Relay Feedback Auto Tuning of PID

Controllers
Antonio Flores T. ∗
Departmento de Ingenierı́a Quı́mica
Universidad Iberoamericana
Prolongación Paseo de la Reforma 880
México DF, 01210, MEXICO
October 4, 2007

∗
E-mail: [email protected], http://200.13.98.241ãntonio, phone/fax: (+52)5
2674279

1
1 Introduction
For a certain class of process plants, the so-called “auto tuning” procedure
for the automatic tuning of PID controllers can be used. Such a procedure
is based on the idea of using an on/off controller (called a relay controller)
whose dynamic behaviour resembles to that shown in Figure 1(a). Starting
from its nominal bias value (denoted as 0 in the Figure) the control action
is increased by an amount denoted by h and later on decreased until a value
denoted by −h.

+h

−h

Time
(a)
Pu

y A

Time
(b)

Figure 1:

The closed-loop response of the plant, subject to the above described ac-
tions of the relay controller, will be similar to that depicted in Figure 1(b).
Initially, the plant oscillates without a definite pattern around the nominal
output value (denoted as 0 in the Figure) until a definite and repeated out-
put response can be easily identified. When we reach this closed-loop plant
response pattern the oscillation period (Pu ) and the amplitude (A) of the
plant response can be measured and used for PID controller tuning. In fact,

2
the ultimate gain can be computed as:
4h
Kcu = (1)
πA
Having determined the ultimate gain Kcu and the oscillation period Pu
the PID controller tuning parameters can be obtained from the following
table:

Kc τI τD
P 0.5Kcu
PI 0.45Kcu Pu /1.2
PID 0.6Kcu Pu /2 Pu /8

Example
Let us consider a process system given by the following transfer function:
6
Gp =
48s3 + 44s2 + 12s + 1
and assume that the plant will be controlled by a PI feedback control system.
Design the control system using the relay auto tuning method.
There are some decisions that ought to be taken before testing the relay
auto tuning procedure:

• Pure gain controller value (Kc ).

• Size of the manipulated variable deviation from the bias value (h).

By the time being let us pick up a small deviation of the manipulated variable
from the bias value: h = ±0.1. The value of the controller gain should be
large enough so that the value of the manipulated variable will lie between
the bounds as represented by h. Therefore after some trials Kc = 100. The
implementation of the relay auto tuning procedure is depicted in Figure 2.
In Figure 3 the dynamic behaviour of both the manipulated (u) and
controlled (y) variables is shown. From both plots the following values can
be easily read:

Pu = 12.5
A = 0.07785

3
 Scope 1
Scope 2
6
100
48s3 +44s2 +12s+1
Step Gain Saturation Transfer Fcn Scope

Figure 2:

hence,
Kcu = 1.6355
and the tuning parameters of the PI controller will given by:

Kc = 0.7360
τi = 10.4167

Similar results are also obtained by using the relay function available in
Simulink whose closed-loop implementation is shown in Figure 4.

Example
The dynamic mathematical model of a reaction train of 3 series connected CSTRs
k
(see Figure 5) where the reaction 2A → B takes place reads as follows:

dCA1 Q 2
= (CAo − CA1 ) − kCA1
dt V
dCA2 Q 2
= (CA1 − CA2 ) − kCA2
dt V
dCA3 Q 2
= (CA2 − CA3 ) − kCA3
dt V
using the following values of the design parameters:

4
 0.2

0.15

0.1

0.05

0
u

−0.05

−0.1

−0.15

−0.2
0 10 20 30 40 50 60 70 80 90 100
Time

(a)
0.25

0.2

0.15

0.1
y

0.05

−0.05

−0.1
0 10 20 30 40 50 60 70 80 90 100
Time

(b)

Figure 3:

Scope 7
Scope 8
Scope 9
6
48s3 +44s2 +12s+1
Step 4 Relay Transfer Fcn 2 Scope 6

Figure 4:

5
 Q1 Q2 Q
Q 3
0 C A1 C A2 C A3
C A0

V1 V2 V3

Figure 5: Flowsheet of the 3 CSTRs isothermal reaction train.

Parameter Value Units

k 5x10−4 L/(mol-min)
CAo 100 mol/L
Qi 50 L/min
Vj 1000 L
the conversion degree of reactant A coming out from the third reactor is around
67% which corresponds to CA3 = 32.5641 mol/L. Using the relay auto tuning
procedure design and implement a closed-loop PI control system capable of rais-
ing the reaction train conversion from 67% up to 80% (which is equivalent to
CA3 = 20 mol/L). The manipulated variable is the reactant A feed stream volu-
metric flow rate sent to the first reactor (Q), which happens to be the same along
the reaction train, whereas the controlled variable is the reactant A concentration
leaving the third reactor CA3 .
The transfer function between the controlled and manipulated variables
reads as follows:
CA3 (s) 0.0106s2 + 0.003105s + 0.0003101
Gp (s) = = 3
Q(s) s + 0.2875s2 + 0.02734s + 0.00086
which can be approximated by the following first order plus time delay trans-
fer function:
0.3603
G(s) = e−4.1845
18.389s + 1
using a Padé first order approximation for the representation of the delay:
−0.3603s + 0.1722
G(s) =
18.39s2 + 9.789s + 0.478

6
 The Simulink implementation of the relay auto tuning procedure is shown
in Figure 6, whereas the closed-loop response obtained from the application
of this automatic tuning procedure is depicted in Figure 7.

time
Clock u
To Workspace1
To Workspace4

Scope 4
Scope 5
−0.3603 s+0.1722
100
18 .39 s2 +9.789 s+0.478
Step 1 Gain 1 Saturation 1 Transfer Fcn 1 Scope 3

To Workspace3

Figure 6:

0.2

0.15

0.1

0.05

0
u

−0.05

−0.1

−0.15

−0.2
0 10 20 30 40 50 60 70 80 90 100
Time

(a)
−3
x 10
8

2
y

−2

−4

−6
0 10 20 30 40 50 60 70 80 90 100
Time

(b)

Figure 7:

7
 Hence,

h = 0.1
A = 5.2x10−3
Pu = 13.4
Kcu = 24.4854

therefore, the PI controller tuning parameters are given as follows:

Kc = 11.0184
τi = 11.1667

The Simulink implementation of the closed-loop feedback control system

is shown in Figure 8, whereas the closed-loop response obtained using a PI
control system is depicted in Figure 9.

time u
Clock To Workspace1 To Workspace4

50

Constant1 Scope 4
Scope 5
0.0106 s2 +0.003105 s+0.0003101
PID
s3 +0.2875 s2 +0.02734 s+0.00086
Step 1 PID Controller Saturation Transfer Fcn 1 Scope 3
y
32.5641
To Workspace3
Constant

Figure 8:

50 34

45 32

40 30

35 28

30 26
Q [L/min]

C [mol/L]

25 24

20 22

15 20

10 18

5 16

0 14
0 20 40 60 80 100 120 140 160 180 200 0 20 40 60 80 100 120 140 160 180 200
Time Time

(a) (b)

Figure 9: