0% found this document useful (0 votes)
2 views38 pages

Lecture 12

The document discusses multivariable control systems (MIMO) and their interactions, using examples such as distillation columns and stirred tank reactors to illustrate how changes in one control loop can affect another. It emphasizes the importance of understanding these interactions for effective control system design, including considerations for minimizing loop interactions and analyzing system stability. Additionally, it introduces the concept of a relative gain array to aid in selecting control loops with minimal interaction.

Uploaded by

deepaanmol2002
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
2 views38 pages

Lecture 12

The document discusses multivariable control systems (MIMO) and their interactions, using examples such as distillation columns and stirred tank reactors to illustrate how changes in one control loop can affect another. It emphasizes the importance of understanding these interactions for effective control system design, including considerations for minimizing loop interactions and analyzing system stability. Additionally, it introduces the concept of a relative gain array to aid in selecting control loops with minimal interaction.

Uploaded by

deepaanmol2002
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 38

INSTR F343: Industrial Instrumentation and Control

Multivariable control systems and loop interactions

BITS Pilani Dr. Sujan Yenuganti (office no: 5699 - O)


yenuganti.sujan@pilani.bits-pilani.ac.in
Pilani Campus
Example process (MIMO)

Distillation column

BITS Pilani, Deemed to be University under Section 3, UGC Act


Example: MIMO control system having
interaction
Loop 1 controls the liquid level by manipulating
the effluent flow rate.
Loop 2 controls the temperature by manipulating
the steam flow rate.

When the inlet flow rate (disturbance load) or the


desired value of the liquid level (set point)
changes, the loop 1 attempts to compensate for
the changes by manipulating the effluent flow
rate.

Stirred tank heater


This inturn, will disturb the temperature of the
liquid in the tank, and loop 2 will compensate by
adjusting the value of the steam flow rate.
BITS Pilani, Deemed to be University under Section 3, UGC Act
Example: MIMO control system having
interaction

On the other hand, when the temperature of the


inlet stream (disturbance load) or the desired
value of the temperature (set point) changes, the
loop 2 attempts to compensate for the changes by
manipulating the steam flow rate.

This will leave the liquid level undisturbed.

Thus, we notice that loop 1 affects loop 2, but not Stirred tank heater
vice versa.
The stirred tank heater has a one-way interaction

BITS Pilani, Deemed to be University under Section 3, UGC Act


Example: MIMO control system having
interaction
Loop 1 controls the effluent concentration by
manipulating the inlet flow rate.
Loop 2 controls the temperature by manipulating the
coolant flow in the jacket.

When a change in the inlet concentration (disturbance


load) or the desired effluent concentration (set point), loop
1 attempts to compensate for the changes by manipulating
the feed flow rate.

However, this change in feed rate also disturbs the


temperature of the liquid in the tank, and loop 2 will Stirred tank reactor system
compensate by adjusting the value of the coolant flow
rate, which inturn affects the effluent concentration.
BITS Pilani, Deemed to be University under Section 3, UGC Act
Example: MIMO control system having
interaction

On the other hand, an attempt to compensate for


changes in feed temperature (disturbance load) or
the desired reactor temperature (set point), loop 1
attempts to compensate for the changes in effluent
concentration by manipulating the feed flow rate,
which in turn disturbs the reactor temperature.

Here the loop 1 interacts with loop 2 in both


directions. Stirred tank reactor system

The stirred tank reactor has interaction in both-ways.

BITS Pilani, Deemed to be University under Section 3, UGC Act


MIMO process

𝑦ത1 s = 𝐻11 𝑠 𝑚
ഥ 1 𝑠 + 𝐻12 𝑠 𝑚
ഥ2 𝑠

𝑦ത2 s = 𝐻21 𝑠 𝑚
ഥ 1 𝑠 + 𝐻22 𝑠 𝑚
ഥ 2 (𝑠)

A change in m1 or m2 will affect both


controlled outputs

A MIMO process
BITS Pilani, Deemed to be University under Section 3, UGC Act
MIMIO control system

A typical control system for MIMO process


BITS Pilani, Deemed to be University under Section 3, UGC Act
Interaction in a MIMO control system…

To understand the nature of interaction between the two control loops, we will study the effects of
input changes on the outputs when

1) One loop is closed, and the other is open


2) Both loops are closed.

One loop is closed:

Assume loop 1 is closed and loop 2 is open. Also, assume m2 is constant and make a change in the set
point y1,SP

BITS Pilani, Deemed to be University under Section 3, UGC Act


Interaction in a MIMO control system…

H11

Interaction with
H12
H21 one loop closed

H22

𝐻11 𝐺𝑐1 𝐻21 𝐺𝑐1


𝑦ത1 s = 𝑦ത1,𝑆𝑃 𝑠 𝑦ത2 s = 𝑦ത1,𝑆𝑃 (𝑠)
1 + 𝐻11 𝐺𝑐1 1 + 𝐻11 𝐺𝑐1

BITS Pilani, Deemed to be University under Section 3, UGC Act


Interaction in MIMO control system…

Both loops closed:

Initially, the process is at steady state with both outputs at desired values.

Consider a change in the set point y1,SP only, and keep the set point of loop 2 the same.

The controller of loop 1 will change the value of m1 in such a way to bring the output y1 to the new
set point value.

This is the direct effect of m1 on y1 through loop 1, and is shown by the dashed line.

BITS Pilani, Deemed to be University under Section 3, UGC Act


Interaction in MIMO control system…

Interaction with both loop closed

BITS Pilani, Deemed to be University under Section 3, UGC Act


Interaction in MIMO control system…

Both loops closed:

The control action of m1 will not only attempt to bring y1 to the new set point, but will also disturb y2
from its steady state value.

Then the controller of loop 2 will attempt to compensate for the variations in y2 by changing the
value of the manipulated variable m2.

But a change in m2, in turn, affects output y1.

This is an indirect effect of m1 on y1 through loop 2, as shown by the dashed line.

BITS Pilani, Deemed to be University under Section 3, UGC Act


Interaction in MIMO control system…

Interaction with both loop closed

BITS Pilani, Deemed to be University under Section 3, UGC Act


MIMO control system design
consideration points

• Choose the manipulation that has a direct and fast effect on a controlled variable.

• Choose the couplings so that there is a little dead time between every manipulation and the
corresponding controlled variable.

• Select the coupling so that the interaction of the control loops is minimal, etc.

BITS Pilani, Deemed to be University under Section 3, UGC Act


Quantitative analysis of interaction
For a two-input two output process, let the relationship be,
𝑦ത1 s = 𝐻11 𝑠 𝑚
ഥ 1 𝑠 + 𝐻12 𝑠 𝑚
ഥ2 𝑠

𝑦ത2 s = 𝐻21 𝑠 𝑚
ഥ 1 𝑠 + 𝐻22 𝑠 𝑚
ഥ 2 (𝑠)
Using closed loop controllers, we can have,
𝑚
ഥ 1 s = 𝐺𝐶1 𝑠 [𝑦ത1,𝑆𝑃 𝑠 − 𝑦ത1 s ]

𝑚
ഥ 2 s = 𝐺𝐶2 𝑠 [𝑦ത2,𝑆𝑃 𝑠 − 𝑦ത2 s ]
Using above relationships, we can have,
1 + 𝐻11 𝐺𝑐1 𝑦ത1 + 𝐻12 𝐺𝑐2 𝑦ത2 = 𝐻11 𝐺𝑐1 𝑦ത1,𝑆𝑃 + H12 Gc2 𝑦ത2,𝑆𝑃

𝐻21 𝐺𝑐1 𝑦ത1 + 1 + 𝐻22 𝐺𝑐2 𝑦ത2 = 𝐻21 𝐺𝑐1 𝑦ത1,𝑆𝑃 + H22 Gc2 𝑦ത2,𝑆𝑃

BITS Pilani, Deemed to be University under Section 3, UGC Act


Quantitative analysis of interaction…
we have,
1 + 𝐻11 𝐺𝑐1 𝑦ത1 + 𝐻12 𝐺𝑐2 𝑦ത2 = 𝐻11 𝐺𝑐1 𝑦ത1,𝑆𝑃 + H12 Gc2 𝑦ത2,𝑆𝑃

𝐻21 𝐺𝑐1 𝑦ത1 + 1 + 𝐻22 𝐺𝑐2 𝑦ത2 = 𝐻21 𝐺𝑐1 𝑦ത1,𝑆𝑃 + H22 Gc2 𝑦ത2,𝑆𝑃

Solving above relationships for finding out input-output relationships,


𝑦ത1 s = 𝑃11 𝑠 𝑦ത1,𝑆𝑃 𝑠 + 𝑃12 𝑠 𝑦ത2,𝑆𝑃 𝑠
𝑦ത2 s = 𝑃21 𝑠 𝑦ത1,𝑆𝑃 𝑠 + 𝑃22 𝑠 𝑦ത2,𝑆𝑃 (𝑠)

Where,
𝐻11 𝐺𝐶1 + 𝐺𝑐1 𝐺𝑐2 𝐻11 𝐻22 − 𝐻12 𝐻21 𝐻12 𝐺𝐶2
𝑃11 𝑠 = 𝑃12 𝑠 =
𝑄 𝑠 𝑄 𝑠

𝐻22 𝐺𝐶2 + 𝐺𝑐1 𝐺𝑐2 𝐻11 𝐻22 − 𝐻12 𝐻21 𝐻21 𝐺𝐶1
𝑃22 𝑠 = 𝑃21 𝑠 =
𝑄 𝑠 𝑄 𝑠
BITS Pilani, Deemed to be University under Section 3, UGC Act
Quantitative analysis of interaction…

𝑦ത1 s = 𝑃11 𝑠 𝑦ത1,𝑆𝑃 𝑠 + 𝑃12 𝑠 𝑦ത2,𝑆𝑃 𝑠


𝑦ത2 s = 𝑃21 𝑠 𝑦ത1,𝑆𝑃 𝑠 + 𝑃22 𝑠 𝑦ത2,𝑆𝑃 (𝑠)
Where,
𝐻11 𝐺𝐶1 + 𝐺𝑐1 𝐺𝑐2 𝐻11 𝐻22 − 𝐻12 𝐻21 𝐻12 𝐺𝐶2
𝑃11 𝑠 = 𝑃12 𝑠 =
𝑄 𝑠 𝑄 𝑠

𝐻22 𝐺𝐶2 + 𝐺𝑐1 𝐺𝑐2 𝐻11 𝐻22 − 𝐻12 𝐻21 𝐻21 𝐺𝐶1
𝑃22 𝑠 = 𝑃21 𝑠 =
𝑄 𝑠 𝑄 𝑠

𝑄 s = 1 + H11 Gc1 1 + H22 Gc2 − H12 H21 Gc1 Gc2

BITS Pilani, Deemed to be University under Section 3, UGC Act


Remarks:

• If we have 𝐻12 = 𝐻21 = 0, there will be no interaction.


• The stability of the closed loop outputs of the two interacting loop is determined by
the roots of the characteristic equation, Q(s).

𝑄 s ≡ 1 + H11 Gc1 1 + H22 Gc2 − H12 H21 Gc1 Gc2 = 0

• If individual loops are stable, they does not guarantee the overall stability, i.e. when
both loops are closed.

BITS Pilani, Deemed to be University under Section 3, UGC Act


Relative gain array and selection of loops

• There can be different ways in which the manipulated variables can be used to control the
controlled process output.

