Control Lecture1 PDF

Biochemical Engineering Process control
Process control
References:
1) Process systems analysis and control by Coughanowr, 3rd edition 2009.
2) Process modeling, simulation and control by Luyben, 1996.
3) Essential of process control by Luyben, 1997.
4) Process control system by Shinskey.
Introduction
Temp.
indicator
Hot water
Water
Heater
Steam
Cold water
(a) Open loop system
(b) Manual Control system (c) Automatic Control system (Closed Loop)
1
TYPICAL FERMENTER MONITORING AND CONTROL SCHEMATIC
2
Control System Objectives

Economic Incentive
Safety
Equipment Protection
Reduce variability
Increase efficiency
Ensure the stability of a process
Elimination of routine
Definitions:
System: it is a combination of components that act together and perform a certain objective.
(disturbance)
Transmitter
Plant: it is the machine of which a particular quantity or condition is to be controlled.
Process: is defined as the changing or refining of raw materials that pass through or remain in a liquid,
gaseous, or slurry state to create end products.
Control: in process industries refers to the regulation of all aspects of the process. Precise control of level, pH,
oxygen, foam, nutrient, temperature, pressure and flow is important in many process applications.
Sensor: a measuring instrument, the most common measurements are of flow (F), temperature (T),
pressure (P), level (L), pH and composition (A, for analyzer).The sensor will detect the value of the
measured variable as a function of time.
Set point: The value at which the controlled parameter is to be maintained.
Controller: The brain of the control system. A device which receives a measurement of the process
variable, compares with a set point representing the desired control point, and adjusts its output to minimize
the error between the measurement and the set point.
Error Signal: The signal resulting from the difference between the set point reference signal and the process
variable feedback signal in a controller.
Feedback: The signal from the sensor, which is fed back to the controller.
3
Feedback Control: A type of control whereby the controller receives a feedback signal representing the
condition of the controlled process variable, compares it to the set point, and adjusts the controller output
accordingly.
Feedforward Control: A type of control which takes corrective action based on disturbances before the
process variable is upset.
Final Control Element: That component of a control system (such as a valve, heater, variec, motor) which
directly changes the manipulated variable.
Steady-State: The condition when all process properties are constant with time, transient responses having
died out.
Transmitter: A device that converts a process measurement (pressure, flow, level, temperature, etc.) into
an electrical or pneumatic signal suitable for use by an indicating or control system.
Controlled variable: process output which is to be maintained at a desired value by adjustment of a process
input.
Manipulated variable: process input which is adjusted to maintain the controlled output at set point.
Disturbance: a process input (other than the manipulated parameter) which affects the controlled parameter.
Process Time Constant ( ): Amount of time counted from the moment the variable starts to respond that it
takes the process variable to reach 63.2% of its total change.
Block diagram: it is a shorthand pictorial representation of the case and effect relationship between the input
and the output of the system. It is easier to visualize the control system in terms of a block diagram.
X(s) Transfer function Y(s)

Input G (s) Output
Block diagram
Transfer Function: it is the ratio of the Laplace transform of output ( response function) to the Laplace
transform of the input ( driving force) under assumption that all initial conditions are zero unless that given
another value.
e.g. the transfer function of the above block diagram is G (s) = Y(s)/X(s)
Closed-loop control system: it is a feedback control system which the output signals has a direct effect upon
the control action.
4
Heater
(Heating) Temperature
Final control element
Transmitter
(Thermocouple)
Advantage: more accurate than the open-loop control system.

Disadvantages: (1) Complex and expensive
(2) The stability is the major problem in closed-loop control system
Open-loop control system: it is a control system in which the output has no effect upon the control
action. (The output is neither measured nor fed back for comparison with the input).
(Timer) Motor
Advantages:
(1) Simple construction and ease of maintenance.
(2) Less expensive than closed-loop control system.
(3) There is no stability problem.
Disadvantages:
(1) Disturbance and change in calibration cause errors; and output may be different from what is
desired.
(2) To maintain the required quality in the output, recalibration is necessary from time to time
Note: any control system which operates on a time basis is open-loop control system, e.g. washing machine,
traffic light etc.
Ex: (1) Draw the block diagram for the following equation:
x3 a1 x1 a2 x2 5 x1 a1
11
+
x2 + x3
a2
5
2) Draw the block diagram for the liquid level control system
Controller
Set point + Controller Control Tank actual level
Desired level valve
-
Float
d 2 x2 dx1
H.W: (1) Draw the block diagram of the equation x3 x1
dt 2 dt
(2) Draw the block diagram of the for the liquid level control system
Controller
6
Forcing functions , Input Variables (Typical test signals):
1- Unit step function
x(t) = 0 , t<0 Denoted by u(t)

1 , t0
1
L u(t) =
s
A
If x(t) = A u(t) L Au(t) = , [ A is the amplitude of the step function ].
s
Step function used for a system subjected to a sudden disturbance.
x(t) x(t)
Step function (A) y(t)

Unit step 1
0 time time
2- Unit impulse; (t) : where the amplitude and the time 0 y(t)
L (t) = 1 x(t)
d
(t) = u (t )
dt
0 t
Used for a system subjected to a shock input. 0

Y ( s)
If the transfer function is G ( s)
X ( s)
Y(s) = X(s)G(s) , [ X(s) is unit impulse ]
Y(s) = G(s)
* If the impulse is A (t) L A(t) = A , [ A is the amplitude of impulse ]
* Also the unit impulse is a unit pulse function with very short time: unit impulse = lim unit pulse
t 0
7
x(t)
y(t)
3- Unit ramp function
x(t) = 0 , t<0
t , t 0 = t u(t) [ unit ramp function ]
1
L t u(t) =
s2
Ramp function used for a system subjected to a gradually change.
x(t)
y(t)
, for unit ramp, K=1

t
4- Sine input function
x(t) = Asin t [ A is the amplitude and is the radian frequency, rad/min ]
A
L x(t) = L Asin t =
s 2
2
1 2
Note: = 2f , f = frequency (Hz) = , Tp = period of oscillation (min/cycle) , = 2f =
Tp Tp
x(t)
8
Mathematical Model of Open Loop Systems (Dynamic Process Without Any Control):
Linear system: it is one in which the equations of the model are linear.
Linear differential Eqn: is a differential equation in which the coefficients are constants or function only of
the independent variable.
Non-linear system: it is one represented by non linear equations.
Linear Eqn Y = 2X + 5 , Non-linear Eqn Y2 = X + 3 or Y = sin X

d2y d 2 y dy 2
Linear Diff. Eqn 3x , Non-linear Diff. Eqn ( ) x A sin x
dx 2 dx 2 dx
Y Y
X X
Linear equation Non-linear equation
First Order Systems:

Ex. 1: Obtain the transfer function of the pressure system shown in figure below.
R P+P
P+Pi
v,c
Assume small deviation in the variable from their steady state
values, then the system consider linear.
Where:
P = gas pressure in the vessel at steady state, N/m2.
Pi = small change in inflow gas pressure, N/m2.
P= in gas pressure in the vessel, N/m2.
V = volume of the vessel, m3 (capacity)
m = mass of gas in the vessel, kg. Pi P
q = mass flow of gas, kg/s = P/R P
= density of gas, kg/m3
C = capacitance, m2 = mass/ pressure (m/p) Time
R = resistance to flow, sec/m2
9
Solution: mass balance: qi qo = dm/dt , qo = 0 ( where qi ,qo & m are the deviations from steady state)
qi = dm/dt but q = P/R
pi p dm
or
dm pi p
but m = pC dm/dt = C dp/dt
R dt dt R
C dp p i p
dt R
dp
CR p pi
dt
Take the Laplace transform: CR s P(s) + P(s) = Pi(s)
P( s )

1
(1st order system)
Pi ( s) CR s 1
Ex. 2: Find the transfer function of thermometer;
- Assume all the resistance to heat transfer

is in the liquid film.
- All Cp is in the mercury.
Solution: Energy balance around the bulb
hA ( x y ) - 0 = m Cp dy/dt
Where: x & y are the temperatures deviation from steady state

m is the mass of mercury , kg
h is the heat transfer Coeff. of mercury, w/m2k.
A is the surface area of bulb, m2.
dy x y ,
dt
mCp
is time constant of thermometer = (sec)
hA
Laplace transform s y(s) + y(s) = x(s)
y( s)

1
x( s ) s 1
H.W : A thermometer having a time constant of 0.1 min is at steady state temp. of 30 oC . at t = 0, the
thermometer is placed in a temperature bath maintained at 50 oC . Determine the time needed for
thermometer to read 45 oC .
10
Ex. 3 : Find the transfer function of the liquid level tank
f = flow rate, m3/s

h = level of water, m
Solution: mass balance R = resistance to flow = h/f2, sec/m2
C = capacitance, m2
f1 - f 2 = Cdh/dt (1)
At steady state f1- f 2 = Cdh/dt = 0 (2)
Eq. 1 Eq. 2
f1* - f 2* = Cdh*/dt
f*, h* are the deviation from steady state, is the density of water and is equal in inflow and outflow
f1* f2* = Cdh*/dt , f2* = h*/R
dh *
RC h * Rf 1* , Laplace transform
dt
RC s H(s) + H(s) =R F1(s)
H ( s) R

F1 ( s) RCs 1
Ex. 4 : Find the transfer function of the mixed reactor , Fo , F & V are constants and Fo= F
Solution: we obtain the process model in terms of deviation variables, Fo, co
. Fo co F c = Vdc/dt , divided by F
dc
+ c = co , = V/F , Laplace transform
dt
s C(s) + C(s) = Co(s)

V, c F, c
C ( s)

1
Co ( s) s 1
H.W: (1) Find the transfer function of above example if the Fo F

(2) Find the transfer function of above example if Fo and co are variables.
11
Ex. 5 : Obtain the T. F. and draw the block diagram of the following thermal system
i , o are steady state temp. of inlet and outlet liquid flow.

o+ o
H is the steady state heat input rate , kJ/sec. H+h hot liquid
Heat is suddenly change to H + h

and temp. of inflow change to i + i ,
so the temp of outflow changed to o+ o i+ i
cold liquid insulator
Solution: heat balance for the system, so the equation which describe the system is:
d o
h + G cp i G cp o = M cp (Assume reference temp. is 0oC)
dt
G is the flow rate of liquid, kg/sec and M is the mass of liquid in tank.
1
Let the thermal resistance, R = ( oC sec/kJ )
G cp
and the thermal capacitance C = M cp ( kJ/oC )
1 1 d
So the Eq. above becomes: h i o C o
R R dt
Laplace transform: R H(s) + i (s) = o(s) + RCs o(s)
R 1
o ( s) H ( s) i ( s)
RCs 1 RCs 1
i(s)
i (s) 1
RCs 1 OR +
H(s) R 1 o (s)
+ + RCs 1
H(s) R o (s)
RCs 1 +
12
Second Order Systems:
Ex. 6: Obtain the transfer function of the following system (no reaction):
F
Where: Fi , ci
F = volumetric flow rate, Fi = F1
c = conc. of solute in stream.
V = liquid volume in tank. Fii cii
V1 c1 F1,c1
Solution: mass balance on concentration; i.e.
In out = accumulation
V2 c2 F2, c2
dc1
Tank 1: Fi ci F1 c1 = V1
dt (non-interacting system)
dc1
1 c1 ci , [1 = V1/F1]
dt
Laplace transform 1 s C1(s) + C1(s) = Ci(s)
C1 ( s)

1
. (1) Ci(s)
1
C1(s)
Ci ( s) 1 s 1 1s 1
dc 2
Tank 2: F1 c1 + Fii cii F2 c2 = V2
dt
dc2
2 c2 K1c1 K 2 cii
dt
Laplace transform 2s C2(s) + C2(s) = K1 C1(s) + K2 Cii(s)
K1 K2
C 2 ( s) C1 ( s) Cii ( s)
2s 1 2s 1
Substitute C1(s) from Eq. (1)
K1 K2
C 2 ( s) Ci ( s) Cii ( s)
( 1 s 1)( 2 s 1) 2s 1
K2
Cii(s) +
Cii(s) K2 + 1
2s 1 + OR K1 + 2s 1 C2(s)
1s 1
C2(s) Ci(s)
Ci(s) K1 +
( 1 s 1)( 2 s 1)
13
Q2 ( s )
Ex. 7: Obtained the T.F. of the liquid level system shown in figure below (i.e. )
Qi ( s )
Where:
q, h = are small deviations from steady state
h1 h2 h
R1 = , R2 = 2
q1 q2
C = capacitance (cross section area)
V = capacity
(Interacting system)
dh
Solution: Tank1 C1 1 qi q1 C1sH1(s) = Qi(s) Q1(s) . (1)
dt
h h2 H1 ( s ) H 2 ( s )
q1 1 Q1 ( s) . (2)
R1 R1
dh2
Tank2 C2 q1 q 2 C2sH2(s) = Q1(s) Q2(s) . (3)
dt
h H ( s)
q2 2 Q2 ( s) 2 . (4)
R2 R2
Substitute Eq.1 and Eq. 4 in Eq. 3 C2sR2Q2(s) = Qi(s) C1sH1(s) Q2(s) . (5)
From Eq. 2 subs. H1 in Eq. 5 C2sR2Q2(s) = Qi(s) C1s(R1Q1(s)+H2(s)) Q2(s) ....(6)
From Eq.3 subs.Q1 in Eq.6 C2sR2Q2(s) =Qi(s) C1sR1C2sH2(s)-C1sR1 Q2(s)-C1sH2(s) Q2(s) . . (7)
Form Eq.4 subs. H2 in Eq.7 C2sR2Q2(s)=Qi(s) C1sR1C2sR2Q2(s)- C1sR1Q2(s)-C1sR2Q2(s) Q2(s) .. (8)
Q2 ( s)

1
=
1
Qi ( s) R1C1 R2 C 2 s ( R1C1 R2 C 2 R2 C1 ) s 1 1 2 s ( 1 2 R2 C1 ) s 1
2 2
H 2 (s)
H.W: Find the T.F. of above system [ ].
Qi ( s )
14
Ex. 8: Obtain the T. F. of the mechanical translation system shown in figure below
k = spring constant, N/m

N
b = viscous friction coefficient,
m
s
Solution: Apply Newtons law for translation system
F=ma
d2y dy d2y dy
u (t ) m 2
b ky(t ) Or m 2
b ky(t ) = u(t)
dt dt dt dt
Taking the Laplace transform
ms2Y(s) + bsY(s) + kY(s) = U(s)
Y ( s) 1
U ( s) ms bs k
2
Eo (s)
Ex. 9: Obtain the T. F. of the L. R. C. circuit shown in figure below ( )
Ei ( s)
L = inductance, Henry
R = resistance, Ohm
C = capacitance, Farad L R
C
ei i eo
Solution: Apply Kirchhoffs law
eL + eR + eC = ei (1)
di 1
C
eL L ..(2) , eR = R i ..(3) eC eo idt ..(4) , substitute in Eqn.1
dt
di
L Ri eo ei , Laplace transform LsI(s) + RI(s) + Eo(s) = Ei(s) (5)
dt
But form Eq. 4 Eo(s) = I(s)/Cs I(s) = Eo(s)Cs , susb. in Eq. 5
LsEo(s)Cs + REo(s)Cs + Eo(s) = Ei(s) Eo ( s)

1
Ei ( s) LCs RCs 1
2
15
Linearization of non-linear function: To convert the non-linear function depends on 2 or more variables:
f f
f(x, y) = f ( x , y ) x, y (x x) x, y ( y y)
x y
e.g. Linearize the following function:
f(x, y) = x y , x 2, y 5
Solution: f(x, y) = x y y ( x x ) x ( y y )
= x y yx y x xy x y
f(x, y) = 5x + 2y -10
H (s) 1/2
Ex.10: Find the T. F. of the following system ; , f2 = K h , K is constant (m2.5/sec)
F (s)
dh
Solution: f1 f2 = C (*)
dt
dh
f1 Kh1/2 = C
dt
by Taylor series expansion:
( H )(h H ) 2
f 2 f 2 ( H ) f ( H )(h H ) f
2
/
2
//
......
2!
Where f 2/ ( H ) is the 1st derivative of f 2 at H , f 2// ( H ) is the 2nd derivative of f 2 at H , etc.
f 2 f 2 ( H ) f 2/ ( H )(h H )
f 2 ( H ) F2 F1 F
1
f 2/ ( H ) KH ( 1 / 2) K1
2
f 2 F K1 (h H ) , substitute in Eq.(*)
dh
f 1 F K1 (h H ) C (take the deviations from steady state i.e. f, h )
dt
dh
C K1 h(t ) f1 (t )
dt
16
C F ( s)
Laplace transform sH(s) + H(s) = 1
K1 K1
H ( s)

K2
, = C/K1 , K2 =1/K1
F1 ( s) s 1
k C ( s)
Ex.11: For the following interacting system, exothermic reactor, where A B , k k o e E / RT . Find A
Ti ( s )
-rA = kcA
Solution: Mass balance on A: qi , CAi , Ti
(In + generation) (out + consumption) = Accm.

T coolant
dc
V A qi c Ai (qc A Vk o c A e E / RT ) (1)
dt
Reaction term(non-linear)
To , coolant V, , CA, T q, CA, T
Where reaction term, -rA= kcA = kocAe-E/RT
Linearize the non-linear term
R R
Assume R k o e E / RT c A = c (t ) T (t )
c A cA ,T A
T c A ,T
E E / RT
= ko e E / RT c A (t ) 2 o
ke c AT (t ) K C c A (t ) KT T (t ) .. (2)
RT
E E / RT
Where: K C k o e E / RT , K T 2 o
k e cA
RT
Substitute Eq.2 in Eq.1
dc A
V qc A (t ) VK C c A (t ) qc Ai (t ) VK T T (t )
dt
V dc A (t ) q VK T
c A (t ) c Ai (t ) T (t )
(q VK C ) dt q VK C q VK C
K1 K2
Laplace transform C A ( s) C Ai ( s) T ( s)
1s 1 1s 1
q VK T
Where: ( K1 , K2 )
q VK C q VK C
Heat balance: Accum = In (out + heat by reaction)
dT
VC P qC P (Ti T ) hA(T To ) Vk o C A (t )e E / RT H r (Hr = -Ve)
dt
17
After linearize the non-linear term
dT
VC P qCPTi (t ) qCPT (t ) hAT (t ) hATo (t ) H rVK C C A (t ) H rVK T T (t )
dt
dT
VC P T (t )[ qCP hA H rVK T ] qCPTi (t ) hATo (t ) HVKC C A (t )
dt
dT
2 T (t ) K 3Ti (t ) K 4To (t ) K 5 C A (t ) , Laplace transform
dt
K3 K4 K5
T ( s) Ti ( s) To ( s) C A ( s)
2s 1 2s 1 2s 1
The block diagram:
G4(s) G1(s)
Ti(s) K3 CAi(s) K1
2s 1 1s 1
G5(s) G2(s)
To(s) T(s) CA(s)
K4 K2
2s 1 1s 1
G3(s)
K5
In general: 2s 1
CA(s) = CAi(s) G1(s) T(s) G2(s) ..(*)
T(s) = Ti(s)G4(s) + To(s)G5(s) + CA(s)G3(s) substitute in Eq.*
CA(s) = CAi(s)G1(s) G2(s)[Ti(s)G4(s) + To(s)G5(s) + CA(s)G3(s)]
G1 ( s) G2 ( s)G4 ( s) G2 ( s)G5 ( s)
C A ( s) C Ai ( s) Ti ( s) To ( s)
1 G2 ( s)G3 ( s) 1 G2 ( s)G3 ( s) 1 G2 ( s)G3 ( s)
K2 K3
C A ( s) G2 ( s)G4 ( s) s 1 2s 1 K2 K3
= 1 =
Ti ( s) 1 G2 ( s)G3 ( s) K2 K5 ( 1 s 1)( 2 s 1) K 2 K 5
1
1s 1 2 s 1
18
C A ( s) C A (s)
If ? , ?
To ( s ) C Ai ( s )
C A ( s)
Note: can be found by another way:
Ti ( s )
G4(s) G2(s)
Ti(s) K3 T(s) K2 CA(s)
2s 1 1s 1
K5
2s 1
G3(s)
C A ( s) G2 ( s)G4 ( s) K2 K3
=
Ti ( s) 1 G2 ( s)G3 ( s) ( 1 s 1)( 2 s 1) K 2 K 5
Block Diagram Reduction
Canonical form of block diagram (feedback):
R(s) C(s)
G(s)
H(s)
19
Rules of Block Diagram Algebra
sequence Block Diagram Equivalent block diagram
1. cascade
2.Parallel
3.feedback
4. Moving a
summing
point ahead
of a block
Z(s)=X(s)G(s) Y(s)
5. Moving a
summing
Point
beyond
a block
Z(s)=G(s)[X(s) Y(s)]
6. Moving
a take-off
point ahead
of a block
Y(s)=X(s)G(s)
7. Moving a
take-off
point
beyond
a block
B B
8. separate
or merge A A+B-C A A+B-C
summing
point C C
20
Ex.1: Draw the block diagram of the following system from its equations and reduce the block diagram.
dh1
Solution: Tank1 C1 qi q1 C1sH1(s) = Qi(s) Q1(s) . (1)
dt
h1 h2 H 1 ( s) H 2 ( s)
q1 Q1 ( s) . (2)
R1 R1
dh2
Tank2 C2 q1 q 2 C2sH2(s) = Q1(s) Q2(s) . (3)
dt
h2 H ( s)
q2 2 Q2 ( s ) . (4)
R2 R2
1
From Eq.1 H 1 ( s) [Qi ( s) Q1 ( s)] Qi(s) 1 H1(s)
C1 s
C1 s
Q1(s)
H 1 ( s) H 2 ( s)
From Eq.2 Q1 ( s) 1
R1
R1
H1(s) Q1(s)
H2(s)
1
From Eq.3 H 2 ( s) [Q1 ( s) Q2 ( s)]
C2 s 1
Q1(s) C2 s H2(s)
Q2(s)
H 2 ( s) 1
From Eq.4 Q2 ( s ) H2(s) Q2(s)
R2 R2
21
By grouping:
H2(s)
Qi(s) 1 H1(s) 1 Q1(s) 1 H2(s) 1 Q2(s)

C1 s R1 C2 s R2
Q1(s) Q2(s)
C1s R2
1 1 1 1 Q2(s)
C1 s R1 C2 s R2
Qi(s) 1 1 Q2(s)
R1C1 s 1 R2 C 2 s 1
R2C1s
Qi(s)
1 Q2(s)
R1C1R2C2 s ( R1C1 R2C2 R2C1 ) s 1
2
22
Ex.2: For the two inputs one output system below find the transfer matrix.
N(s)
R(s) G1(s) G2(s) C(s)
H(s)
Solution: step1: eliminate from N(s)
C R ( s) G1 ( s)G2 ( s)

R( s) 1 G1 ( s)G2 ( s) H ( s)
G1 ( s)G2 ( s)
C R (s) R( s )
1 G1 ( s)G2 ( s) H ( s)
R(s) G1(s) G2(s) C(s)
Step2: eliminate from R(s)
C N ( s) G2 ( s )
H(s)
N ( s) 1 G1 ( s)G2 ( s) H ( s)
G2 ( s )
C N ( s) N ( s) N(s) C(s)
1 G1 ( s)G2 ( s) H ( s) G2(s)
G1(s) H(s)
C(s) = CR(s) + CN(s)
G1 ( s)G2 ( s) G2 ( s )
R( s ) + N ( s)
1 G1 ( s)G2 ( s) H ( s) 1 G1 ( s)G2 ( s) H ( s)
G1 ( s)G2 ( s) G2 ( s) R(s)
C(s) =
1 G1 ( s)G2 ( s) H ( s) 1 G1 ( s)G2 ( s) H ( s) N(s)
Transfer matrix
23
Time Domain Analysis of Control System
Transient response and steady state response: The time response of a control system consists of two parts:
Transient response: that which goes from initial state to final state (the time required to reach a new steady
state).
And
Steady state response: the manner in which the system output behaves as time approaches infinity.
Let the c(t) is the response, in general:
c(t) = ct(t) + css(t)
Where ct(t): transient response

css(t): st. st. response.
if c(t) = - e-t
The steady state response is the part of the final total response which dose not approaches zero as time
approaches infinity.
Then css(t) = [ since lim (- e-t ) = 0 ]

t
The transient response is the part of the total response which approaches zero as time approaches .
Then ct(t) = - e-t
First Order Systems: A first order systems such as thermometer, liquid level (one tank) etc.
The input-output relation is:
Y ( s)

K
..(*) X(s) Y(s)
s
X ( s) s 1
Where K is the st. st. gain of system & is time constant.
Note: all systems having the same transfer function will exhibit the same output in response to the same
input.
1. Unit step response of 1st order system
If the [x(t)] is a unit step input, then the Laplace transform of x(t) is 1/s [X(s) = 1/s ]

Y(s) = 1 K 1
= K( ) By partial fraction:
1 K a
=
b
s s 1 s s 1
s s 1 s s 1
1 K
a=[ s] s0 = K
-t/
y(t) = K(1-e ) , if K=1 s s 1
1 K
-t/ b=[ (s+1)]s -1/ = -K
y(t) = (1-e ) s s 1
24
-t/
Characteristics of 1st order system for unit step input [y(t) = (1-e )]:
y(t) = 1 , (t )
y(t) = 0 , (t 0)
At t = y(t) = 0.632 or y(t) = 63.2% of its total change,
From the figure below:
dy 1 1
t 0 e t / t 0 is found
dt
Time elapsed 1 2 3 4
y(t)% of final 63.2 86.5 95 98
value
dy 1
t 0 slop
dt
Another method to find the time constant, :
y(t) = x [1- e-t/ ]
x y (t )
e t /
x
0.632
x y (t )
Let =Z
x
y(t)
ln Z = -t/ is found from slop = 1/
0 t
ln Z 1 2 3 4 5
Time
If step change has amplitude = A, K 1
A K
Y ( s)
s s 1
y(t) = AK(1 e-t/ )
25
2. Unit ramp response of 1st order system
1
Since the Laplace transform of a unit ramp is substitute in eqn (*) Note: if
s2
1 K 1 2
Y(s) = 2 = K[ 2 ]
s s 1 s s s 1
y(t) = K(t + e-t/ ) , if K = 1 ,
y(t) = (t + e-t/ ) for t 0
s=-1
3. Unit impulse response of 1st order system

1/
Laplace transform of a unit impulse is 1 [i.e. X(s) = 1]
y(t) = K
, if K=1 y(t) =
1
e t /
s 1
1
y(t) = y(t)
s 1
1
y(t) = e t / for t 0

An important property of linear system:
For a unit ramp input, the output is
y(t) = (t + e-t/ ) for t 0
For the unit step input which is the derivative of unit-ramp input, the output is
dy (t )
y(t)step = ramp
dt
-t/
y(t)step = 1-e for t 0
Also for unit impulse, y(t)impulse=

1
y(t)impulse = e t / for t 0

Note: the non-linear systems do not posses this property.
26
Ex: 1st order system subjected to a step change of magnitude 3. The following differential equation
represents the dynamics of the process
d o (t )
3 o (t ) 4
dt
Find: (1) st. st. gain
(2) time constant
(3) T. F.
(4) o(t)
d o (t )
Solution: the standard form of above system is
o (t ) K i (t )
dt
1 d o (t ) 4
The equation of the problem can be written as o (t ) ..(1)
3 dt 3
By comparing = 1/3
Ki(t) = 4/3 but i(t) = 3
K = 4/9 (st. st. gain)
Or at st. st. there is no diff. Eq.
o (t )
o(t) = 4/3 compare with standard form K or o(t) = K i (t)
i (t )
K i (t) =4/3 , i (t)= 3
K = 4/9
1 d o (t ) 4
Then Eq.1 becomes o (t ) i (t ) (2)
3 dt 9
1 4
Laplace transform of Eq.2 s o ( s) o ( s) i ( s)
3 9
o (s) 4 / 9

i ( s) 1 s 1
3
4/3
i (s) =3/s , o ( s)
1
s( s 1)
3
o(t) = 4/3(1 e-3t) [standard form: o(t) = AK(1 et/)]
C ( s)
Note: if K , this is zero order system, and K is the st. st. gain
R( s)
27
Dead time, time lag, transport lag:
L
f(t) u f(t-D)
s
L output L f (t D ) F ( s )e D
G( s) = =
L input L f (t ) F (s)
D s L
G(s) = e , D (time i.e. min, sec, ..etc)
u
Note: In case of thermal pipes in addition of dead time there is a heat loss
a wall Temp
( s 1) o ( s) a ( s) s /
T. F. G(s) = =e D e D i
( s 1) i ( s) a ( s)
x=0 x=L
o
o
At constant wall temp. (a = 0): e s e
D D /
i
L , = time constant of heat loss = mcP c P D , D = diameter of pipe
D
u UA 4U
o 1
If i is constant i.e. = 0 : ( )(1 e s e
D D /
)
a s 1
Ex: consider a mixing process of 1 m3 volume. If the system is subjected to a step change in feed
concentration of 10 kg/m3 , find the system response co(t) as a function of time.
ci(t) Fi=0.01
Solution: F=0.2m3/s
M. B. on the tank Pipe length = 2m

Cross section area = 0.21
In Out = Accm mc2 / (t )
o
A co(t)
/
dc (t )
Fc i (t ) F1co/ (t ) V
o
L = 2m
dt
dco/ (t )
co/ (t ) Kci (t ) , = V/F1 & K = F/F1
dt
C o/ ( s) K

Ci ( s) s 1
28
Co ( s)
For pipe G(s) = /
e D s
Co ( s)
Co ( s) Co/ ( s) Co ( s) Ke D s
(*)
Ci ( s) Ci ( s) Co/ ( s) s 1
D = L/u , u = F1/A = 0.21/0.21 = 1m/s
D = 2/1 = 2sec , =V/F = 1/0.2 = 5sec K = F/F1 = 0.2/0.21 1 substitute in Eq.(*)
2 s
Co ( s) e
co(t)
C i ( s ) 5s 1
10
10
ci(t) = 10 Ci(s) =
s
t 2
10 e 2 s
Co ( s) , co(t) = 10 (1 e 5
)
s 5s 1
D=2 t
Note: (1) if there is no time lag co(t)
Co ( s) 1 10 1
Co(s) = 10
Ci ( s) 5s 1 s 5s 1
co (t) = 10 (1 e-t/5 )
0 t
(2) If there is a pure delay only (pipe only)
Co ( s)
e D s
Ci ( s)
Ci(s) = 10/s
10 D s
Co(s) = e
s
Co(t) = 10 (t-D) or
Co(t- D) = 10
29
Second Order Systems: The following differential equation describes the 2nd order system
d 2 y(t ) dy (t )
a2 2
a1 ao y (t ) bx(t )
dt dt
d 2 y(t ) dy (t )
2
2 n n2 y (t ) K n2 x(t )
dt dt
b
Where: K n2
a2
ao
n , n natural frequency
a2
= damping factor
Damping = is the progressive reduction of oscillation

n = attenuation
a1
2 n
a2
Y ( s) K n2
So the transfer function is 2 (general form of T. F. of 2nd order system)
X ( s) s 2 n s n2
Y ( s) K 1
OR 2 2 (T = process time constant or characteristic time = )
X ( s) T s 2Ts 1 n
Response of second order

A
1- Step response of second order system (i.e. X(s) = )
s
AK n2
Y(s) = s( s 2 n s n )
2 2
There are fife different cases according to value depend on it the roots of characteristic Eq.
(s2 + 2ns + n2 = 0)
30
(1) 0 < < 1 ; s1 & s2 = n jn 1 2 , under damped case , when overshoot occurs.
(2) = 1 ; s1 & s2 = -n , critically damped case, when the time response is as fast as possible without
overshoot.
(3) > 1 ; s1 & s2 = n n 2 1 , over damped case , when response is lower then critical.
(4) = 0 ; s1 & s2 = n , undamped case , oscillatory
(5) < 0 ; s1 & s2 = n j n 1 2 , negatively damped case
oscillatory
Y(t)/K
under damped case
Critically
damped
over damped case
31
For case 1 (under damped case 0 < < 1 ):
AK n2 ( s 2 n )
AK
1
Y(s) = = 2
s( s 2 2 n s n2 ) s s 2 n s n2

y(t) = AK 1 e nt cos d t sin d t t0
1 2

1 1
2
Or in terms of (T =
1
), y(t) = AK 1
1
e t / T
sin 1 2 t
tan
n 2 T
1
For = 0 , and for step input ; y(t) = AK (1 cosnt)
-nt
For = 1 , and for step input ; y(t) = AK [1 e (1 + nt )]
For > 1 , and for step input ;
1 ( 2 1) nt 1 2 1 )nt

y(t) = AK 1 e e (
2 2 1( 2 1) 2 2 1 ( 2 1)
Step Response for 0 < < 1
Definitions of transient response specifications (Dynamic terms) for 0 < < 1 :
The transient response of a practical control system often exhibits damped oscillations before
reaching steady state. The transient response characteristics of a control system to a unit-step input are:
1. Delay time, td : it is the time required for the response to reach 50% of its final value.
2. Rise time, tr : it is the time required for the response to rise from (10 90 %), (5 95%) or
(0 100%) of its final value. For under damped (0 100%) is used, for over damped (10 90%) is
used.
d
tr , = tan-1 (rad) , d n 1 2 (d = damped natural frequency)
d
3. Peak time, tp : it is the time required for the response to reach the 1st peak of the overshoot.

