Time Domain Response Analysis

When model of a system found, the model needs to simulate in closed loop form in order to
obtain its output response.

R(s) E(s) G(s) C(s)

Output response of a second order system is illustrated as follow

From the output response we have 2 main areas of response: transient response and steady
state, that need to analyze before we design a suitable controller for the system. The response
is illustrated in time domain (t). Analysis on the response is viewed in 3 categories:
1. Transient response performance, includes rise time (tr), overshoot maximum (Mp), peak
time (tp), and settling time (ts)
2. Steady state errors (ess)
Steady state errors is the error that exist between output C(s) and input R(s), when the
response enter its steady state area.
3. Stability
Stability is defined as how the output C(s) remain tracks the input R(s) in the transient
area and steady state area.

However, in case the response is seem unstable, analysis on category 1 and 2 no need to do,
because steady state condition will never reach along the test.
Moreover, analysis on the model can be done precisely in mathematic for those 3 kinds of
analysis. The analysis can be started from by expressing the model in closed loop transfer
function (T).

𝐶(𝑠) 𝐺(𝑠)
𝑇(𝑠) = =
𝑅(𝑠) 1 + 𝐺(𝑠)𝐻(𝑠)

Example:

R(s) 50
C(s)
s ( s  10)

where:

50
𝐺(𝑠) = 𝑠(𝑠+10) and 𝐻(𝑠) = 1

50 50
𝐶(𝑠) 𝑠(𝑠+10) 𝑠(𝑠+10) 50 50
Such that, 𝑇(𝑠) = 𝑅(𝑠) = 50 = 𝑠(𝑠+10)+50 = 𝑠(𝑠+10)+50 = 𝑠2 +10𝑠+50
1+( )
𝑠(𝑠+10) 𝑠(𝑠+10)

From the T(s) also can be defined characteristic equation of the system in polynomial form:

𝑠 2 + 10𝑠 + 50

Then, compare the closed loop transfer function of the system with general form of second
order system:

50 𝑛 𝜔 2
= 𝑠2 +2𝜁𝜔𝑠+𝜔
𝑠2 +10𝑠+50 𝑛
2

From the equation it can be calculated that

𝜔𝑛 = √50 = 7.07

2𝜁𝜔𝑛 = 10
10 10
𝜁 = 2𝜔 = 2𝑥7.07 = 0.7
𝑛

where, ζ (zeta) and 𝜔𝑛 are defined as damping ration and natural frequency, respectively.

When ζ and 𝜔𝑛 are calculated, analysis on the system response can be started for transient
response performance analysis. Performance of the transient response specifications such as
rise time (tr), overshoot maximum (Mp), peak time (tp), and settling time (ts) will depend on the
value of ζ and 𝜔𝑛 .
However, before we calculate the transient response specification, we need understand the
impact the change of ζ and 𝜔𝑛 on the transient response performance. The impact of ζ can be
clearly seen in the graph below.

The graph clearly illustrate that a smaller ζ gives a smaller damping on the output response
(higher overshoot). The best choice for the response is having ζ to result small overshoot
(around 0.7).

Meanwhile, influence of 𝜔𝑛 on the output response can be seen from the following s-plane for
complex function s.

A better transient response performance and stability can be reach when the roots of s of the
system characteristic equation are placed as negative as possible far from zero. Such that, a
system with a bigger 𝜔𝑛 will have a better transient response performance and have more
guaranteed stability.
Details of mathematical calculation for the 3 categories of the response analysis are explained
as follow:

A. Transient Response Specification Performances

1. Rise Time
𝜋−𝛽
𝑡𝑟 =
𝜔𝑑
where:
√1 − 𝜁 2
𝛽 = 𝑎𝑟𝑐𝑡𝑎𝑛
𝜁
𝜔𝑑 = 𝜔𝑛 √1 − 𝜁 2 )

A bigger 𝜔𝑛 is needed to result a smaller 𝑡𝑟 .

2. Overshoot Maximum

𝜁
− 𝜋
𝑀𝑝 = 𝑒 √1−𝜁 2

3. Settling Time

4
𝑡𝑠 =
𝜁𝜔𝑛

B. Steady state errors (ess)

Steady state error can be calculated from the following diagram block:

𝐶(𝑠) 𝐺(𝑠)
𝑇(𝑠) = =
𝑅(𝑠) 1 + 𝐺(𝑠)𝐻(𝑠)

𝐸(𝑠) = 𝑅(𝑠) − 𝐶(𝑠)𝐻(𝑠)

𝐸(𝑠) 𝐶(𝑠)𝐻(𝑠)
=1−
𝑅(𝑠) 𝑅(𝑠)
𝐶(𝑠)
Substitute 𝑅(𝑠) and we have
𝐸(𝑠) 𝐺(𝑠)𝐻(𝑠) 1 + 𝐺(𝑠)𝐻(𝑠) − 𝐺(𝑠)𝐻(𝑠) 1
=1− = =
𝑅(𝑠) 1 + 𝐺(𝑠)𝐻(𝑠) 1 + 𝐺(𝑠)𝐻(𝑠) 1 + 𝐺(𝑠)𝐻(𝑠)
 1
𝐸(𝑠) = 𝑅(𝑠)
1 + 𝐺(𝑠)𝐻(𝑠)

where, R(s) is input of the system in 3 variants of input:

1
- Step input, 𝑠
1
- Ramp input, 𝑠2
1
- Parabolic input, 𝑠3

Hence, it can be concluded that error of the system is depend on type of input and model of the
system.

Then, steady state errors of the system can be calculated as

𝑠𝑅(𝑠)
𝑒𝑠𝑠 = lim 𝑒(𝑡) = lim
𝑠→∞ 𝑠→0 1 + 𝐺(𝑠)𝐻(𝑠)

Note: Go in deep with Exampel 5-4 (Ermanu A. Hakim, Sistem Kontrol, page 78).

C. Stability Test

The stability is tested by using Routh-Hurwitz criterion. Stability test on the model can go
thought with characteristic equation of the system. Polynomial equation of the system can be
expressed as follow

𝑄(𝑠) = 𝑎0 𝑠 𝑛 + 𝑎1 𝑠 𝑛−1 + Λ + 𝑎𝑛−1 𝑠 + 𝑎𝑛

The Routh-Hurwitz is defined as below

𝑠𝑛 𝑎0 𝑎𝑛−1

𝑠 𝑛−1 𝑎1 𝑎𝑛
𝑎1 𝑎𝑛−1 −𝑎0 𝑎𝑛
𝑠1 =𝐴
𝑎1

𝐴𝑎𝑛 −𝑎1 (0)

𝑠0 =𝐵
𝐴

Example:

Given the characteristic equation of a system as

𝑄(𝑠) = 𝑠 3 − 4𝑠 2 + 2𝑠 + 6

Please test stability of the system by using Routh-Hurwitz criterion whether the system stable
or unstable.
Solution:

By using Routh-Hurwitz table we solve:

𝑠3 1 2

𝑠2 −4 6

(−4)(2)−(1)(6)
𝑠1 = 3.5
−4

(3.5)(6)−(−4)(0)
𝑠0 =6
3.5

Conclusion:
In the first column of the table there is two changes of sign of the roots, from 1 to -4, and from
-4 to 3.5. It means that the system is unstable.

Mid Test:

1. Obtain a state space model of a system from its differential equations.

2. Changed the state space model into a transfer function representation.
3. Obtain characteristic equation of the system and calculate ζ and 𝜔𝑛 .
4. Obtain transient response performance and steady state errors, and test stability of the
system.