STABILITY

Today’s lecture will discuss about one of the most important

characteristics of the dynamic behavior of a system i.e. stability. So let me
first define what is stability? and why it is important to have a stable
system?

In system analysis, there are many ways to mathematically define stability

of a system and for Linear time invariant (LTI) system is define as

Stable: An LTI system is stable if and only if its natural response approaches
zero as time → ∞.


Unstable: An LTI system is unstable if and only if its natural response grows
without bound as time → ∞.

Marginally stable: An LTI system is marginally stable if and only if its
natural response neither grows nor approaches zero as time → ∞ (for
example, a sinusoidal response).
What is a natural response of a system?
The natural response of a system is the response which is not due to any
input, but only that which is due to initial conditions.
Let me make it more clear with the help of an example
Consider a first-order system:
𝑥 + 𝑎𝑥 = 𝑏𝑟

Taking Laplace transform on both sides,

𝑠𝑋 𝑠 − 𝑥 0 + 𝑎𝑋 𝑠 = 𝑏𝑅(𝑠)

Let r = 0, then R(s) =0.

𝑥(0)
𝑋 𝑠 =
𝑠+𝑎
 ⇒ 𝑥 𝑡 = 𝑥 0 𝑒 234

𝑎>0 𝑎<0
lim 𝑥 𝑡 = 0 lim 𝑥 𝑡 → ∞
4→: 4→:
Stable Unstable

Figure 1: Stable and unstable natural response of the example system

This is the natural response of the system.

Another Example:
Consider the following system, when the input r(t) is zero.

𝑥 + 𝜔> 𝑥 = 𝑏𝑟 = 0

⇒ 𝑠 > 𝑋 𝑠 − 𝑥 0 − 𝑠𝑥 0 + 𝜔> 𝑋 𝑠 = 0

𝑠𝑥(0) 𝑥(0)
⇒𝑋 𝑠 = +
𝑠> + 𝜔> 𝑠> + 𝜔>

so
𝑥(0)
𝑥 𝑡 = 𝑥 0 𝑐𝑜𝑠𝜔𝑡 + 𝑠𝑖𝑛𝜔𝑡
𝜔
 Figure 2: Marginally stable natural response of the example system

The system has bounded response, but is marginally stable.

A system is BIBO stable if for a bounded input it has a bounded

output (response).


A bounded input may be, for example, a step signal, or a sinusoidal signal.
A ramp or an impulse are not bounded and cannot be used for stability
analysis using this criterion.

The impulse response can also be used as a basis for stability analysis. A
linear system is stable if and only if the absolute value of its impulse
response, integrated over an infinite range, is finite. A consequence of this
definition is that the impulse response of a stable system will approach
zero as time approaches infinity.

Consider a system with an input r(t) = δ(t) (ideal impulse) and an output
c(t). The responses of a stable, marginally stable or unstable system are
presented in Figure 3.
 a) Stable b) Unstable c) Marginally Stable

• For a stable system the steady-state value of the impulse response

is zero (Figure 3 a): lim c(t) = 0
t→∞

• For an unstable system the steady-state value of the impulse

response is not limited, (Figure 3 b):
lim c(t) = ∞
t→∞

• For a marginally (critically) stable system the steady-state value of

the impulse response is different from zero, but it is limited,
(Figure 3 c):
0 < lim c(t) < ∞
t→∞

For linear systems the stability is related to the location of the roots of the
characteristic equation. Whether a linear system is stable or unstable is a
property of the system itself and does not depend on the input or driving
function of the system. The input contributes only to the steady-state
response terms in the solution.

The location of the poles in the s-plane indicate the system response. The
poles in the left-hand portion of the s-plane result in a decreasing response
for disturbance inputs. Similarly, poles on the jω axis and in the right-hand
plane result in a neutral and an increasing response, respectively, for an
impulse input.

This division of the s-plane is shown in Figure 4 where the location of the
poles and the corresponding impulse response are represented with the
same color.
 Figure 4: Stability in the s-plane

• Real positive poles (𝑝D ) and complex poles with positive real parts
(𝛼F ) contribute in the system response with terms of the type 𝑒 GH 4 or
𝑒 IJ 4 𝑠𝑖𝑛𝜔F 𝑡 that increase towards infinity.

