ME 461

Control Engineering

Semester: July 2021

Dr. Sumon Saha

Professor
Department of Mechanical Engineering
Bangladesh University of Engineering and Technology
Monday, January 10, 2022
 Control Tentative Course Plan
Engineering
➢ Course Overview
➢ Introduction to Control Systems
➢ Mathematical Modelling and Transfer Function
➢ System Response (First and Second Order)
➢ Block Diagram and Transfer Function
➢ Poles and Zeros
➢ Stability Analysis
➢ Closed-loop Controllers
➢ Steady-State Analysis
➢ Applied Problems on Control Engineering
➢ Frequency Response
➢ Root Locus Plot
➢ Bode Plot
 Control Today’s topic
Engineering

Closed Loop Controllers

 Control Closed loop Control system
Engineering
 Control Control Action
Engineering
Question: Write different types of control modes/action.

D controller
 Control Two Position Controller
Engineering
➢ Simplest version of a discontinuous controller.
➢ Two different output states (ON and OFF), for instance pmin
and pmax.

➢ Cheap and used in

residential heating and
domestic refrigerators.

➢ Limited use in process One

control due to discontinuous. controller
switch
point
➢ Controller is essentially just a switch which is activated by the
error signal. Oscillations occurs about the set-point.
 Control Two Position Controller
Engineering
➢ In virtually any practical implementation, there is a neutral zone
where no change in the controller output occurs.
➢ To prevent noise from switching the controller unnecessarily
when the output is nearly
the set point.

➢ Rapid switching occurs for

too narrow hysteresis.

➢ Cycling of controlled
variable  excessive wear on Practical case (dead band/ hysteresis)
control valve.

Two controller switch points

 Control Three Position Controller
Engineering
➢ Three different switching states.
➢ In a temperature control system, these states are not only OFF
and Heating as in a two-position controller, but also Cooling.
➢ So a three-position controller fulfills the function of two coupled
two-position controllers that switch at different states.

xd dead band
xdg differential gap

➢In the field of control valve technology, three-position controllers

are frequently used in combination with electric actuators.
 Control Two Position Controller
Engineering
For the liquid-level control system in the following figure, write
down their differential equations and the corresponding transfer
function both during filling up and emptying the tank. Also draw
their corresponding response curve h (liquid level height) versus t
(time).
 Control Two Position Controller
Engineering
For the liquid-level control system in the following figure, write
down their differential equations and the corresponding transfer
function both during filling up and emptying the tank. Also draw
their corresponding response curve h (liquid level height) versus t
(time).
dh
RC + h = Rqi ( t )
dt
R
G (s) =
RCs + 1
dh
During filling up RC + h = RQ
dt
dh
During emptying RC +h=0
dt
 Control Two Position Controller
Engineering
For the liquid-level control system in the following figure, write
down their differential equations and the corresponding transfer
function both during filling up and emptying the tank. Also draw
their corresponding response curve h (liquid level height) versus t
(time).
dh
RC + h = RQ
dt

h ( t ) = h2 e − t / RC

h ( t ) = h1e − t / RC + RQ (1 − e − t / RC )
 Control Proportional (P) Controller
Engineering
Describe proportional controller with its dynamic behavior.
 Control Proportional (P) Controller
Engineering
 Control Proportional (P) Controller
Engineering
In feedback control, the objective is to reduce the error signal to zero
where
e ( t ) = ysp ( t ) − ym ( t ) (1)
and
e ( t ) = error signal
ysp ( t ) = set point
ym ( t ) = measured value of the controlled/process variable
(or equivalent signal from the sensor/transmitter)
Although Eq. (1) indicates that the set point can be time-varying, in
many process control problems it is kept constant for long periods of
time.
What is proportional control action? Why is it impossible to
completely eliminate the steady-state error of a system using
proportional control alone?
 Control Proportional (P) Controller
Engineering
For proportional control, the controller output is proportional to the error
signal, p (t ) = p + K pe (t ) (2)
where:
p ( t ) = controller output
p = bias (steady-state) value
K p = proportional controller gain (usually dimensionless)
 p − p ( s ) = K p E ( s ) (3)
Transfer function: G ( s ) = K p
Advantage: Easy to implement.

