Lecture #8

2024. 04. 04(Thr.)

Hyun-Seok Cho, Ph.D.

Chemical and Biomolecular Engineering

Sogang University

CBE3006-Spring 2024, _032424_K303

1. Steady-State Gain
The steady-state of a transfer function can be used to calculate the steady-state
change in an output due to a steady-state change in the input. For example,
suppose we know two steady states for an input, u, and an output, y. Then we can
calculate the steady-state gain, K, from:

𝑦lj 2 − 𝑦lj 1
𝐾=
𝑢lj 2 − 𝑢lj 1

If a transfer function model has a steady-state gain, then: 𝐾 = lim 𝐺 𝑠

𝑠→0

2
2. Order of a transfer function model
Consider a general n-th order, linear ODE:

𝑑𝑛 𝑦 𝑑𝑦 𝑛−1 𝑑𝑦 𝑑𝑚 𝑢
𝑎𝑛 𝑛 + 𝑎𝑛−1 𝑛−1 + ⋯ 𝑎1 + 𝑎0 𝑦 = 𝑏𝑚 𝑚 +
𝑑𝑡 𝑑𝑡 𝑑𝑡 𝑑𝑡
𝑚−1
𝑑 𝑢 𝑑𝑢
𝑏𝑚−1 𝑚−1 + ⋯ + 𝑏1 + 𝑏0 𝑢
𝑑𝑡 𝑑𝑡

Take Laplace transfer, assuming the initial conditions are all zero.

𝑌 𝑠 σ𝑚
𝑖=0 𝑏𝑖 𝑠
𝑖
𝐺 𝑠 = = 𝑛
𝑈 𝑠 σ𝑖=0 𝑎𝑖 𝑠 𝑖

3
3. Additive property
Suppose that an output is influenced by two inputs and that the transfer functions
are known:
𝑌 𝑠 𝑌 𝑠
= 𝐺1 𝑠 and = 𝐺2 𝑠
𝑈1 𝑠 𝑈2 𝑠

Then the response to changes in both U1 and U2 can be written as:

𝑌 𝑠 = 𝐺1 𝑠 𝑈1 𝑠 + 𝐺2 𝑠 𝑈2 𝑠

The block diagram is:

4
4. Multiplicative property
Suppose that :
𝑌 𝑠 𝑈2 𝑠
= 𝐺2 𝑠 and = 𝐺3 𝑠
𝑈2 𝑠 𝑈3 𝑠

Then, 𝑌 𝑠 = 𝐺2 𝑠 𝑈2 𝑠 𝑎𝑛𝑑 𝑈2 𝑠 = 𝐺3 𝑠 𝑈3 𝑠

𝑌 𝑠 = 𝐺2 𝑠 𝐺3 𝑠 𝑈3 𝑠

5
numerator = 1;
denominator = [2,3,1];
sys = tf(numerator,denominator);
yyaxis left
plot(step(sys));
yyaxis right
plot(impulse(sys));

6
numerator = 0.1;
denominator = [1,0.02]; %define the first order transfer function
sys = tf(numerator,denominator); % transfer function
step (sys) % plotting step response
impulse (sys) % plotting impulse response
ramp(sys); % plotting ramp response

transfer_function = tf(zpk(zero, pole, gain)) = tf(zpk(2,1,3))

transfer_function =

3s-6
-------
s-1

7
function [y,t]=impresp(num,den,tO,dt,tf )

%Program for calculation of impulse response of strictly proper SISO systems

%num = numerator polynomial coefficients of transfer function
%den = denominator polynomial coefficients of transfer function
%(Coeff icients of 'num' and 'den' are specified as a row vector, in
%decreasing powers of 's')
%tO = time at which unit impulse input is applied
%dt = time-step (should be smaller than 1/ (largest natural freq.))
%tf = final time for impulse response calculation
%y = impulse response; t= vector of time points

8
#7. The controller is in the forward path, and the closed-loop transfer function

Error signal
𝐸𝑎 𝑠
+ 𝑠+1 𝑈 𝑠 1
𝑌 𝑠
𝑅 𝑠 𝐺𝑐 𝑠 = 𝐺 𝑠 =
- 𝑠+2 500𝑠2
Reference Controller Process
input
Unity feedback

#8. The controller is in the feedback path, and the closed-loop transfer function
𝐸𝑎 𝑠
+ 1
𝑅 𝑠 𝐺 𝑠 = 𝑌 𝑠
- 500𝑠2
Process

𝑠+1
𝐻 𝑠 =
𝑠+2
Controller
9
Negative feedback control system

Controller Actuator Process

𝐸𝑎 𝑠 𝑍 𝑠 𝑈 𝑠
+ 𝐺𝑐 𝑠
𝑅 𝑠 𝐺𝑎 𝑠 𝐺 𝑠 𝑌 𝑠
-

𝐻 𝑠

Controller 𝑌(𝑠) G(s)Ga(s) Gc(s)