• Real negative poles 𝑝D and complex poles with negative real parts
(𝛼F )contribute in the system response with terms of the type 𝑒 GH 4
or 𝑒 IJ 4 𝑠𝑖𝑛𝜔F 𝑡 that approach zero as time approaches infinity.


• Complex poles located on the imaginary axis ±𝑗𝜔F give an
undamped sinusoidal term in the system output.


• One pole at the origin (𝑝D = 0) contributes with a constant term in
the system response.


S-plane interpretation of stability

 Pole location Natural Response Stability
Left half s-plane Exponential decay Stable
or damped oscillations
Right half s-plane Exponential increase or Unstable
Oscillations with exponentially
Increasing magnitude
On imaginary axis

Multiplicity = 1 Sinusoidal response Marginally stable

Multiplicity > 1 Response of the type Unstable
𝑡𝑠𝑖𝑛(𝜔𝑡 + 𝜙)

Hurwitz Criterion
The Hurwitz stability test is the most useful for testing the stability of
large-order LTI systems.

Let,

The system given by G(s) is stable iff (if and only if) all the roots of its
characteristic equation D(s) = 0 are on the left hand side of the s-plane.
But it is not always easy to find the roots, especially for high order
polynomials.

A necessary condition for stability.

If the system is stable then all the coefficients a0, a1, . . . , an are positive.

Note that the above is not a sufficient condition. So,

Stability ⇒ All coefficients of characteristic polynomial are positive.


All coefficients of characteristic polynomial are positive ⇏Stability.

Some coefficients of characteristic polynomial are negative or zero ⇒
Unstable.

However, although necessary, these requirements are not sufficient. That

is, we immediately know the system is unstable if they are not satisfied,
yet if they are satisfied, we must proceed to ascertain the stability of the
system.

Example:

𝐷 𝑠 = 𝑠 P + 𝑠 > + 11𝑠 + 51

= 𝑠 + 3 𝑠 > − 2𝑠 + 17
So,

𝑝V = −3, 𝑝>,P = 1 ± 𝑗4

which means that the system is unstable.

𝐷 𝑠 = 𝑠P + 𝑠> + 1

The coefficient of s1 is zero.

which means that the system is unstable.

Routh’s Stability Criterion

So we need a criterion which provide sufficient condition for stability.

This criterion proposed by Routh, the coefficients of the polynomial are
arranging in particular pattern what is known as Routh table or Routh
array.

This criterion states that a necessary and sufficient condition for

stability is that all of the elements in the first column of the Routh
array be positive.
The best way to demonstrate the formation of the Routh table is through
an example. Consider a sixth order system.

D s = 𝑎[ 𝑠 \ + 𝑎V 𝑠 ] + 𝑎> 𝑠 ^ + 𝑎P 𝑠 P + 𝑎^ 𝑠 > + 𝑎] 𝑠 V + 𝑎\
The Routh table or Routh array is formed as follows:

This is called the Routh array.

The figure below shows the operation in the first two rows.
Now, look at the sign of the numbers in the first column of the Routh
array.


1. If all the elements of the first column of the Routh array are of the same
sign (usually positive) then all the roots are on the left hand side of the s-
plane.


2. If there are changes in sign, then the number of sign changes indicates
the number of roots with positive real parts.

Example:

𝐷 𝑠 = 𝑠 ^ + 2𝑠 P + 3𝑠 > + 4𝑠 + 5

𝑠 ^ ∶ 1 3 5
𝑠 P ∶ 2 4 0
2∙3−1∙4 2∙5−1∙0
𝑠 > ∶ = 1 = 5
2 2
1∙4−2∙5
𝑠 V ∶ = −6
1
−6 ∙ 5 − 1 ∙ 0
𝑠 [ ∶ = 5
−6

In this example, the number of changes in sign of the coefficients in the

first column is (1 → −6 𝑎𝑛𝑑 − 6 → 5). This means that there are two
roots with positive real parts. 𝑝V,> = 0.2878 ± 1.416𝑗 𝑝P,^ =
−1.2878 ± 0.8579𝑗