Disadvantage: Always involves an offset in the output response

causing difference between the set-point and the actual output.
 Control Proportional (P) Controller
Engineering

Dynamic Behavior of a P Controller

 Control Proportional (P) Controller
Engineering
Some controllers have a proportional band setting instead of a
controller gain. Proportional band is the range of errors over which
the linear relationship between controller output and error tends to
exist. Given that a controller output signal ranges from a minimum
(COmin) to maximum (COmax) value, then:
COmax − COmin
PB =
Kp
The proportional band PB (in
%) is defined as
100%
PB = (4)
Kp

Disadvantage: Overshoot problem that arises when a proportional

controller is used at high gain.
 Control Problem
Engineering
Problem: Consider a proportional controller which is to be
used to control the height of water in a tank where the water
level can vary from zero to 9.0m. What is the proportional band
and transfer function that will be required if the required height
of water is 5.0m? The controller is to fully close a valve when
the water rises to 5.5m and fully open it when the water falls to
4.5m.

9.0m

5.0m 5.5m (close)

4.5m (open)
 Control Problem
Engineering
Problem: Consider a proportional controller which is to be
used to control the height of water in a tank where the water
level can vary from zero to 9.0m. What is the proportional band
and transfer function that will be required if the required height
of water is 5.0m? The controller is to fully close a valve when
the water rises to 5.5m and fully open it when the water falls to
4.5m.

PB = (0.5+0.5)/9
= 11.2% 9.0m

5.0m 5.5m (close)

Kp = 100/PB 4.5m (open)
= 8.9
 Control Example: P Controller
Engineering
Suppose we have a simple spring-mass-damper system,

The mathematical model of this system is

d 2x dx
m 2 + b + kx = F
dt dt
The transfer function is then obtained as
X (s) 1
G (s) = =
F (s) ms 2 + bs + k
 Control Example: P Controller
Engineering
Let, m = 1 kg, b = 10 Ns/m, k = 20 N/m, f = 1 N

By plugging these values in the transfer function:

X (s) 1
G (s) = = 2
F ( s ) s + 10s + 20

Now, n = 20 = 4.47 and ( )

 = 10 / 2 20 = 1.12  1

So it is a over-damped system.
Our goal is to investigate the influence of different controllers to
obtain:
➢ Fast rise time
➢ Minimum overshoot
➢ No steady-state error.
 Control Example: P Controller
Engineering
No Controller

1
F(s) X(s)
s + 10 s + 20
2

The (open) loop transfer function for unit step input is given by
X (s) 1
G (s) = = 2
F ( s ) s + 10s + 20
1
and for unit step input F ( s ) =
s
The steady-state value for the output is
1
xss = lim x ( t ) = lim sX ( s ) = lim sG ( s ) F ( s ) =
t → s →0 s →0 20
 Control Example: P Controller
Engineering
No Controller: Open-loop step response

➢ 1/20 = 0.05 is the final value of the output to an unit step input.
➢ This corresponds to a steady-state error of 95%, quite large!
➢ The settling and the rise time (10% to 90% for over-damped
system) are about 1.6 and 0.88 sec respectively.
 Control Example: P Controller
Engineering
Proportional Controller (negative unity feedback system)

F(s)
1 X(s)
+ Kp
s 2 + 10 s + 20
−

The closed-loop transfer function for unit step input is given by

Kp
X (s) s 2
+ 10 s + 20 Kp
= = 2
F (s) Kp s + 10 s + ( 20 + K p )
1+ 2
s + 10 s + 20
The steady-state value for the output is
X (s) Kp
xss = lim x ( t ) = lim sX ( s ) = lim s F (s) =
t → s →0 s →0 F (s) 20 + K p
 Control Example: P Controller
Engineering
Proportional Controller: Closed-loop step response
X (s) 300
Kp = 300 = 2
F ( s ) s + 10 s + 320

n = 17.9  = 0.28  1

>> g = tf([300],[1 10 320])

>> stepinfo(g,'Risetimelimits',[0 1])
>> damp(g)
>> step(g)

The above plot shows that the proportional controller reduces both the
rise time (Tr = 0.108 sec for underdamped system) and the steady-state
error, increases the overshoot, and decreases the settling time (Ts =
0.77 sec) by small amount.
 Control Change in gain in P
Engineering controller

• Increase in gain:
→ Upgrade both steady-state and transient responses
→ Reduce steady-state error → Reduce stability!
 Control P controller with High gain
Engineering
 Control Home Work: P controller
Engineering
For the liquid-level control system in the following figure, write
down their differential equations and the corresponding transfer
function. Also draw their corresponding response curve h (liquid
level height) versus t (time).

