5.

PID Controller
(Proportional, Integrator,
Derivative)

Hamid

1
Session Outlines & Objectives
Outlines
 The PID control algorithms
 The practical aspects

Objectives
 Understand PID control
 Understand function of every term of PID control (P&I&D)
 Can choose good combination and can tune PID for many
applied process control.
 Know additional feature installed in controller for
appllication in practice

2
What is PID Control?
 The PID stands for Proportional - Integrator
- Derivative
 Also known as three-term control
 It’s implemented as a computer program

today
◦ The controller comes in many different forms
◦ PID control is often combined with logic, sequential functions,
selectors, and simple function blocks to build the complicated
automation systems

3
Why PID Control?
 The PID algorithm is simple, easy to
understand, and relatively easier to tune
than the other controller
◦ It became the standard tool when process control emerged in the
1940s
◦ In process control today, more than 90% of the control loops are
of PID type

4
Key Concepts (1)
 Process Action
◦ Defines the relationship between changes in the valve and change in the
measurement
DIRECT Increase in valve position causes an increase in the measurement
REVERSE Increase in valve position causes a decrease in the measurement

 Controller Action
◦ Defines the relationship between changes in the measured variable and
change in the controller output
DIRECT Increase in measured variable causes an increase in the output
REVERSE Increase in measured variable causes a decrease in the output

The controller action must be the opposite of the process action

5
Key Concepts (2)
 The PID algorithm doesn’t know the correct
output to bring the process variable to the
setpoint
◦ The algorithm must have process measurement to perform
◦ It merely continues to move the output in the direction which should move
the process toward the setpoint

 The PID algorithm must be ‘tuned’ for the

particular process loop
◦ Each of the terms of the PID equation must be understood
◦ The tuning is based on the dynamics of the process response

6
Key Concepts (3)
 Manual & Auto modes
◦ Manual mode
 The human operator adjusts the output to operate the
process
 Manual Mode is very useful when unusual conditions
exist:
 Plant start-up
 Plant shut-down
 Emergencies
◦ Auto mode
 The control algorithm manipulates the output to hold
the process measurements at their setpoints
 Should be the most common mode for normal operation

7
PID Control Algorithm
 Comprises three elements:
◦ Proportional – also known as Proportional Gain or simply Gain
◦ Integral – also known as Automatic Reset or simply Reset
◦ Derivative – also known as Rate or Pre-Act

 Available in several combinations of these

elements:
◦ Proportional only (P)
◦ Proportional and Integral (PI) (most common)
◦ Proportional, Integral, and Derivative (PID)
◦ Proportional and Derivative (PD)

8
Proportional Mode (1)
CO = P .e + b
b

SP e + CO
+ P +
P = proportional gain (dimensionless)
CO = controller output (%) PVm
b = bias (%) (also known as manual reset)
e = ( SP – PVm ) (%)  “reverse action”, or
e = – ( SP – PVm ) (%)  “direct action”

 Some manufacturers use Proportional Band (PB) instead of

proportional gain
◦ PB is the % change in the input which a 100% change in the output

100
PB = P CO = P . b=
e+ 10
e+b
0
PB 9
Proportional Mode (2)
 Proportional-only control can produces an
offset
◦ The offset can be reduced by increasing the controller gain (or
decreasing the proportional band). But one cannot make the
controller gain arbitrarily large since too high a gain induces
oscillation PV
and/or
m
instability SP

Large P

Small P

time

10
Proportional Mode (2)
◦ To remove offset, the human operator has to “reset” the controller
manually by adjusting the value of the manual reset (the “b” term)

PVm
SP

Manual reset
P constant

Manual reset

time

11
Integral (Automatic Reset) Mode (1)
1 e
CO = T
m ò e dt SP
+ 1
Tm ò
dt
CO

Tm = integral time (minute per repeat or PVm

second per repeat)
CO = controller output (%)
e = ( SP – PVm ) (%)  “reverse action” or
e = – ( SP – PVm ) (%)  “direct action”
 Some manufacturers use repeat per minute (or repeat per second)
instead of minute per repeat (or second per repeat)
◦ Repeat per minute (or second) is the time it takes the reset (or integral) element
to repeat (reset) the action of the proportional element

1
Im=repeat minute per repeat (Tm) CO = Trò e dt
per minute (Tr) =
12
Integral (Automatic Reset) Mode (2)
 As long as error exists, the controller will
change its output; hence it is capable of
driving the error to zero
 Speed of response is reduced (compared to

P-only mode) PV
m
SP

P-only

I-only

Small P

time
13
Proportional-Integral Mode (1)
SP e CO
+ 1
Tm ò
dt

PVm

1
CO = P e( + T
m ò e dt )

14
Proportional-Integral Mode (2)
 Response of PI controller toP step
Slope =
change in
Output response
T for various
error values of T
m

m
Controller Output (CO)

Response equal
in magnitude to
Proportional
response

Response due to
Proportional P
control action

Tm time
Error

0
time

15
Proportional-Integral Mode (3)
 It combines the best features of the
proportional and integral modes
◦ The proportional offset is eliminated with little loss of response
speed
PVm PI
SP

P-only

I-only

Small P

time

16
Derivative Mode (1)
de SP e CO
CO = D dt + d
D dt

PVm
D = derivative time (minute or second)
CO = controller output (%)
e = ( SP – PVm ) (%)  “reverse action” or
e = – ( SP – PVm ) (%)  “direct action”

 Speed of response is increased (compared to P-

only mode)
 Hypersensitive to noise and other high-

frequency disturbances
17
Derivative Mode (2)
 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 not used in practice
 They are usually found in combination with
other control elements, mostly in
combination with proportional control

18
Proportional-Derivative Mode (1)

d b
D dt
SP e +
+
P CO
+
+
+
PVm

de
)
CO = P (e + D dt + b

19
Proportional-Derivative Mode (2)
 Response of PD controller to Response
ramp due change
to in
Proportional and
error Response due to
Proportional and
Controller Output (CO) Derivative modes
Derivative modes
Theoretical
Actual

Response due to
Proportional mode
only
D

time

Ramp
Error

Generator OFF
Ramp
Generator ON time

20
Proportional-Derivative Mode (3)
 PD control can produces an offset
 To avoid proportional offset, the bias “b”

should be set when the PVm is at setpoint

 Commonly found in slow-response process
control, e.g. temperature, pH, composition
controls PV
m
SP

PD

P-only

Small P

time
21
Proportional-Integral-Derivative Mode
(1)
 Combine the best feature of P, I, and D terms
I
ò dt
SP e + CO
+ +
P
+
d
D dt

PVm

de
CO =( P e + I ò
e dt + D dt )
22
Proportional-Integral-Derivative Mode
(2)

P-only – notice the offset

P and I – offset gone

P, I and D – best

time

23
Additional PID Concepts
Interactive vs. Non-interactive PID
Algorithm
 Refer to interaction between the reset and
derivative terms
 Also known as ‘series’ or ‘parallel’
 Almost all analog controllers are interactive
 Many digital controller are non-interactive,
some are interactive
 The only difference is in the tuning of
controller with derivative

24
Reset Windup (1)
 All actuators have limitations:
◦ Ex.: A motor has limited speed, a valve cannot be more than fully opened or
fully closed
 Windup phenomena is caused by the interaction of
integral action and saturations
 When this happens the feedback loop is broken and
the system runs as an open loop because the
actuator will remain at its limit independently of the
process output
◦ If a controller with integrating action is used, the error will continue to be
integrated. This means that the integral term may become very large or,
colloquially, it “winds up”
◦ It is then required that the error has opposite sign for a long period before
things return to normal
◦ The consequence is that any controller with integral action may give large
transients when the actuator saturates

25
Reset Windup (2)
yPV
m
ysp
SP
A

c
CO

Time

 Note that controller output saturates causing area “A” to accumulate

by the integral action
 After the disturbance returns to its normal level, the controller output
remains saturated for a period of time causing an upset in PVm

26
Anti-Reset Windup

ySPsp y
PVm

c
CO

Time
 When the manipulated variable saturates, the integral is not
allowed to accumulate
 When control returns, the controller takes immediate action
and the process returns smoothly to the setpoint
27
Methods for Anti-Reset Windup
 Turn off the integral when a valve saturates
or a control loop is not in use.
 Clamp the controller output to be greater

than 0% and less than 100%.

 Apply internal reset feedback
 Apply external reset feedback

28
Bumpless Transfer (1)
 Practically all controllers can be run in two modes:
manual or automatic
 When the system is in manual mode, the control

algorithm produces a control signal that may be different

from the manually generated control signal, or vice
versa. It is necessary to make sure that the two outputs
coincide at the time of switching. This is called bumpless
transfer
 With bumpless transfer, an internal setpoint is used for

the controller and the internal setpoint is ramped at a

slow rate from the initial conditions to the actual desired
setpoint to order to provide a smooth startup of a
control loop

29
Bumpless Transfer (2)
 Comparison of true and internal setpoint

True Setpoint

Internal Setpoint

Time

30
Bumpless Transfer (3)
 Control Performance with and without
bumpless transfer
w/o bumpless transfer

w/ bumpless transfer

Time

31
Derivative on Process Rather than Error (1)

The Facts:
◦ A step change in the set point results in a step change in the process
◦ The derivative term acts on the rate of change of the error
◦ The rate of change of a step change is very large
◦ An operator step change of the setpoint would causes a very large
change in the output, upsetting the process

Derivative
Spike
MV

SP

PVm
time
Fig. Process variable and valve response
to a setpoint change using standard PID

32
Derivative on Process Rather than Error (2)

Solution: Let derivative act only on process

rather than error 1
T ò
dt
m
SP e + CO
+ +
P

d
P .D dt

PVm

1 . dPVm
CO = P (e + T
m ò )e -P D dt
33
Derivative on Process Rather than Error (3)

 Process variable and valve response to a

setpoint change

MV

SP

PVm
time
Fig. Process variable and valve response
to a setpoint change using
“Derivative on Process Measurement” PID

34
Derivative on Filtered Process
Rather than Process (1)

The Fact:
◦ The derivative mode is hypersensitive to noise

35
Gain on Process Rather than Error (1)
The Facts:
◦ In applications with high gain, a step change can result in a
sudden, large movement in the valve
◦ Not as severe as derivative effect, but still can upset the process

MV

Gain
response
SP

PVm
time
Fig. Process variable and valve response
to a setpoint change using
“Gain on error and Derivative on Process Measurement” PID

36
Gain on Process Rather than Error (2)
Solution: Let gain act only on process rather
SP e
than error+ P
T ò
dt
m
+ CO
P

d
P .D dt

PVm

P dPVm
CO = T
m
e dt + P( PVm +
ò dt )
37
Gain on Process Rather than Error (3)
 Process variable and valve response to a
setpoint change

MV

SP

PVm
time
Fig. Process variable and valve response
to a setpoint change using
“Gain and Derivative on Process Measurement” PID

38
 Digital PID Algorithms (1)
 Data acquisition concepts
Sampled signal Original signal

Continuous signal
Dt
sample

Dt Sampled signal Original signal

Dt = sampling time
2Dt

Shannon’s sampling theorem:

The sampling frequency must be greater or equal to two times of the highest
frequency occurring in the signal to be sampled

39
 Digital PID Algorithms (2)
 Integral  Derivative

de ei – ei-1
ò e dt @ Dt S ei dt @ Dt
.

1 de
CO = P e( + T
Continuous form:
ò D )
e dt + dt
m

Dt D
CO
Digital form: = P ei[+ Tm S ei +Dt (ei – e]i-1)

i = sampling instant
40
 Digital PID Algorithms (3)
 Two forms of digital PID algorithms:
◦ Positional form

Dt D
COi = P ei[+ Tm S ei dt D
+t (ei – ]ei-1)

◦ Velocity form
 Inherently have anti reset windup
D feature
COi = COi-1 + P [(ei – ei-1) + Dt eDi t+ (ei – 2e]i-1+ei-2)
Tm

41
Guidelines for Common Control Loops (1)

Flow and liquid pressure control

 Fast response with no time delay (no

pipe/transportation)
 Usually with small high-frequency noise
 PI controller with intermediate controller gain

Liquid level control

 Noisy due to splashing and turbulence
 High gain, low integral action of PI controller for

integrating process
 Conservative setting for averaging control when it is

used for damping the fluctuation of the inlet stream

42
Guidelines for Common Control Loops (2)

Gas pressure control

 Usually fast and self regulating
 PI controller with small integral action (large

reset time)

Temperature control
 Wide variety of the process nature
 Usually slow response with time delay
 Use PID controller to speed up the response

43
Guidelines for Common Control Loops (3)

Composition control
 Similar to temperature control usually with

larger noise and more time delay

 Effectiveness of derivative action is limited
 Temperature and composition controls are

the prime candidates for advance control

strategies due to its importance and difficulty
of control

44
Session Summary
 PID control, which is the most widely used
control algorithm in process control
application, comes in different forms and
terms
 Each of the terms of the PID equation must be
understood to obtain a right combination of
the PID control elements for various process
control application objectives

45