CHE 510: Process Control & Instrumentation

MODULE 3:
PID Controller Design, Tuning and Troubleshooting

Ayorinde Bamimore, Ph.D.

Department of Chemical Engineering
Obafemi Awolowo University
 Module Objectives
After completing this module, students will be able to:

I. Design PID controllers for any type of system,

II. Use PID tuning rules for quick parameterization of controller,
III. Tune PID controllers to achieve trade-off between performance and
robustness,
IV. Calculate performance indices and use,
V. Use MATLAB/SIMULINK PID controller auto-tuner,
VI. Simulate closed-loop system using MATLAB/SIMULINK,
VII. Troubleshoot nonperforming control loop

Obafemi Awolowo University 2

Chemical Engineering
 Controller Design Principles

Controller design principle entails the following:

(1) Selection of the best controller type to use for the process:
These controller types can be P, PI, PID or PD
(2) Chose the magnitude of the controller parameters, , , so
as to achieve the best possible overall closed-loop system response.

The controller design problem statement can then be stated as:

Chose the controller type, and the parameters for the
chosen controller, so that some objective criteria of closed-
loop performance are satisfied.

Obafemi Awolowo University 3

Chemical Engineering
 Performance Criteria for Closed-Loop System

The performance of a closed-loop system can be assessed

using the following criteria:
1. The closed-loop system must be stable
2. The effects of disturbances are minimized, providing good
disturbance rejection
3. Rapid, smooth responses to set-point changes are
obtained, that is, good set-point tracking
4. Steady-state error (offset) is eliminated
5. Excessive control action is avoided
6. The control system is robust, that is, insensitive to changes
in process conditions and to inaccuracies in the process model.

Obafemi Awolowo University 4

Chemical Engineering
 Tradeoffs between Performance and Robustness

 In typical control applications, it is not possible to achieve all of these

goals simultaneously, because they involve inherent conflicts and
tradeoffs.
 The tradeoffs must balance two important objectives, performance
and robustness.
 A control system exhibits a high degree of performance if it provides
rapid and smooth responses to disturbances and set-point changes
with little, if any, oscillation.
 A control system is robust if it provides satisfactory performance for a
wide range of process conditions and for a reasonable degree of
model inaccuracy.

Obafemi Awolowo University 5

Chemical Engineering
 Performance and Robustness Trade-off
 Robustness can be achieved by choosing conservative controller
settings (typically, small values of Kc and large values of τI), but this
choice tends to result in poor performance.
 Thus, conservative controller settings sacrifice performance in order to
achieve robustness.
 Robustness analysis for PID controllers is considered in Appendix J of
Seborg et al. (2017) and in the control engineering
literature (Åström and Hägglund, 2006; Kristiansson and Lennartson,
2006).
 Detailed analyses of performance-robustness tradeoffs for PID
controllers have recently been published (Garpinger and Hägglund,
2015).

Obafemi Awolowo University 6

Chemical Engineering
Performance Trade-off between Setpoint Tracking and
Disturbance Rejection
 A second type of tradeoff occurs because PID controller settings that
provide excellent disturbance rejection can produce large overshoots
for set-point changes.
 On the other hand, if the controller settings are specified to provide
excellent set-point tracking, the disturbance responses can be very
sluggish.
 Thus, a tradeoff between set-point tracking and disturbance
rejection occurs for standard PID controller.

Obafemi Awolowo University 7

Chemical Engineering
 Dynamic Response Performance Criteria

 Minimum rise time, tr

 Minimum settling time, ts
 Specified maximum overshoot, a1<aspec
 For oscillatory responses, a specified maximum decay ratio,
a2/a1<dr, to ensure sufficient rapid settling of oscillations
 Time integral performance criteria
1. Integral Absolute Error (IAE)
2. Integral Squared Error (ISE)
3. Integral Time-weighted Absolute Error (ITAE)
4. Integral Time-weighted Squared Error (ITSE)

Obafemi Awolowo University 8

Chemical Engineering
 Time Integral Performance Criteria

1. Integral Absolute Error (IAE)

(3.1a)

2. Integral Squared Error (ISE)

(3.1b)

3. Integral Time-weighted Absolute Error (ITAE)

(3.1c)

4. Integral Time-weighted Squared Error (ITSE)

(3.1d)
Obafemi Awolowo University 9
Chemical Engineering
 Example 3.1: Effect of Controller Tuning for the
Three Tank System
For PI only controller, investigate the effect of the tuning
parameters, and on the closed-loop system response.

yd +
+ y
+
_

Fig.3.1: Three-tank system under PID Control

Obafemi Awolowo University 10

Chemical Engineering
 Example 3.1

Recall, 1.17c)

The closed-loop transfer function is given as:

(3.2a)

Simplifying Eq.3.2a, we have:

(3.2b)

Substituting in Eq.3.2b, , we have:

(3.3)

Obafemi Awolowo University 11

Chemical Engineering
 Example 3.1

 Note that for small,

one must use higher
controller gains to eliminate
offset in a timely fashion.
 This provides short rise
times, but poor decay ratio
performance (poorly
damped oscillations) and
significant overshoot
Fig.3.2a: Three-tank system under
PID Control

Obafemi Awolowo University 12

Chemical Engineering
 Example 3.1

 For larger values of , smaller

controller gains may be used, and this
results in larger rise times, but less
overshoot.
 It is generally difficult to have very
short rise times and small overshoot.
 However, if settling time was the only
criterion (e.g., time to stay with ± 10%
of the set-point change), then a wide
range of controller parameters with
different types of responses would
Fig.3.2b: Three-tank system provide settling times of about 30
under PID Control minutes.

Obafemi Awolowo University 13

Chemical Engineering
 Example 3.2

For PI only controller, investigate the effect of the tuning

parameters, and on the closed-loop system response.
d

yd +
+ y
+
_

Fig.3.3: First order system under PI control

Obafemi Awolowo University 14

Chemical Engineering
 Example 3.2

Fig.3.4a indicates that increasing

or decreasing tends to speed
up the response.

In addition, the response becomes

more oscillatory as either or
decreases.

But in general, closed-loop

responses become more oscillatory
as is increased
Fig.3.4a: Effect of controller
parameters on disturbance rejection

Obafemi Awolowo University 15

Chemical Engineering
 Example 3.2

Fig.3.4b indicates that increasing

or decreasing tends to speed up
the response.

In addition, the response becomes

more oscillatory as either or
decreases.

But in general, closed-loop

responses become more oscillatory
Fig.3.4a: Effect of controller as is increased
parameters on disturbance
rejection

Obafemi Awolowo University 16

Chemical Engineering
 Controller Tuning Methods

1. Controller tuning with fundamental process models

Optimization of Time-Integral Criteria
Model Following PID Controller designs
Direct synthesis
Internal Model Control (IMC)
Use of Stability Margins
Ziegler-Nichols
Tyreous-Luyben
2. Controller Tuning using Approximate Models
Ziegler-Nichols, AMIGO, Direct synthesis, IMC, ITAE
3. Controller Tuning using frequency response models
Nyquist, Bode, Gain Margin (GM), Phase Margin (PM)
4. Controller Tuning without a model
Continuous Cycling Method, Relay Auto-tuning, Step Test Method
Obafemi Awolowo University 17
Chemical Engineering
 Controller Tuning by Optimization of
Time-Integral Criteria
Controller tuning by optimization of time-integral criteria entails finding
the controller parameters which minimizes the time-
integral criterion. This can be posed as in Eq.3.4:

(3.4a)
If the in Eq.3.4a is an ITSE, the problem can be posed as in
Eq.3.4b:

(3.4b)
The above optimization problem can easily be implemented in
MATLAB/SIMULINK.

Obafemi Awolowo University 18

Chemical Engineering
 Example 3.3: Controller Tuning by Optimization

Determine the optimal PI controller settings and the optimal closed-

loop response for each of the integral-time criteria in Eq.3.1a-d.

yd +
+ y
+
_

Fig.3.1: Three-tank system under PID Control

Obafemi Awolowo University 19

Chemical Engineering
 Example 3.3: Controller Tuning by Optimization
Step 1: Develop a SIMULINK model of the closed-loop system

Fig.3.5a: SIMULINK Model of the Three-tank System under PI Control

Obafemi Awolowo University 20

Chemical Engineering
 Example 3.3: Controller Tuning by Optimization

Fig.3.5b: SIMULINK model of the subsystem “performance indices”

Obafemi Awolowo University 21

Chemical Engineering
 Example 3.3: Controller Tuning by Optimization

Step 2: Create a function m-file called “fun152”

function y=fun152(x)
assignin('base','P1',x(1))
assignin('base','I1',x(2))
[t,xx,yy]=sim('model152.mdl',100); y=yy(end,1)

Step 4: Create an ordinary m-file and run

xo=[1,1]
[x,fval] = fminsearch(@fun152,xo,optimset('MaxIter',100))
P1=x(1)
I1=x(2)
Results: P1 = 0.5593, I1 = 0.0236 , 0.0422

Obafemi Awolowo University 22

Chemical Engineering
 Example 3.3: Controller Tuning by Optimization

Optimal Controller parameters

obtained:

Fig.3.6: Simulation Results

Obafemi Awolowo University 23

Chemical Engineering
 Controller Design by Direct Synthesis Method

+ y
yd +
+
_

Fig.3.7: Block diagram for a standard feedback control system

Obafemi Awolowo University 24

Chemical Engineering
 Controller Design by Direct Synthesis Method
Consider the block diagram in Fig.3.7, the closed-loop transfer function
for set-point changes is given by:
(3.5)
For simplicity, let and assume that , then Eq.3.5
becomes:
(3.6)
where is the desired closed-loop transfer function
By making the subject of the formula (3.6), we have:
(3.7a)

N.B.: The actual process, is often differentiated from the model using

Obafemi Awolowo University 25

Chemical Engineering
 Controller Design by Direct Synthesis Method

(3.7b)

 The performance of the DS controller often depends on the specified

closed-loop transfer function.
 Ideally, we would like to have = 1 so that the controlled
variable tracks set-point changes instantaneously without any error.
However, this ideal situation, called perfect control, cannot be
achieved by feedback control because the controller does not respond
until e ≠ 0.

Obafemi Awolowo University 26

Chemical Engineering
 Controller Design by Direct Synthesis Method

Case I: Processes without time delays

For processes without time delays, the first-order model
in Eq.3.8 is a more reasonable choice:
(3.8)

Where is the desired closed-loop time constant. The first order model in
Eq.3.8 has a settling time of .
Substituting Eq.3.8 into 3.7b, we have:
(3.9)
N.B.: The term provides integral control action and thus eliminates offset. Design
parameter provides a convenient controller tuning parameter that can be used to
make the controller more aggressive (small ) or less
aggressive (large ).

Obafemi Awolowo University 27

Chemical Engineering
 Controller Design by Direct Synthesis Method

Case I: Processes with time delays

If the process transfer function contains a known time delay, a
reasonable choice for the desired closed-loop transfer function, q is
(3.10)
Substituting (3.8) into (3.7b), we have:
(3.11)
N.B.:
Although the controller (3.11) is not in the standard PID form, it is
physically realizable.
Eq.3.11 can be used to derive PID controller for simple process model.

Obafemi Awolowo University 28

Chemical Engineering
 Controller Design by Direct Synthesis Method

Recall the Taylor’s series: (3.11*)

By substituting (3.11*) into the denominator of Eq.3.11, we have:
(3.12)
N.B.: This controller contains integral action.
Process example I: FOPTD model
Given (3.13)
Derive PI controller for this process using DS.
Solution
By substituting (3.13) into (3.12), we have:
(3.14)

where and

Obafemi Awolowo University 29

Chemical Engineering
 Controller Design by Direct Synthesis Method

Process example II: SOPTD model

Given (3.15)
Derive PID controller for this process using DS.
Solution
By substituting (3.15) into (3.12), we have:
(3.16)
By rearranging Eq.3.14 into the standard PID controller,
(3.17a)
Where:
, , (3.17b)

Obafemi Awolowo University 30

Chemical Engineering
 Example 3.4:PID Controller Design by Direct Synthesis

Use the DS method to calculate PID controller settings for the

process:
(3.18a)
Consider three values of the desired closed-loop time constant:
= 1, 3, and 10. Evaluate the controllers for unit step changes in
both the set point and the disturbance, assuming that .
Repeat the evaluation for two cases:
(a) The process model is perfect ( ).
(b) The model gain is incorrect, = 0.9, instead of the actual
value,
(3.18b)

Obafemi Awolowo University 31

Chemical Engineering
 Example 3.4:PID Controller Design by Direct Synthesis

By using the controller settings obtained in Eq.3.17, the PID controller

parameters obtained are summarized in Table 3.1.

Table 3.1: PID Controller Settings for Example 3.4

Obafemi Awolowo University 32

Chemical Engineering
Example 3.4: PID Controller Design by Direct Synthesis

Fig.3.7: SIMULINK Diagram for Implementing Example 3.4

Obafemi Awolowo University 33

Chemical Engineering
 Example 3.4:PID Controller Design by Direct Synthesis
Discussion of results
As increases, the
responses become more
sluggish, and the maximum
deviation is larger after the
disturbance occurs
at t = 80.

Fig.3.8a: Simulation results with a unit

setpoint change introduced at t=0 and a
unit disturbance change introduced at
t=80

Obafemi Awolowo University 34

Chemical Engineering
 Example 3.4:PID Controller Design by Direct Synthesis
Discussion of results
* When the model gain is 0.9, about
50% too low, the closed-loop response
for = 1 in Fig.3.8b is excessively
oscillatory and would even become
unstable if K = 0.8 had been
considered.
* The disturbance responses for =3
and = 10 in Fig.3.8b are actually
better than the corresponding
responses in Fig.3.8a because the
Fig.3.8b: Simulation results with a former have shorter settling times and
unit setpoint change introduced at smaller maximum deviations.
t=0 and a unit disturbance change * This improvement is due to the larger
introduced at t=80
values of Kc for case (b).

Obafemi Awolowo University 35

Chemical Engineering
 Controller Design by Internal Model Control (IMC)

IMC is more comprehensive model-based design method. It was

developed by Morari and co-workers .The IMC method, like the DS
method, is based on an assumed process model and leads to analytical
expressions for the controller settings.
IMC and DS are closely related and produce identical controller if the
design parameters are specified in a consistent manner. However, the
IMC approach has the advantage that it allows model uncertainty and
tradeoffs between performance and robustness to be considered in a
more systematic fashion.

Obafemi Awolowo University 36

Chemical Engineering
 PID Controller Design by IMC

The Internal Model Control (IMC) philosophy relies on the Internal Model
Principle, which states that “Control can be achieved only if the
control system encapsulates either implicitly or explicitly, some
representation of the process to be controlled”.
In particular, if the control scheme has been developed based on an
exact model of the process, then perfect control is theoretically possible.

Obafemi Awolowo University 37

Chemical Engineering
 PID Controller Design by IMC

Set-point Output

yd y
gIMC(s) g(s)

Fig.3.9: Open loop Control Strategy

Consider the process g(s) given above, suppose is a model of the

process, By setting gIMC to be the inverse of the model of the process, we
have:

(3.19)

Obafemi Awolowo University 38

Chemical Engineering
 PID Controller Design by IMC
If , (the model is an exact representation of the process):
(3.20a)

(3.20b)

Then, the output will always be equal to the setpoint.

N.B:
 This ideal control performance is achieved without the feedback. What
this tells us is that if we have complete knowledge about the process (as
encapsulated in the process model) being controlled, we can achieve
perfect control.

Obafemi Awolowo University 39

Chemical Engineering
 PID Controller Design by IMC
 It also tells us that feedback control is necessary only when knowledge
about the process is inaccurate or incomplete.
The IMC strategy
 In practice, process-model mismatch is common; the process model
may not be invertible and the system is often affected by unknown
disturbances. Thus, the above open loop control arrangement will not
be able to maintain the output at the setpoint.
 The IMC method is based on the simplified block diagram shown in the
Fig.3.10.

Obafemi Awolowo University 40

Chemical Engineering
 PID Controller Design by IMC
In the diagram, d(s) is an
unknown disturbance
affecting the system. The
yd(s) +
g(s) + + manipulated input u(s) is
introduced to both the
process and its model. The
process output, y(s), is
+ compared with the output of
the model, resulting in a
signal, . That is:

Fig.3.10a: The IMC Control Structure A (3.21)

Obafemi Awolowo University 41

Chemical Engineering
 PID Controller Design by IMC
(3.21)
 If is zero for example, then is a measure of the difference in
behavior between the process and its model.
 If , then is equal to the unknown disturbance, thus,
may be regarded as the information that is missing in the model,
and can therefore be used to improve control. This is done by
subtracting from the setpoint, yd(s), which is very similar to
effecting a setpoint trim.
 The resulting control signal is given by:

Obafemi Awolowo University 42

Chemical Engineering
 PID Controller Design by IMC
(3.22a)

Substituting Eq.3.21 into 3.22a, we have:

(3.22b)
Making u(s) the subject of formula Eq.3.22b, we have:
(3.22c)
But from Fig.3.10a:
(3.22d)
Substituting Eq.3.22c into 3.22d, we have:
(3.22d)
Eq.3.22d can be further simplified into:
(3.22e)

Obafemi Awolowo University 43

Chemical Engineering
 PID Controller Design by IMC
N.B.:
 From this closed-loop expression (Eq.3.22e), we can see that if
and then perfect setpoint tracking and
disturbance rejection is achieved.
 Notice that, theoretically, even if , perfect disturbance
rejection can be still be realized provided
 Additionally, to improve robustness, the effects of process model
mismatch should be minimized.

Obafemi Awolowo University 44

Chemical Engineering
 PID Controller Design by IMC
Since discrepancies between process and model behavior usually
occur at high frequency end of the system’s frequency response, a
low pass filter is usually added to attenuate the effects of
process-model mismatch. Thus, the internal model controller is
usually designed as the inverse of the process model in series with
a low-pass filter, that is,
(3.23)
The order of the filter is usually chosen such that is
proper, to prevent excessive differential control action.
Obafemi Awolowo University 45
Chemical Engineering
 PID Controller Design by IMC

yd(s) + + g(s) + +
+

Fig.3.10b: The IMC control structure B

Obafemi Awolowo University 46

Chemical Engineering
 PID Controller Design by IMC

yd(s) +
g(s) + +

Fig.3.11: Conventional Feedback control structure

By comparing Fig.11 with Fig.3.10b, we have

(3.24)

Obafemi Awolowo University 47

Chemical Engineering
 PID Controller Design by IMC

The IMC controller is designed in four steps:

(1) The process model is factorized into invertible(good guy) and
non-invertible(bad guy) – (like time delays and RHP) elements
i.e.,
For example:

(A.)