Kp

qo
 Control Integral (I) Controller
Engineering
For integral control, the rate of change of the controller output is
proportional to the error signal,
dp ( t )
= K I e (t )
dt
p ( t ) = p + K I  e ( t ) dt
where:
p ( t ) = controller output
p = bias (steady-state) value
KI = integral controller gain (usually dimensionless)
 p − p ( s ) = K I E ( s ) / s
KI
Transfer function: G (s) =
s
Advantage: Improve the steady-state performance.
 Control Integral (I) Controller
Engineering
➢ Integral control action is used to fully correct system deviations at any
operating point.
➢ As long as the error is nonzero, the integral action will cause the value
of the manipulated variable to change.
➢ The higher the integral action coefficient KI, the greater the integral
action of an I controller.
✓ No steady-state error
✓ Sluggish response at low KI
✓ At high KI, the control loop tends to oscillate/may become instable
Explain what happens when integral controller used alone?
The integral controller may lead to oscillatory output and hence
has poor stability, although it improves the steady state
performance.
 Control Integral (I) Controller
Engineering

Dynamic Behavior of an I Controller

 Control Example: I Controller
Engineering
Integral Controller

F(s)
KI 1 X(s)
+ s s 2 + 10s + 20
−

The closed-loop transfer function for unit step input is given by

KI
X (s)
=
s ( s 2
+ 10 s + 20 )
=
KI
F (s) 1+ KI s ( s 2 + 10 s + 20 ) + K I
s ( s 2 + 10 s + 20 )
3rd order system
The steady-state value for the output is
X (s)
xss = lim x ( t ) = lim sX ( s ) = lim s F (s) = 1
t → s →0 s →0 F (s)
 Control Example: I Controller
Engineering
Integral Controller: Closed-loop step response

KI = 1

X (s) 1
= 3
F ( s ) s + 10 s 2 + 20 s + 1

The above plot shows that the integral controller eliminates the steady-
state error, and increases the settling time (Ts = 76.7 sec).
 Control Change in gain in I controller
Engineering

• Increase in gain:→ Upgrade only transient responses.

→ appear oscillatory response→ Reduce stability!
 Control I controller with High gain
Engineering

KI = 200
 Control PI Controller
Engineering
For PI control, the equation describing its action can be written as,

p ( t ) = p + K p e ( t ) + K I  e ( t ) dt

where, Transfer function:

KI
G (s) = K p +
s
 Control PI Controller
Engineering
Effect of reducing the rise time (never eliminate
P the steady state error)
Effect of eliminating the steady state error (may
I make the transient response worse)
➢ No offset error.
➢ In this combination, one P and one I controller are connected in
parallel.
➢ If properly designed, they combine the advantages of both
controller types (stability and rapidity; no steady-state error), so
that their disadvantages are compensated for at the same time.
Explain why integral controller is not usually used independently
rather it is generally used with proportional controller.
 Control PI Controller
Engineering

Dynamic Behavior of an PI Controller

 Control Example: PI Controller
Engineering
Proportional-Integral Controller

F(s)
KI 1 X(s)
Kp +
+ s s 2 + 10s + 20
−

The closed-loop transfer function for unit step input is given by

K p + KI / s
X (s) s 2
+ 10 s + 20 K ps + KI
= = 3
F (s) K + K / s s + 10 s 2 + ( 20 + K p ) s + K I
1+ 2 p I

s + 10 s + 20
3rd order system
The steady-state value for the output is
X (s)
xss = lim x ( t ) = lim sX ( s ) = lim s F (s) = 1
t → s →0 s →0 F (s)
 Control Example: PI Controller
Engineering
Proportional-Integral Controller: Closed-loop step response

Kp = 30
KI = 70
X (s) 30 s + 70
= 3
F ( s ) s + 10 s 2 + 50 s + 70

The above plot shows that the PI controller eliminates the steady-state
error, reduces rise time and increases the overshoot.
 Control Change in gain in PI
Engineering controller

• Increase in gain:→ Upgrade only transient response.

→ increase settling time slightly → increase oscillation & overshoot!
 Control Derivative (D) Controller
Engineering
For derivative control, the controller output is proportional to the
derivative of the error signal,
de ( t )
p (t ) = p + Kd
where: dt
p ( t ) = controller output
p = bias (steady-state) value
K d = derivative controller gain (usually dimensionless)

 p − p  ( s ) = K d sE ( s )
Transfer function: G ( s ) = sK d
 Control Derivative (D) Controller
Engineering
➢ Derivative controllers generate the manipulated variable from the
rate of change of the error and not –as P controllers –from their
amplitudes.
➢ These react much faster than P controllers: even if the error is
small, derivative controllers generate – by anticipation – large control
amplitudes as soon as a change in amplitude occurs.
➢ A steady-state error signal, however, is not recognized by D
controllers, because regardless of how big the error, its rate of change
is zero.
➢ Therefore, derivative-only controllers are rarely used in practice.
They are usually found in combination with other control elements,
mostly in combination with proportional control.
 Control Derivative (D) Controller
Engineering

Dynamic Behavior of a D Controller

 Control Example: D Controller
Engineering
Derivative Controller

F(s)
1 X(s)
sK D
+ s 2 + 10 s + 20
−

The closed-loop transfer function for unit step input is given by

sK D
X (s) s 2
+ 10 s + 20 sK D
= = 2
F (s) 1+ sK D s + (10 + K D ) s + 20
s 2 + 10 s + 20
The steady-state value for the output is
X (s)
xss = lim x ( t ) = lim sX ( s ) = lim s F (s) = 0
t → s →0 s →0 F (s)
 Control Example: D Controller
Engineering
Derivative Controller: Closed-loop step response

KD = 10

X (s) 10 s
= 3
F ( s ) s + 20 s 2 + 20

The above plot shows that the derivative controller alone never
eliminates the steady-state error, rather makes it worse.
 Control PD Controller
Engineering
For PD control, the equation describing its action can be written as,
de ( t )
p (t ) = p + K pe (t ) + Kd
dt
where, Transfer function:
G ( s ) = K p + sK d
 Control PD Controller
Engineering
Effect of reducing the rise time (never eliminate
P the steady state error)
Effect of increasing stability, reducing overshoot
D and improving transient response (no response
for steady state error)

• The control response for steady-state error in PD controllers is

just as it occurs in P controllers.
• Due to the immediate control action whenever there is a change
in the error signal, the control dynamics is faster than with P
controllers.
• Disadvantage: A change in set value will require an offset error.
 Control PD Controller
Engineering

Dynamic Behavior of an PD Controller

 Control Example: PD Controller
Engineering
Proportional-Derivative Controller

F(s)
1 X(s)
K p + sK D
+ s 2 + 10s + 20
−

The closed-loop transfer function for unit step input is given by

K p + sK D
X (s) s 2
+ 10 s + 20 K p + sK D
= = 2
F (s) K + sK s + (10 + K D ) s + 20 + K p
1+ 2 p D

s + 10 s + 20
The steady-state value for the output is
X (s) Kp
xss = lim x ( t ) = lim sX ( s ) = lim s F (s) =
t → s →0 s →0 F (s) 20 + K p
 Control Example: PD Controller
Engineering
Proportional-Derivative Controller: Closed-loop step response

Kp = 300
KD = 10
X (s) 10 s + 300
= 2
F ( s ) s + 20 s + 320

The above plot shows that the PD controller reduces both overshoot
and settling time (Ts = 0.29 sec) and small change of steady-state error
for using P controller.
 Control Change in gain in PD
Engineering controller

• Increase in gain:→ Upgrade only transient response.

→ Decrease rise time and overshoot → steady-state error remains!
 Control PID Controller
Engineering
For PID control, the equation describing its action can be written as,
de ( t )
p ( t ) = p + K p e ( t ) + K I  e ( t ) dt + K d
dt
where, Transfer function:
KI
G (s) = K p + + sK d
s
 Control PID Controller
Engineering
In the time domain:

de ( t )
p ( t ) − p = K p e ( t ) + K I  e ( t ) dt + K d
dt
 1 de ( t ) 
= K p e ( t ) +  e ( t ) dt + Td 
 Ti dt 
Kp
Integral time constant: Ti =
Ki

Kd
Derivative time constant: Td =
Kp

 1 
G ( s ) = K p 1 + + sTd 
 sTi 
 Control PID Controller
Engineering
Effect of reducing the rise time (never eliminate
P the steady state error)
Effect of eliminating the steady state error (may
I make the transient response worse)
Effect of increasing stability, reducing overshoot
D and improving transient response (no response
for steady state error)
• It is also known as three-mode controller.
• No offset error.
• Tendency for oscillations is reduced.
• Requires proper tuning of PID parameters in accordance with
the system dynamics and response requirement.
• Tuning is inexpensive, but the implementation is time-
consuming.
 Control PID Controller
Engineering

