Dynamic Response Characteristics of

More Complicated Systems

Textbook: Dale E Seborg Chapter-6

U D Dwivedi
1. Poles and Zeros and Their Effects on System Response.
2. Time Delays.
3. Approximation of Higher-Order Systems.
4. Interacting and Noninteracting Processes.
Poles and Zeros and Their Effect on System Response
K
Example) G ( s ) = (6.1)
s (τ 1 s + 1)(τ 22 s 2 + 2ζτ 2 s + 1)
Where 0 ≤ ζ < 1 .

Use partial fraction and inverse LT.

 The response of the system to any input contains following:
1. A constant term due to factor, s

2. exp(− t τ 1 ) term due to factor τ 1s + 1

3. exp ( − ζ t τ 2 ) sin ( 1 − ζ 2
⋅t τ2)
terms due to factor τ 22 s 2 + 2ζτ 2 s + 1
and exp( −ζ t τ )cos( 1 − ζ 2 ⋅ t τ )
2 2
  Additional terms determined by the specific input forcing term will
also appear in the response, but the intrinsic dynamic features of the
process, the so-called response models or natural modes, are
determined by the process itself.
Each of the above response modes is determined from the factors
of the denominator polynomial(roots of the characteristic equation);
 Poles !

s1 = 0
1
s2 = −
τ1 (6.2)

ζ 1−ζ 2
s3 = − + j
τ2 τ2
ζ 1−ζ 2
s4 = − − j
τ2 τ2
Figure 6.1. Roots of the denominator of G (s )
plotted in the complex s plane.
 • Pole The roots of the characteristic polynomial.
decide the stability and the swiftness of the response and oscillations.
• The effect of pole for the process.
– Pole at origin : the presence of an integrating element.
– Pole closer to the imaginary axis : slower response mode.

Two poles: s1 (= − 1 τ 1 ) > s2 (= − 1 τ 2 ).

Response modes: M 1 = exp(− t τ 1 ) andM 2 = exp(− t τ 2 ) .
M 1 decays slower than M 2 .
– Pole to the right of the imaginary axis : unstable response mode.
Pole: s = 1 τ .
Response mode: M = exp(t τ ) .
M grows without bound as t becomes large, characteristic of
unstable systems..
– Complex Pole( always appear as part of a conjugate) : oscillatory mode.
 • Zero ; The roots of the numerator of the transfer function.

• The effect of zero for the process.

– It has no effect on the number and location of the poles and their
associated response modes unless there is an exact cancellation of a
pole by a zero  It has no effect on the stability of the process.
– It exert a profound effect on the coefficients of the response modes
in the system response. Such coefficients are found by partial
fraction expansion  It decide the weighting factors of each
response modes.
Example. Calculate the response of the lead-lag element(6.3)
to a step change of magnitude M in its input.
K (τ a s + 1)
G (s) = (6.3)
τ1s + 1
Solution Y ( s ) = KM (τ a s + 1) (6.4)
s (τ 1 s + 1)
(6.3) can be expanded into partial fractions.
1 τ a − τ1
Y ( s ) = KM ( + ) (6.5)
s τ 1s + 1
Pole determines the response mode!
 τa 
y (t ) = KM 1 − (1 − ) exp(− t τ 1 ) (6.6)
 τ1 

Zero determines the weighting factor of the response mode!

 If τ a = τ 1 , the transfer function simplifies to K as a result cancellation
of numerator and denominator elements  Pole-Zero Cancellation.
 K (τ a s + 1)
G (s) =
τ1s + 1

Figure 6.2. Step response of G (s ) with various τ a

• Case a: 0 < τ 1 < τ a

 The presence of a zero in the first-order
• Case b: 0 < τ a < τ 1
system causes a jump discontinuity in y(t)
• Case c: τ a < 0 < τ 1
at t=0, when the step input is applied.
Example. For the case of a single zero in an overdamped
second-order transfer function,
K (τ a s + 1)
G (s) = (6.7)
(τ 1 s + 1)(τ 2 s + 1)
calculate the response to a step change of magnitude M .

Solution The response of this system is

 τ a − τ1 τa −τ2 
y (t ) = KM 1 + exp(− t τ 1 ) + exp(− t τ 2 ) (6.8)
 τ1 − τ 2 τ 2 − τ1 

Note y (t → ∞) = KM as expected; hence, the effect of including

the single zero dose not change the final value nor dose it change
the number or location of the response modes.
But the zero does affect how the response mode are weighted in
the solution, (6.8).
 • Case a: τ a > τ 1
• Case b: 0 < τ a ≤ τ 1
• Case c: τ a < 0

Figure 6.3. Step response of G (s ) (6.7) with various τ a .

