1
a 90 0 90 0
s j
a a
a s j a a
or, in dB,
1
20 log M 20 log 20 log 20 log
a a
Figure 3
Notice from the middle term that the high-frequency approximation equals the low-frequency
approximation when a , and decreases for a . This result is similar to equation
20log M 20log a 20log 20log , except the slope is negative rather than positive.
a
The Bode log-magnitude diagram will decrease at a rate of 20 dB/decade rather than increase at a
rate of 20 dB/decade after the break frequency.
The phase plot is the negative of the previous example sine the function is the inverse. The phase
begins at 00 and reaches -900 at high frequencies, going through -450 at the break frequency. Both
the Bode normalized and scaled log-magnitude and phase plot are shown in Figure 4 (d).
Our next function, G(s) = s, has only a high-frequency asymptote. Letting s=jω, the magnitude is
20 log ω, which is the same as equation 20log M 20log a 20log 20log . Hence, the
a
Bode magnitude lot is a straight line drawn with a +20 dB/decade slope passing through zero dB
when ω = 1. The phase plot, which is a constant +900, is shown with the magnitude plot in
Figure 4(a).
The frequency response of the inverse of the preceding function, G(s) = 1/s, is shown in Figure
3(b) and is a straight line with a -20 dB/decade slope passing through the zero dB at ω = 1. The
Bode phase plot is equal to a constant -900.
Figure 4
Problem: Draw the Bode plots for the system shown in figure below, where
Gs K s 3 /ss 1s 2
Solution:
Use the normalized plot in order to determine the cut-off frequency easier.
The magnitude plot should begin a decade below ( 0.1 ) the lowest freq break and extend a
decade above the highest break freq ( 100 ).
Figure 5 (magnitude)
Figure 5 (Phase)
At low frequency:
Gs n n 0 0
2 2
At high frequency:
Gs s 2 2 2 180 0
The log-magnitude:
20 log 2 40 log
Magnitude-phase plots;
Figure 6
For G s 1/ s2 2ns n 2
© Career Avenues A 112
IPC
Magnitude-phase plots:
Solution
Figure 7
DETERMINING STABILITY
The specification below should be get in order to ensure the stability of the system using Bode
Plots. The closed loop system will be stable if the frequency response has a gain less than unity
when phase is 1800.
Ex 4:
Use Bode Plots to determine the range of K within which the unity feedback system is stable. Let
Gs
K
s 2s 4s 5
Solution:
Plots,
Figure 8 (Ex 4)
At this point magnitude plot is at -20db. Then the system is stable. The magnitude can move
until +20db in order ensure the stability of the system.
Then, at this point the gain K = 40 * 10 = 400. So, range of K that make the system stable is
0K 400
Evaluating Gain Margin and Phase Margin. Refer Figure 9 for gain margin and phase margin
concept;
Ex 5:
Solution:
Plots the Bode plots for magnitude and phase using the same procedure as Ex 4
Then, 20 log 5 = 13.98db. Means bode plots in Figure 8 should increase 13.98db in magnitude.
Therefore, to find gain margin look at phase plot and the freq when phase is 180 o. At this freq,
determine from the magnitude plot how much the gain can be increased before reaching 0dB.
From Figure 8, the phase angle is 180o at approximately 7 rad/s. On the magnitude plot, the gain
is – 20 dB + 13.98 dB = – 6.02 dB.
For phase margin, refer freq at magnitude plot where the gain is 0 dB. At this freq, look on the
phase plot to find the difference between the phase and 180 o. The difference is phase margin.
From Figure 8, remember that the magnitude plot is 13.98 dB lower than actual plot, the 0 dB
crossing (-13.98 dB for the normalized plot shown in Figure 8) occurs at 5.5 rad/s.
At this freq, the phase angle is – 165o. Thus, the phase margin is – 165 – (– 180) = 15o
A relationship exists between the peak value of the closed-loop magnitude response and the
damping ratio is given by equation below,
1
Mp
2 1 2
p
This condition happens at a frequency , of
p n 1 2 2
GATE 2002
A first order system with unity gain and time constant is subjected to a sinusoidal input of
frequency 1 / .
GATE 2002
The frequency response of a first order system, has a phase shift with lower and upper bounds
given by
a] , b] , c] , 0 d] 0,
2 2 2 2 2
Correct answer is [d]
GATE 2000
GATE 1995
Bode diagrams are generated from output response of the system subjected to which of the
following input –
a] Impulse b] step c] sinusoidal d] ramp
GATE 2010
The transfer function, G(s), whose asymptotic Bode diagram is shown below, is
a] 10s + 1 b] s – 10 c] s + 10 d] 10s – 1
s + 10 correct answer
s
10 10
10
N K s 1 / 1
1 1 K G
wc D 10 0.15 1
wc
s 10
0.1
GATE 1995
According to Bode stability criterion. A system is unstable if the open loop frequency response
exhibits an amplitude ratio exceeding unity at frequency for which phase lag is –
a] 00 b] 450 c] 900 d] 1800
© Career Avenues A 119
IPC
GATE 2004
1
Consider a system with open loop transfer function G s
s 1 2s 1 5s 1
Match the range of (frequency) in Group I with the slope of the asymptote of the log AR
(amplitude ratio) versus log plot in Group II.
Group I Group II
P. 0 < < 0.2 1. – 5
Q. > 0 2. – 3
3. – 2
4. – 1
5. zero
a] P – 5, Q – 2 b] P – 4, Q – 2 c] P – 5, Q – 3 d] P – 4, Q – 1
80 Answer
GATE 2008
If the control loop is to operate at a gain margin of 2.0, the gain of the proportional controller
must equal
a] 0.85 b] 2.87 c] 3.39 d] 11.50
Solution:
2K c
AR = = 0.1745 Kc
5.39 0.62 1
2
1 1
G.M.
AR 0.1745K c
1
2.0 =
0.1745K c
KC = 2.864 2.87
SUMMARY
The frquency response of a process is the response of the process to sinusoidal forcing functions.
Amplitude ratio (AR) and Phase angle (Φ):
1
AR = tan 1
1
2 2
1
AR = tan 1 1 tan 1 2
1 r 1
2 2
1
2 2
2
1 2
AR = tan 1
1 2 1
2
2 2 2 2
Transportation lag:
Φ = Φ1 + Φ2 + Φ3 + …… + Φn
The frequency at which both the low frequency and high frequency asymptotes intersect is
ωc = 1/τ
In the literature of contro theory the amplitude ratios are reported in decibels db as,
ZN SETTINGS
Ziegler-Nichols Controller Settings
Consider selection of a controller Gc for the general control system of Fig. 17.5. We first plot the
Bode diagram for the final control element, the process, and the measuring element in series,
G1G2H(jω). It should be emphasized that the controller is omitted from this plot. Suppose the
diagram appears as in Fig. 17.6. As noted on the figure, the crossover frequency for these three
components in series is ωco. At the crossover frequency, the overall gain is A, as indicated.
According to the Bode criterion, then, the gain of a proportional controller which would cause
the system of Fig. 17.5 to be on the verge of instability is 1/A. We define this quantity to be the
ultimate gain K. Thus
1
Ku = (17.3)
A
The ultimate period Pu is defined as the period of the sustained cycling that would occur if a
proportional controller with gain Ku were used. From the discussion α Fig.17.3, we know this to
be
2
Pu = time/cycle (17.3a)
co
The factor of 2π appears, so that Pu will be in units of time per cycle rather than time per radian.
It should be emphasized that Ku and Pu are easily determined from the Bode diagram of Fig.
17.6.
The Ziegler-Nichols settings for controllers are determined directly from Ku and Pu according to
the rules summarized in Table 17.1. Unfortunately, specifications of K c and D for PD control
cannot be made using only Ku and Pu. In general, the values 0.6 Ku and Pu/8, which correspond to
the limiting case of no integral action in a three-mode controller, are too conservative. That is,
the
Table 17.1
Ziegler-Nichols Controller Settings
Type of Control Gc(s) Kc l D
Proportional Kc 0.5 Ku
1 Pu
Proportional-integral (PI) K c 1 0.45 Ku
1s 1.2
1 Pu Pu
Proportional-integral-derivative (PID) K c 1 Ds 0.6 Ku
1s 2 8
resulting system will be too stable. There exist methods for this case which are in principle no
more difficult to use than the Ziegler-Nichols rules. One of these is selection of D for maximum
Kc at 300 phase margin, which was discussed above. Another method, which utilizes the step
response and avoids trial and error, is presented in Chap. 19.
The reasoning behind the Ziegler-Nichols selection of values of Kc is relatively clear. In the case
of proportional control only, a gain margin of 2 is established. The addition of integral action
introduces more phase lag at all frequencies (see Fig. 16.10); hence a lower value of K c is
required to maintain roughly the same gain margin. Adding derivative action introduces phase
lead. Hence, more gain may be tolerated. This was demonstrated in Example 17.2. However, by
and large the Ziegler-Nichols settings are based on experience with typical processes and should
be regarded as first estimates.
Example 17.3. Using the Ziegler-Nichols rules, determine Kc and l for the control
system shown in Fig. 17.10.
For this problem, the computation will be done without plotting a Bode diagram;
however, the reader may wish to do the problem with such a diagram. We first obtain the
crossover frequency by applying the Bode stability criterion:
– 1800 = – tan – ‘(0) – 57.3(1.02) (ω)
The value 57.3 converts radians to degrees. Solving this equation by trial and error gives
for the crossover frequency. ωco = 2 rad/min. The amplitude ratio (AR) at the crossover
frequency for the open loop can be written
1K c K
AR = 1 c
1 2 2.24
where we have used Eq. (16.16) for the first-order system and the fact that the amplitude
ratio for a transport lag is one. According to the Bode criterion, the AR is 1.0 at the crossover
frequency when the system is on the verge of instability. Inserting AR = 1 into the above
equation and solving for Kc gives Kcu = 2.24. From the Ziegler-Nichols rules of Table 17.1, we
obtain
Kc = 0.45 Kcu = (0.45) (2.24) = 1.01
and
1 Pu / 1.2 [2 / co ] / 1.2 [2 / 2]1.2 2.62min.
Figure 17-10
Block diagram for Example 17.3
Ans: a
3. The value of ultimate period of oscillation P u is 3 minutes, and that of the ultimate
controller gain Kcu is 2. Select the correct set of tuning parameters (controller gain K c,
the derivative time constant τD in minutes, and the integral time constant τ1 in minutes)
for a PID controller using Ziegler-Nichols controller settings.
