Sunteți pe pagina 1din 18

Feedback and Control Systems

STABILITY AND STEADY-STATE ERRORS

5.1. Introduction

In the first part of this course, it is stated that feedback control system design has three main objectives:

• Produce the desired transient response.


• Ensure stability of the system.
• Minimize steady-state error.

In the previous discussions, transient response performance of first- and second-order systems was
presented. It will be revisited when methods for designing feedback control systems will be discussed. In
this section, the second and third design objectives will be presented.

5.2. Stability

Stability is the most important system specification. If a system is unstable, transient response and steady-
state errors are moot points. An unstable system cannot be designed for specific transient response
performance or steady-state error specifications. Stability can be defined in many ways, depending on the
kind of system being analyzed or the point of view of the definition.

For linear, time-invariant systems, stability can be defined in terms of the natural response. Recall that the
total response of the system ct is the sum of two responses: the forced response and the natural
response. Also, it is said that if a system is controlled the steady-state output consists only of the forced
response. Thus, it can be concluded that:

• An LTI system is stable if the natural response approaches zero as time approaches infinity.
• An LTI system is unstable if the natural response grows without bound as time approaches infinity.
• An LTI system is marginally stable if the natural response neither decays nor grows but remains
constant or oscillates as time approaches infinity.

These definitions rely on natural response, but this perspective is difficult to look into when considering the
total response. For the total response, the following points can be raised:

• If the input is bounded and the total response does not approach infinity as time approaches
infinity, then the natural response is not approaching infinity; thus the system is stable.
• If the input is unbounded, the total response becomes unbounded also, but it cannot be determined
whether it is the natural response that becomes unbounded, or it is the forced response that
becomes unbounded.

Stability and Steady-State Errors Page 1


Feedback and Control Systems

With this, alternate definitions, one that looks into the input and the total response, is in order. Thus,

• A system is stable if every bounded input yields a bounded output (called the BIBO stability
requirement).
• A system is unstable if any bounded input yields an unbounded output.
• A system is marginally stable if it is stable for some inputs and unstable for others.

How can a system be concluded as stable? Looking into the natural response, which is required to
approach zero as time approaches infinity, it can be seen that the system poles that yield the natural
response should be negative if they are real or have negative real parts if they are complex. Thus,

• Stable systems have closed-loop transfer functions with poles only in the left half-plane of the
complex s-plane.
• Those systems which have closed-loop poles that are on the right half-plane of the complex s-
plane are unstable.
• Systems whose closed-loop poles that are on the jω-axis and multiplicity of one are marginally
stable systems. When at least one of the closed-loop poles on the jω-axis have a multiplicity of two
or more, the system is considered unstable.

Step responses of unity feedback system and their closed-loop poles plotted on the complex s-plane are
shown in the figure below. The first system is considered to be stable, the second one unstable.

Figure 6.1. Step response and pole plot of stable and unstable systems.

Stability and Steady-State Errors Page 2


Feedback and Control Systems

Determining the stability of feedback control systems is not a simple task however. Some systems will have
higher order poles when reduced to an equivalent closed-loop transfer function, which requires factoring
and solving for the roots of the denominator of the transfer function. But without actually solving for the
poles, the following observations can be used to determine whether the system is stable or not:

• Since all of the poles must be negative if real or have negative real part if complex, the
denominator of the transfer function can be factored into s  a
 where a
real and positive, or
complex with positive real part. The product of all such terms is a polynomial with all coefficients
having the same sign.
• No term in the polynomial can be missing, since this will imply cancellation between positive and
negative coefficients or imaginary axis roots in the factor, which is not the case.

Thus a system is unstable (or at best, marginally stable) if at least one of the coefficients of the
denominator of the closed-loop transfer function has a sign that is different from the others, and/or at least
one term in the polynomial is missing. These observations can be used to determine outright whether the
system is stable or not. However, when both conditions exist (all coefficients have the same sign and no
power of s missing) other methods, like computer-aided methods, of solving for the roots are employed.

5.3. Routh-Hurwitz Criterion for Stability

There is a method by which the stability of feedback control systems can be determined without actually
solving for the roots. Using the Routh-Hurwitz criterion for stability, one can tell how many poles are in the
left half-plane (LHP), right half-plane (RHP) and on the jω-axis.

The method requires two steps:

• Generate a table called a Routh table.


• Interpret the Routh table to tell how many closed-loop system poles are in the LHP, RHP and jω-
axis.

Generating a Basic Routh Table. For the closed-loop transfer function shown in figure 5.2, the initial
layout for the Routh table is shown in figure 5.3, and a compled of the table in figure 5.4.

Figure 5.2. Closed-loop transfer function to be used in the generation of Routh table.

Stability and Steady-State Errors Page 3


Feedback and Control Systems

Figure 5.3. Initial layout of the Routh table.

Figure 5.4. Completed Routh table.

Example 5.1
Make a Routh table for the system shown below.

Answer:

Stability and Steady-State Errors Page 4


Feedback and Control Systems

Interpreting a Basic Routh Table. The basic Routh table applies to systems with poles in LHP and RHP.
Systems with imaginary poles and the kind of the Routh table that results will be discussed next. 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.

Example 5.2
Based on the completed basic Routh table in Example 5.1, determine whether the system is stable.

Answer:
The system is unstable since there are two poles in the RHP of the pole map of the system, as shown by
two sign changes in the first column of the Routh table.

Example 5.3
Make a Routh table for the following polynomial and tell how many roots are in the RHP and LHP.

Ps 3s  9s  6s  4s  7s  8s  2s  6

Answer:

Since there are four sign changes in the first column, there are four poles in the RHP and three poles in the LHP.

Routh-Hurwitz Special Case – Zero Only in the First Column. 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 zero in the first column. The value  is then allowed to approach zero from either the
positive or the negative side, after which the signs of the entries in the first column can be determined.

Example 5.4
Determine the stability of the closed-loop transfer function
10
Ts 
s  2s  3s  6s  5s  3
 

Stability and Steady-State Errors Page 5


Feedback and Control Systems

Answer:
The Routh table is completed as follows

If  is taken to approach 0 and 0! , the following sign changes results:

Thus, with two sign changes in either cases, there are two poles in the RHP and the system is unstable.

Another method that can be used when a zero appears only in the first column of a row is by obtaining a
polynomial that has roots reciprocal that of the original polynomial which, in this case will not arrive at a
Routh table with a zero in the first column.

Example 5.5
Redo Example 5.4 by the use of the method described above.

Answer:
If the reciprocal polynomial is used, the following Routh table results.

Thus, just like in the previous example, there are two sign changes and therefore there are two poles in the
RHP. The system is unstable.
Stability and Steady-State Errors Page 6
Feedback and Control Systems

Routh-Hurwitz Special Case – Entire Row Zero. Proceeding first with an example, the procedure for the
case when an entire row becomes zero is demonstrated.

Example 5.6
Determine the number of right half-plane poles in the closed-loop transfer function
10
Ts 
s  7s  6s  42s  8s  56
 

Answer:
There will be an entire row of zeros in this case. However, using the procedure described when handling
this case, the resulting Routh table will be:

The table shows no sign changes in the first column; hence the system has no poles in the RHP.

An entire row of zeros will appear in the Routh table when a purely even or a purely odd polynomial is a
factor of the original polynomial. Even polynomials only have roots that are symmetrical about the origin.
They can be, as shown in figure 5.5,

• symmetrical and real (A).


• symmetrical and imaginary (B).
• quadrantal (C).

Figure 5.5. Roots of an even polynomial.

Stability and Steady-State Errors Page 7


Feedback and Control Systems

Another characteristic of the Routh table for the case in question is that the row previous to the row of zeros
contains the even polynomial that is a factor of the original polynomial. Finally, everything from the row
containing the even polynomial down to the end of the Routh table is a test of only the even polynomial.

Example 5.7
For the transfer function
20
Ts
s"  s  12s  22s  39s  59s  48s  38s  20

tell how many poles are in the right half-plane, in the left half-plane and on the jω-axis.

Answer:

Example 5.8
Use the Routh-Hurwitz criterion to find how many poles of the following closed-loop system Ts are in the
RHP, LHP and on the jω-axis.
s  7s # 21s  10
Ts 
s  s  # 6s  0s # s  # s  6

Answer:
Two RHP poles, two LHP and two jω.

Example 5.9
Find the number of poles in the LHP, RHP and on the jω-axis for the system shown below.

Answer:
Two in RHP, two in LHP

Stability and Steady-State Errors Page 8


Feedback and Control Systems

Example 5.10
Find the number of poles in LHP, RHP and on the jω-axis for the system shown

Answer:
Three poles in the LHP and two poles in RHP.

Example 5.11
Find the number of poles in LHP, RHP and on the jω-axis for the system shown

Answer:
Two in RHP, 4 in LHP and 2 in jω-axis

The Routh-Hurwitz criterion gives vivid proof that changes in the gain of a feedback control system results
in differences in transient response because of the changes in closed-loop pole locations.

Example 5.12
Find the range of gain K for the system shown below that will cause the system to be stable, unstable and
marginally stable.

Answer:
When K % 1386, the system is stable. When K & 1386, the system is unstable. When K 1386, the
system is marginally stable.