 Case a shows that overshoot can occur if the zero( − 1 τ a ) is
large enough , that is , if τ a is sufficiently large.
 Case b looks much like an ordinary first-order process response
as a result of the single zero.
Case c exhibits a so-called inverse response initially.
 6.2Time Delays
 Whenever material or energy is physically moved in a
process or plant there is a time delay associated with the
movement.
Example A fluid is transported through a pipe in plug flow.

Figure 6.4. Transportation of fluid in a pipe.

The transportation time between points 1 and 2 is given by
length of pipe volume of pipe
θ= = (6.9)
fluid velocity volumetric flowrate
If x is some fluid property at point 1 and y is the same property at
point 2, then y and x are related by a simple time delay θ .
0 t <θ
y (t ) =  (6.10)
 x(t − θ ) t ≥ θ
Thus the output y (t ) is simply the same input function shifted back in
time by the amount of the delay.

Y (s)
= G=( s ) exp(−θ s )
X (s)

Figure 6.5. The effect of a pure delay is a translation of the function in time.

The transfer function of a time delay of θ units can be obtained by

applying the Laplace transform to (6.10).
 • First-Order Plus Time Delay(FOPTD) transfer function.

K ⋅ exp(−θs )
G ( s) = (6.12)
τs + 1
 Polynomial approximation technique !
– This transfer function can be used to approximate various
overdamped system.

–The simple exponential form is some what deceptive since is it a

non-rational transfer function  The presence of a time delay in a
process means that we cannot factor the process transfer function in
terms of simple poles and zeros only.
 • Polynomial approximation to exp(−θs ) .
1. Taylor series expansion.
θ 2s2 θ 3s3 θ 4s4 θ 5s5
exp(−θs ) = 1 − θs + − + − +  (6.13)
2! 3! 4! 5!
Only good for θ is small.
2. Padé approximation. Long division
θ θ
exp( − s ) 1 − s θ 2s2
exp( −θ s ) = 2 = 2 =1−θ s + + (6.14)
θ θ 2
exp( s ) 1+ s
2 2 1st order approximation.

θ θ2
1 − s + s2
= 2 12 (6.15)
θ θ 2
1 + s + s2 2nd order approximation.
2 12
 6.3Approximation of Higher-Order Systems
•As discussed earlier, a time delay can be used to approximate
high-order model dynamics.
• Consider the step response of a hypothetical nth-order system with
n equal time constants.
K
Gn ( s ) = (6.16)
τ
( s + 1) n
n
 This transfer function can represent a series of n stages, each
described by a first-order transfer function, with a total residence
time τ equally divided among the stages.
• The system response to a step of magnitude M.
 n −1
(nt τ ) i 
y (t ) = KM 1 − exp(− nt τ )∑  (6.17)
 i =0 i! 
(6.17) is well approximated by a time delay of magnitude τ , as
shown in Figure 6.6.
Figure 6.6. A pure time delay element can approximate a large
number of first-order systems in series.

lim Gn ( s ) = K exp(−τs ) (6.18)

n →∞
• Many processes that do not contain an explicit time delay
consist of a large number of process units connected in
series( for example trays in a distillation or adsorption
column).
 A time-delay term can be used in a process model as a
approximation for a number of small time constants.
Example nth-order process( G (s ) ) composed of first-order
processes in series ( τ 1 ,τ 2 ,,τ n) is dominated by two of these
process( τ 1 and τ 2 ), then an approximation transfer function
G (s ) for the system is K
G (s) =
(τ 1 s + 1)(τ 2 s + 1)(τ n s + 1)
K exp(−θs )
≈ G (s) = (6.19)
(τ 1 s + 1)(τ 2 s + 1)
n
Where θ = ∑τ i (6.20)
i =3
 Approximation of Higher-Order TF
approximating high-order transfer function
models with lower-order models that have similar
dynamic and steady-state characteristics.
In previous slide it is showed that
the transfer function for a time delay can be
expressed as a Taylor series expansion.
For small values of s, −θ 0 s
e ≈ 1 − θ0 s

18
 • An alternative first-order approximation consists of the transfer
function,

−θ 0 s 1 1
e= θ0 s
≈ (6-58)
1 + θ0 s
Chapter 6

where the time constant has a value of θ 0 .

• These expressions can be used to approximate the pole or zero
term in a transfer function.

