STABILITY

CHAPTER 13
• The overall response of the control system was no higher than
second-order.
• Hence, the system is inherently stable.
• In this chapter we consider the problem of stability in a control
system only slightly more complicated than any studied previously.
• This system might represent proportional control of two stirred-
tank heaters with measuring lag.
• In this discussion, only set point changes are to be considered.
For a unit-step change in R, the transform of the response is

To obtain the transient response C(t), it is necessary to find the

inverse of Eq.
This requires obtaining the roots of the denominator of Eq.,
which is third-order.
We can no longer find these roots as easily as we did for the
second-order systems by use of the quadratic formula.
It is apparent that the roots of the denominator depend upon
the particular values of the time constants and Kc .
These roots determine the nature of the transient response,
according to the rules presented in Fig.
It is of interest to examine the nature of the response for the
control system of Fig. as K c is varied
• Assuming the time constants 1 , 2 , and 3 to be fixed.
• To be specific, consider the step response for several values of
Kc

• it is seen that as K c increases, the system response becomes

more oscillatory.
• In fact, beyond a certain value of K c , the successive amplitudes
of the response grow rather than decay;
• This type of response is called unstable
• In this chapter, the focus is on developing a clearer
understanding of the concept of stability.
• In addition, we develop a quick test for detecting roots having
positive real parts,
DEFINITION OF STABILITY (LINEAR SYSTEMS)

• A bounded input function is a function of time that always

falls within certain bounds during the course of time.
• For example, the step function and sinusoidal function are
bounded inputs.
• The function f (t)= t is obviously unbounded.
• A linear mathematical model (set of linear differential
equations describing the system) from which stability
information is obtained is meaningful only over a certain
range of variables.
• For example, a linear control valve gives a linear relation
between flow and valve-top pressure only over the range of
pressure (or flow) corresponding to values between which
the valve is shut tight or wide open.
• When the valve is wide open, for example, further change
in pressure to the diaphragm will not increase the flow.
• We often describe such a limitation by the term saturation.
• A physical system, when unstable, may not follow the
response of its linear mathematical model beyond certain
physical bounds but rather may saturate.
• However, the prediction of stability by the linear model is
of utmost importance in a real control system since
operation with the valve shut tight or wide open is clearly
unsatisfactory control.
 STABILITY CRITERION
• Translate the stability definition into a simpler criterion,

CHARACTERISTIC EQUATION

Suppose a unit-step change in set point is applied.

Then

where r 1 , r 2 , . . . , r n are the n roots of the equation

F (s) is a function that arises in the rearrangement to the right-
hand form
• If there are any of the roots r 1 , r 2 , . . . , r n in the right half of the complex plane,
the response C ( t ) will contain a term that grows exponentially in time and the
system is unstable.
• If there are one or more roots of the characteristic equation at the origin, there is an
s m in the denominator of Eq (where m ≥ 2) and the response is again unbounded,
growing as a polynomial in time.
• This condition specifies m as greater than or equal to 2, not 1, term will invert to a
term of the form C1 t, which is unbounded.)
• Additionally, if there is a pair of conjugate roots of the characteristic equation on the
imaginary axis, the contribution to the overall step response is a pure sinusoid, which
is bounded.
• However, if the bounded input is taken as sin ω t, where ω is the imaginary part of
the conjugate roots, the contribution to the overall response is a sinusoid with an
amplitude that increases as a polynomial in time [the response will have a term of
the form C 1 t sin(ω t + )].
• Thus, if a root lies on the imaginary axis, there is the potential for repeating the root
of a bounded input (such as a step input or a sinusoid input), and the response will
be unstable.
• Therefore, the right-half plane, including the imaginary axis, is the unstable region
for location of roots of the characteristic equation.
 • characteristic equation of a control system, which determines
its stability, is the same for set point or load changes.
• The stability depends only upon G (s), the open-loop transfer
function.
• Furthermore, although the rules derived above were based on
a step input, they are applicable to any input.
which is equivalent to

Solving by the quadratic formula gives

The characteristic equation is therefore

Since the real part of s 1 and s 2 is negative (3/2), the system is stable.
 ROUTH TEST FOR STABILITY
• The Routh test is a purely algebraic method for
determining how many roots of the characteristic
equation have positive real parts.
• from this it can also be determined whether the system
is stable.
• The test is limited to systems that have polynomial
characteristic equations.
• This means that it cannot be used to test the stability of a
control system containing a transportation lag
(exponential function) .
• The procedure for application of the Routh test is
presented without proof.
• The proof is available elsewhere (Routh, 1905) and is
mathematically beyond the scope of this text.
• The procedure for examining the roots is to write the
characteristic equation in the form

• where a0 is positive. (If a0 is originally negative, both sides are

multiplied by -1.)
• In this form, it is necessary that all the coefficients be positive
if all the roots are to lie in the left half-plane.

• If any coefficient is negative, the system is definitely unstable,

and the Routh test is not needed to answer the question of
stability. (However, in this case, the Routh test will tell us the
number of roots in the right half-plane.)
• If all the coefficients are positive, the system may be stable or
unstable.
• It is then necessary to apply the following procedure to
determine stability.
 Routh Array
Arrange the coefficients of Eq. into the first two rows of the
Routh array as follows:

The array has been filled in for n= 7 to simplify the discussion.

For any other value of n, the array is prepared in the same manner.
In general, there are n+ 1 rows.
For n even, the first row has one more element than the second row.
The elements in the remaining rows are found from the formulas

• The elements for the other rows are found from formulas that
correspond to those just given.
• The elements in any row are always derived from the elements
of the two preceding rows.
• During the computation of the Routh array, any row can be
divided by a positive constant without changing the results of
the test.
(The application of this rule often simplifies the arithmetic.)
• Having obtained the Routh array, we can apply the following
theorems to determine stability.
 THEOREMS OF THE ROUTH TEST
• Theorem 1.The necessary and sufficient condition for all the roots
of the characteristic equation to have negative real parts (stable
system) is that all elements of the first column of the Routh array (
a 0 , a 1 , b 1 , c 1 , etc.) be positive and nonzero.

• Theorem 2. If some of the elements in the first column are

negative, the number of roots with a positive real part (in the right
half-plane) is equal to the number of sign changes in the first
column.

• Theorem 13.3. If one pair of roots is on the imaginary axis,

equidistant from the origin, and all other roots are in the left half-
plane, then all the elements of the n th row will vanish and none of
the elements of the preceding row will vanish. The location of the
pair of imaginary roots can be found by solving the equation

where the coefficients C and D are the elements of the array in the ( n – 1)st row
as read from left to right, respectively.
• The algebraic method for determining stability is limited in its
usefulness in that all we can learn from it is whether a system
is stable.
• It does not give us any idea of the degree of stability or the
roots of the characteristic equation.

Since there is no change in sign in the

first column, there are no roots having
positive real parts, and the system is
stable.
Determine the values of K c for which the control system is stable.
For the value of K c for which the system is on the threshold of instability,
determine the roots of the characteristic equation.
The characteristic equation 1+ G (s) =0 becomes

Rearrangement of this equation for use in the Routh test gives

It is concluded that the

system will be stable only
if Kc  10

Since the proportional sensitivity of the

controller Kc is a positive quantity, we see
that the fourth entry in the first column,
6(1 + K c ), is positive.
• At Kc =10, the system is on the verge of instability, and the
element in the n th (third) row of the array is zero.
• According to Theorem 3, the location of the imaginary roots is
obtained by solving

Therefore, two of the roots on the imaginary axis are

located at
The third root can be found by expressing Eq. (13.11) in factored form

where s1 , s2 , and s 3 are the roots. Introducing the two imaginary roots
( s1 = j 11 and s2=-j 11 ) into Eq. and multiplying out the terms give
• Comparing this equation with Eq. we see that s3 = - 6.
Let Using the new parameters given above in this equation leads to

Notice that the order of the characteristic equation has increased from
three to four as a result of adding integral action to the controller.
The Routh array becomes

Because there are two sign changes in the first

column, we know from Theorem 2 of
the Routh test that two roots have positive real
parts.
• In next chapter we will discuss for obtaining more
information about the actual location of the roots of the
characteristic equation.
• This will enable us to predict the form of the curves for
various values of K c .
• The advantage of these tools is that they are graphical and
are easy to apply compared with standard algebraic
solutions of the characteristic equation.
• There are two distinct approaches to this problem: root locus
methods and frequency-response methods.
• Root locus methods allow rapid determination of the location
of the roots of the characteristic equation as functions of
parameters such as Kc.
• However, they are difficult to apply to systems containing
transportation lags.
• Also, they require a reasonably accurate knowledge of the
theoretical process transfer function.