tp
d
4. Maximum (percent) overshoot, Mp (=A/B) : it is the maximum peak value of the response curve
measured from unity. If the final steady state value of the response differs form unity, then it is
common to use the maximum percent overshoot.
y (t p ) y ()
Maximum percent overshoot = 100%
y ( )
( / 1 2 ) ( / d )
Mp e =e
32
TP
0.5 or 0.2
td
ts
tP
5. Settling time, ts : it is the time required for the response curve to reach and stay within a range
about the final value of 5% or 2%.
4 3
t s 4T (for 2% criteria) , t s (for 5% criteria)

C
6. Decay ratio: the ratio of the sizes of successive peaks and is given by
A
2 / 1 2
Decay ratio = e = (Mp)2
7. Period of oscillation, TP : is the time elapsed between two peaks
2T
TP =
1 2
Notes:
1. The maximum overshoot and rise time conflict with each other. If one is made smaller, the other becomes
larger.
2. For rapid response n must be large.
3. The performance of is 0.3 < < 0.5
33
Impulse Response for 0 < < 1:
AK
X(s) = A Y(s) =
T s 2Ts 1
2 2
1 t / T t
y(t) = AK[ e sin 1 2 ]
1 2 T T
d
Note: Y ( s) impulse sY ( s) step and y(t ) impulse ( y (t ) step )
dt
Ex: For the system shown in figure below, = 0.6 and n = 5 rad/sec. Obtain rise time, tr, peak time, tp,
maximum overshoot, Mp and settling time, ts
when the system is subjected
to a unit-step input. R(s) n2 C(s)
s( s 2 n )
Solution:

(1) t r
d
d n 1 2 5 1 0.6 2 4 rad/sec
d
-1
= tan , = n = 0.6 5 = 3
4
= tan-1 3 = 0.93 rad
0.93
t r = 0.55 sec
4

(2) t p = = 0.79 sec
d 4
( / d ) ( 3 / 4 )
(3) Mp= e =e = 0.095 and maximum percent overshoot = 9.5%
4
(4) t s = 4/3 = 1.33 sec (for 2% criteria),

3
ts = 3/3 = 1 sec (for 5% criteria)

34
Notes: (1) The second order system may be combined from two1st order systems as follows:
K
G( s) , this is for non-interacting systems
( 1 s 1)( 2 s 1)
For interacting system
K
G( s) , A1 = capacitance of the 1st system, R2 = resistance in 2nd system
1 2 s ( 1 2 A1 R2 ) s 1
2
K
Or G( s) , a b = 12 , a + b = 1 + 2 + A1R2
( a s 1)( b s 1)
Where a & b are effective time constants
(2)
2
(3) Effect of KC or K on closed loop
Damping Factor (z ') 1.5
overdamped
1
0.5 underdamped
0
unstable
-0.5
0 2 4 6 8
Kc
35
Higher order systems: If there is multi first order systems in series as shown in multi reactors below
The transfer function of the system is
C1 ( s) K1

Co ( s) 1 s 1
63%
C2 ( s) K2
.etc
C1 ( s) 2 s 1
Cn ( s)

K1 K 2 . . .K n
Co ( s) ( 1 s 1)( 2 s 1)....( n s 1)
t1
D
To simplify the transfer function, there are two methods:
1) If the time constants are equal, then the T. F. becomes
Cn ( s) K K e D s
G( s) , K = K1K2 .Kn , = t1 - D
Co ( s) ( s 1) n s 1
2) If the time constants of each system are not equal then the T. F. becomes
Cn ( s) K e D s
G( s)
Co ( s) ( 1 s 1)( 2 s 1)
Where: 1 & 2 are dominant time constants and D
36
Ex: Find the simplest lower order approximation of the following transfer function
Solution: may approximate the full order function as
Where 1.6 is the sum of dead times 0.1, 0.5, and 1; i.e. 1.6 = 0.1 + 0.5 + 1
If X representing a unit step input, the response of the full order function (solid curve) and that of the first
order with dead time approximation (dotted curve) are shown in
Higher order system may combined from 1st order system and 2nd order system as follows
K
G( s)
( s 1)(T s 2Ts 1)
2
37
Control Action
It is the manner, in which the automatic controller compares the actual value of the process output with the
actual desired value, determines the deviations and produce a control signal which will reduce the deviation
to zero or to small value.
Classification of industrial automatic controller: they are classified according to their control
action as:
1) Two position or on-off controller.

2) Proportional controller (P)
3) Integral controller (I)
4) Proportional plus Integral controller (PI)
5) Proportional plus Derivative controller (PD)
6) Proportional plus Integral plus Derivative controller (PID)
The automatic controller may be classified according to the kind of power employed in the operation,
such as pneumatic controller, hydraulic controller or electronic controller.
Elements of industrial automatic controller
Final element
control or Process
actuator
38
Self operated controller: In this controller the measuring element (sensor) and the actuator in one unit.
It is widely used for the water and gas pressure control.
1) Two position or on-off controller:

It is a special case of proportional controller with very high gain. It is simple and inexpensive such
as relay or open-shut type valve, so it is widely use in both industrial and domestic control systems.
Let m(t) is the output signal from controller [manipulated signal]

e(t) is the actuating error signal [input signal to controller]
so m(t) = M1 for e(t) > 0

M2 for e(t) 0
M1 = on position (maximum)
M2 = off position (minimum) Constants
The two-position controller is used extensively in home heating and cooling systems, refrigerators,
hot water tanks (boiler), air compressors (air conditioner), and other applications where the cost of
more precise control is not justified
To reduce the number of cycling (on-off switching) a controller with differential gap is used to increase
the useful life of components.
39
E M1
M
Set point M1
E M
M2
M2
Without differential gap With differential gap
Differential gap: is a range through which the actuating error signal must move before the switching occurs.
e.g.
Mercury
Without differential gap

With differential gap
Electromagnetic valve
40
2) Proportional control action (P):

For a controller with a proportional control action, the relationship between the output of the controller, m(t),
and the actuating error signal (input to controller) is
m(t) = m + KC e(t)
M(s)
KC
In terms of deviation: m(t) = KC e(t) set point
M ( s)
Laplace transform K C = GC (s)
E ( s)
Output signal
Proportional mode has a continuous linear relationship between the error signal and the controller output.
KC is the controller gain i.e. is the amount by which the error is multiplied to obtain the output. Or called
(Proportional sensitivity).
m is output of controller at steady state.
m(t) output of controller (manipulated variable).
e(t) is the error (difference between the set point and controlled variable)
The proportional controller is essentially an amplifier with an adjustable gain.
Proportional band: It is a term used to define the gain or sensitivity of the proportional controller, so it is
the error required to move the output of the controller from its lowest to its highest value.
Error (expressed in percentage of the range of the controlled variable)
E
PB% = 100
M
1
PB% = 100 That is, the controller output is scaled 0-100% and the error from set point is
KC
scaled 0-100%. Where KC in this equation is dimensionless
Notes:
The primary benefit of proportional action is that it speedup the response of the process.
The smaller proportional band corresponds to high gain or high proportional sensitivity.
Disadvantage of proportional controller is giving a steady state error (off-set).
Proportional controller is often the best type for level control.
Ex: Convert from PB to Kc, where Proportional band is equal to 200%, the range of the error from set
point is 200 psi and the controller output range is 0 to 100%.
Solution:
100% 100% , K cD refers to dimensionless gain
K cD 0.5
PB 200%
100%
K c 0.5 0.25 % / psi
200 psi
41
3) Integral control action (I):

The controller output, m(t), is changed proportional to the time integral of the deviation or error, e(t).
m(t) = m + KI e(t )dt

0
set point KI M(s)
s
In terms of deviation:
t
m(t) = KI e(t )dt

0
M ( s) K I
Laplace transform output signal
E ( s) s
Notes:
For zero actuating error, the value of m(t) remains stationary.
The integral control action is sometimes called reset control.
The basic purpose of integral action is to derive the process back to its set point when it has been
disturbed.
Integral action is needed if it is a desirable to have zeroed off-set.
It is limited in application to processes of small capacitance and fast response which are subject to
large change in load such as are found in some flow systems.
Integral mode is rarely used alone since it adds lag to the system without the benefits of proportional
mode so it is almost used with proportional mode (reduces stability).
4) Proportional plus integral control action (PI):

t
KC
It is defined by the equation: m(t ) m K C e(t )
i e(t )dt
0
t
KC
In terms of deviation: m(t ) K C e(t )
i e(t )dt
0
M ( s) 1
Laplace transform K C (1 ) (Note: for optimum setting let i )
E ( s) is
KC is gain Adjustable
i is integral time
KC
= KI
i
1
is called reset rate, which is the no. of times per minute that the proportional part of the control action is
i
duplicated. Reset rate measured by repeats per minute.
42
M(s)
K C (1 i s) M(s) Step change=A
AKC
is 2AKC
i
A
AKC
(b) Diagram of depicting a step input;

(c) Diagram of controller output.
Ex: PI controller has output initially 50%, Gain = 2, reset = 2 minutes per repeat. The controller is subjected
to a sustained error of 5%, what is the output after 4 minutes?
Solution: Proportional Response = KC e = 2x5=10%

Integral Action in 4 minutes the control will go through 2 repeats.
Integral action = 2 x 10 = 20%
Total output change is proportional + integral = 30%
M ( s) 1 5% 5% K C 5% K C
Or K C (1 ) , E (s) M(s) =
E ( s) is s s is2
m(t) = 5% 2 + 5%2t/i =10% + 20% = 30%
5) Proportional plus derivative control action (PD):
de(t )
It is defined by the equation: m(t) = m +KC e(t) + KC d
dt
(So the controller output, m(t), is proportional to both size of error and rate of change of error)
M ( s)
Laplace transform K C (1 d s)
E ( s)
KC is gain
d is derivative time (rate time)
Derivative time: the time base over which the error change is measured (the time interval by which the rate
action advances the effect of the proportional control action).
The derivative control action sometimes called rate control is where the magnitude of controller output,
m(t), is proportional to the rate of change of the actuating error signal, e(t).
It speed up the control action by anticipating where it measures the rate of change the error and applying a
control action proportional to the rate of change.
Derivative mode is never used alone since its output is zero except when the error signal is changing (ramp
function). So the PD used in slow process such temperature control in reactor.
So its advantages: (1) Fast action (dose not used for fast process e.g. pH, pressure).
(2) Add stability to a closed loop control system.
Disadvantages:
1. Effective only during transient response, so that can never be used alone.
2. If the steady state error of a system is constant with respect to time, the time derivative of this error is
zero and the derivative control has no effect on the steady state error.
3. Amplifies noise signal.
43
m(t)
KC(1+ds) M(s)
Derivative only
(a) Block diagram of a proportional-plus-derivative controller; (b) Diagram depicting a unit ramp input
(c) Diagram depicting the controller output.
6) Proportional-plus-Integral-plus-Derivative control action (PID):
The equation of this combined action is
t
KC de(t )
m(t ) m K C e(t )
i e(t )dt K
0
C d
dt
M ( s) 1
The transfer function is K C (1 d s)
E ( s) is
m(t) M(s)
m
m(t)
K C (1 i s i d s 2 ) M(s)
is
PI
(a) Block diagram of PID controller; (b) Diagram depicting a unit ramp input; (c) Diagram depicting the
controller output.
44
PID controllers are used where signals are not noisy and where tight dynamic response is important.
e.g. temperature controllers in reactors are usually PID.
In bioprocess environment, where the lag of a control loop is usually not a major concern, the most common
form of control is the PI.
Notes:
For large KC more accuracy but approaches instability.
For small i more sensitive but more oscillatory.
For large d increasing in speed of response.
Typical controller outputs to Various Process Inputs and Disturbances
45
Simple Rules for Controller Selection

The proportional controller can be used if nonzero steady-state error is acceptable. P-control is the principle
method of control and should do most of the work. In process control its typical use is in pressure control or
level control.
PI controller can be used if the offset is unacceptable and dynamics of the controlled system is simple. Its
advantage is the zero steady-state error but at oscillatory behaviour and addition time to reach to the desired
value.
PD controller grace to derivative action does not exhibit large overshoot and settles fairly quickly.
However, without integral action, the nonzero steady-state error results with the same value as that of the P
controller. This mode is used when there are excessive lag or inertia problems in the process.
PID controller is suitable in remaining cases - it is robust, forecasts future process behavior. Its drawback is
an increased sensitivity to measurement noise.
Ex: A unit-step change in error is introduced into a PID controller. If Kc = 10, i = 1, and d = 0.5, plot the
response of the controller, m(t).
Solution: The equation of PID controller is
M ( s) 1
K C (1 d s) m(t)
E ( s) is
1
E(s) = slop = 10
s
10 1
M ( s) (1 0.5s) 5
s s
10 10
M ( s) 5 10
s s2
m(t) = 10 + 10t + 5(t) 0 t
Note: in general there are two types of control action:

(1) Reverse control action. (2) Direct control action
Reverse action: Error = SP PV; K = + Ve Direct action: Error = PV SP; K = - Ve
Where the PV or H increases (Fo decrease) the OP decreases Where as the PV increases (tank filling) the OP increases
(closing the inlet valve more) to reduce the filling rate. (opening the outlet valve more) to drain the tank faster .
46
Final control Elements:
Control valve, Heater, Variec, Motor, pump, damper, louver, . etc.
Control valve: that can control the rate of flow of a fluid in proportion to the amplitude of a pressure
(electrical) signal from the controller. From experiments conducted on pneumatic valves, the relationship
between flow and valve-top pressure for a linear valve can often be represented by a first-order transfer
function:
Air supply Air supply
Control valve (Air to close) Control valve (Air to open)
Q( s ) KV

P( s) V s 1
Where: Kv: steady-state gain i.e., the constant of proportionality between steady-state flow rate and valve-
top pressure.
v: time constant of the valve and is very small compared with the time constants of other
components of the control system.
So the T. F. of the valve is approximated by:
Q( s )
KV So, the valve is said to contribute negligible dynamic lag.
P( s)
EX: Consider the 1st order T. F. of the process with control valve
Valve process
P(s) KV KP Y(s)
If we assume no interaction; V s 1 s 1
The T. F. from P(s) to Y(s) is
Y ( s) KV K p
For a unit step input in p
P( s) ( v s 1)( s 1)
1 KV K p 1 t / v 1 t /
Y(s) = y(t) = KvKp 1 v e e
s ( v s 1)( s 1) v v
47
Y ( s) K v K p
If >> v the T. F. is For a unit step input in p
P( s) (s 1)
y(t) = KvKp (1 e-t/ )
Notes:
(1) A typical pneumatic valve has a time constant of the order of 1 sec. Many industrial processes
behave as first-order systems or as a series of first-order systems having time constants that may range from a
minute to an hour. So the lag of the valve is negligible.
(2) The time constant of lag valve depends on the size of valve, air supply characteristics, whether a valve
positioner is used, etc.
Transmitter, Transducer and sensor:
Ex: Find the output if the input is 10 mW.
Solution:
o = i G = 10 50 = 500 mw
Ex: Find the output if the input is 50oC.
Solution:
V
o = i G = 2oC 50 o
= 100 V
C
H.W Determine the current when the temperature is 120oC.
48
Closed Loop Control System

mi Ti
Ex.1: Continuous stirred tank heater
transducer
controller m T
Set point
mi Ti Q
Tm
Tm
(A) (B)
Flow control system
Pressure control system
Block diagram of closed loop control system GL(s)

L(s)
Load
Comparator
Td(s) E(s) M(s) Final control Q(s) T(s)
Controller Process
Set point element
GC(s) GV(s) GP(s)
Tm(s) Measuring
device
Gm(s)
49
Components of control loop:

1. Process (Heating)
2. Controlled variable (Temperature)
3. Measuring device (Thermocouple)
4. Transmitter or Transducer
5. Final control element [Control valve (as in system B) or Heater (as in system A)]
6. Controller
Where:
T = controlled variable
Tm = measured variable
Td = desired variable = set point (TSP)
E = error = Td Tm
Over all closed- loop control system:

G( s ) G ( s)G P ( s)
T ( s) TSP ( s) L L ( s)
1 Gm ( s)G(s) 1 Gm ( s)G(s)
G(s) = GC(s) GV(s) GP(s)
There are two types of control systems:
(1) Servo systems: In these systems the controlled output followed the changing in set point Such as aircraft
and satellite systems. The T.F. of closed loop system of this type is
G(s)
T ( s) TSP ( s)
1 Gm ( s)G( s)
Servomechanism system
50
(2) Regulatory Systems: In these systems the set point (desired value) is constant and the change occurring
in the load such as chemical and biochemical systems. The T.F. of closed loop control system of this type is
G L ( s)G P ( s)
T ( s) L ( s)
1 Gm ( s)G( s)
Ex: Regular loop with the following elements
3
GP(s) = (process)
10s 1
1
GL(s) = (Load)
10s 1
Gm(s) = 1 (measuring device) if not given take 1
GC(s) = 2 (controller)
GV(s) = 1.5 (valve)

GL(s)
Determine the system response for a unit step in load L(s) 1
10s 1
SP(s) o(s)
2 1.5 3
Set point
10s 1
GC(s) GV(s) GP(s)
1
Solution:
Gm(s)
(s) GL ( s )
Regulator loop: o
L ( s) 1 Gm ( s)G( s)
3 9
G(s) = GC(s) GV(s) GP(s) = 21.5 =
10s 1 10s 1
1
o (s)
10s 1 =
1
o(t)
L ( s) 1 9 10s 10
10s 1
1
L(s) = 1/10
s
1
o(s) = 10 o(t) = 1/10 (1 e-t)

s ( s 1)
At t = 0 , o(t) = 0
At t = , o(t) = 1/10 = off-set = st.st. error off-set
Set point
51
EX: Determine the transfer functions C/R, C/U1, and B/U2 for the system show in Fig. below. Also
determine an expression for C in terms of R and U1 for the situation when both set-point change and load
change occur simultaneously.
Solution:
, G = G c G1 G2 G3 H 1 H2
For separate changes in R and U1, we may obtain the response C
C= ,
If both R and U1 occur simultaneously,
Steady state Error (off-set):
It is the difference between the set point (desired point) and the measured value when the time
approaches .
L(s)
Error = e(t) = r(t) b(t) R(s) E(s) C(s)

G1(s) G2(s)
Or E(s) = R(s) B(s) B(s)
In case of the change in setpoint H(s)
B(s) = C(s) H(s)
C ( s) H ( s)
E(s) = 1 R( s)
R ( s )
C ( s) G( s)
But , where G(s) = G1(s)G2(s)
R( s ) 1 G ( s ) H ( s )
52
G( s) H ( s)
E(s) = 1 R( s)
1 G( s) H ( s)
1
E (s) R( s )
1 G( s) H ( s)
s
Steady state error = e(t) = lim sE(s) = lim [ R( s ) ]
t s 0 s 0 1 G( s) H ( s) s 0
In case of the change in load
E(s) = - B(s) = - C(s) H(s)
Steady state error = e(t) = lim sE(s) = [- C(s) H(s)]

t s 0 s 0
Ex: For the following system find the steady state error for
a. Unit step input.

b. Unit ramp input.
c. Unit impulse input. R(s) s2 4 C(s)
s4 s ( s 1)
Solution:
St. st. Error = e(t ) = lim sE (s )

t s0
1
E(s) = R( s) , H(s) = 1
1 G( s) H ( s)
R( s) s( s 1)( s 4) R( s) s( s 1)(s 4)
E(s) = = = 3 R( s )
s2 4 s( s 1)( s 4) 4( s 2) s 5s 2 8s 8
1
s 4 s ( s 1)
1
a. For unit step input, R =
s
s( s 1)(s 4) 1
e() = lim s 3 =0
s0 s 5s 2 8s 8 s
1
b. For unit ramp input, R =
s2
s( s 1)(s 4) 1 4 1
e() = lim s =
s0 s 5s 8 s 8 s
3 2 2
8 2
c. For unit impulse input, R(s) = 1
s( s 1)(s 4)
e() = lim s 1 = 0
s0 s 5s 2 8 s 8
3
53
Construct the controller with the Process

Consider the control system for the heated stirred tank that is represented in the figure (A), the block diagram
of the system is shown in figure (B). We assumed that the transfer function of the valve is constant (KV = 1)
also assumed that the transfer function of measuring element = 1
(A) (B)
Proportional Control for Set-Point Change (Servo Problem)
For proportional control, GC = K, the overall transfer function is
T ( s) KC A /( s 1) KC A

TR 1 KC A /( s 1) s 1 KC A
T ( s) A1

TR ( s) 1 s 1
Where ,
According to this result, the response of the tank temperature to change in set point is first-order, and
If the unit step input in setpoint, the response becomes
T (t ) = A1(1 e-t/1 ) and the final value of T (t ) is <1
Thus, the ultimate value of the temperature T/ () does not match the desired change. This discrepancy is
called offset
Set point
Offset = TR - T ()
54
Proportional Control for Load Change (Regulator Problem)

The same control system shown in Fig. (B) is to be considered. This time the setpoint remains fixed; that is,
TR = 0.
T ( s) A1 A /( s 1) 1

Ti( s) 1 K C A /( s 1) s 1 K C A
T ( s) A2

Ti( s) 1 s 1
Where
The response of the system to a unit-step change in inlet temperature Ti is
T (t ) = A2(1 e-t/1 )
Load
And the final response is T () =
The offset becomes: T (t )
Offset = 0 - T () = 0 =
Setpoint
Proportional-Integral Control for Load Change

In this case, we replace Gc in Fig. (B) by KC(l + l/is) as in figure (C). The overall transfer function for load
change is therefore
T ( s) AA 1 /( s 1) Ti

Ti( s) 1 K C A(1 1 / i s) /( s 1)
Rearranging this gives A-1

TR A T
T ( s) is K C (1 1/ i s)
s 1
Ti( s) ( s 1)( i s) K C A( i s 1)
T ( s) is
Or figure (C)
Ti( s) i s ( K C A i i ) s K C A
2
55
The transfer function may be written in the standard quadratic form
T ( s) ( i / K C A) s

Ti( s) ( i / K C A) s 2 i (1 1 / K C A) s 1
T ( s) As
Or 2 2 1 . (*)
Ti( s) T1 s 2T1 s 1
i 1 i 1 KC A
Where: A1 = i/KCA, T1 = ,
KC A 2 KC A
For a unit-step change in load,
A1
T ( s)
T s 2T1 s 1
1
2 2
This eqn is equivalent to the response of second order system to an impulse function of magnitude A1
T (t ) =
=1
A=1
KC = 3.5
Setpoint
Setpoint
Offset = TR - T () = 0 0 = 0
Proportional-Integral Control for Set-Point Change

Again, the controller transfer function is KC(l + l/is), the transfer function from Fig. (C) is
T ( s) K C A(1 1 / i s) /( s 1)

TR ( s) 1 K C A(1 1 / i s) /( s 1)
This equation may be reduced to the standard quadratic form to give
T ( s) s 1
2 2 i
TR ( s) T1 s 2T1 s 1
Where T1 and are the same functions of the parameters as in Eq. (*).
For unit step change in setpoint ( TR (s)= 1/s)
56
1 is 1 i 1
T ( s) =
s T12 s 2 2T1 s 1 T1 s 2T1s 1 s(T1 s 2T1s 1)
2 2 2 2
i t / T t 1
2 t 1 2

T (t ) = e sin 1 2

+ 1 e t / T
sin 1 tan 1
1 2 T T 2 T
1
Unit-step response for set point change (PI control).
Again notice that the integral action in the controller has eliminated the offset.
Offset = TR - T () = 1 1 = 0
Proportional Control of System with Measurement Lag

We now consider the same control system, the stirred-tank heater of Fig. (A), with a first order measuring
element having a transfer function l/(ms + 1) and proportional controller
The block diagram for the modified system is now shown in Fig. (D)
The T. F. for the set point is
A-1
T ( s) A ( s 1)
2 12 m
TR ( s) T2 s 2T2 s 1
Where:
T2 =
=
Figure (D)
The effect of controller gain and measuring lag on system response for unit-step change in set point are
shown in the figures below
Setpoint
57
Over all control system
Ex: Consider the two tank reactor control system shown in figure below. The volumetric flow rate
( F = 100 ft3/min) contains reactant A ( co = 0.1 lbmol/ft3) which decomposes with 1st order reaction A B
by rate (-rA = kc) where k is rate constant, a function of temperature. k1 = 1/6 min-1 & k2 = 2/3 min-1 where
T2 > T1. The purpose of the control system is to maintain c2 constant by adding pure A to tank 1 through
valve. The sample from tank 2 is withdrawn a rate of 0.1 ft3/min. Another data are:
Mw of A = 100 lb/lbmol, A = 0.8 lbmol/ft3, m = 1 lbmol/min, V1 = V2 = 300 ft3.
Control valve: flow of A through the valve varies 0 2 ft3/min as the valve top pressure varies 3 15 psig.
Neglect v.
Measuring element: the measuring element converts concentration of A to pneumatic signal. The output of
the measuring element varies from 3 -15 psig as the concentration of A varies from 0.01 to 0.05 lb mole/ft3.
Neglect lag (m = 0). The controller is proportional type.
L=50 ft, A=0.001 ft2
Solution: Reactor T. F.
M.B on A; tank 1:
dc1 m
V Fc o m ( F )c1 Vk1c1 . (1)
dt A
m m
<< F F + F
A A
dc1
Eqn (1) becomes V ( F Vk1 )c1 Fco m
dt
dc 1 1 V
1 1 c1 co m , 1 = 2 min
dt k1V k1V F k1V
1 F (1 )
F F
Laplace transform . (2)
c1 0.0733lbmol / ft 3 (This value is evaluated at s = 0)

dc
Tank 2: V 2 Fc1 Fc 2 Vk 2 c2 (3)
dt
By the same way . (4)
58
c2 0.0244 lbmol/ft3
M / A 2 0
Control valve: Gc(s) = KV = = 1/6 ft3/min/psi or
P 15 3
Normal operating pressure, p (st.st.) is
m / A 1/ 0.8
p = pmin + 3 10.5 psi
KV 1/ 6
m m
The Eqn for the valve: KV ( p p ) m(t) = m KV ( p(t ) p) A
A A
50
Transportation lag: d 0.5 min
0.1 / 0.001
d s
T. F. of sample line is e = e0.5 s
B( s ) 15 3
Measuring element: Gm(s) = e-0.5s Km = 300 psi/(lbmol/ft3)
C 2 ( s) 0.05 0.01
Normal pen reading, b = bmin + Km ( c2 c2 min ) = 3 + 300 (0.0244 0.01) = 7.32(t 0.5) psig
The Eqn for measuring device is b(t - 0.5) = 7.32 + Km(c2(t) 0.0244)
P( s)
Controller: GC(s) = KC =
E (s)
The Eqn for controller: p(t) = p + KC e(t),
e(t) = cR b , cR is desired pneumatic signal or set point, psig.
/
At st. st. c R = b = 7.32 [e(t) = 0]
Block diagram Co(s)
C/R(s)
CR(s) + E(s) M(s) + C1(s) C2(s)
P(s)
Km KC KVA
- +
controller Control
valve
B(s)
Km e 0.5 s
OR Co(s)
C/R(s) C2(s)
CR(s) + E(s) P(s) M(s) 1 + C1(s)
Km KC KVA
- F M+
controller Control F
B(s) valve
Km e 0.5 s
59
Stability
A stable system is one where the controlled variable will always settle near the set point. An unstable
system is one where, under some conditions, the controlled variable drifts away from the set point or breaks
into oscillations that get larger and larger until the system saturates on each side.
Stable system unstable system
Mathematically: The control system is stable if the R(s) C(s)

denominator of closed loop transfer function has no
roots with positive real parts.
C ( s) G( s)

R( s ) 1 G ( s ) H ( s )
Absolute Stability: that is whether the system is stable or unstable. A control system is in equilibrium if (in
the absence of any disturbance or input) the output stays in the same state.
In the terms of impulse: the system is stable if its impulse response approaches zero as time approaches
infinity.
Relative stability: it is a degree of stability. That is, if the control system is stable, how close is it to being
unstable.
Rouths stability criterion: It tells us whether or not there are positive roots in a polynomial equation
without actually solving for them. It can also be determined if the system is stable (if there are no roots with
positive real parts). The test limited to systems which have polynomial characteristic Eqn but without
transportation lag.
C ( s) bo s m b1 s m1 .....bm
The closed loop T.F.
R( s) ao s n a1 s n1 .....am
Where aosn+a1sn-1+..an = 0 is characteristic Eqn .(1)
If this Eqn has some roots with positive real parts, then the system is unstable, or some roots equal to zero,
the system is marginally stable (oscillatory), therefore it is unstable.
Then for stability the roots of characteristic Eqn must have negative
real parts (closed-loop poles lie in the left half s-plane). j
Negative part positive part
S-plane
60
Routh array:
Note: if any coefficient in the characteristic Eqn is negative or zero, then the system is unstable, if we
interested in the absolute stability.
If all coefficients are positive then arrange the coefficients are as following:
sn ao a2 a4 .0
sn-1 a1 a3 a5 .0
sn-2 b1 b2 b3 .0
sn-3 c1 c2 c3 .0
s0 ...
a1a2 ao a3 a1a4 ao a5
Where: b1 , b2 , .etc
a1 a1
b1a3 a1b2 b a a b
c1 , c2 1 5 1 3 , etc
b1 b1
If all terms in 1 column have positive signs, the system is stable, if there is change in signs of the 1 st column
st
terms, then the system unstable and the number of roots which have positive real parts are equal to the
number of changes in the signs.
Ex: Apply the Rouths stability criterion to the equation:
s4 + 2s3 +3s2 + 4s + 5 = 0 and determine the number of positive real roots.
Solution:
s4 1 3 5 0
s3 2 4 0
s2 1 5 0
s -6 0
s0 5 0
Hence: the system is unstable and there are two positive real parts because there are two changes in
sign of the coefficients in the 1st column.
Special cases: (1) If any coefficient in the 1st column is zero, but the others are not, then replace it by (very
small number) and continue.
Ex: Apply Rouths criterion to the equation: s3 + 2s2 + s + 2 = 0

And determine the number of positive real roots.
61
Solution: s3 1 1
s2 2 2
s 0
so 2
if the sign of the coefficient above the zero ( ) is the same as the below it, it indicates that there are a pair of
imaginary roots . If the sign of the coefficient above the zero ( ) is opposite that below it, that means one
sign change.
(2) if all the coefficients in any derived row are zero, the evaluation of the rest of the array can be continued
by forming an auxiliary polynomial with the coefficients of the last row and by using the coefficients of the
derivative of this polynomial in the row.
Ex: Determine the number of positive real roots to the equation: s5 + 2s4 +24s3 + 48s2 25s 50 = 0
Solution: The array of the coefficients is
s5 1 24 -25
s4 2 48 -50
s3 0 0 the auxiliary polynomial Eqn P(s) = 2s4 + 48s2 50
dp( s)
8s 3 96s
ds
The terms in the s3 row are
s3 8 96
2
s 24 -50
s 112.7 0
s0 -50
There is 1 positive root with real part.

Application of Rouths stability criterion to control system analysis
Ex: Determine the range of K for stability of the following system
Solution: the closed loop T.F. is R(s) + K C(s)

- s( s s 1)( s 2)
2
C (s) K

R( s ) s( s s 1)(s 2) K
2
The characteristic Eqn is
s4 + 3s3 + 3s2 + 2s + K = 0 [Or characteristic Eqn is 1+ G(s)H(s) = 0 where G(s)H(s) is open loop T.F.]
s4 1 3 K According to 1st column for stability, K must be positive (K > 0)

14
s3 3 2 0 and 2-9K/7 > 0 >K>0
9
s2 7/3 K 0
s 2-9K/7 0
s0 K 0
62
Frequency Response Analysis
It is defined as the steady state behavior of the system when forced by a sinusoidal input. It is the most
important methods for stability analysis and used for design purposes control system.
Suppose the input to the process is sinusoidal signal [x(t) = A sin t]
Where: A is amplitude
1 2
is frequency (Hertz) = or (rad/time) =
T T
T is period of one complete cycle (time)
The output [y(t)] will also be sine wave with the same frequency but different amplitude and phase as shown
below.
Input = A sin t
A B
T Output = B sin (t- )
The frequency response characteristics of a system can be obtained by the sinusoidal T.F. in which s
is replaced by j.
e.g. G(s) s = j G(j) , G(j) is a complex number consists of amplitude and phase.
G(j) = G( j ) e j
Amplitude ratio = AR = G( j ) = Re G( j )2 Im G( j )2 Im (G)

Im G( j )
Phase = = = tan 1 G(j)
Re G( j )
Ex: Find the frequency response of the 1st order of the process with T.F.
Kp
G( s) and the input x(t) = A sin t Arg G(j) Re(G)
s 1
Solution: y(t) = x(t) G(t)
Kp
Substitute s = j in G(s) G( j ) multiply by denominator
1 j
Kp 1 j K p (1 j ) Kp ( K p )
G( j ) . = = j
1 j 1 j 1 2 2 1 2 2 1 2 2
Kp K p
Re [G(j)] = , Im [G(j)] =
1 2 2 1 2 2
63
K p K p K p K p 2 2
2 2 2 2
Kp
AR = G( j ) =
2 2
=
2 2
=
1 1 (1 )
2 2 2
1 2 2
Kp Kp
So G( j ) = =
1 j 1 2 2
K p

1 1 = tan-1 (-) = - tan-1
2 2
Phase = tan
Kp

1 2 2
Hence y(t) = x(t)G(t) = A G( j ) sin(t tan-1) X Kp Y
AK p s 1
y(t) = sin(t tan-1)
1 2 2
1 1 1 1
T.F. s s 1
s 1 s s2 s 2 s 1
2 2
1 1 1 1
AR 2 2 1
1
2 2
2 (1 2 2 ) 2 (2 ) 2
2
-tan-1 -90 -290 90 tan-1 -tan -1
1 2 2
Ex: The stirred-tank heater of figure below has a capacity of 15 gal. Water is entering and leaving the tank at
the constant rate of 600 lb/min. The heated water that leaves the tank enters a well-insulated section of 6-in.-
ID pipe. Two feet from the tank, a thermocouple is placed in this line for recording the tank temperature. The
electrical heat input is held constant at 1,000 kw. If the inlet temperature is varied according to the relation
Where Ti is in degrees Fahrenheit and t is in minutes, find the eventual behavior of the thermocouple reading
Tm. Compare this with the behavior of the tank temperature T. Ignore m for thermocouple. Steady state
Ti = 75oF.

Solution: The deviation variable Ti is
The steady state Tm is found by steady state heat balance

q mc p (Tm Ti )
64
q
Ti =
mc p
Hence Tm = Tm 170
dT
Heat balance on tank: mcp (T - Ti ) =Vcp
dt
T ( s ) 1 V
= = G1(s) , 1 0.202 min
Ti ( s ) 1 s 1 m
T 2
For pipe m e D = G2(s) , D = L/U = = 0.0396 min
T 600 /( 60.2 0.197)
s
Tm ( s ) e D e 0.0396s
= G(s) = G1(s) G2(s)
Ti( s ) 1 s 1 0.202s 1
To find AR and phase, replace s = j
G(j) = G( j )
G( j ) = G1 ( j ) G2 ( j )
= G1 ( j ) G2 ( j )
1 1
G1 ( j ) = = = 0.17
1 1
2 2
(0.202) 2 (46) 2 1
D
G2 ( j ) = e = e 0.039646 = cos(0.0396 46) j sin(0.0396 46)
= cos 2 (0.0396 46) sin 2 (0.0396 46) =1

G1 ( j) tan 1 1 tan 1 0.202 46 1.464 rad 84o
G2 ( j ) D - 0.039646 = -1.822 rad = - 104o
G( j ) 84 104 188o
Tm (t ) Ti G( j ) 0.17 5 sin(46t 188) 0.535 sin(46t 188)
T (t ) Ti G1 ( j ) 0.535 sin(46t 84)
Final values: Tm (t ) 170 0.535 sin(46t 188)

T (t ) 170 0.535 sin(46t 84)
65
Bode Plot
It is a graphical representation of open loop T.F. [G(j)H(j)] to illustrate the relative stability. Bode
plots consist of two graphs: Bode magnitude plot and Bode phase angle plot, where
G( j) H ( j) = G( j ) H ( j ) G( j ) H ( j )
Log magnitude of G(j)H(j) or log magnitude of G(j) in decibel units:
G(j in decibels = 20 log G(j
A- Gain; G(j) = K
L = 20 log K the curve is horizontal line in Bode plot.

If K > 1 , L is positive , if K < 1 L is negative and if K = 1 L is zero. Increasing K moves L curve up and
decreasing K moves L curve down.
Phase angle of G(j) = K is arg G(j) = 0
1
B- First order factor ; G(j) = or G(j) = 1 + j
1 j
1 1 1
For : L = 20 log = 20 log
1 j 1 j 1 2 2
Where L = 0 at << 1 and L = - 20 log at >> 1
arg G(j) = -tan-1 = 0o at << 1 and arg G(j) = -tan-1 = -90o at >> 1
For 1 + j : L = log 1 2 2 where L = 0 at << 1 and L = 20 log at >> 1
arg G(j) = tan-1 = 0o at << 1 and arg G(j) = tan-1 = 90o at >> 1
66
L (dB)
Slope = 20 dB/decade
1 + j
1 Slope = -20 dB/decade
1 j
Log
= -tan-1
tan-1 0 0 0
-tan-1 1/10 1/10 -5.7
1/2 1/2 -26.6
1 1/ -45
Log
2 2/ -63
Bode plot of first order factors
1 -90
For G(j) =
1 j
When 0 , 0
When , -90
1
C- nth power of s or j i.e G(j) = or G(j) = (j)n [integral or derivative]
( j ) n
1 1
For G(j) = : L = 20 log = -20 n log and arg G(j) = -n90o
( j ) n
n
For G(j) = (j)n : L = 20 log n = 20 n log and arg G(j) = n90o
67
log
Bode plots of 1/sn and sn
D
D- dead time; G(j) = e j
G(j = e j D = cos D j sin D = cos 2 D sin 2 D = 1
L = 20 log 1 = 0 and arg G(j) = - D [note here in radian]
L, dB
Log
D
Bode plots of dead time e j
68
1 1
E- Quadratic factor (second order); G(s) = or G(j) =
s 2 s 1
2 2
j 2 1
2 2
1
L = 20 log
(1 ) (2 ) 2
2 2 2
2
arg G(j) = -tan -1
1 2 2
Regardless of value
When << 1 L 0
When >> 1 L 20 log 1/22 = - 40 log
i.e. the slope = -40dB/decade or -12/octave
When = 0 = 0o
When = 1 = -90o
````````````````````````````````````````````````````````````````````` When = = -180o
Log
69
Relative Stability (Degree of stability):

The relative stability indicators gain margin and phase margin are easily determined from Bode plots of
G(j)H(j) as shown:
=1
stable
Phase crossover frequency, : is the frequency at which the phase angle first reaches -180o
G(j)H(j) = -180o
Gain crossover frequency, 1: is the frequency at which the open-loop gain first reaches the value 1.
G( j ) H ( j ) = 1
Gain margin, Kg: is the amount of gain in dB that can be allowed to increase in the loop before the closed
loop system reaches instability.
1
It is given by Kg = or in the terms of dB Kg, dB = -20 log G( j ) H ( j )
G( j ) H ( j )
Phase margin, : is that amount of additional phase lag at which 1 required to bring the system to the verge
of instability. It is given by
= 180 + , where is phase angle of G(j)H(j)
Notes:
1- For stable system the gain margin, Kg, and phase margin, , must be positive.
2- alone or Kg does not give a sufficient indication of the relative stability.
3- For satisfactory performance should be 30 60o and Kg should be greater than 6 dB.
4- For stable system, Kg indicates how much the gain (K) can increased before the system becomes
unstable. For an unstable system Kg is indicates of how much the gain (K) must be decreased to make
the system stable.
70
Tuning of controller
Optimum setting of controller is the final step in prepare of control system to good perform. The best
setting when the decay ratio C/B = 1/4
Ziegler-Nichols method: The Ziegler-Nichols (ZN) controller setting is pseudo standards in the
control field. They are easy to find and to use and give reasonable performance on some loops, and used to
compare with other controller settings. They are often used as first guesses, but they tend to be too
underdamped.
Ex: Find the Ziegler-Nichols settings of the process which has
1 1
GP(s) = (process), H(s) = (measuring device)
(5s 1)(2s 1) 10s 1
Solution: Using proportional controller with KC = 1, the phase crossover frequency can be found by the Eq.
-180o = G(j)H(j)
-180o = - tan-1 5 tan-1 2 tan-1 10
By trial and error = 0.415 rad/min
1
Ku = [Ku is ultimate value of gain]
AR
1
But AR = G( j ) H ( j ) = = 0.08
25(0.415) 1 4(0.415) 2 1 100(0.415) 2 1
2
Ku =
1
= 12.6
0.08
The ultimate value period (Pu) is
71
2 2 Pu
Pu = = = 15.14 min/cycle
0.415
Thus the Ziegler-Nichols settings are
For a proportional controller, KC = Ku/2 = 12.6/2 = 6.3
For a PI controller, KC = Ku/2.2 = 5.7
and i = Pu/1.2 = 15.14/1.2 = 12.62 min
(Note the KC in PI controller becomes less because I mode make the system less stable, so reduce KC to
increase stabilty)
For a PID controller KC = Ku/1.7 = 7.4 , i = Pu/2 = 7.57 min , d = Pu/8 = 1.89 min
(Note KC in PID controller was increased and i was decreased because D mode helps to stabilize the
system. If PD alone were used the same values are taken as in PID mode)
Pu : The period of the resulting oscillation which would occur if a proportional controller with Ku is used.
Note: can be obtained from Bode diagram but without introduce the controller T.F. in OL T.F.
The Ziegler-Nichols settings can be tabulated as following
Controller KC i d
P Ku/2 -------- --------
PI Ku/2.2 Pu/1.2 --------
PID Ku/1.7 Pu/2 Pu/8
Ex: The following block diagram represents two-tank reactor system. Using the Ziegler-Nichols rules,
determine controller settings for various modes of control depending on Bode diagram.
Solution: 1
Open loop T.F. without controller ( s 1)(2s 1)
e 1/ 2 s
G(s)H(s) =
( s 1)(2s 1)
e 1/ 2 s
e 1 / 2 j
G(j)H(j) =
( j 1)(2 j 1)
L = L1 + L2 + L3
72
1 1
L1 = 0 , L2 = 20 log , L3 = 20 log
2 1 4 2 1
Corners: = 1, =
arg G(j)H(j) = -1/2 tan-1 tan-1 2
20
10
-10
-20
-30
-40
-50
-60
-70
-80
0.1 0.5 1 5 10 100 1000
From the phase plot = 1.66 rad/min
from the magnitude plot L = -16.6 dB
AR = log-1 -16.6/20 = 0.148
Ku = 1/AR = 1/0.148 = 6.75
Pu = 2 / = 3.78 min/cycle
Controller KC i (min) d (min)

P Ku/2= 3.38 -------- --------
PI Ku/2.2= 3.1 Pu/1.2=3.15 --------
PID Ku/1.7= 4 Pu/2=1.9 Pu/8=0.47
73
Advance Control
Cascade Control: Cascade control is especially useful in reducing the effect of a load disturbance that
moves through the control system slowly. The inner loop has the effect of reducing the lag in the outer loop,
with the result that the cascade system responds more quickly with a higher frequency of oscillation.
Feed-back (conventional) control Cascade control
(a) Feed-back (conventional) control (b) Cascade control
To prevent the sluggish response of kettle temperature to a disturbance in oil-supply temperature, a

cascade control system shown in figure (b) is proposed. In this system, which includes two controllers and
two measuring elements, the output of the primary controller (master) is used to adjust the set point of a
secondary controller (slave), which is used to control the jacket temperature. Under these conditions, the
primary controller adjusts indirectly the jacket temperature.
74
Analysis of cascade control system
The block diagram of kettle control system is shown in the figure (1):
U
R
GC1 Gc2 GV Gjacket Gkettle To
GP2 GP1
Gm2
Gm1
Figure (1)
To determine the transfer function To/R, the inner loop is reduced to one block as shown in figure below
To GC1Ga GP1

R 1 GC1Ga GP1Gm1 GC 2GV GP 2 GP1
P2 To
1 GC 2GV GP 2Gm 2
To obtain the transfer function relating
output to load, To/U, the block diagram
of figure (1) is reduced to the figure below
Gm1
To GP1Gb

U 1 GC 2 GV Gb GC1Gm1GP1
GP2 GP1 To
GP 2
Gb
GP 2 GC 2 GV Gm 2 GV Gm2
GV Gm1
Ex: Compare conventional control with cascade control, in the following figures.
75
Solution: The conventional control with cascade control were simulated on a computer. The values of KC
and i were chosen by trial and error to give the response to a step change in set point shown as Curve I of
Fig. (A). this response, which has a decay ratio of about 4, was obtained with KC = 2.84 and i = 5. The
Ziegler-Nichols settings (Kc = 3.65 and i = 3.0 ) gave a set-point response that was too oscillatory.
The response of the system to a step change in U of 4 units is shown as Curve II of Fig. (B). The load
response for no control (i.e. KC = 0) is also shown as Curve I for comparison.
The controller gain KC2 of the inner loop was chosen arbitrarily to be 10.0. This value was chosen to be high
in order to obtain a fast-responding inner loop, a desirable situation for cascade control. Because of the
introduction of the inner loop, the dynamics of the control system have changed and it is necessary to tune
the primary controller parameters for a good response to a step change in set point. By trial and error,
primary controller settings of KC1 = 1.0 and i = 0.63
Figure (A)
above
Figure (B)
above
Ex: Show that the cascade control system gives a better response than conventional control.
76
Solution: The equivalent single-loop control system of the cascade system, shown in Fig. (c)
Comparing Fig. (a) with Fig. (c) shows that

the use of cascade control has replaced a
second-order critically damped system
represented by the first two blocks of the
plant [1/(s+1)2] with the following
underdamped second-order system:
1 1
Where: K = 10/11 , = ,
11 11
This second-order underdamped system, for which and are small, responds much faster than the critically
damped second-order transfer function of the first two blocks of the open-loop system.
Feedforward Control:
When the disturbances are large, or where the process dynamics are sluggish, feedback control results
in significant and sustained errors. The basic notation of Feedforward control is to detect disturbances as they
enter the process and make adjustment in manipulated variables before the output variable (controlled
variable) deviates from the set point.
77
Consider a process with several inputs and an output that is to be controlled:
The equivalent Laplace domain equation is
GP
GP
Using Feedforward control
The Laplace domain equation is
GP
To specify the feedforward controller

transfer function Gff to minimize the
effect of xd2 on y. Ideally, we want
y*(s) = 0, and this requires that:
GP
78
If the transfer function of the process and the disturbance are first order plus dead time (FODT):
1 m
K d d
i.e. GP e , Gd 2 e
Ps 1 d s 1
And if the sensor gain is Km and valve gain is KV
K d P s 1 (d m )
G ff e
K m KV d s 1
For more vigorous manipulated variable response is in order from the start, Tlead must be larger than Tlag.
Note: Feed forward is good, but not sufficient: Of course, there will always be another disturbance xd1 that
is not compensated by feedforward control.
So it may be retain a primary loop feedback controller to ensure that y* remains under control.
Tuning Rules for feedforward controller parameters:
lead s 1
The T.F. of the feedforward controller is: G f K f
lag s 1
1- Removing the feedback controller action (GC) by setting the controller to manual.
2- Set the feedforward controller to the computed steady-state gain (Kf) necessary to compensate ultimately
for a step change in Ci . This means that the dynamic portion of Gf (s) will be removed and only the constant
term (Kf) will remain.
3- Make a step change in Ci and observe the open-loop transient of C. The general shapes of the response to
be expected are shown in figure below.
4- If the response shown in Fig. (a) occurs, lead must predominate in Gf(s) [i.e. Tlead > Tlag]. If the response of
Fig. (b) occurs, lag must predominate in Gf(s) [i.e. Tlead < Tlag].
tP = peak time
79
Tuning parameters for feedforward controller

Predominant mode Tlead Tlag
Lead 1.5tP 0.7tP
Lag 0.7tP 1.5tP
Ex: Apply the feedforward tuning rules

to the system in Figure below.
Solution:
After removing the feedback controller

action [GC(s)] and let Gf = Kf = -1 the
equivalent diagram is shown in the
figure (a):
A unit-step change in Ci produces the

transient for C shown as Curve I in
Figure (b). It can be seen that lead
must predominate in Gf (s). The peak
value occurs at tP = 2.
Applying the tuning rules of the feedforward

Controller:
Tlead = 1.5tP = 3
Tlag = 0.7tP = 1.4
The feedforward controller transfer function is Figure (a)
3s 1
Gf (1)
1.4s 1
The result for a unit-step change in Ci

is shown as Curve II in figure (b) for
system in figure (a) with feedforward
controller of T.F. of Eq. (1)
Figure (b)
80
Ratio Control: An important control problem in an industry is the combining of two or more streams to
provide a mixture having a desired ratio of components. Examples of this mixing operation include the
blending of reactants entering a reactor or for the injection of a fuel-air mixture into a furnace.
The figure below is shown a control system for blending two liquid streams A and B with ratio of Kr.
Fluid A qA
Flow-measuring
element
Controller
Kr
Flow-measuring
element
Fluid B qB
PB = Supply pressure
From the block diagram, the flow of B may be written:
Gm1 K r GC GV GP
QB QA + PB
1 GC GV Gm 2 1 GC GV Gm 2
The control action for a flow-control system is usually PI. Derivative action is usually avoided in flow
control because the signal from a flow-measuring element is inherently noisy and the derivative action would
amplify the noise and give poor control
81

Control Lecture1 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Control Lecture1 PDF

Uploaded by

Copyright:

Available Formats

Biochemical Engineering Process control

(a) Open loop system

TYPICAL FERMENTER MONITORING AND CONTROL SCHEMATIC

Control System Objectives

Plant: it is the machine of which a particular quantity or condition is to be controlled.

X(s) Transfer function Y(s)

Advantage: more accurate than the open-loop control system.

Forcing functions , Input Variables (Typical test signals):

1- Unit step function

x(t) = 0 , t<0 Denoted by u(t)

Step function (A) y(t)

Used for a system subjected to a shock input. 0

* If the impulse is A (t) L A(t) = A , [ A is the amplitude of impulse ]

3- Unit ramp function

Ramp function used for a system subjected to a gradually change.

, for unit ramp, K=1

4- Sine input function

x(t) = Asin t [ A is the amplitude and is the radian frequency, rad/min ]

Non-linear system: it is one represented by non linear equations.

Linear Eqn Y = 2X + 5 , Non-linear Eqn Y2 = X + 3 or Y = sin X

First Order Systems:

qi = dm/dt but q = P/R

Ex. 2: Find the transfer function of thermometer;

- Assume all the resistance to heat transfer

Solution: Energy balance around the bulb

Where: x & y are the temperatures deviation from steady state

Laplace transform s y(s) + y(s) = x(s)

Ex. 3 : Find the transfer function of the liquid level tank

f = flow rate, m3/s

At steady state f1- f 2 = Cdh/dt = 0 (2)

RC s H(s) + H(s) =R F1(s)

Solution: we obtain the process model in terms of deviation variables, Fo, co

s C(s) + C(s) = Co(s)

H.W: (1) Find the transfer function of above example if the Fo F

i , o are steady state temp. of inlet and outlet liquid flow.

Heat is suddenly change to H + h

cold liquid insulator

Second Order Systems:

From Eq. 2 subs. H1 in Eq. 5 C2sR2Q2(s) = Qi(s) C1s(R1Q1(s)+H2(s)) Q2(s) ....(6)

Form Eq.4 subs. H2 in Eq.7 C2sR2Q2(s)=Qi(s) C1sR1C2sR2Q2(s)- C1sR1Q2(s)-C1sR2Q2(s) Q2(s) .. (8)

k = spring constant, N/m

Solution: Apply Newtons law for translation system

Taking the Laplace transform

ms2Y(s) + bsY(s) + kY(s) = U(s)

But form Eq. 4 Eo(s) = I(s)/Cs I(s) = Eo(s)Cs , susb. in Eq. 5

LsEo(s)Cs + REo(s)Cs + Eo(s) = Ei(s) Eo ( s)

e.g. Linearize the following function:

by Taylor series expansion:

Where f 2/ ( H ) is the 1st derivative of f 2 at H , f 2// ( H ) is the 2nd derivative of f 2 at H , etc.

Solution: Mass balance on A: qi , CAi , Ti

(In + generation) (out + consumption) = Accm.

After linearize the non-linear term

The block diagram:

T(s) = Ti(s)G4(s) + To(s)G5(s) + CA(s)G3(s) substitute in Eq.*

CA(s) = CAi(s)G1(s) G2(s)[Ti(s)G4(s) + To(s)G5(s) + CA(s)G3(s)]

Block Diagram Reduction

Canonical form of block diagram (feedback):

Rules of Block Diagram Algebra

sequence Block Diagram Equivalent block diagram

Qi(s) 1 H1(s) 1 Q1(s) 1 H2(s) 1 Q2(s)

R(s) G1(s) G2(s) C(s)

Time Domain Analysis of Control System

Let the c(t) is the response, in general: