0% found this document useful (0 votes)
51 views6 pages

Modeling of The Modified Quadruple-Tank Process: A. Numsomran, V. Tipsuwanporn, K. Tirasesth

Download as pdf or txt
Download as pdf or txt
Download as pdf or txt
You are on page 1/ 6

- 818 -

Modeling of the Modified Quadruple-Tank Process



A. Numsomran, V. Tipsuwanporn, K. Tirasesth

Department of Instrumentation Engineering, Faculty of Engineering,
King Mongkuts Institute of Technology Ladkrabang, Bangkok 10520, Thailand
(Tel: +66 81925 0356 ; E-mail: knarjin@kmitl.ac.th)

Abstract: This paper presents Modeling of the Modified Quadruple-Tank Process. The new general form of plant model is
developed for illustrating the highly flexible plant structure which can be adjusted to many styles that are used for the propose of
giving control system engineers experience with multivariable control system design. In this paper described clearly about
structure and physical properties of Modified Quadruple-Tank Process, Mathematical plant modeling, Analysis of plant transfer
matrix characteristic, Transformation of general form of plant model to several different styles of water level plant. The value of
General form modeling concept for Modified Quadruple-Tank process is ability in transform its transfer matrix to several
different styles just only assign the value of inlet and outlet valve ratio. By the several model of transformed Modified Quadruple-
Tank , it can be used to teach students in the skills of understanding control limitation due to interactions, model uncertainties,
nonminimum phase behavior, and unpredictable time variations, Design decentralized controllers, and understanding their
limitations, Implementing decouplers to reduce the effect of interactions, and understanding their limitations.

Keywords: Modified Quadruple-Tank Process

1. INTRODUCTION
Most of industrial control problems are nonlinear and have
multiple controlled variables that are common properties for
the models of industrial processes to have significant
uncertainties, strong interactions, and nonminimum phase
behavior so it is important for control system engineer,
chemical engineer to understand the nonidealities of industrial
processes by carrying out experiments with a good laboratory
apparatus. A Modified quadruple tank process was designed
and constructed to give control system engineers laboratory
experience with key multivariable control concepts. The
general form of plant model creating can keep the all
properties of existing quadruple tank about multivariable zero
locations and their directions of transfer matrix G which have
intuitive physical interpretations in terms of how the valves
1
J
and
2
J
are set. Except above explain, General form
modeling concept for Modified Quadruple-Tank process is
ability in transform its transfer matrix to several different
styles such as quadruple-tank interacting process, quadruple-
tank non interacting process, couple-tank interacting process,
couple-tank non interacting process, 2 SISO Loop Second
order non interacting process, 1 SISO Loop Second order
interacting process and 1 SISO Loop Third order interacting
process just only assign the value of inlet and outlet valve
ratio. By the several model of transformed Modified
Quadruple-Tank , it can be used to teach students in the skills
of Understanding control limitation due to interactions, model
uncertainties, nonminimum phase behavior, and unpredictable
time variations, Design decentralized controllers, and
understanding their limitations, Implementing decouplers to
reduce the effect of interactions, and understanding their
limitations. The details of modeling of the modified
quadruple-tank process are described in this paper clearly as
the following section. In section 2 has the details of Structure
and Modeling of Modified Quadruple-Tank Process. . In
section 3 has the details of further analysis of system which is
about Relative Gain Array (RGA) and zeros of the system. In
section 4 has the detail of the modified quadruple-tank process
transformation and the conclusion is explained in section 5.

2. MODIFIED QUADRUPLE-TANK PROCESS
A. Structure of Modified Quadruple-Tank Process
In this section the physical model of the Modified
quadruple-tank is presented. From the basic equations (2) that
hold for each of the tanks and for the two pumps. Then they
are put together to obtain the model of the whole system as the
following describe.
Modified Quadruple Tank Process which has a flexible
structure design is constructed by four cylindrical tanks which
are mounted vertically on an acrylic board and are arranged in
a symmetric tank sequence as shown in Fig.1. The capacity of
all 4 tanks is 1192.5 cm
3
.Their outlet holes are connected to
outlet valve which function as flow resistance for water outlet
flowing. A small hole is drilled at the bottom of bottom two
tank to channel the water from each tank to a differential
pressure transmitter via an 8-mm tubing diameter for water
level measurement. The 32,000-cm
3
water reservoir tank is
used for store and recycle water for the experiment. Two API
Instrument gear pumps are used.
A Y-junction is used to divide the flow such that water is
channeled to a bottom level tank and the upper level tank
diagonal to it. This arrangement makes both levels in the
bottom two tanks a function of both pump flow rates. By
adjusting the valve knob, the process can be operated so that
one of the multivariable transition zeros is in the right half
plane, the left half plane, or switches between the two planes
in a stochastic time-varying manner.

SICE Annual Conference 2008
August 20-22, 2008, The University Electro-Communications, Japan
PR0001/08/0000-0818 400 2008 SICE
- 819 -


Figure 1. Schematic diagram of the modified quadruple-
tank process

A connected valve in fig.1 is the part that we design for
increasing laboratory plant flexibility. It combines the water
flow path of tank1 with tank2. When the connected valve ratio
x
| take the value over 0 , it will create the interacting channel
between water process in tank1 and tank2. By the interacting
structure, we can assess the performance of control system
design in the interacting condition.
The visual programming control interface used in the
laboratory was modified for use with this apparatus. The
interface enables students with a minimum background in
computer programming to make changes in the control
structures.
The target is to control the level in lower tanks with two
pumps. The process inputs are ( )
1
u t and ( )
2
u t (input voltages
to the pumps) and the outputs are ( )
1
h t and ( )
2
h t (water level
in tank1 and tank2)
B. Modeling of Modified Quadruple-Tank Process
Mass balances and Bernoullis law yield non-linear plant
equations as following.
Non Linear System
( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( ) ( )
( )
3 3 1 1 1
1 3
1 1
1 2 1 2 1
2 2 2 4 4
2 4
2 2
1 2 1 2 2
2 2
3 3 3
3 2
4 4 4
4
( )
2 ( ) 2 ( )
sgn 2 ( )
( )
2 ( ) 2 ( )
sgn 2 ( )
1
( )
2 ( ) ( )
( )
2 (
p
x x
p
x x
p
a dh t a
gh t gh t
dt A A
k
a
h t h t g h t h t u t
A A
dh t a a
gh t gh t
dt A A
k
a
h t h t g h t h t u t
A A
k
dh t a
gh t u t
dt A A
dh t a
gh
dt A
| |

|
| |

|
|
= +
+
= +
+ +

= +
=
( )
1 1
1
1
) ( )
p
k
t u t
A

+
(1)

Where
: A Cross section area of tank
2
( ) cm
:
i
a Cross section area of the outlet hole
2
( ) cm
:
x
a Cross section area of the connection hole between tank1
and tank2
2
( ) cm
( ) :
i
h t Water level ( ) cm
:
j
u Voltage input of pump ( ) volt
:
i
| Outlet valve ratio
:
x
| Connected valve ratio
:
j
Inlet valve ratio
:
pi
k Gain of pump
3
( / / sec) cm volt
: g Specific gravity
2
(981 / ) cm s
The pumps generate a flow proportional to the applied
voltage: ( ) ( )
, pump j pi j
q t k u t = . The flow that split up by the
valves to tank1 is ( )
1 1 1 p
k u t , tank2 is ( )
2 2 2 p
k u t , tank3 is
( ) ( )
2 2 2
1
p
k u t and tank4 is ( ) ( )
1 1 1
1
p
k u t .
Linearization
We represent the system by a state space form:
, x Ax Bu y Cx Du = + = + .From the equation (2), that the
system as ( ) , h f h u =

, where f is non-linear function of


water level h and the pump voltage u so we have to linearize
the system around the steady state ( , ) h u for the state space
representation. From the aid of Taylor series expansion,
Linearizing the non-linear system has represented by a state
space system as following.
x Ax Bu = + , y Cx Du = +

( )
( )
1 1 1 1
1 1
0
0
1 3
1 1 1 1
1 1 2 2
0
0
2 2 4 2 1
1 1
3 3 2 2 2
0 0 0
0
3
4 4
1
1
0 0 0 1 1
0
4
k
p
T T T T
A
x x
x x k
p
x T T T T x u
A
x x
x x k u
p
T
x x A
k
p
T
A

(
(
(
(
(
(
(
( ( (
(
( ( (
(
( ( ( (
( = + ( ( ( (
(
( ( ( (

(
( ( (
(
( ( (
(
(

(
(
(
(

< <


(2)
where the time constant
i
T such that
1
2
i i
i
i
a g
T A
h
|
= and
1 2
1
,
2
x x
x
a g
s
T A
h h
|
=


Transfer Matrix
The Laplace transform of (2) yields to the transfer matrix
of the four tank system.
( ) ( )
( )
( )
( )
( )
( )
1 1 1
,
2 2
Y s U s
G s C sI A B G s
Y s U s
( (

( ( = =
( (


( )
( ) ( )
( ) ( )
11 12
21 22
g s g s
G s
g s g s
(
( =
(


- 820 -

( )
( )
( ) ( ) ( )
( ) ( )
( ) ( )
1 /
1 1 1 1 1 2 1
1 1 /
4 2 2
11
2
/
1 2
1 /
1 1
1 /
2 2
T k T T k T A
p p x
A
T s T s T T
x
g s
T T T
x
T s T T
x
T s T T
x

| |

|
+
|
+ + +
|
\ .
=
| |
|
+ +
|
+ +
|
\ .


( )
( )
( )
1 /
2 2 2 2 2
2
1
2
1
3
12
2 1 2
1 1
2 1
1
T k k A
T
p p
T s T
x
A T
T s
x
g s
T T T T
x
T s T s
T T T T
x x x

| |

| |
|
|
+ + +
|
|
+
|
\ .
\ .
=
| | | | | |
| | |
+ + + +
| | |
\ . \ . \ .
< <
( )
( )
( )
1 /
1 1 1 1 2 1
1
1
1
1
2
4
21
1 2 1
1 1
1 2
2
T k T k A
T T
p p
x
T s
A T T
T s
x
g s
T T T T
x
T s T s
T T T T
x x x

| | | |

| |
| |
|
+ + +
| |
|
+
| |
\ .
\ . \ .
=
| | | | | |
| | |
+ + + +
| | |
\ . \ . \ .
< <
( )
( )
( )
1 /
2 2 2 2 1 2 2
1
1 1
3 1
22
2
/
1 2 2
1
2
1
1
1
T k T T k T A
p p x
A T
T s T s
T
x
g s
T T T T
x
T s
T T
x
T s
T
x

| |
|

|
|
+
| | |
| | + + +
| |
\ . \ .
=
| |
|
| | |
| |
+ +
| | | |
\ .
| | + +
| |
\ . \ .

(3)

3. Further Analysis of the System
A. Relative Gain Array (RGA)
For 2x2 systems, the relative gain array show how inputs
and outputs of the system are coupled. It has the following
form.
With
( ) ( )
( ) ( ) ( ) ( )
11 22
11 22 12 21
0 0
0 0 0 0
g g
m
g g g g

=

(4)
With equation (4) m can be found as:

( ) ( )
( ) ( )
2 1 1 2
1 2 1 2
1
x x
x x
T T T T
m
T T T T


+ +
=
+ + +
(5)
For big value of m , the dominating elements of the
transfer matrix are the diagonal elements. Output 1 is
effected mostly by input1, output2 by input2. If m is small,
output 1 depends mainly on input2, and output 2 on input
1. To know which input mainly affects which output is
important for design of PI controllers for the MIMO
system.
B. Zeros of the system
The zeros of the transfer matrix(3) are defined as the zeros
of ( ) det G s
( ) G s has four zeros. To find them we have to find the
zeros from the solutions of numerator of ( ) G s is equal zero.
From the fig. 3 explain the relation between the zeros position
on s-plane and valve ratio value
1 2
, that take the value of
1 2
+ in ( ) 0, 2 .


Figure 2. Zero of ( ) G s related with flow ratio
1 2
0 2 < + <

The zeros of ( ) G s can be located either in the left or in the
right half-plane. The system is minimum phase (both zeros are
in the left half-plane) for
1 2
1 2 < + < and the system is non-
minimum phase for
1 2
0 1 < + < .We found that the zero at
origin occur when
1 2
1 + = .

4. MODIFIED QUADRUPLE-TANK PROCESS
TRANSFORMATION

As we explain that the modified quadruple tank apparatus is
created for adapting as several process styles which can help
students for understanding control limitations due to
interactions, model uncertainties, non minimum phase
behavior, unpredictable time variations, understanding their
limitation of controller design and implementing a fully
multivariable control system. In this section will explain about
conditions and transfer matrix for each modified quadruple
tank process transformation styles.
- 821 -
A. Quadruple-tank non interacting process
Condition If
1 2
0 , 1 < < and 0
x
| = then plant model will be
a Quadruple-tank non interacting process
Transfer Matrix It has the schematic diagram of the model as
Fig. 3 and plant transfer matrix as equation (6).The plant
transfer matrix has 2 zeros and 4 poles. It is in the mode of
minimum phase when
1 2
+ take the value in ( ) 1, 2 . It is in
the mode of non minimum phase when
1 2
+ take the value
in ( ) 0,1 .
( )
( )
( ) ( )
( ) ( )
( )
( )
1 11 12 1
2 21 22 2
H s G s G s U s
H s G s G s U s
( ( (
=
( ( (
( ( (

<
(6)
where
( )
( )
1 1
11
1
1
/
1
p
T k A
G s
T s

=
+
( )
( )
( ) ( )
1 3
1 2
12
2
1 1
/ 1
p
T s T s
T k A
G s

+ +
=

<

( )
( )
( ) ( )
2 1
21
4 2
1
/
1 1
1
p
T k A
G s
T s T s

=
+ +

<
( )
( )
2
2 2 2
22
1
/
p
T s
T k A
G s

+
=


1
u
2
u
1
T
2
T
3
T
4
T
1
R
2
R
3
R
4
R
1
(1 )
1
2

2
(1 )

Figure 3 Schematic diagram of the quadruple-tank process.

B. Couple-tank interacting process
Condition If
1 2
, 1 = and 0 1
x
| < < then plant model will
be a Couple-tank interacting process
Transfer Matrix It has the schematic diagram of the model as
Fig. 4 and plant transfer matrix as equation (7).The plant
transfer matrix has 1 zeros and 4 poles.
( )
( )
( ) ( )
( ) ( )
( )
( )
1 11 12 1
2 21 22 2
H s G s G s U s
H s G s G s U s
( ( (
=
( ( (
( ( (

< (7)
where
( )
1 1
11
2
1 2 1
1
2
2
/
/
1
1
p
x
x
x
Tk A
G s
TT T T
T s
T T
T s
T
=
+ +
+ +

| |
|
| |
|
|
| | |
\ .
| |
\ . \ .

( )
2 2
12
2 1 2
2 1
1
/
1 1
p
x x x
T k A
G s
T T T T
T s T s
T T T T
=
+ + + +
| | | | | |
| | |
\ \ . \ . .
< <

( )
1 1
21
1 2 1
1 2
2
/
1 1
p
x x x
Tk A
G s
T T T T
T s T s
T T T T
=
+ + + +
| | | | | |
| | |
\ . \ . \ .
< <
( )
2
1 2 2
2
1
1
2 2
22
/
1
1
/
x
x
x
p
TT T T
T s
T T
T s
T
T k A
G s
+ +
+ +
=
| |
|
| |
|
|
| | |
\ .
| |
\ . \ .

1
u
2
u
1
T
2
T
3
T
4
T
1
R
2
R
3
R
4
R
1
(1 )
1
2

2
(1 )
x
R

Figure 4 Schematic diagram of the couple-tank interacting
process.

C. Couple-tank non interacting process
Condition If
1 2
, 1 = and 0
x
| = then plant model will be a
Couple-tank non interacting process
Transfer Matrix It has the schematic diagram of the model as
Fig. 5 and plant transfer matrix as equation (8).The plant
transfer matrix has 0 zeros and 2 poles.
1
u
2
u
1
T
2
T
3
T
4
T
1
R
2
R
3
R
4
R
1
(1 )
1

2
(1 )

Fig. 5. Schematic diagram of the couple-tank non interacting
process.


( )
( )
( )
( )
( )
( )
1 11 1
2 22 2
0
0
H s G s U s
H s G s U s
( ( (
=
( ( (
( ( (

< (8)
where
- 822 -
( )
( )
1 1
11
1
/
1
p
T k A
G s
T s
=
+

( )
( )
2 2
22
2
/
1
p
T k A
G s
T s
=
+

D. SISO Loop Second Order Non Interacting Process
Condition If
2
0 u = ,

1
0 = and 0
x
| = then plant model
will be a SISOLoopSecondOrderNonInteracting
Transfer Matrix It has the schematic diagram of the
model as Fig. 6 and plant transfer matrix as equation
(9).The plant transfer matrix has 0 zeros and 2 poles.

1
u
2
T
3
T
2
R
3
R
1
(1 )
1

4
R
4
T
1
T
1
R
x
R
2
(1 )
2

2
u

Fig. 6. Schematic diagram of the 2-tank second order non
interacting process.

( )
( ) ( )
( )
1 1
21 2
0 0
0
0
H s U s
G s H s
( ( (
=
( ( (
(

< (9)
where
( )
( )( )
2 1
21
2 4
/
1 1
p
T k A
G s
T s T s
=
+ +

E. SISO Loop Second Order Interacting Process
Condition If
2
0 u = ,
1
1 = and 0 1
x
| < < then plant
model will be a SISO Loop Second Order Interacting
Process
Transfer Matrix It has the schematic diagram of the
model as Fig. 7 and plant transfer matrix as equation
(10).The plant transfer matrix has 0 zeros and 2
poles.(G21)

Fig. 7. Schematic diagram of the couple-tank non interacting
process.

