Multivariable control
Multivariable process
(Seborg Chapter 18 )
Control Loop Interaction
Single input and single output
Control of Multiple-Input, Multiple-
Output Processes
Multiloop controllers
Modeling the interactions
Relative Gain Array (RGA)
Singular Value Analysis (SVA)
Decoupling strategies
Control of multivariable processes
In practical control problems there typically are a number of
process variables which must be controlled and a number
which can be manipulated
Several simple physical examples are shown in Fig. 18.1.
Note "process interactions" between controlled and
manipulated variables.
Chapter 18
Controlled Variables: xD , xB , P, hD , and hB
Manipulated Variables: D, B, R , Q D , and Q B
Chapter 18
Characterizing process interactions and selecting an appropriate
multiloop control configuration.
If process interactions are significant, even the best multiloop control
system may not provide satisfactory control.
In these situations there are incentives for considering multivariable
control strategies
Chapter 18
Definitions:
Multiloop control: Each manipulated variable depends on only a
single controlled variable, i.e., a set of conventional feedback
controllers.
Multivariable Control: Each manipulated variable can depend on two
or more of the controlled variables.
Examples: decoupling control, model predictive control
Multiloop Control Strategy
Typical industrial approach
Consists of using n standard FB controllers (e.g. PID), one for each
controlled variable.
Control system design
1. Select controlled and manipulated variables.
2. Select pairing of controlled and manipulated variables.
3. Specify types of FB controllers.
Chapter 18
Example: 2 x 2 system
Two possible controller pairings:
U1 with Y1, U2 with Y2 or
U1 with Y2, U2 with Y1
Note: For n x n system, n! possible pairing configurations.
Chapter 18
Transfer Function Model (2 x 2 system)
Two controlled variables and two manipulated variables
(4 transfer functions required)
Y1 ( s ) Y1 ( s )
Chapter 18
GP11 ( s ), GP12 ( s )
U1 ( s) U 2 (s)
Y2 ( s ) Y2 ( s )
GP 21 ( s), GP 22 ( s )
U1 ( s) U 2 ( s)
Thus, the input-output relations for the process can be
written as:
Y1 ( s ) GP11 ( s )U1 ( s ) GP12 ( s )U 2 ( s )
Y2 ( s ) GP 21 ( s )U1 ( s ) GP 22 ( s )U 2 ( s )
Or in vector-matrix notation as,
Y ( s ) GP ( s )U ( s )
where Y(s) and U(s) are vectors,
Chapter 18
Y1 ( s ) U 1 ( s )
Y (s) , U (s)
Y (
2 s ) U
2 ( s )
And Gp(s) is the transfer function matrix for the process
G P11 (s) G P12 (s)
G P (s )
G
P 21 (s ) G P 22 ( s )
Control-loop interactions
Process interactions may induce undesirable interactions
between two or more control loops.
Example: 2 x 2 system
Chapter 18
Control loop interactions are due to the presence of a
third feedback loop.
Problems arising from control loop interactions
i) Closed -loop system may become destabilized.
ii) Controller tuning becomes more difficult
Block Diagram Analysis
For the multiloop control configuration the transfer function
between a controlled and a manipulated variable depends
on whether the other feedback control loops are open or
closed.
Chapter 18
Example: 2 x 2 system, 1-1/2 -2 pairing, From block
diagram algebra we can show
Y1 ( s ) (second loop open)
GP11 ( s),
U1 ( s)
Y1 ( s ) GP12GP 21GC 2 (second loop closed)
GP11
U1 (s) 1 GC 2GP 22
Note that the last expression contains GC2 .
Chapter 18
Relative gain array
Provides two useful types of information:
1) Measure of process interactions
2) Recommendation about best pairing of controlled and
Chapter 18
manipulated variables.
Requires knowledge of steady state gains but not process
dynamics.
Example of RGA Analysis: 2 x 2 system
Steady-state process model,
Y1 K11U1 K12U 2
Y2 K 21U1 K 22U 2
Chapter 18
The RGA is defined as:
11 12
RGA
21 22
where the relative gain, ij, relates the ith controlled
variable and the jth manipulated variable
open - loop gain
ij
closed - loop gain
Scaling Properties:
i) ij is dimensionless
ii) ij ij 1.0
i j
Chapter 18
For 2 x 2 system,
1
11 , 12 1 11 21
K K
1 12 21
K11K 22
Recommended Controller Pairing
Corresponds to the ij which has the largest positive value.
In general:
1. Pairings which correspond to negative pairings should not be
selected.
2. Otherwise, choose the pairing which has ij closest To one.
Examples:
Process Gain Relative Gain
Matrix, : Array, :
Chapter 18
K11 0 1 0
0
K 22
0 1
0 K12
0 1
K
21 0 1 0
K11
0
K12
K 22
1 0
0 1
K11 0
K
21 K 22 1 0
0 1
Recall, for 2X2 systems...
1
Y1 K11U 1 K12U 2 11 , 12 1 11 21
K K
1 12 21
Y2 K 21U1 K 22U 2 K11K 22
EXAMPLE:
K K12 2 1.5
K 11
K 21 K 22 1.5 2
Chapter 18
2.29 1.29
Recommended pairing is Y1
and U1,Y2 and U2.
1.29 2.29
EXAMPLE:
2 1.5 0.64 0.36
K 0.36 0.64
1 . 5 2
Recommended pairing isY with U ,Y with U .