(B.) and

(2) Form the idealized IMC controller,

Obafemi Awolowo University 48
Chemical Engineering
 PID Controller Design by IMC

(3)Add a filter to make the controller proper and

typically has the form where = tuning

parameter and r is chosen to make the controller proper or semi

proper.
(4)Adjust to vary the speed of response of the closed-loop
system. If is small the closed-loop system is fast, if is large
the closed-loop system is more robust (insensitive to model error)

Obafemi Awolowo University 49

Chemical Engineering
 Example 3.5

Consider a FOPTD process with the following model:

(3.25)

(1) Design three controllers:

(a) IMC ( =2sec) (b.) IMC ( =5sec) (c.) IMC ( =10sec)

(2) Evaluate the three controllers by comparing their performance for unit
step changes in both set-point and disturbance. Assume perfect model
(i.e., ) and

Obafemi Awolowo University 50

Chemical Engineering
 Example 3.5
Solution

Step1:

Step2:

Step3: ∙ = with

Case (a.) when =2,

(3.26)

(3.27)

Obafemi Awolowo University 51

Chemical Engineering
 Example 3.5

Use First order Taylor series approximation

; ;

(3.28)

Obafemi Awolowo University 52

Chemical Engineering
 Example 3.6

Given a FOPTD model

Use the IMC design method to parameterize for the model:

(a.) PI controller, using first-order Taylor series approximation for
the delay term, i.e.,
(b.) PID controller, using first-order (1/1) Pade approximation for

the delay term, i.e.,

Obafemi Awolowo University 53

Chemical Engineering
 Example 3.6

Solution

(a.) By substituting for we have,

Step 1: Factor this model as

where

Step 2: Idealized IMC, ,

Step 3: Add a filter

where r=1,

Obafemi Awolowo University 54

Chemical Engineering
 Example 3.6

Step 4: Equivalent conventional controller is

(3.29a)

(3.29b)

(3.29c)

(3.29d)

Obafemi Awolowo University 55

Chemical Engineering
 Example 3.6

By comparing with , we have:

(b.) By substituting for (3.29e)

Step 1: (3.29f)

Step 2: (3.29g)

Obafemi Awolowo University 56

Chemical Engineering
 Example 3.6

Step 3: (3.29h)

Step 4: (3.29i)

(3.29j)

(3.29k)

Obafemi Awolowo University 57

Chemical Engineering
 Example 3.6

= (3.29l)

= (3.29m)

= (3.29n)

= (3.29o)

Obafemi Awolowo University 58

Chemical Engineering
 Example 3.6

Comparing with (3.29p)

, , (3.29q)

Obafemi Awolowo University 59

Chemical Engineering
 IMC Tuning Relations

Table 3.4: IMC Controller settings for PID controllers

Chien and Fruehauf (1990)

Obafemi Awolowo University 60

Chemical Engineering
 IMC Tuning Relations

Table 3.4: IMC Controller settings for PID controllers

Obafemi Awolowo University 61

Chemical Engineering
 IMC Tuning Relations

Table 3.4: IMC Controller settings for PID controllers

Obafemi Awolowo University 62

Chemical Engineering
 Guidelines for Selecting

Table 3.5a: Guidelines for choosing τc for the IMC and SIMC
tuning methods and a FOPTD model

Obafemi Awolowo University 63

Chemical Engineering
 Example 3.7

A process model for a liquid storage system is given by Chien and

Fruehauf (1990):
(3.30)
Use Table 3.4 to calculate PI and PID controller settings
for K=0.2 and τc=8. Repeat for τc=15 and do the following:
(a) Compare the four controllers for unit step changes in the set
point and disturbance, assuming that
(b) In order to characterize the robustness of each controller of
part (a), determine Kmax, the largest value of K that results in a
stable closed-loop system for each controller

Obafemi Awolowo University 64

Chemical Engineering
 Example 3.7

By using Cases M and N from Table 3.4, which IMC-based PID controller
settings for integrating systems with delay, the obtained parameters are
summarized in Table 3.5b

Table 3.5b: IMC-Based PID Controllers for Example 3.7

Obafemi Awolowo University 65

Chemical Engineering
 Example 3.7

 The closed-loop
responses in Fig. 3.12a
are more sluggish and
less oscillatory for
τc = 15 than they are for
τc = 8.
 Also, for τc = 15 the
overshoot is smaller for
the set-point change,
Fig.3.12a: Simulation results with PI and the maximum
controller. Setpoint change is introduced
at t=0 and disturbance at t=150 deviation is larger after
the disturbance.
Obafemi Awolowo University 66
Chemical Engineering
 Example 3.7

The PID controller provides

a better disturbance
response than the PI
controller
with a smaller maximum
deviation. In addition, the
PID controller has a very
short settling time for τc =
8, which gives it the best
performance of the four
controllers considered.