Drill Problems 5.1


1. Tell how many roots of the following polynomial are in the right half-plane, left half-plane and on the
jω-axis:
a. Ps s   3s  5s  4s  s  3
b. Ps s   6s  5s  8s  20

Stability and Steady-State Errors Page 9


Feedback and Control Systems

2. Using the Routh table tell how many poles of the following closed-loop transfer functions are in the
RHP, LHP, and on the jω-axis.
' "
a. Ts ( ) * +
' !' ' !' '!

'* '+ ' ,


b. Ts
'( !') '* !'+ '!

K/'+ ,0
3. Determine if the unity feedback system of the figure below when Gs ' ,' 
.

K'!1
4. In the system of in the previous item, let Gs ''!2
. Find the range of K for closed-loop stability
when
a. a % 0, b % 0
b. a % 0, b & 0
c. a & 0, b % 0
d. a & 0, b & 0

K' 1
5. Given the unity feedback system of item number 3 with Gs '' ,.' 
, find the following
a. The range of K that keeps the system stable.
b. The value of K that makes the system oscillate.
c. The frequency of oscillation when K is set to the value that makes the system oscillate.

Stability and Steady-State Errors Page 10


Feedback and Control Systems

5.3. Steady-State Errors

Steady-state error is the difference between the input and the output for a prescribed test input as t 5 ∞.
Test inputs used for analysis and design are summarized in the table below.

Figure 5.6 will help explain how steady-state errors are evaluated.

Figure 5.6. Steady-state error for step and ramp inputs.

Stability and Steady-State Errors Page 11


Feedback and Control Systems

For the step input in figure 5.6a, output 1 has no steady-state error (e, ∞ 0) while output 2 has a finite
steady-state error (e ∞ % ∞) as measured vertically between the input and output 2 after the transients
have died down. For the ramp input in figure 5.6b, again, output 1 has no steady-state error while that of
output 2 has finite steady-state error. For this input, another possibility exists, such as when an output 3
has a different slope than that of the ramp input; which in that case, output 3 has an infinite steady-state
error (e ∞ ∞.

Note that the error is the difference between the input and the output, a closed-loop system is formed as
shown in figure 5.7a. Note that this system uses the closed-loop transfer function Ts in the forward path.
For figure 5.7b, this feedback system uses the open-loop transfer function Gs.

Figure 5.7. Measurement of steady-state error.

Many steady-state errors in control systems arise from nonlinear sources. However, the discussion of
steady-state errors for our purpose are errors that arise from the configuration of the system itself and the
type of applied input.

Figure 5.8. System with a finite and zero steady-state error for a step inout.

For a system such that in figure 5.8a, where Rs is the input and Cs is the output, Es Cs #
Rs is the error signal. Note that in this case when there is only a pure gain K, for ct to be of finite
value and nonzero, an error signal et must always be present. From this system, it can be said that

1
e'' c (5.1)
K ''

where e'' is the steady-state error and c'' is the steady-state output. Based on 5.1, it can be generalized
that although the steady-state error does not become zero, it diminishes as K, the value of the gain is
increased.

Stability and Steady-State Errors Page 12


Feedback and Control Systems

If the forward path gain is replaced by an integrator, such as the one shown in figure 5.8b, there will be zero
error in the steady-state for a step input. As ct increases, et decreases, because ct rt # ct.
This decrease will continue until et is zero, in which the integrator continues to have an output. Thus,
systems like that of figure 5.8b have zero steady-state output for a step input.

5.4. Steady-State Error for Unity Feedback Systems

Steady-state error can be calculated from a system’s closed-loop transfer function, Ts or the open-loop
transfer function Gs for unity feedback systems.

Steady-State Error in Terms of <=. Consider figure 5.7a. Using the final value theorem of Laplace
transform, the steady-state error in terms of the closed-loop transfer function is

e∞ lim s RsB1 # TsC (5.2)


'5A

Example 5.13

Find the steady-state error for the system of figure 5.7a if Ts '+ ' ,A
and the input is a unit step.

Answer:
1
e ∞ 
2

Steady-State Error in Terms of D=. More insights can be deduced when the steady-state error is
evaluated using the open-loop transfer function Gs. Using the final value theorem, and ensuring that the
closed-loop system is stable, the steady-state error is

s Rs
e∞ lim (5.3)
'5A 1  Gs

For the step input, 5.3 becomes

1
e'EFG ∞ (5.4)
1  lim Gs
'5A

The term lim'5A Gs is called the dc gain of the forward transfer function. In order to have zero steady-
state error, the limit lim'5A Gs must approach infinity. For it to happen, the open-loop transfer function
must have at least one pole in the origin, or, there should be at least a pure integration of the open-loop
transfer function. If there are no integrations, the lim'5A Gs attains a finite value and therefore a finite
steady-state error will result.

Stability and Steady-State Errors Page 13


Feedback and Control Systems

For the ramp input, 5.3 becomes

1
eH1IG ∞ (5.5)
lim sGs
'5A

For 5.5 to become zero, lim'5A sGs must approach infinity. Therefore, there should be at least two
integrations of the open-loop transfer function. If there is only one integration, the steady-state error
becomes a finite value. If there are no integrations in the forward path, this will result to an infinite steady-
state error.

For the parabolic input, 5.3 becomes

1
eG1H1 ∞ (5.6)
lim sGs
'5A

If there are three integrations, the steady-state error is zero. If there are only two, a finite steady-state error
results. Finally, if there is at most one integration, the steady-state error is infinite.

Example 5.14
Find the steady-state errors for inputs 5 ut, 5t ut and 5t  ut to the system shown below.

Answer:
5
e∞ e'EFG ∞
21

e∞ eH1IG ∞ ∞

e∞ eG1H1 ∞ ∞

Stability and Steady-State Errors Page 14


Feedback and Control Systems

Example 5.15
Find the steady-state errors for inputs 5 ut, 5t ut and 5t  ut to the system shown below.

Answer:
e∞ e'EFG ∞ 0

1
e∞ eH1IG ∞
20

e∞ eG1H1 ∞ ∞

Example 5.16
A unity feedback system has the following forward transfer function:

10s  20s  30


Gs
ss  25s  35

Find the steady-state error for the following inputs: 15 ut, 15t ut and 15t  ut.

Answer:
e∞ e'EFG ∞ 0

e∞ eH1IG ∞ 2.1875

e∞ eG1H1 ∞ ∞

Example 5.17
A unity feedback system has the following forward transfer function:

10s  20s  30


Gs
s  s 25s  35s  50

Find the steady-state error for the following inputs: 15 ut, 15t ut and 15t  ut.

Answer:
The closed-loop system is unstable. Calculations cannot be made.

Stability and Steady-State Errors Page 15


Feedback and Control Systems

5.5. Static Error Constants and System Type

In this section, the parameters specifying steady-state error performance of unity negative feedback
systems will be defined. These steady-state error performance specifications are called static error
constants.

The static error constants are

• The position constant, K G

K G lim Gs (5.7)


'5A

• The velocity constant, K K

K K lim sGs (5.8)


'5A

• The acceleration constant, K 1

K 1 lim s Gs (5.9)


'5A

These quantities can assume the value of zero, finite constant or infinity depending on the form of Gs.
Also, the steady state error decreases when the static error constant increases.

Example 5.18
For each of the system shown below, evaluate the static error constants and find the expected error for the
standard step, ramp and parabolic inputs.

Stability and Steady-State Errors Page 16


Feedback and Control Systems

Answers:
For system (a) the static error constants are K G 5.208, K K 0 and K 1 0. The expected errors are
e'EFG ∞ 0.161, eH1IG ∞ ∞ and eG1H1 ∞ ∞.

For system (b) the static error constants are K G ∞, K K 31.25 and K 1 0. The expected errors are
e'EFG ∞ 0, eH1IG ∞ 0.032 and eG1H1 ∞ ∞.

For system (c) the static error constants are K G ∞, K K ∞ and K 1 875. The expected errors are
e'EFG ∞ 0, eH1IG ∞ 0 and eG1H1 ∞ 1.14 L 10! .

System Type. The following table summarizes the system type of systems, which depends on the number
of integrations of the forward transfer function.

5.6. Steady-State Error Specifications

Static error constants can be used to specify the steady-state error characteristics of control systems, such
as what has been done with the damping ratio, natural frequency, rise, peak and settling time and percent
overshoot specifying transient response performance of systems.

Example 5.19
What information can be deduced from the system whose static error constant is K K 1000?

Answer:
1. The system is stable.
2. The system is Type 1.
3. A ramp is the test input signal.
,
4. The steady-state error is e∞ ,AAA 0.001

Stability and Steady-State Errors Page 17


Feedback and Control Systems

Example 5.20
What information is obtained in the specification K G 1000?

Answer:
1. The system is stable.
2. The system is type 0.
3. The input test signal is a step.
,
4. The steady-state error is e∞ .
,AA,

Example 5.21
Given the control system below, find the value of K so that there is 10% error in the steady-state.

Answer:
K 672

Example 5.22
A unity feedback system has the following forward transfer function
Ks  12
Gs
s  14s  18

Find the value of K to yield a 10% error in the steady-state.

Answer:
K 189

REFERENCES:

N. Nise. (2008). Control Systems Engineering 6th Edition. United States of America: John Wiley & Sons.

R. Dorf & R. Bishop. (2008). Modern Control Systems 12th Edition. New Jersey: Prentice Hall.

Stability and Steady-State Errors Page 18

S-ar putea să vă placă și