( )
( )
( )
( )
( )
1 11 1
2 21
0
0 0
H s G s U s
H s G s
( ( (
=
( ( (
( (

< (10)
where
( )
1 1
11
1
2
1 2 1
2
2
/
1
/
1
p
x
x
x
T k A
G s
T s
TT T T
T
T
T s
T
=
+ +
| |

|
| |
\ .
+ +
|
\ .
( )
1 1
21
1 2
1 2 1
2
/
1 1
p
x
x x x
T k A
G s
T s T s
T T T T
T T T T
=
+ + + +
| | | |

| |
\ . \ .
< <

F. SISO Loop Third Order Interacting Process
Condition If
1
0 u = ,
2
0 = and 0 1
x
| < < then plant model
will be a SISO Loop Third Order Interacting Process
Transfer Matrix It has the schematic diagram of the model
as Fig. 8 and plant transfer matrix as equation (11).The plant
transfer matrix has 0 zeros and 3 poles. (G22)


Fig. 8. Schematic diagram of the couple-tank non interacting
process.

( )
( )
( )
( ) ( )
1 12
2 2 22
0 0
0
H s G s
U s H s G s
( ( (
=
( ( (
( (

< (11)
where
( )
( )
2 2
2
3
2 1 2
2 1
1
12
/
1
1
1 1
p
x
x
x x
x
x
k A T
T s T
T T s
T T T T
T s T s
T T T T
G s
+ +
+
+ + + +
=
| |
|
\ .
| | | | | |
| | |
\ . \ . \ .
< <
- 823 -
( )
( )
2
1 2 2
2
1
1
1 2 2
1
3 1
22
/
1
1
/
1 1
x
x
x
p x
x
TT T T
T s
T T
T s
T
TT k T A
T
T s T s
T
G s
+ +
+ +

| |
+ + +
|
\ .
| |
|
| |
|
|
| | |
\ .
| |
|
\ . \ .
=


5. CONCLUSION

The value of General form modeling concept for Modified
Quadruple-Tank process is ability in transform its transfer
matrix to several different styles just only assign the value of
inlet and outlet valve ratio. By the several model of
transformed Modified Quadruple-Tank , it can be used to
teach students in the skills of Understanding control limitation
due to interactions, model uncertainties, nonminimum phase
behavior, and unpredictable time variations, Design
decentralized controllers, and understanding their limitations,
Implementing decouplers to reduce the effect of interactions,
and understanding their limitations. Selecting the best control
structure, based on the characteristics o the multivariable
process.
REFERENCES
[1] G.C.Goodwin, Control system design, New Jersey,
Prentice Hall, 2001.
[2] M. Grebeck, A Comparison of controllers for the
quadruple tank system, Technical report, Department
of Automatic control, Lund Institute of Technology,
Lund, Sweden, 1998.
[3] W.L. Luyben, Process Modeling simulation and control
for chemical engineers, New York, McGraw-Hill,1990.
[4] R.C.Dorf, Modern Control systems, California,
Addison-Wesley, 2001.
[5] Shreesha C., "MISO Structure Based Control-Relevant
Identification of MIMO Systems," Proc. of the American
Control Conference, Arlington, VA, pp. 1184-1189,
June 2001.
[6] K. H. Johansson, "The Quadruple-Tank Process a
multivariable process with an adjustable zero," IEEE
Transactions on Control Systems Technology, vol. 1, pp.
456-465, May 2000.
[7] M. Sato, "Theory of Hyperfunctions," Tokyo Journal,
Sect I.8 (1959), pp. 139-193,1959.
[8] I.M.Gelfand and G.E. Shilov, Generalized
Functions,Academic Press, 1964.
[9] T. Trisuwannawat, N. Komine, and M. Iida, "Optimal
H2 control of Oscillations via Derivative State
Constraints, Proc. of the American Control Conference,
San Diego, CA, pp. 2305-2309, 1999.
[10] T. Trisuwannawat K. Tirasesth, M. Iida, N. Komine and
Y. Ochiai,"Derivative State Constrained Optimal H2
Control for Oscillatory Systems and Its Application,"
IEEJ Transactions on Industry Application Society; IAS,
vol. 120-D, No. 6 , pp. 775-781, 2000.

You might also like