𝑇 𝑠 = =
𝑅(𝑠) 1+ G(s)Ga(s) Gc(s) H(s)

𝐸𝑎 𝑠 = R(s) – B(s) = R(s) – H(s)Y(s)

𝑌 𝑠 = G(s)U(s) = G(s)Ga(s)Z(s) = G(s)Ga(s) Gc(s) Ea(s)

𝑌 𝑠 = G(s)Ga(s) Gc(s) [Ea(s)] = G(s)Ga(s) Gc(s) [R(s) – H(s)Y(s)]

𝑌 𝑠 [1+ G(s)Ga(s) Gc(s) H(s)]= G(s)Ga(s) Gc(s)R(s)

10
Assumptions:
1. Perfect mixing; thus, the exit temperature T is also the temperature of the tank contents.
2. The liquid volume V is constant because the inlet and outlet flow rates are equal.
3. The density and heat capacity C of the liquid are assumed to be constant.
Thus, their temperature dependence is neglected.
4. Heat losses are negligible.
11
 Suppose the process is initially at steady state:
Define steady-state values as below

ሜ 𝑇𝑖 0 = 𝑇ሜ𝑖 , 𝑄 0 = 𝑄ሜ
𝑇 0 = 𝑇,

For steady state conditions:

0 = 𝑤𝐶 𝑇ሜ𝑖 − 𝑇ሜ + 𝑄ሜ
dT
V C = wC (Ti − T ) + Q (2-36)
Subtraction:
dt
𝑑𝑇
𝑉𝜌𝐶 = 𝑤𝐶 𝑇𝑖 − 𝑇ሜ𝑖 − 𝑇 − 𝑇ሜ + 𝑄 − 𝑄ሜ
𝑑𝑡

12
 𝑑𝑇
𝑉𝜌𝐶 = 𝑤𝐶 𝑇𝑖 − 𝑇ሜ𝑖 − 𝑇 − 𝑇ሜ + 𝑄 − 𝑄ሜ
𝑑𝑡

ሜ 𝑇𝑖 0 = 𝑇ሜ𝑖 , 𝑄 0 = 𝑄ሜ
𝑇 0 = 𝑇,

ሜ
𝑇 ′ ≜ 𝑇 − 𝑇, 𝑇𝑖′ ≜ 𝑇𝑖 − 𝑇ሜ𝑖 , 𝑄′ ≜ 𝑄 − 𝑄ሜ

where we have introduced the following

“deviation variables”, also called
“perturbation variables”:

The advantage of using deviation 𝑑𝑇 ′

variables is that the initial condition 𝑉𝜌𝐶 = 𝑤𝐶 𝑇𝑖′ − 𝑇 ′ + 𝑄′
𝑑𝑡
term becomes zero.
Laplace transform:

𝑉𝜌𝐶 𝑠𝑇 ′ 𝑠 − 𝑇 ′ 𝑡 = 0 = 𝑤𝐶 𝑇𝑖′ 𝑠 − 𝑇 ′ 𝑠 − 𝑄′ 𝑠

13
𝑉𝜌𝐶 𝑠𝑇 ′ 𝑠 − 𝑇 ′ 𝑡 = 0 = 𝑤𝐶 𝑇𝑖′ 𝑠 − 𝑇 ′ 𝑠 − 𝑄′ 𝑠

𝑇′ 𝑡 = 0 𝑇 ′ ≜ 𝑇 − 𝑇ሜ

𝑇 ′ 0 = 𝑇 0 − 𝑇ሜ

Initially steady state

𝑇 0 = 𝑇ሜ 𝑇′ 0 = 0

𝐾 1
𝑇′ 𝑠 = 𝑄′ 𝑠 + 𝑇𝑖′ 𝑠
𝜏𝑠 + 1 𝜏𝑠 + 1

1 𝑉𝜌
𝐾≜ and 𝜏 ≜
𝑤𝐶 𝑤

14
 𝐾 1 1 𝑉𝜌
𝑇′ 𝑠 = 𝑄′ 𝑠 + 𝑇𝑖′ 𝑠 𝐾≜ and 𝜏 ≜
𝜏𝑠 + 1 𝜏𝑠 + 1 𝑤𝐶 𝑤

1) Suppose Ti(t) is constant 2) Suppose Q(t) is constant

at the steady-state value at the steady-state value

𝑇𝑖 𝑡 = 𝑇ሜ𝑖 ⇒ 𝑇𝑖′ 𝑡 = 0 ⇒ 𝑇𝑖′ 𝑠 = 0 𝑄 𝑡 = 𝑄ሜ ⇒ 𝑄′ 𝑡 = 0 ⇒ 𝑄′ 𝑠 = 0

𝑇′ 𝑠 𝐾 𝑇′ 𝑠 1
= =
𝑄′ 𝑠 𝜏𝑠 + 1 𝑇𝑖′ 𝑠 𝜏𝑠 + 1

Transfer function Transfer function

The two transfer functions show the individual effects of Q and Ti on T.

What about simultaneous changes in both Q and Ti ?

15
Linear models which can be transformed into Transfer Function models.

Most physical processes and physical models are nonlinear.

Over a small range of operating conditions, the behavior may be approximately linear.

Linear approximation can be useful.

The approximate linear models can be obtained analytically by a “Taylor Series Expansion” of
a nonlinear function.

𝑑𝑦 𝜕𝑓 𝜕𝑓 𝑑𝑦 ′ 𝜕𝑓 ′
𝜕𝑓
= 𝑓 𝑦, 𝑢 𝑓 𝑢, 𝑦 = 𝑓 𝑢,lj 𝑦lj + ′
ቚ𝑠 𝑢 + ቚ𝑠 𝑦 ′ = ቚ𝑠 𝑢 + ቚ𝑠 𝑦 ′
𝑑𝑡 𝜕𝑢 𝜕𝑦 𝑑𝑡 𝜕𝑢 𝜕𝑦

where 𝑢′ = 𝑢 − 𝑢lj , 𝑦 ′ = 𝑦 − 𝑦lj and subscript s denotes the steady state,

Note that the partial derivative terms are actually constants because they have been
evaluated at the nominal operating point, s.

16
 𝑑ℎ
Mass balance: 𝐴 = 𝑞𝑖 − 𝑞
𝑑𝑡
Valve relation:
A = area, Cv = constant 𝑞 = 𝐶𝑣 ℎ

𝑑ℎ 1 𝐶𝑣
= 𝑞𝑖 − ℎ = 𝑓(ℎ, 𝑞𝑖 )
𝑑𝑡 𝐴 𝐴

with 𝑦 = ℎ and 𝑢 = 𝑞𝑖 .

Thus, we can utilize to linearize around ℎ = ℎሜ and 𝑞𝑖 = 𝑞lj 𝑖 ,

𝑑ℎ′ 𝜕𝑓 𝜕𝑓
= ℎ′ + 𝑞𝑖′ 𝑑𝑦 ′ 𝜕𝑓 𝜕𝑓
𝑑𝑡 𝜕ℎ 𝑠 𝜕𝑞𝑖 𝑠 ′
= ቚ𝑠 𝑢 + ቚ𝑠 𝑦 ′
𝑑𝑡 𝜕𝑢 𝜕𝑦

17
 𝑑ℎ′ 𝜕𝑓 𝜕𝑓
= ℎ′ + 𝑞𝑖′
𝑑𝑡 𝜕ℎ 𝑠
𝜕𝑞𝑖 𝑠

where:
𝜕𝑓 𝐶𝑣 𝜕𝑓 1
=− =
𝜕ℎ 𝑠 2𝐴 ℎሜ 𝜕𝑞𝑖 𝑠
𝐴

𝑑ℎ′ 1 ′ 𝐶𝑣
= 𝑞𝑖 − ℎ′
𝑑𝑡 𝐴 2𝐴 ℎሜ

This model can be expressed in terms of a valve resistance, 𝑅

𝑑ℎ′ 1 ′ 1 ′ 2 ℎሜ
= 𝑞𝑖 − ℎ with 𝑅 ≜
𝑑𝑡 𝐴 𝐴𝑅 𝐶𝑣

18
Consider a transfer function:

𝑌(𝑠) 𝑑
=
𝑈(𝑠) 𝑏𝑠 + 𝑐

(a)What is the steady-state gain?

(b)For a step change of magnitude M in input, describe the output response.

19
Consider the following transfer function:
_
𝑠
𝑌(𝑠) 3𝑒
G(s)= =
𝑈(𝑠) 10𝑠+1

(a) What is the steady-state gain?

(b) What is the time constant (t)?

(c) If U(s)=4/s, what is the value of the output y(t) when t goes to infinity?
_
1−𝑒 𝑠
(d) If U(s)= 𝑠 , the unit rectangular pulse, what is the output when t goes to infinity?

(e) If u(t) = d(t), the unit impulse at t=0, what is the output when t goes to infinity?

(f) If u(t)=5sin2t, what is the value of the output when t goes to infinity?

20
The contents of the stirred-tank heating system shown in below figure are heated
at a constant rate of Q (J/s) using a gas fired heater. The flow rate w(g/s) and
volume V (cm3) are constant, but the heat loss to the surroundings QL(J/s) varies
with the wind velocity v ( cm/s) according t the expressions

U(t)=Ū+bv2(t)
QL=UA(T-Ta)

where Ū, A, b, and Ta are constants. Derive the transfer function between exit
temperature T and wind velocity v. List any additional assumptions that you
make.
21