Dynamic Behavior of an PID Controller

 Control Example: PID Controller
Engineering
Proportional-Derivative Controller

F(s)
KI 1 X(s)
Kp + + sK D
+ s s 2 + 10s + 20
−

The closed-loop transfer function for unit step input is given by

K p + K I / s + sK D
X (s) s 2
+ 10 s + 20 K s 2
+ K ps + KI
= = 3
D

F (s) K + K / s + sK s + (10 + K D ) s 2 + ( 20 + K p ) s + K I
1+ p I D

s 2 + 10 s + 20
The steady-state value for the output is
X (s)
xss = lim x ( t ) = lim sX ( s ) = lim s F (s) = 1
t → s →0 s →0 F (s)
 Control Example: PID Controller
Engineering
PID Controller: Closed-loop step response

Kp = 350
KD = 5500
KI = 300
X (s) 5500 s 2 + 350 s + 300
= 3
F ( s ) s + 5510 s 2 + 370 s + 300

The above plot shows that the PID controller results no overshoot, no
steady-state error and fast rise time.
 Control Change in gain in PID
Engineering controller
 Control Response Characteristics:
Engineering PID Controllers
 Control Comparison: PID Controller
Engineering
 Control Tuning PID Controller
Engineering
Ziegler–Nichols tuning rules

F(s)
 1  X(s)
K p 1 + + sTd  Plant
+  Ts
 i


−

For determining values of the proportional gain Kp, integral time Ti

and derivative time Td on the transient response characteristics of
a given plant.

There are two methods called Ziegler–Nichols tuning rules: the

first method and the second method. Here the second method will
be discussed with example.
 Control Example: Tuning PID Controller
Engineering
Ziegler–Nichols tuning rules

F(s)
 1  1 X(s)
K p 1 + + sTd 
+  Ts
 i

 s ( s + 1)( s + 5 )
−

First set Ti = ∞ and Td = 0. We obtain the closed-loop transfer

function as
X (s) Kp
=
F ( s ) s ( s + 1)( s + 5 ) + K p

Characteristics Equation:
s ( s + 1)( s + 5 ) + K p = 0
s 3 + 6 s 2 + 5s + K P = 0
 Control Example: Tuning PID Controller
Engineering
Ziegler–Nichols tuning rules
The value of Kp that makes the system marginally stable so that
sustained oscillation occurs can be obtained by use of Routh’s
stability criterion.

Sustained oscillation with period Pcr (is measured in sec.)

 Control Example: Tuning PID Controller
Engineering
Ziegler–Nichols tuning rules
Characteristics Equation: s 3 + 6 s 2 + 5s + K P = 0
Routh Table

So K P = K cr = 30
Characteristics Equation: s 3 + 6 s 2 + 5s + 30 = 0
Marginal stability: 6 s 2 + 30 = 0  s 2 + 5 = 0  s =  j 5

Frequency of sustained oscillation is  = 5 = 2.24 rad/sec

 Control Example: Tuning PID Controller
Engineering
Ziegler–Nichols tuning rules
22
Period of sustained oscillation is Pcr = = = 2.8sec
 2.24

K p = 0.6 K cr = 0.6  30 = 18
Ti = 0.5 Pcr = 0.5  2.8 = 1.4
Td = 0.125 Pcr = 0.125  2.8 = 0.35
 Control Example: Tuning PID Controller
Engineering
Ziegler–Nichols tuning rules K p = 0.6 K cr = 0.6  30 = 18
Ti = 0.5 Pcr = 0.5  2.8 = 1.4
Unit Response Curve
Td = 0.125Pcr = 0.125  2.8 = 0.35
 Control Example: Tuning PID Controller
Engineering
Ziegler–Nichols tuning rule has provided a starting point for fine
tuning.

K p = 18
Ti = 3.077
Td = 0.7692

Unit Response Curve

 Control Example: Tuning PID Controller
Engineering
Ziegler–Nichols tuning rule has provided a starting point for fine
tuning.

K p = 39.42
Ti = 3.077
Td = 0.7692

Unit Response Curve

 Control Conclusion
Engineering

➢ Increasing the proportional feedback gain reduces steady-

state errors, but high gains almost always destabilize the
system.

➢ Integral control provides robust reduction in steady-state

errors, but often makes the system less stable.

➢ Derivative control usually increases damping and improves

stability, but has almost no effect on the steady state error.

➢ These 3 kinds of control combined from the classical PID

controller