Vous êtes sur la page 1sur 27

1

E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Enhanced Single-Loop Control Strategies
(Advanced Control)
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Cascade Control
Time-Delay Compensation
Inferential Control
Selective and Override Control
2
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Cascade Control
A disadvantage of conventional feedback control is that
corrective action for disturbance does not begin until
after the controlled variables deviates from the set-point.
Disturbance:
Oil flow rate
Cold oil temperature
satisfactory response
Disturbance:
Fuel gas supply pressure
(Fuel gas flow)
sluggish response
Example: Furnace temperature control
Conventional feedback control
3
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Feedforward control: disturbances have to be measured and a
model is required to calculate the controller output
Cascade control: use a secondary measurement and a secondary
feedback controller
Another Example: Temperature control of tanks in series
How to improve control performance ?
Disturbance
4
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Cascade control: a primary control loop (TT1 and TC1) and
a secondary control loop (TT2 and TC2)
Set-point
5
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Cascade Control of Exothermic Chemical Reactor
6
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Cascade Control
Distinguishing features:
1. Two FB controllers but only a single control valve
(two controlled variables, two sensors, and one
manipulated variable).
2. Output signal of the "master" controller is the set-
point for slave" controller.
3. Two FB control loops are "nested" with the "slave"
(or "secondary") control loop inside the "master" (or
"primary") control loop.
Terminology:
slave vs. master
secondary vs. primary
inner vs. outer
7
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Primary (outer)
control loop
Secondary (inner)
Control loop
Master
controller
Slave
controller
A Furnace Cascade Temperature Control
8
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Block Diagram of Cascade Control System
1
2
1
2
= hot oil temperature
= fuel gas pressure
= cold oil temperature
(or cold oil flow rate)
= supply pressure of gas fuel
Y
Y
D
D
1 1
2 2
1 1
2
2
= measured value of
= measured value of
= set point for
= set point for
m
m
sp
sp
Y Y
Y Y
Y Y
Y Y
(Furnace Example)
1 : primary loop
2 : secondary loop
9
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Design Considerations for Cascade Control
For a cascade control system to function properly, the
secondary control loop must response faster than the
primary loop.
The secondary controller is normally a P or PI controller.
The primary controller is usually PI or PID.
First, design G
c2
for inner loop.
(inside-out)
2 2
2
2 2 2 2
1
=
+
c v p
sp c v p m
G G G
Y
Y G G G G
10
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Then, design G
c1
for outer loop.
1 2
1
2 2 2 2 1 2 2 1 1
(16 5)
1
=
+ +
p d
c v p m c c v p p m
G G
D G G G G G G G G G G
Y
( )
1 2 2 2
1
1 2 2 2 1 2 2 1 1
1
(16 8)
1
d c v p m
c v p m c c v p p m
G G G G G
D G G G G G G G G G G
Y
+
=
+ +
11
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Example
Consider the block diagram with the following transfer
functions:
( )( )
1 2
1 2 2 1
5 4
1
1 4 1 2 1
1
1 0.05 0.2
3 1
v p p
m m d d
G G G
s s s
G G G G
s
= = =
+ + +
= = = =
+
Great improvement
D
1
step response
Primary loop: PI
Secondary loop: P
Slight improvement
D
2
step response
12
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
The presence of time delay in a process limits the
performance of a conventional feedback control system.
A time delay will adversely affects closed-loop stability.
Time Delay Compensation (Smith Predictor)
Example 16.2
Compare the set-point responses for a second-order process with a time
delay and without the delay. The transfer function is
Assume . Use the following PI controllers.
For , .
For
(the controller gain must be reduced to meet stability requirements)
( ) ( )
( )
5 1 3 1

=
+ +
s
p
e
G s
s s
1
m v
G G = =
0, = 3.02 =
c
K
6.5 =
I
2, = , 7. 1.23 0. = =
I c
K
13
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Closed-loop set-point responses
The response of the closed-loop system will be sluggish compared to
that of the control loop with no time delay. (not only 2 min.)
14
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
For this ideal case the controller responds to the error signal that would
occur if no time delay were present.
Model
If the process model is perfect and the disturbance is zero, then
( )
( )
1 1 2
16 19
sp
E E Y Y Y Y Y = =

'
2
Y Y =

( )
1
16 20
sp
E' Y Y =

Block Diagram of Smith Predictor


* - s
G=G e
Without time delay
15
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s

=
* - s
G= G G e

Assuming there is no model error the inner loop has the


effective transfer function
( )
( )
16 21
1 * 1

= =
+
c
c
s
c
G P
G
E
G G e
'
( )
, =

G G
Equivalent Structure for Smith Predictor
c
G '
16
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
For no model error:
By contrast, for conventional feedback control
1 1 1

'
= = =
'
+ + +
* s * s
c c c
* s * *
sp
c c c
G G e G G e G G Y
Y
G G e G G G G
( )
*
*
16 23
1
1

= =
+
+
s
c c
s
sp c
c
G G G G e Y
Y G G
G G e
The Smith predictor has the advantage of eliminating the
time delay from the characteristic equation. Unfortunately,
this advantage is lost if the process model is inaccurate.
(It can still provide improvement if the model parameters
are within about of the actual values) 30%
17
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Closed-loop set-point responses
Smith predictor (SP) has the same controller settings with conventional
feedback PI controller. ( )
The responses are identical, except for the initial time delay.
SP ( ) 2 =
3.02 =
c
K
6.5 =
I
PI ( ) 0 =
18
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
How about the disturbance response?
The Smith predictor generally is beneficial for handling
disturbance. However, under certain conditions, a
conventional PI controller can provide better regulatory
control than SP.
( )
1 1
1
d
* s
c
*
c
G G G e
Y
D
G G

(
+

=
+
2 =
19
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Inferential Control
Problem: Controlled variable cannot be measured or has
large sampling period.
Possible solutions:
1. Control a related variable (e.g., temperature instead
of composition).
2. Inferential control: Control is based on an estimate
of the controlled variable.
The estimate is based on available measurements.
Examples: empirical relation, neural network
Modern term: soft sensor
20
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Inferential Control with Fast and Slow
Measured Variables
Soft sensor block diagram
( Note: X and/or Y can be used for control )
21
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Selective Control & Override Control
For every controlled variable, it is very desirable that
there be at least one manipulated variable.
But for some applications,
N
C
> N
M
where:
N
C
= number of controlled variables
N
M
= number of manipulated variables
Solution: Use a selective control system or an override.
( )
C M
N N =
22
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Low selector:
High selector:
Median selector:
The output, Z, is the median of an odd number of inputs
Selectors are used to improve the control performance as well as
to protect equipment from unsafe operating conditions.
Selectors
(useful when redundant sensors are used to measure a single process variable)
23
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
multiple measurements
one controller
one final control element
Example: High Selector Control System
Control of a reactor hotspot temperature
Determine the hotspot temperature
24
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Overrides
An override is a special case of a selective control system
One of the inputs is a numerical value, a limit.
Used when it is desirable to limit the value of a signal (e.g.,
a controller output).
Temperature control of a heater
for safety
25
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
2 measurements, 2 controllers,
1 final control element
A selective control system to handle a sand/water slurry
(regulate the level and exit flow rate)
Keep flow rate above
its minimum value
FC
26
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Block Diagram for the Selective Control Loop
faster
27
E
n
h
a
n
c
e
d

S
i
n
g
l
e
-
L
o
o
p

C
o
n
t
r
o
l

S
t
r
a
t
e
g
i
e
s
Split-Range Control
Multiple final control elements or multiple controllers
Reactor temperature control (both heating and cooling are used)