Fig.3.12b: Simulation results with PID

controller. Setpoint change is introduced at
t=0 and disturbance at t=150

Obafemi Awolowo University 67

Chemical Engineering
 Example 3.7

Table 3.6: Kmax for Example 3.7 * The numerical value of Kmax can be
obtained from a stability analysis
using the Direct Substitution Method.
* The numerical results shown in
Table 3.6 indicate that K can increase
significantly from its nominal value of
0.2 before the closed-loop system
becomes unstable.
* Thus, these IMC controllers are
quite robust and become even more
so as τc increases.
* The approximate values of Kmax
were obtained by using Taylor’s
series for the time-delay.

Obafemi Awolowo University 68

Chemical Engineering
PID Controller Tuning using Approximate Model

Table 3.7:
Controller Design
Relations Based on
the ITAE
Performance Index
and a First-Order-
plus-Time-Delay
Model (Lipták, 2006).

Obafemi Awolowo University 69

Chemical Engineering
 PID Controller Tuning using Approximate Model

Table 3.8a: AMIGO Tuning Rules for PI Controllers (Åström and

Hägglund, 2006)

Obafemi Awolowo University 70

Chemical Engineering
 PID Controller Tuning using Approximate Model

Table 3.8b: AMIGO Tuning Rules for PI Controllers (Åström and

Hägglund, 2006)

Obafemi Awolowo University 71

Chemical Engineering
 Example 3.8

A blending system with a measurement time delay can be

modeled as

(3.31)

Calculate PI controller settings using the following tuning

relations:
(a) IMC (τc = τ/3)
(b) IMC (τc = θ)
(c) ITAE (disturbance)
(d) AMIGO

Obafemi Awolowo University 72

Chemical Engineering
 Example 3.8

Table 3.9: PI controller parameters for example 3.8

Comments
It appears that the ITAE (disturbance) settings are the most
aggressive, and the AMIGO settings are the least aggressive. But this
assertion should be checked by simulation.

Obafemi Awolowo University 73

Chemical Engineering
 Equivalent PID Controller Settings for the
Parallel and Series Forms
Table 3.9: Equivalent PID Controller Settings for the
Parallel and Series Forms

Obafemi Awolowo University 74

Chemical Engineering
 Tuning for Lag-Dominant Models

 First-or second-order models with relatively small time delays

are referred to as lag-dominant models.

 The IMC and DS methods provide satisfactory set-point
responses, but very slow disturbance responses, because the
value of is very large.
 Fortunately, this problem can be solved in three different ways.
Method 1: Integrator Approximation

Approximate by , where

 Then we can use the IMC tuning rules (Rule M or N) to specify the controller
settings.
Obafemi Awolowo University 75
Chemical Engineering
 Tuning for Lag-Dominant Models
Method 2: Limit the Value of
• For lag-dominant models, the standard IMC controllers for first order
and second-order models provide sluggish disturbance responses because
is very large.
• For example, Controller G in Table 3.4 has where is very large.
• As a remedy, Skogestad (2003) has proposed limiting the value
of (often called Skogestad IMC, SIMC):
(3.32)
where is the largest time constant (if there are two)

Obafemi Awolowo University 76

Chemical Engineering
 Tuning for Lag-Dominant Models

Method 3: Design the Controller for Disturbances, Rather

Set-point Changes
• The desired CLTF is expressed in terms of (y/d)des, rather than (y/yd)des
• Reference: Chen & Seborg (2002)

Obafemi Awolowo University 77

Chemical Engineering
 Example 3.9

Consider a lag-dominant model with (θ/ ) = 0.01:

(3.33)

Design four PI controllers:

a)IMC
b)IMC based on the integrator approximation
c)IMC with Skogestad’s modification (Eq.3.32)
d)Direct Synthesis method for disturbance rejection (Chen and
Seborg, 2002): The controller settings are Kc = 0.551 and

Obafemi Awolowo University 78

Chemical Engineering
 Example 3.9
Evaluate the four controllers by comparing their performance for
unit step changes in both set point and disturbance. Assume that
the model is perfect and that gd(s) =g(s).
Solution: The PI controller settings are:
Table 3.10: PI controller settings

Obafemi Awolowo University 79

Chemical Engineering
 Example 3.9

Fig. 3.13: Comparison of set-

point responses (top) and
disturbance responses (bottom)
for Example 3.9.

The responses for the Chen and

Seborg and integrator
approximation methods are
essentially identical

Obafemi Awolowo University 80

Chemical Engineering
 PID Controller Tuning without Process Model
(Online Controller Tuning)

The three popular online controller tuning

methods are:

1. Continuous cycling method

2. Relay Auto-Tuning
3. Step Test Method

Obafemi Awolowo University 81

Chemical Engineering
 Continuous Cycling Method
 In 1942, Ziegler and Nichols published a classic paper that
introduced their continuous cycling method for on-line controller
tuning (Ziegler and Nichols, 1942).
 This paper and its Ziegler–Nichols (Z-N) tuning relations have had
an enormous impact on both process control research and practice.
 Despite serious shortcomings, the Z-N relations and (various
modifications) have been widely used as a benchmark in
comparisons of tuning relations.
 The Z-N continuous cycling method is based on experimentally
determining the closed-loop stability limit for proportional-only
control.

Obafemi Awolowo University 82

Chemical Engineering
 Continuous Cycling Method
 The following trial and error procedure is used:
Step 1. After the process has reached steady state (at least approximately),
introduce proportional-only control by eliminating the integral and derivative
modes (i.e., set τD = 0 and τI = ∞).
Step 2. Introduce a small, momentary set-point change so that the
controlled variable moves away from the set point. Gradually increase Kc (in
absolute value) in small increments until continuous cycling occurs. The term
continuous cycling refers to a sustained oscillation with a constant amplitude.
The value of Kc that produces continuous cycling (for proportional-only
control) is defined to be the ultimate gain, Kcu. The period of this sustained
oscillation is the ultimate period, Pu.
After Kcu and Pu have been determined, the controller settings can be
calculated using the Z-N tuning relations in Table 3.11.

Obafemi Awolowo University 83

Chemical Engineering
 Continuous Cycling Method
Table 3.11: PID controller settings using ZN method

Obafemi Awolowo University 84

Chemical Engineering
 Continuous Cycling Method

The continuous cycling method has two major disadvantages:

1. The trial-and-error determination of Kcu and Pu can be quite
time consuming if the process dynamics are slow.

2. For many applications, continuous cycling is objectionable

because the process is pushed to a stability limit.
Consequently, if external disturbances or process changes
occur during the test, unstable operation or a hazardous
situation could result (e.g., a “runaway” chemical reaction).

Obafemi Awolowo University 85

Chemical Engineering
 Relay Auto Tuning
 Åström and Hägglund (1984) developed an attractive
alternative to the continuous cycling method.
 In their relay auto-tuning method, a simple experimental test
is used to determine Kcu and Pu.
 For this test, the feedback controller is temporarily replaced
by an on–off controller (or relay).
 After the control loop is closed, the controlled variable
exhibits a sustained oscillation that is characteristic of on–off
control.

Obafemi Awolowo University 86

Chemical Engineering
 Relay Auto Tuning

 The operation of the relay auto-tuner includes a dead zone as

shown in Fig.3.14.
 The dead band is used to avoid frequent switching caused by
measurement noise.
 The ultimate gain and the ultimate period can easily be
obtained from Fig.3.14.
 The ultimate period Pu is equal to the period of oscillation for
the process output.

Obafemi Awolowo University 87

Chemical Engineering
 Relay Auto Tuning

Fig.3.14: Auto-tuning using a relay controller

Obafemi Awolowo University 88

Chemical Engineering
 Relay Auto Tuning
Advantages of relay auto-tuning over continuous cycling
method:
1. Only a single experimental test is required instead
of a trial-and-error procedure.
2. The amplitude of the process output a can be
restricted by adjusting relay amplitude d.
3. The process is not forced to a stability limit.
4. The experimental test is easily automated.

Disadvantages of relay auto tuning

For slow processes, it may not be acceptable to subject the process
to the two to four cycles of oscillation required to complete the test.
Thus, the Step Test Method may be preferred for slow processes
Obafemi Awolowo University 89
Chemical Engineering
 Step Test Method
 In their classic paper, Ziegler and Nichols (1942) proposed a second on-
line tuning technique based on a single step test.
 The experimental procedure is quite simple.
 After the process has reached steady state (at least approximately), the
controller is placed in the manual mode.
 Then a small step change in the controller output (e.g., 3 to 5%) is
introduced.
 The controller settings are based on the process reaction curve
(as discussed in CHE509), the open-loop step response.

Obafemi Awolowo University 90

Chemical Engineering
 Step Test Method
 Consequently, this on-line tuning technique is referred to as the
step test method or the process reaction curve method.
 An appropriate transfer function model can be obtained from the
step response by using the parameter estimation methods of
CHE509.
 Typical process reaction curves after step test are shown in Fig.3.15

Obafemi Awolowo University 91

Chemical Engineering
 Step Test Method

Fig.3.15: Typical process reaction curves:

(a) non-self regulating process, (b) self regulating process
Obafemi Awolowo University 92
Chemical Engineering
 Step Test Method
Advantage of Step Test Method
The chief advantage of the step test method is that only a single
experimental test is necessary.
Disadvantages of Step Test Method
But the method does have four disadvantages:
1. The experimental test is performed under open-loop conditions.
Thus, if a significant disturbance occurs during the test, no corrective
action is taken. Consequently, the process can be upset, and the test
results may be misleading.

Obafemi Awolowo University 93

Chemical Engineering
 Step Test Method
2. For a nonlinear process, the test results can be sensitive to the
magnitude and direction of the step change. If the magnitude of
the step change is too large, process nonlinearities can influence the
result. But if the step magnitude is too small, the step response may be
difficult to distinguish from the usual fluctuations due to noise and
disturbances. The direction of the step change (positive or negative)
should be chosen so that the controlled variable will not violate a
constraint.

Obafemi Awolowo University 94

Chemical Engineering
 Step Test Method
3. The method is not applicable to open-loop unstable processes.
4. For analog controllers, the method tends to be sensitive to controller
calibration errors. By contrast, the continuous cycling method is less
sensitive to calibration errors in Kc because it is adjusted during the
experimental test.

Obafemi Awolowo University 95

Chemical Engineering
 Example 3.10:Ziegler-Nichols Tuning
Consider a third-order system. Design P, PI and PID controller for the
system using Ziegler-Nichols tuning rules d

yd +
+ y
+
_
Fig. 1.17

Obafemi Awolowo University 96

Chemical Engineering
 Example 3.10
Recall the characteristics equation, given by:

(1.65)

(1.74)

which simplifies to:

(1.75)

Substituting
(1.81)

Rearranging and collecting like terms, give:

(1.82)

Obafemi Awolowo University 97

Chemical Engineering
 Example 3.10
For Eq.1.82 to hold, the real and the imaginary parts have to be identically zero:
(1.83a)
(1.83b)

Eqs. 1.83a and b have to be solve simultaneously for and

From Eq.1.83a: or

From Eq.1.83b: (1.84)

When ,
When ,

Obafemi Awolowo University 98

Chemical Engineering
 Example 3.10
Maximum kc for which the system is on the verge of instability
is

Frequency of sustained oscillation at that point is

The ultimate period is obtained as:
(3.34)

Thus,
(3.35)

Obafemi Awolowo University 99

Chemical Engineering
 Example 3.10

Table 3.11: PID controller settings using ZN method

Table 3.12: PID controller settings using ZN method

Obafemi Awolowo University 100

Chemical Engineering
 Example 3.10

Comments: The response to a

set-point change for each type of
controller is seen in Fig.3.16.
Note that the P controller has some
offset while the PI controller has
large overshoot and slowly damped
oscillations. By contrast, the PID
controller provides a rather good
response except that there is about
40 percent overshoot. Clearly, for
the PI controller, the earlier, more
rigorous controller designs are
Fig.3.16: Closed-loop response under P,
much better than the Ziegler-
Nichols designs based on rules of
PI, and PID control using Z-N settings. thumb.
Obafemi Awolowo University 101
Chemical Engineering
 Troubleshooting Control Loops
 If a control loop is not performing satisfactorily, then troubleshooting is
necessary to identify the source of the problem.
 Based on experience in the chemical industry, he has observed that a
control loop that once operated satisfactorily can become either
unstable or excessively sluggish for a variety of reasons that include:
(a) Changing process conditions, usually changes in throughput rate.
Sticking control valve stem,
(b) Plugged line in a pressure or differential pressure transmitter.
Fouled heat exchangers, especially reboilers for distillation columns.

Obafemi Awolowo University 102

Chemical Engineering
 Troubleshooting Control Loops
(c) Cavitating pumps (usually caused by a suction pressure that is too
low).

The starting point for troubleshooting is to obtain enough background

information to clearly define the problem. Many questions need to be
answered:
(1) What is the process being controlled?
(2) What is the controlled variable?
(3) What are the control objectives?
(4) Are closed-loop response data available?

Obafemi Awolowo University 103

Chemical Engineering
 Troubleshooting Control Loops
(5) Is the controller in the manual or automatic mode? Is it reverse or
direct acting?
(6) If the process is cycling, what is the cycling frequency?
(7) What control algorithm is used? What are the controller settings?
(8) Is the process open-loop stable?
(9) What additional documentation is available, such as
control loop summary sheets, piping and instrumentation
diagrams, etc.?

Obafemi Awolowo University 104

Chemical Engineering