• For a process with N controlled variables and N manipulated variables, there can be N ! ways to
form the loops.

• Relative gain array provides a way to find out the pairs of manipulated variables and controlled
variables, which can be chosen to have minimal interaction between loops.

BITS Pilani, Deemed to be University under Section 3, UGC Act


Relative gain array and selection of
loops…

Assume that m2 remains constant. Introduce a step change


in input m1 of magnitude m1 and then record the new
steady state value of y1.

Let y1 be the change from the previous steady state value.

The open-loop steady state gain between y1 and m1 when


m2 is kept constant is given by

Δ𝑦1
Δ𝑚1 𝑚2 =𝑐𝑜𝑛𝑠𝑡𝑛𝑎𝑡

BITS Pilani, Deemed to be University under Section 3, UGC Act


Relative gain array and selection of
loops…
Find,
Δ𝑦1
Δ𝑚1 𝑚2 =𝑐𝑜𝑛𝑠𝑡𝑛𝑎𝑡

The open-loop steady-state gain between y1


and m1 when y2 is kept constant is given by

Δ𝑦1 ′
Δ𝑚1 𝑦2 =𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡

BITS Pilani, Deemed to be University under Section 3, UGC Act


Relative gain array and selection of
loops…
Δ𝑦1 /Δ𝑚1 𝑚2
Now, the relative gain becomes, 𝜆11 =
Δ𝑦1 ′/Δ𝑚1 𝑦2

Inferences from this relative gain index,


• If 𝜆11 = 0, then 𝑦1 does not respond to 𝑚1 and 𝑚1 should not be used to control 𝑦1 .

• If 𝜆11 = 1, then 𝑚2 does not affect 𝑦1 and the control loop between 𝑦1 and 𝑚1 does not interact
with the loop of 𝑦2 and 𝑚2 . In this case we have completely decoupled loops.

• If 0 < 𝜆11 < 1, then an interaction exists and as m2 varies, it affects the steady state values of y1.
The smaller the value of 𝜆11, the larger the interaction becomes.

• If 𝜆11 < 0, then m2 causes a strong effect on y1 and in the opposite direction from that caused by
m1. In this case, the interaction effect is very dangerous for stability of the system.
BITS Pilani, Deemed to be University under Section 3, UGC Act
Relative gain array and selection of
loops…
In a similar manner, we can also calculate the relative gains for other pairs as well,

Δ𝑦1 /Δ𝑚1 𝑚2
𝜆11 = relative gain between y1 and m1
Δ𝑦1 ′/Δ𝑚1 𝑦2

Δ𝑦1 /Δ𝑚2 𝑚1
𝜆12 = relative gain between y1 and m2
Δ𝑦1 ′/Δ𝑚2 𝑦2

Δ𝑦2 /Δ𝑚1 𝑚2
𝜆21 = relative gain between y2 and m1
Δ𝑦2 ′/Δ𝑚1 𝑦1

Δ𝑦2 /Δ𝑚2 𝑚1
𝜆22 = relative gain between y2 and m2
Δ𝑦2 ′/Δ𝑚2 𝑦1

BITS Pilani, Deemed to be University under Section 3, UGC Act


Relative gain array and selection of
loops…
The relative gain array then becomes,
𝑚1 𝑚2
𝜆11 𝜆12 𝑦1
Λ=
𝜆21 𝜆22 𝑦2
Also,

It can be shown that the sum of the relative gains in any row or column of the array is equal to 1.

If we know one of the four relative gains, the other three can be easily computed

BITS Pilani, Deemed to be University under Section 3, UGC Act


Further remarks on relative gain array

• Relative gains provide a measure of interaction based on steady state considerations. Nothing
is said about dynamic interactions.

• The relative gain array is a square matrix, signifying that the no. of manipulated variables is
equal to number of controlled outputs.

• One can compute the relative gains using two methods: one is a computational approach
based on the mathematical model of the system and the second one is experimental
approach.

BITS Pilani, Deemed to be University under Section 3, UGC Act


Further remarks on relative gain array…

• Relative gain array concept is not limited to two-input-two-output systems only. It can be
extended to general processes with more than two controlled variables.

Δ𝑦𝑖 /Δ𝑚𝑗
𝑚
𝜆𝑖𝑗 =
Δ𝑦𝑖 /Δ𝑚𝑗
𝑦

𝑁 𝑁

෍ 𝜆𝑖𝑗 = 1, 𝑓𝑜𝑟 𝑗 = 1,2, … , 𝑁 ෍ 𝜆𝑖𝑗 = 1, 𝑓𝑜𝑟 𝑖 = 1,2, … , 𝑁


𝑖=1 𝑗=1

BITS Pilani, Deemed to be University under Section 3, UGC Act


Numerical

Select the loops using relative gain array for the following process,
1 1
𝑦ത1 = 𝑚ഥ1 + 𝑚
ഥ2
𝑠+1 0.1𝑠 + 1

−0.2 0.8
𝑦ത2 = 𝑚
ഥ1 + 𝑚
ഥ2
0.5𝑠 + 1 𝑠+1

BITS Pilani, Deemed to be University under Section 3, UGC Act


Solution

1 1
𝑦ത1 = 𝑚
ഥ1 + 𝑚
ഥ2
𝑠+1 0.1𝑠 + 1

−0.2 0.8
𝑦ത2 = 𝑚
ഥ + 𝑚

0.5𝑠 + 1 1 𝑠 + 1 2
Let us compute the relative gains 𝜆𝑖𝑗 ,
Δ𝑦1 Make a unit step change in m1 (i.e. , 𝑚
ഥ1 =
= ?
Δ𝑚1 𝑚2 =𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
1/s), and m2 as constant (i.e. 𝑚
ഥ 2 = 0)

1 1 1 1
𝑦ത1 = 𝑚
ഥ 1 ⇒ 𝑦ത1 = . Δ𝑦1 /Δ𝑚1 𝑚2 = =1
𝑠+1 𝑠+1 𝑠 1

BITS Pilani, Deemed to be University under Section 3, UGC Act


Contd..

1 1
𝑦ത1 = 𝑚
ഥ1 + 𝑚
ഥ2
𝑠+1 0.1𝑠 + 1

−0.2 0.8
𝑦ത2 = 𝑚
ഥ + 𝑚

0.5𝑠 + 1 1 𝑠 + 1 2

Δ𝑦1 Make a unit step change in m1 (i.e. , 𝑚 ഥ1 =


= ?
Δ𝑚1 𝑦2 =𝑐𝑜𝑛𝑠𝑡𝑛𝑎𝑡
1/s), and y2 as constant (i.e. 𝑦ത2 = 0)

−0.2 0.8 −0.2 0.8


𝑦ത2 = 𝑚
ഥ1 + 𝑚
ഥ2 ⇒ 0= 𝑚
ഥ1 + 𝑚
ഥ2
0.5𝑠 + 1 𝑠+1 0.5𝑠 + 1 𝑠+1

BITS Pilani, Deemed to be University under Section 3, UGC Act


Contd…

𝑠+1
𝑚
ഥ 2 = 0.25 𝑚

0.5𝑠 + 1 1
Using above equation to find 𝑦ത1 ,
1 1
𝑦ത1 = 𝑚
ഥ1 + 𝑚
ഥ2
𝑠+1 0.1𝑠 + 1
1 1 𝑠+1
𝑦ത1 = 𝑚
ഥ1 + × 0.25 𝑚
ഥ1
⇒ 𝑠+1 0.1𝑠 + 1 0.5𝑠 + 1
1 1 1 𝑠+1 1
Δ𝑦1 = lim[𝑠 𝑦ത1 ] = lim 𝑠 + × 0.25 = 1.25
𝑠→0 𝑠→0 𝑠 + 1 𝑠 0.1𝑠 + 1 0.5𝑠 + 1 𝑠

1.25
Δ𝑦1 = 1.25 ⇒ Δ𝑦1 /Δ𝑚1 𝑦2 = = 1.25
1
BITS Pilani, Deemed to be University under Section 3, UGC Act
Contd..

Δ𝑦1 /Δ𝑚1 𝑚2 1
𝜆11 = = = 0.8
Δ𝑦1 /Δ𝑚1 𝑦2 1.25

Also,

𝑚1 𝑚2
0.8 0.2 𝑦1
Λ=
0.2 0.8 𝑦2

The relative gain array indicates that m1 should be used to control y1


and m2 should be used to control y2.

BITS Pilani, Deemed to be University under Section 3, UGC Act


Numerical

• For a refinery distillation column, we have four product stream with four flow
rates as manipulated variables. The RGA for this process is shown below.
Recommend pairing of manipulated variables to controlled variables
(compositions, yi).

Hint: Select the control loops by pairing the controlled outputs yi with the
manipulated variables mi in such a way that the relative gains λij are positive and
as close as possible to unity.
BITS Pilani, Deemed to be University under Section 3, UGC Act
Solution

𝑦1 𝑤𝑖𝑡ℎ 𝑢1 𝑦2 𝑤𝑖𝑡ℎ 𝑢4 𝑦3 𝑤𝑖𝑡ℎ 𝑢2 𝑦4 𝑤𝑖𝑡ℎ 𝑢3

BITS Pilani, Deemed to be University under Section 3, UGC Act


Numerical

Figure shown below depicts two interacting tanks with equal areas of 1 m2 with
m1 flowing in tank-1 and m1 flowing in tank-2. Interconnecting pipe and
outflow pipe have resistances R1 and R2, respectively. Both R1 and R2 are
equal to 0.5 sec/m2. Find 𝜆11 and comment upon the possible pairings.

BITS Pilani, Deemed to be University under Section 3, UGC Act


Solution

 dh1 
 dt     h1     m1 
 dh     h     m 
 2   2    2 
 dt 

dh 1 (h 1  h 2 )
A1  m1 
dt R2
dh 2 (h1  h 2 ) h2
A2   m2 
dt R1 R2

BITS Pilani, Deemed to be University under Section 3, UGC Act


Contd…

𝑆𝑖𝑛𝑐𝑒, 𝐴1 = 𝐴2 = 1 𝑚2 and 𝑅1 = 𝑅2 = 0.5 𝑠𝑒𝑐/𝑚2


𝑠𝑜,
𝑑ℎ1
𝑑𝑡 = −2 2 ℎ1 + 1 0 𝑚1
𝑑ℎ2 2 −4 ℎ2 0 1 𝑚2
𝑑𝑡
𝑑ℎ1
𝑑𝑡 = 𝐴𝐻 + 𝐵𝑀; 𝐴 = −2 2 ; 𝐵 = 1 0
𝑑ℎ2 2 −4 0 1
𝑑𝑡

BITS Pilani, Deemed to be University under Section 3, UGC Act


Contd…

s  4 2 
 2  
Gp   2 
s 2
s  6s  4
1 0.5
𝑆𝑡𝑒𝑎𝑑𝑦 𝑠𝑡𝑎𝑡𝑒 𝑔𝑎𝑖𝑛 𝑚𝑎𝑡𝑟𝑖𝑥 𝑖𝑠 𝐾 =
0.5 0.5
1
𝜆11 = =2
𝐾12 𝐾21
1−
𝐾11 𝐾22
Highly undesirable interactions
2 −1
𝑅𝐺𝐴(Λ) = Need cross controller/decouplers
−1 2

BITS Pilani, Deemed to be University under Section 3, UGC Act

You might also like