1 1 2 2
EXAMPLE:Thermal Mixing System
The RGA can be expressed in terms of the manipulated variables:
Chapter 18
Wh Wc
W Wc Wh
W W Wc Wh
c h
Wh Wc
T W W Wc Wh
c h
Note that each relative gain is between 0 and 1. Recommended
controller pairing depends on nominal values of W,T, Th, and Tc.
See Exercise 18.16
EXAMPLE: Ill-conditioned Gain Matrix
y = Ku
2 x 2 process y 1 = 5 u1 + 8 u2
y2 = 10 u1 + 15.8 u2
Chapter 18
specify operating point y, solve for u
Adj K
-1
u=K y = y
det K
K11K 22 K11K 22
RGA : 11
K11K 22 K12K 21 det K
effect of det K 0 ?
RGA for Higher-Order Systems:
For and n x n system,
U1 U2 Un
Y1 11 12 1n
Y2 21 22 2 n
Chapter 18
Yn n1 n1 nn
Each ij can be calculated from the relation
ij K ijH ij
Where Kij is the (i,j) -element in the steady-state gain matrix, K:
Y KU
And Hij is the (i,j) -element of the .
H K 1 T
Note that,
KH
EXAMPLE 18.4 : Hydrocracker
The RGA for a hydrocracker has been reported as,
U1 U2 U3 U4
Chapter 18
Y1 0.931 0.150 0.080 0.164
Y2 0.011 0.429 0.286 1.154
Y3 0.135 3.314 0.270 1.910
Y4 0.215 2.030 0.900 1.919
Recommended controller pairing?
Singular Value Analysis
K = W S VT
S is diagonal matrix of singular values
(s1, s2, , sr)
The singular values are the positive square roots of the
eigenvalues of
KTK (r = rank of KTK)
W,V are input and output singular vectors Columns of W and
V are orthonormal. Also
WWT = I
VVT = I
Calculate S, W, V using MATLAB (svd = singular value
decomposition)
Condition number (CN) is the ratio of the largest to the
smallest singular value and indicates if K is ill-conditioned.
CN is a measure of sensitivity of the matrix properties to changes in
a specific element.
Consider 1 0
K
10 1
(RGA) = 1.0
If K12 changes from 0 to 0.1, then K becomes a singular matrix,
which corresponds to a process that is hard to control.
Chapter 18
RGA and SVA used together can indicate whether a process is easy
(or hard) to control.
10.1 0
(K ) = CN = 101
0 0.1
K is poorly conditioned when CN is a large number (e.g., > 10).
Hence small changes in the model for this process can make it very
difficult to control.
Selection of Inputs and Outputs
Arrange the singular values in order of largest
to smallest and look for any i/i-1 > 10; then
one or more inputs (or outputs) can be deleted.
Chapter 18
Delete one row and one column of K at a time
and evaluate the properties of the reduced gain
matrix.
Example 18.7:
0 .4 8 0 .9 0 0 .0 0 6
K 0 .5 2 0 .9 5 0 .0 0 8
0 .9 0 0 .9 5 0 .0 2 0
0.5714 0.3766 0.7292
W 0.6035 0.4093 0.6843
0.5561 0.8311 0.0066
1.618 0 0
0 1.143 0
0 0 0.0097
0.0541 0.9984 0.0151
V 0.9985 0.0540 0.0068
18
Chapter18
0.0060 0.0154 0.9999
Chapter
CN = 166.5 (1/3)
The RGA is as follows:
2.4376 3.0241 0.4135
1.2211 0.7617 0.5407
2.2165 1.2623 0.0458
Preliminary pairing: y1-u2, y2-u3,y3-u1.
CN suggests only two output variables can be controlled. Eliminate one input and one output
(3x32x2).
Chapter 18
Chapter 18
Alternative Strategies for Dealing with Undesirable
Control Loop Interactions
1. "Detune" one or more FB controllers.
2. Select different manipulated or controlled variables.
e.g., nonlinear functions of original variables
3. Use a decoupling control scheme.
4. Use some other type of multivariable control scheme.
Chapter 18
Decoupling Control Systems
Basic Idea: Use additional controllers to compensate for process
interactions and thus reduce control loop interactions
Ideally, decoupling control allows setpoint changes to affect only the
desired controlled variables.
Typically, decoupling controllers are designed using a simple process
model (e.g. steady state model or transfer function model)
Chapter 18
Design Equations:
We want cross-controller, GC12, to cancel out the effect of U2 on Y1.
Thus, we would like,
T12GP11U 2 GP12U 2 0
Since U2 0 (in general), then
Chapter 18
GP12
T12
GP11
Similarly, we want G21 to cancel the effect of M1 on C2. Thus, we
require that...
T21GP 22U1 GP 21U1 0
GP 21
T21
GP 22
cf. with design equations for FF control based on block diagram
analysis
Alternatives to Complete Decoupling
Static Decoupling (use SS gains)
Partial Decoupling (either GC12 or GC21 is set equal to zero)
Chapter 18
Process Interaction
Corrective Action (via cross-controller or decoupler).
Ideal Decouplers:
GP12 ( s )
T12 ( s )
GP11 ( s )
GP 21 ( s )
T21 ( s )
GP 22 ( s )
Variations on a Theme:
Partial Decoupling:
Use only one cross-controller.
Static Decoupling:
Design to eliminate SS interactions
Chapter 18
Ideal decouplers are merely gains:
K P12
T12
K P11
K P 21
T21
K P 22
Nonlinear Decoupling
Appropriate for nonlinear processes.