19
 Skogestad’s “half rule”
• Skogestad (2002) has proposed an approximation method
for higher-order models that contain multiple time
constants.
• He approximates the largest neglected time constant in the
following manner.
- One half of its value is added to the existing time delay
(if any) and the other half is added to the smallest retained
time constant.
- Time constants that are smaller than the “largest
neglected time constant” are approximated as time delays
using (6-58).
20
 Example 6.4
Consider a transfer function:

K ( −0.1s + 1)
G (s) = (6-59)
( 5s + 1)( 3s + 1)( 0.5s + 1)
Chapter 6

Derive an approximate first-order-plus-time-delay model,

− θs
Ke
G ( s ) = (6-60)
τs + 1
using two methods:
(a) The Taylor series expansions of Eqs. 6-57 and 6-58.
(b) Skogestad’s half rule

Compare the normalized responses of G(s) and the approximate

21
models for a unit step input.
 Solution

(a) The dominant time constant (5) is retained. Applying

the approximations in (6-57) and (6-58) gives:

−0.1s + 1 ≈ e −0.1s (6-61)

Chapter 6

and
1 1
≈ e−3s ≈ e−0.5 s (6-62)
3s + 1 0.5s + 1

Substitution into (6-59) gives the Taylor series

approximation, GTS ( s ) :

−0.1s −3s −0.5 s −3.6 s

Ke e e Ke
=GTS ( s ) = (6-63)
5s + 1 5s + 1
22
 (b) To use Skogestad’s method, we note that the largest neglected
time constant in (6-59) has a value of three.

• According to his “half rule”, half of this value is added to the

next largest time constant to generate a new time constant
τ= 5 + 0.5(3) = 6.5.
• The other half provides a new time delay of 0.5(3) = 1.5.
Chapter 6

• The approximation of the RHP zero in (6-61) provides an

additional time delay of 0.1.
• Approximating the smallest time constant of 0.5 in (6-59) by
(6-58) produces an additional time delay of 0.5.
• Thus the total time delay in (6-60) is,

θ = 1.5 + 0.1 + 0.5 = 2.1

23
 and G(s) can be approximated as:
−2.1s
Ke
G Sk ( s ) = (6-64)
6.5s + 1

The normalized step responses for G(s) and the two approximate
models are shown in Fig. 6.10. Skogestad’s method provides
Chapter 6

better agreement with the actual response.

Figure 6.10
Comparison of the
actual and
approximate models
for Example 6.4.

24
 6.4 Interacting and Noninteracting Processes
• Processes with variables that interact with each other or that
contain internal feedback of material or energy(recycle
streams) will exhibit so-called interacting behavior.
• Noninteracting Processes.
H1 ' (s) K1 Q1 ' ( s ) 1
= (6.21) = (6.22)
Qi ' ( s ) τ 1 s + 1 H 1 ' ( s ) K1

H 2 ' (s) K2 Q ' (s) 1

= (6.23) 2 = (6.24)
Q1 ' ( s ) τ 2 s + 1 H 2 ' (s) K 2

H 2 ' ( s ) H 2 ' ( s ) Q1 ' ( s ) H 1 ' ( s )

=
Qi ' ( s ) Q1 ' ( s ) H 1 ' ( s ) Qi ' ( s )
K2
= (6.25)
Figure 6.7. A series configuration of (τ 1 s + 1)(τ 2 s + 1)
two noninteracting tanks.
• Interacting Processes.
 Tank 1 level depends on
tank 2 level ( and vice
versa ) as a result of the
interconnecting stream with
flow rate q1.
1
Figure 6.8. Two tanks in series q1 = (h1 − h2 ) (6.26)
whose liquid levels interact. R1

R1 R2 A2
( R1 + R2 )( s + 1)
H1 ' (s) R1 + R2 K1 ' (τ a s + 1)
= = 2 2 (6.27)
Qi ' ( s ) R1 R2 A1 A2 s + ( R2 A2 + R1 A1 + R2 A1 ) s + 1 τ s + 2ζτs + 1
2

Much more complicated expression than (6-21)

R2
H 2 ' (s) R1 + R2 H 2 ' (s) R2
= (6.28)  = 2 2 (6.29)
H 1 ' ( s ) R1 R2 A2 s + 1 Qi ' ( s ) τ s + 2ζτs + 1
R1 + R2
6.5 Multiple-Input, Multiple-Output(MIMO) Processes
• MIMO system is an industrial process control
applications involve a number of input(manipulated) and
output (controlled) variables. ↔ SISO system.
• Example)

Figure 6.9. A multi-input, multi-output mixing process.

• Process inputs - flow rates of hot and cold streams( wh and wc ).
• Process outputs - the level h in the tank and temperature T .
• Potential disturbances(or load variables) - the temperatures of the
inlet streams Th and Tc .