A) Kc = 1.1; τ1 = 2.1, τD = 1.31 B) Kc = 1.5; τ1 = 1.8, τD = 0.51
C) Kc = 15; τ1 = 1.8, τD = 0.51 D) Kc = 1.2; τ1 = 1.5, τD = 0.38
Solution
GATE 2008
Math the following
Group 1 Group 2
P. Ziegler Nichols (1) Process Reaction Curve
Q. Under damped response (2) Decay ratio
R. Feed-forward Control (3) Frequency Response
(4) Disturbance measurement
(A) P – 3, Q – 2, R – 4 (B) P – 1, Q – 2, R – 3
(C) P – 3, Q – 4, R – 2 (D) P – 1, Q – 4, R – 2
GATE 2008
Statement for Linked Answer Questions
The cross-over frequency associated with a feedback loop employing a proportional controller to
control the process represented by the transfer function
2e s
Gp(s) = , (units of time are minutes)
s 1
2
is found to be 0.6 rad/min. Assume that the measurement and valve transfer functions are
unity.
Q. The time constant, (in minutes) is
(A) 1.14 (B) 1.92 (C) 3.23 (D) 5.39
Solution:
Gc = Kc
2K c e S
∴ GOL =
S 1
2
Put S = jω
2K c e j
GOL =
j 1
2
2K c
AR
22 1
2tan 1
At ωCO = 0.6 rad/min = – 1800
– 1800 = – 0.6 – 2 tan-1 ω = 0.6
ωCO → crossover frequency
= 5.39 min.
SUMMARY
According to bode stability criterion “A control system is unstable if the open loop frequency
response exhibits an AR exceeding unity at the frequency for which the phase angle is 180 o”.
Gain margin:
1
Gain margin =
A
Typical specifications for the designs are that the gain margin should be greater than 1.7. A gain
margin of unity or less indicates an unstable system.
Phase margin: It is the difference between 180o and the phase lag at the frequency for which the
gain is unity
Typical specifications are that the phase margin must be greater than 30 o.
'
x d1(s)
Gd1(s)
'
x d2(s)
Gd2(s)
'
x m (s) y' (s)
Gm(s)
y' (s) Gm x'm (s) Gd 2 x'd 2 (s) Gd1x'd1 (s) (9.1-1)
We imagine a case in which process Gm(s) could be divided into two parts, connected by a
measurable intermediate variable xi: this could be as simple as two tanks in series, as in Lesson
4. Having specified some of the interior structure of G m, we consider xd2 to be typical of
disturbances that affect the process further upstream and xd1 to affect the process downstream,
after the intermediate variable.
'
x d1(s)
Gd1(s)
'
x d2(s)
Gd2a(s)
'
x m (s) x'i (s) y' (s)
Gm2(s) Gm1(s)
and
G G G (9.1-3)
d2 d2a m1
Also, the intermediate variable is given by xi' (s) Gm2 x'm (s) Gd 2a x'd 2 (s)
(9.1-4)
RESPONSE TO DISTURBANCES
Suppose, for illustration, that we let each of these transfer functions be first order. (9.1-5)
Then the responses of xi and y to a step in xd2 are shown in Figure 9.2-1.
The controlled variable is y. The manipulated variable affects the controlled variable through the
transfer function. These assignments are familiar from previous Lessons. However, we have
some new assignments to make:
By looking in more detail at the composition of this process, we identify an intermediate
variable that influences the controlled variable, responds to disturbances before the controlled
variable does, and responds to the manipulated variable, as well. We can use this extra
information in a scheme called cascade control.
Furthermore, suppose we can measure a disturbance variable that frequently disturbs the
process. By this means, we can forecast when the controlled variable is about to be altered
and forestall it in a scheme called feedforward control.
The idea is to insert a secondary feedback control loop between the controlled variable y and
manipulated variable xm. The secondary loop controls intermediate variable xi. This variable
must hold several qualifications:
it must respond to important disturbances (those that significantly affect the controlled
variable)
it must also convey the effects of such disturbances to the controlled variable
it must respond to the manipulated variable
Gd1(s)
'
x d2(s)
Gd2a(s)
'
x m(s) x'i (s) y' (s)
Gm2(s) Gm1(s)
x'co(s) - -
Gc2(s) Gc1(s)
y'sp(s)
Gsp(s)
The secondary loop controls the intermediate (secondary) variable xi by adjusting the
manipulated variable xm. The primary loop controls the controlled variable y by manipulating
the set point of the secondary controller xo1. Thus we have the same controlled variable and set
point as before, but the valve has been augmented by an inner control loop.
Disturbances xd2 ′ are rejected by the secondary loop before they affect the full process, and thus
response is quicker and the impact on y′ less. The primary loop is necessary to handle the other
disturbances, such as xd1′, that always exist. The extra layer of control does not degrade the
response to xd1′, because the process is usually much slower than the controller.
Cascade control is still feedback control, performed with conventional PID control algorithms.
The improvement comes because we're looking inside the process, discriminating among
disturbances, and applying feedback with increased deftness.
We get closer to the root of the problem if we react directly to the disturbance, predicting what
the manipulated variable should do, not waiting for a process response. This is the topic of feed
forward control. We contrast simple feedback control in Figure 9.6-1 with feedforward control in
Figure 9.6-2:
'
x d1(s)
Gd1(s)
'
x d2(s)
Gd2(s)
'
x m(s) y' (s)
Gm(s)
Gv(s) Gs(s)
-
Gc(s)
y'sp(s)
Gsp(s)
Figure 9.6-1 Feedback control diagram
In feedback control, disturbance xd2 proceeds through the process (Gd2) to affect controlled
variable y. The controller reacts to the resulting error and adjusts the manipulated variable; the
change in manipulated variable proceeds through the process (G m) after the fact to reduce the
error.
'
x d1(s)
Gd1(s)
'
x d2(s)
Gd2(s)
'
x m(s) y' (s)
Gm(s)
Gsf(s) Gv(s)
Gff(s)
In feedforward control, disturbance xd2 proceeds in parallel through the process (Gd2) and
through the feedforward controller (Gff). The controller adjusts the manipulated variable to
counteract the disturbance, so that disturbance and manipulated variable affect output variable y
together. In the very best of cases, the manipulated variable would compensate the disturbance
step for step, so that the controlled variable would never be affected!
We also contrast feedforward with the cascade structure of Section 9.5. Feedforward control also
adds another sensor and controller. However, the concept differs from that of cascade in that the
disturbance is measured, but the manipulated variable does not affect it - there is no feedback.
Feedforward is thus more specific than cascade control: it is designed to head off a particular
disturbance. However, it cannot measure how well it did, nor can it respond to other
disturbances, such as xd1, that might affect the controlled variable. Hence feedforward is to be
applied in conjunction with a conventional feedback loop. Both the feedback and feedforward
controllers adjust the manipulated variable in Figure 9.6-3.
'
x d1(s)
Gd1(s)
'
x d2(s)
Gd2(s)
'
x m(s) y' (s)
Gm(s)
G (s) G (s)
sf v Gs(s)
-
Gff(s) Gc(s)
Gsp(s)
y'sp(s)
The feedforward algorithm is not conventional PID. Rather it is specific to the process and the
disturbance. We wish to specify the feedforward controller transfer function G ff to minimize the
effect of xd2 on y. Ideally, we want y′(s) = 0, and from Figure 9.6-2 or 9.6-3 this requires that
G −Gd 2
ff
G GG (9.6-1)
sf v m
The parallel path through the feedforward controller makes use of advance warning about the
disturbance. Given perfect process models, plus the ability to render those in transfer function
Gff, the compensation can completely negate the effect of xd2. Of course, perfection is unlikely,
as we will see later.
We should conclude the comparison with a clarification of concept: think of “cascade” and
“feedforward” as ways to arrange controllers, not confined to the specific arrangements given
above. That is, it is possible.
To have a feedforward controller that adjusts the set point of a feedback controller. In this
circumstance, we have both feedforward control and a cascade structure; we might say that the
feedforward controller “cascades” to a secondary controller. We will show an example in the
next section.
similarities
primary controller attempts to correct deviations from set point
primary controller set point unchanged
manipulated variable unchanged
add extra sensor and controller
differences cascade (Figure 9.5-1) feedforward (Figure 9.6-3)
extra measurement: intermediate variable disturbance variable
characteristics of the new intermediate variable measured disturbance
measured variable: is affected by disturbances is not affected by
and manipulated variable manipulated variable
affects controlled variable affects controlled variable
algorithm of added controller: PID specific to process model
manipulated variable: directed by secondary directed by both feedforward
controller and feedback controllers
primary controller action: varies set point of secondary varies manipulated variable, as
controller before
helps by: reducing the degree of anticipating effects of a
disturbance that reaches the particular disturbance and
controlled variable responding to compensate
A feed stream is pre-heated using a heating oil in a shell-and-tube exchanger. The outlet
temperature is controlled by manipulating the flow rate of the oil.
T F T
F T
The process is subject to several disturbances: the flow rate and inlet temperature of both process
(feed) and service (heating oil) streams may vary. The latter disturbances are particularly
troublesome; because the heating oil is supplied from a header that feeds other, larger users,
swing in the supply pressure and temperature are frequent. We propose employing additional
measurements and control loops, arranged as shown in Figure 9.8-2.
T F T
F T
Flow control of the service stream is now affected by a secondary loop. Two primary controllers
cascade to this new secondary loop. The first is the original feedback controller for the process
temperature. The second is a feedforward controller from the oil temperature.
Should the oil supply pressure decrease, the secondary flow controller will respond by opening
the supply valve. This will allow the oil flow rate to return to its desired value. The cascade
structure responds to the pressure disturbance before the controlled variable is affected.
Should the oil supply temperature decrease, the feedforward controller will respond by directing
the secondary controller to increase the flow rate. The higher flow rate of heating oil will tend to
maintain the heat duty in the heat exchanger even as the supply temperature falls.
The temperature controller on the oil requires a feedforward algorithm because the manipulated
variable (heating oil flow) does not affect the measured variable (heating oil temperature). The
temperature controller on the process stream is a feedback controller that responds to any error in
that temperature, such as may arise from disturbances in the process stream flow and
temperature. Both controllers cascade to a secondary controller.
K d ms 1 m
G ff e d
(9.9-1)
K m d s 1
We generalize this form by associating a controller gain with the ratio of process gains, a lead
time with the xm process path, a lag time with the xd2 process path, and a dead time with the
difference between disturbance and manipulated process dead times. These four parameters may
then be tuned, as with PID parameters, to improve controlled variable response to
T s 1 ff
G ff K ff lead e (9.9-2)
Tlags 1
The effect of gain is to amplify the controller response to an input. The gain satisfies the steady
state relationship between disturbance and manipulated variables. The effect of dead time is to
delay the controller's response, so that it will not affect the controlled variable prematurely.
Such a step is appropriate if the disturbance dead time exceeds that of the process. However, if
the disturbance dead time is less, perfect control requires a negative θff, which implies predicting
the onset of future disturbances! Such a time-machine would be very useful, for many purposes,
but we are not likely to find one - thus the parameter θff would be set to zero.
We cab illustrate (9.9-2) by calculating the output of the controller upon receiving a step input of
magnitude A. Applying the step change and inverting we find
Tlead Tlag t
x co ' AK ff 1
/ Tlag
ff
e (9.9-3)
Tlag
The lead/lag elements shape the development of the response, as shown in Figure 9.9 -1. For
example, if Tlag is set to be greater than T lead, the controller output increases with time. This is
appropriate if the disturbance propagates more slowly through the process than does the
manipulated variable. If the reverse is true, then a more vigorous manipulated variable response
is in order from the start, and so Tlead is set larger than Tlag.
Each controller specifies, at any time, how much the valve should be moved from its previous
position. These two outputs are summed and then sent to the valve transducer.
It may be that the two controllers will oppose each other, so that the sum of the outputs is a
smaller movement than each individual controller would have directed. This is not necessarily a
problem: each controller responds in its own area of expertise (feedback responding to present
error; feedforward forecasting the effects of present disturbance) and the sum of the outputs
addresses both concerns.
Of course, the sum of the controller outputs could fall outside the physical 0 - 100% range of the
valve. In this case, the valve can move only to its limit in response. If this is a frequent
occurrence, the manipulated variable may be insufficiently strong to overcome disturbances.
FLOW CONTROL
As we have already seen in examples, a very common application of a cascade scheme is to use a
flow controller as an inner cascade loop. This allows the flow to be less affected by, for example,
pressure differences, so that it can be a more reliable manipulated variable in the outer loops.
We examine the flow control loop. The equipment comprises a valve, a flowmeter, and
connecting pipe.
If we regard the valve as the final control element and the flowmeter as the sensor, there is really
no process left. Hence a block diagram will show the process as a unity transfer function: the
The valve affects the flow through the stem position, as described in transfer function G v. The
diagram also identifies two disturbances that affect flow: a change in the pressure difference
across the valve (as might result from variations in sources and sink conditions) and changes in
physical properties of the flowing fluid. All these transfer functions come, as in other examples,
from a linearized model of flow through a valve.
After our emphasis on distinguishing manipulated and controlled variables, having the
manipulated variable be identical to the controlled variable may seem peculiar. An alternative
point of view is to consider the manipulated variable to be the valve stem position, as shown in
Figure 9.11-3.
In effect, the transfer function that describes how valve stem position affects flow has moved
from Gv in Figure 9.11- 2 to Gm in Figure 9.11-3. In this series of Lessons, we have preferred the
point of view of Figure 9.11-2; that is, a controller acts through a final control element to
produce a flow, which may then be used to manipulate a process output.
CASCADE PERFORMANCE
From Figure 9.5-1, we derive the transfer function for the closed loop cascade structure. With
nested loops in block diagrams, it is best to begin with the inner loop.
X i' (S) = Gd2 X 'd 2 (S) + Gm2 X 'm (S)
= Gd2 X 'd 2 (S) + Gm2GvGc2 X'col S Gs2Xi' S
so that
G d2 G GG
X i' (S) = x 'd2 s m2 v c2 x col
'
s
1 G L2 1 G L2
Where the transfer function around the secondary loop is
G\L2 = Gm2GvGc2Gs2
G mlG d 2 ' G G GG
= G dl x 'dl s x d 2 s ml m2 v c2 x'col s
1 G L2 1 G L2
1 G L2 1 G L2
Solving for y’
G 1 G L2 ' G mlG d2 G G GG G G
y’(s) = dl x d1 s '
x d2 s ml m2 v c2 cl sp ysp' s
1 G L2 G L1 1 G L2 G L1 1 G L2 G L1
where the transfer function around the primary loop is
GL1 = Gm1Gm2GvGc2Gc1Gs1
Above Equation is general for a two-loop cascade structure with the disturbances arranged as in
Figure 9.5-1. It can be specialized by substituting particular transfer functions for the
components in the cascade.
CASCADE TUNING
Marlin (2000) suggests tuning the inner loop first, with the outer loop in manual setting. Then
tune the outer loop with the inner loop in automatic setting.
y' (s) Gd1x'd1 (s) Gd 2 x'd 2 (s) Gm x'm (s)
Gd1x'd1 (s) Gd 2 x'd 2 (s) GmGv x cof (s) x co (s)
' '
(9.15-1)
Gd1x'd1 (s) Gd 2 x'd 2 (s) G mG vGff Gsf x d2 (s) GmGv Gc Gsp ysp (s) −Gsp y (s)
' ' '
GL = GmGvGcGs (9.15-3)
Notice that the feedforward controller affects only the transfer function for disturbance x d2; other
disturbances and the set point have the usual feedback loop transfer functions. By (9.6-1) we try
to make the xd2 transfer function zero. However, if it is not, the feedback controller is also
available, through the transfer function denominator, to respond to xd2, as it does to other
disturbances.
Because the characteristic equation obtained from (9.15-2) does not depend on the feedforward
controller, adding a feedforward loop has no effect on tuning the feedback controller for closed
loop stability. Even so, if the feedforward controller itself is unstable, the ensemble is likely to be
inoperable.
FEEDFORWARD TUNING
Ideally all the other disturbances would subside so that the feedback loop could be put in manual
and the feedforward loop tuned to respond to its particular disturbance. One may not be so
fortunate, however.
CONCLUSION
By going to the trouble and expense of extra measurements, and obtaining deeper knowledge of
the process, we are enabled to improve on the performance of the single-loop PID controller.
This is not to say that cascade or feedforward enhancements are always to be recommended -
they must be technically feasible (the measurements and character of the new variables being
appropriate) and economically justified. This of course, is a familiar story to engineers.
Feedforward and cascade point the way to further control schemes that make use of process
models. Various forms of model-based control can offer advantages in single control loops, but
really come into their own when we consider how individual control loops may interact - that is,
when realistic process models are MIMO - multiple input/multiple output.
GATE 2003
Water is flowing through a series of four tanks and getting heated as shown in Figure. It is
desired to design a cascade control scheme for controlling the temperature of water leaving the
Tank 4 as thee is a disturbance in the temperature of a second stream entering the Tank 2. Select
the best place to take secondary measurement for the secondary loop.
GATE 2004
The process and disturbance transfer functions for a system are given by
y s 2 y s 1
GP s Gd s
m s 2s 1 5s 1 d s 2s 1 5s 1
The feed forward controller transfer function that will keep the process output constant
for changes in disturbance is
2s 1 5s 1
2 2
2
A) B)
2s 1 5s 1
2 2
2
C)1/2 D) (2s + 1) (5s + 1)
Answer
G p Gd
2s 1 Ss 1
2
2
2s 1 Ss 1
2
2
2s 1 s 1
2
2
1
Correct answer is (B)
GATE 2005
Cascade control comes under the control configuration, which uses
A) one measurement and one manipulated variable
B) more than one measurement and one manipulated variable
C) one measurement and more than one manipulated variable
D) more than one measurement and more than one manipulated variable
GATE 2011
The following diagram shows a CSTR with two control loops. A liquid phase, endothermic
reaction is taking place in the CSTR, and the system is initially at steady state. Assume that the
changes in physical properties of the system are negligible.
GATE 2007
Consider the following instrumentation diagram for a chemical reactor, C sp represents a
concentration setpoint.
Column A Column B
P. control strategy 1) feed forward control
Q. primary control variable 2) cascade control
R. slowest controller 3) concentration in the reactor
S. fastest controller 4) reactor temperature
5) jacket temperature
6) concentration controller
7) reactor temperature controller
8) jacket temperature controller
9) flow controller
A) P – 2, Q – 3, R – 6, S – 9 B) P – 1, Q – 4, R – 8, S – 7
C) P – 10, Q – 7, R – 9, S – 6 D) P – 1, Q – 8, R – 5, S – 9
GATE 2003
In the case of a feed forward control scheme, which of the following is NOT true?
1. It is insensitive to modeling errors
2. Cannot cope with unmeasured disturbances
3. It waits until the effect of the disturbance has been felt by the system before control
action is taken
4. Requires good knowledge of the process model
5. Requires identification of all possible disturbances and their direct measurement
SUMMARY
Cascade controls system is an advanced process control where there are multiple measuring
elements as well as controllers. Usually the response or output from one controller is used to
adjust the set point of the other controller. Under these conditions the primary controller can
operates the final control element (control valve) of the secondary controller. Cascade control
system will have a quick response towards the changes in the processes, and are recommended in
the place of normal controllers of sluggish response.
Combined feed forward plus feedback control can significantly improve performance over
simple feedback control whenever there is a major disturbance that can be measured before
it affects the process output.
In the most ideal situation, feed forward control can entirely eliminate the effect of the measured
disturbance on the process output.
Even when there are modeling errors, feed forward control can often reduce the effect of the
measured disturbance on the output better than that achievable by feedback control alone. Feed
forward control is always used along with feedback control because a feedback control system is
required to track set point changes and to suppress unmeasured disturbances that are always
present in any real process.
CONTROL VALVES
INTRODUCTION
One of the basic components of any control system is the final control element, which comes in a
variety of forms depending on the specific control application. The most common type of final
control element in chemical processing is the pneumatic control value, which regulates the flow
of fluids. Some other types include the variable speed pump and the power controller (used in
electrical heating).
Since the pneumatic control valve is so widely used in chemical processing, this chapter will be
devoted to the description, selection, and sizing of control valves.
Valve Sizing
In order to specify the size of a value in terms of its capacity to provide flow when fully open,
the following equation is used:
p v
q = Cv (20.1)
G
Valve Characteristics
The function of a control valve is to vary the flow of fluid through the valve by means of a
change of pressure to the valve top. The relation between the flow through the valve and the
valve stem position (or lift) is called the valve characteristic, which can be conveniently
described by means of a graph as shown in Fig. 20.3 where three types of characteristics are
illustrated.
Figure 20.3
Inherent valve characteristics (pressure drop across valve is constant) I linear, II
increasing sensitivity (e.g. equal percentage valve), III decreasing sensitivity.
Where 9max is the maximum flow when the valve stem is at its maximum life
Lmax (valve is full-open)
x is the fraction of maximum lift
m is the fraction of maximum flow
Equation (20.3) may now be written
m = q/qmax = f (L/Lmax)
The types of valve characteristics can be defined in terms of the sensitivity of the valve, which is
simply the fractional change in flow to the fractional change in stem position for fixed upstream
and downstream pressures; mathematically, sensitivity may be written
sensitivity = dmldx
Valve characteristic curves, such as the one shown in Fig. 20.3, can be obtained experimentally
for any valve by measuring the flow through the valve as a function of life (or valve-top
pressure) under conditions of constant upstream and downstream pressures. Two types of valves
that are widely used are the linear valve and the logarithmic (or equal percentage) valve. The
linear valve is one for which the sensitivity is constant and the relation between flow and life is
linear. The equal percentage valve is of the increasing sensitivity type.
It is useful to derive mathematical expressions for these types of valves. For the linear valve,
dmldx = (20.5)
Where is a constant.
Assuming that the valve is shut tight when the lift is at lowest position, we have that m =
0 at x = 0. For a single-seated valve that is not badly worn, the valve can be shut off for x = 0.
Integrating Eq. (20.5) and introducing the limits m = 0 at x = 0 and m = 1 at x = 1 gives
0 dm I0 dx
1 '
where m0 is the flow at x = 0. Equation (20.9) shows that a plot of m versus x on semi-log paper
gives a straight line. A convenient way to determine if a valve is of the equal percentage type is
to plot the flow versus life on semi-log paper. The relation expressed by Eq. (20.9) is the basis
for calling the valve characteristic logarithmic. The basis for calling the valve characteristic
equal percentage can be seen by rearranging Eq. (20.7) in the form
dmlm = βdx or hmlm = β∆x
In this form it can be seen that an equal fractional (or percentage) change in flow (Am/m) occurs
for a specified increment of change in stem position (Ax), regardless of where the change in stem
position occurs along the characteristic curve.
The term β can be expressed in terms of ma by inserting m = 1 at x = 1 into Eq. (20.9). The result
is
β = ln(1/m0)
Solving Eq. (20.9) for m gives
This section describes a couple of important points which are necessary to understand prior to
implementing practical control systems.
Control valves come in two sorts: air to open; and air to close. Air to open valves are normally
held closed by the spring and require air pressure (a control signal) to open them - they open
progressively as the air pressure increases. Air to close valves are valves which are held open by
the valve spring and require air pressure to move them towards the closed position. The reason
for the two types of valves is to allow failsafe operation. In the event of a plant instrument air
failure it is important that all control valves fail in a safe position (e.g. an exothermic reactor's
feed valves (or, perhaps, just one of the valves) should fail closed (air to open) and its coolant
system valves fail open (air to close)). The type of valve used obviously impacts on what a
controller has to do - changing the type of valve would mean that the controller would need to
move the manipulation in the opposite direction. To simplify things in this course we shall
assume that we are always using air to open valves - an increase in control action will cause the
valve to open and the flow through it to increase.
DIRECT/INDIRECT CONTROLLERS
In this process I have connected a level controller to the bottom valve. For this configuration the
controller needs to increase its signal (and hence the flow) when the level in the tank increases.
Move the cursor over the diagram to see and alternative configuration. In this case the controller
needs to reduce the flow when the level in the tank increases. Both configurations are equally
capable of controlling the level, but they require the controller to do entirely opposite things.
This is what direction of control action involves. A direct acting controller is one whose output
tends to increase as the measurement signal increases. A reverse acting controller is one whose
output tends to decrease as the measurement signal increases. It isn't necessary to buy separate
controllers - all commercial systems include a switch (usually marked DIR/REV) which allows
the controller to be switched from direct acting to reverse acting.
It is important to get the correct direction of control action. If things are set-up correctly a
feedback control system will experience negative feedback, which means that the system will act
to reduce errors in its output. If you get the direction of control action wrong the system will
undergo positive feedback and will act to reinforce output errors - this is very likely to cause the
system to go unstable. If you are having problems in setting up a stable controller the first thing
to check is that you have set the correct direction of control action!
GATE 2006
The control valve characteristics for three types of control valves (P, Q and R) are given in the
fig. below. Match the control valve with its characteristics.
Solution:
P – Quick opening
Q – Linear
R – Equal percentage
GATE 2006
A liquid level control system is configured as shown in the Fig. below. If the level transmitter
(LT) is direct acting and the pneumatic control valve is air-to- open. what kind of control action
should the controller (LC) have and why?
Correct Answer: 1
Solution:
Ans, (1)
Direct acting since the control valve is direct acting.
GATE 2009
Temperature control of a exothermic chemical reaction taking place in a CSTR is done with the
help of cooling water flowing in a jacket around the reactor. The types of valve and controller
action to be recommended are
A) air to open valve with the controller direct acting
B) air to close valve with the controller indirect acting
C) air to open valve with the controller indirect acting
D) air to close valve with the controller direct acting
GATE 2005
The control valve characteristic is selected such that the product of process gain and the valve
gain.
A) is a linearly increasing function of the manipulated variable
B) is a linearly decreasing function of the manipulated variable
C) remains constant as the value of the manipulated variable changes
D) is an exponentially increasing function of the manipulated variable
GATE 2007
The bias values for the two controllers, so that no offset occurs in either controller are
A) Pressure controller: 40%; Flow controller: 60%
B) Pressure controller: 33%; Flow controller: 67%
C) Pressure controller: 67%; Flow controller: 33%
D) pressure controller: 60%; Flow controller: 40%
GATE 2007
Given that the actual tank pressure is 4 bar(g) and a proportional controller is employed for
pressure control, the proportional band setting of the pressure controller required to obtain a set
point to the flow controller equal to 54 nm3/hr is
A) 50% B) 100% C) 150% D) 187%
GATE 2007
Common data for Questions
Q. A cascade control system for pressure control is shown in the figure given below. The
pressure transmitter has a range of 0 to 6 bar(g) and the flow transmitter range is 0 to 81 nm3/hr.
The normal flow rate through the valve is 32.4 nm3/hr corresponding to the value of set point for
pressure = 1 bar(g) and to give the flow, the valve must be 40% opened. The control valve has
linear characteristics and is fail-open (air to close). Error, set point and control variable are
expressed in percentage transmitter output (% TO). Proportional gain is expressed in the units of
% controller output (CO/%TO).
GATE 1999
Each item given in the left-hand column is closely associated with a specific characteristic listed
in the right-hand column. Match each of the items with the corresponding characteristic.
I) Transportation lag A) Increase in gain margin
II) Control valve B) Phase lag is proportional to frequency
C) Hyperbolic
D) Unstable response
E) Increase in phase margin
F) Phase angle is - 900
Answers are: I – B
II – A
Summary
In many process control systems the final control element is a pneumatic valve that opens and
closes as air pressure on the diaphragm changes.
Air to open valve requires a continuous supply of air to keep the valve open, in case of failure the
valve gets closed and there will be no flow through the valve.
Air to close valve requires a continuous supply of air to keep the valve in closed position, in case
of failure the valve gets fully opened and flow will be maximum through the valve.
A direct acting controller is one whose output tends to increase as the measurement signal
increases.
A reverse acting controller is one whose output tends to decrease as the measurement signal
increases.
PROCESS INSTRUMENTATION
McLeod gauge
Aneroid gauges
Magnetic
Piezoelectric
Optical
Potentiometric
Resonant
THERMAL Two-wire
CONDUCTIVITY Pirani (one wire)
Thermocouple gauge
thermistor gauge
Helical gear
Woltmann meter
Pelton wheel
Current mete
Vortex flowmeters
Ultrasonic (Doppler,
transit time) flow
meters
Temperature Thermocouples
measurement
Thermistors
Resistance
Temperature
Detector (RTD)
Pyrometer
Infrared
HPLC
Refractive index
GATE 2008
Match following
Group I Group II
P. Temperature 1. Hot wire anemometry
Q. Pressure 2. Strain Gauge
R. Flow 3] Chromatographic analyzer
4] Pyrometer
A) P – 1, Q – 2, R – 3 B) P – 4, Q – 1, R – 3
C) P – 1, Q – 2, R – 4 D) P – 4, Q – 2, R – 1
GATE 2011
Match the process parameters in Group I, with the measuring instruments in Group II.
Group I Group II
P. Flame temperature I. Thermocouple
Q. Composition of LPG II. Radiation pyrometer
R. Liquid air temperature III. Gas chromatograph
A] P – III, Q – I, R – II B) P – I, Q – III, R – II
C) P – II, Q – III, R – 1 D)P – II, Q – I, R – III
GATE 2011
The range of standard current signal in process instruments is 4 to 20 mA. Which ONE of the
following is the reason for choosing the minimum signal as 4 mA instead of 0 mA?
A) To minimize resistive heating in instruments
B) To distinguish between signal failure and minimum signal condition
C) To ensure a smaller difference between maximum and minimum signal
D) To ensure compatibility with other instruments
GATE 2005
Match the process variables (Group – I) given below with the measuring devices (Group – II)
Group I Group II
P. High Temperature 1. Orifice meter
Q. Flow 2. Chromatograph
R. Composition 3. Radiation Pyrometer
4. Bi-metallic Thermometer
A) P – 1, Q – 2, R – 3 B) P – 1, Q – 3, R – 2
C) P – 3, Q – 1, R – 2 D) P – 4, Q – 2, R – 1
© Career Avenues A 159
Heat Transfer
GATE 2003
Match the measured process variables with the list of measuring devices given below:
Measured process variables with the list of measuring devices given below:
GATE 2003
Match the type of controller given in Group 2 that is most suitable for each application given in
Group I
Group I Group 2
P. Distillation column bottom level to the controlled with 1 P control
bottom flow
Q. Distillation column pressure to be controlled by 2 P – I control
manipulating vapour flow from the top place
R. Flow control of a liquid from a pump by positioning 3 P – I – D control
valve in the line
S. Control of temperature of a CSTR with coolant flow
in the jacket
A) P – 1, Q – 1, R – 2, S – 3 B) P – 2, Q – 2, R – 3, S – 3
C) P – 2, Q – 2, R – 1, S – 1 D) P – 2, Q – 3, R – 2, S – 3
GATE 1995
Gas chromatorgraphy is used for measurement of
A) temperature B) pressure C) concentration D) flow rate