Lecture 6 - The Basic Feedback Loop 1.

Introduction
K. J. Åström • A nice collection of tools have been developed
• We have looked at a few examples
1. Introduction
• We will now investigate a typical control problem
2. Controllers with two degrees of freedom
• A basis for control system design
3. The Gangs of Four and Six
• How to judge a control system
4. The sensitivity functions
• New concepts and insight
5. Summary
– Sensitivity functions
Theme: Understanding the basic feedback loop. Systems with – Deeper understanding of feedback
two degrees of freedom. The gangs of four and six. Sensitivity
• A basis for a serious look at the design problem
functions
• How to capture a complex reality in tractable mathematics

A Basic Control System Key Issues

d n Find a controller that
r e u x y A: Reduces effects of load disturbances
F Σ C Σ P Σ
B: Does not inject to much measurement noise into the
system
C: Makes the closed loop insensitive to variations in the
−1
process
D: Makes output follow command signals
Ingredients:
Convenient to use a controller with two degrees of freedom,
• Controller: feedback C, feedforward F
i.e. separate signal transmission from y to u and from r to u.
• Load disturbance d: Drives the system from desired state This gives a complete separation of the problem: Use feedback
• Measurement noise n: Corrupts information about x to deal with A, B, and C. Use feedforward to deal with D!
• Process variable x should follow reference r
&
c K. J. Åström August, 2001 1
 2. System with Two Degrees of Freedom Designing System with Two Degrees of Freedom
d n Design procedure:
r e u x y • Design the feedback C to achieve
F Σ C Σ P Σ
– Small sensitivity to load disturbances d
– Low injection of measurement noise n
– High robustness to process variations
−1
• Then design the feedforward F to achieve desired re-
sponse to command signals r
The controller has two degrees of freedom (2DOF) because
the transfer function from reference r to control u is different For many problems in process control the load disturbance re-
from the transfer function from y to u. sponse is much more important than the set point response.
The set point response is more important in motion con-
We have already encountered this in PI control
trol. Few textbooks and papers show more than set point re-
Z t
sponses.
u(t) = k(br(t) − y(t)) + (r(τ ) − y(τ ))dτ
0

Many Versions of 2DOF 3. Relations between signals

r u y d n
F Σ C P
r e u x y
F Σ C Σ P Σ
−1

Mu um
r −1
ym u y
My Σ C Σ P
P PC PCF
X = D− N+ R
−1 1 + PC 1 + PC 1 + PC
P 1 PCF
Y= D+ N+ R
For linear systems all 2DOF configurations have the same 1 + PC 1 + PC 1 + PC
PC C CF
properties. For the systems above we have U =− D− N+ R
1 + PC 1 + PC 1 + PC
C F = Mu + CM y
&
c K. J. Åström August, 2001 2
 Some Observations A Possible Choice
• A system based on error feedback is characterized by four Six transfer functions are required to show the properties of a
transfer functions (The Gang of Four) basic feedback loop. Four characterize the response to load
• The system with a controller having two degrees of disturbances and measurement noise.
freedom is characterized by six transfer function (The PC P
Gang of Six) 1 + PC 1 + PC
C 1
• To fully understand a system it is necessary to look at all 1 + PC 1 + PC
transfer functions
Two more are required to describe the response to set point
• It may be strongly misleading to only show properties of changes.
a few systems for example the response of the output to PC F CF
command signals. This is a common error in the literature. 1 + PC 1 + PC
• The properties of the different transfer functions can be
illustrated by their transient or frequency responses.

Amplitude Curves of Frequency Responses Step Responses

PI control k = 0.775, Ti = 2.05 of P(s) = (s + 1)−4 with PI control k = 0.775, Ti = 2.05 of P(s) = (s + 1)−4 with
M (s) = (0.5s + 1)−4 M (s) = (0.5s + 1)−4

PC F /(1 + PC) PC/(1 + PC) P/(1 + PC) PC F /(1 + PC) PC/(1 + PC) P/(1 + PC)
1.5 1.5 1.5

0 0 0
10 10 10 1 1 1

0.5 0.5 0.5

0 0 0
−1 −1 −1
10 10 10 0 10 20 30 0 10 20 30 0 10 20 30
C F /(1 + PC) C/(1 + PC) 1/(1 + PC)
−1 0 1 −1 0 1 −1 0 1
10
C F /(110+ PC)10 10
C/(1 10+ PC) 10 1
10
1/(1 + PC)
10 10
1.5 1.5 1.5
10
1 1
10 10
1 1 1

0.5 0.5 0.5

0
0
10
0
10 10 0 0 0

−1 0 1 −1 0 1 −1 0 1 0 10 20 30 0 10 20 30 0 10 20 30
10 10 10 10 10 10 10 10 10

&
c K. J. Åström August, 2001 3
 An Alternative A Warning!
Show the responses in the output and the control signal to
a step change in the reference signal for system with pure Please remember to always look at all responses
error feedback and with feedforward. Keep the reference signal when you are dealing with control systems. The step
constant and make a unit step in the process input. Show the response below looks fine but ...
response of the output and the control signal.
Response of y to step in r
1.5 1
1

0.5

0
0 10 20 30 40 50 60
0.5
2

1.5

0.5

0 0
−0.5
0 10 20 30 40 50 60
0 1 2 3 4 5

Four Responses The System

1
Response of y to step in r Response of y to step in d Process P(s) =
1 80 s−1
60 s−1
Controller C(s) =
0.5 40 s
20 Response of y to reference r
0 0 Y ( s) PC 1
0 1 2
Response of u to step in r
3 4 5 0 1 2 3
Response of u to step in d
4 5
= =
1 0 R( s ) 1 + PC s+1
0.5 Response of y to step in disturbance d
0 −0.5
Y ( s) P s s
= = 2 =
−0.5
D ( s) 1 + PC s −1 (s + 1)(s − 1)
−1 −1
0 1 2 3 4 5 0 1 2 3 4 5

What is going on?

&
c K. J. Åström August, 2001 4
 Focus on Feedback The Feedback Problem
d n d n

r=0 e u x y r=0 e u x y
Σ C Σ P Σ Σ C Σ P Σ

−1 −1

The signals have the following relations. Notice that there are
• Neglect following of reference signals (the feedforward
only four transfer functions - The Gang of Four.
problem).
P PC
• Focus on on the feedback problem X = D− N
1 + PC 1 + PC
– Load disturbances P 1
Y= D+ N
– Measurement noise 1 + PC 1 + PC
PC C
– Model uncertainty U =− D− N
1 + PC 1 + PC

The Loop Transfer Function L(s) = P(s) C (s) 4. The Sensitivity Functions
Tells a lot about the system. The transfer functions
1 1
• Sensitivity function S =
1

=
10

1 + PC 1+ L
0
10
PC L
• Complementary sensitivity function T = =
10
−1 1 + PC 1+ L
are called sensitivity functions. They have interesting properties
−2
10
−1
10
0
10 and useful physical interpretations. We have
0 • The functions S and T only depend on the loop transfer
−100
function L

−200
• S+T =1
−300
• Typically S (0) small and S (∞) = 1 and consequently
T (0) = 1 and T (∞) small
−400
−1 0
10 10

&
c K. J. Åström August, 2001 5
 Quiz Disturbance Reduction
d n
Look at the block diagram
d n r=0 e u x y
Σ C Σ P Σ
r e u x y
F Σ C Σ P Σ
−1

−1 Output without control Y = Yol (s) = N (s) + P(s) D (s)

Output with feedback control
Find all relations where the signal transmissions are equal to 1
1
either the sensitivity function or the complementary sensitivity Ycl = N + PL = Yol = SYol = SYol
1 + PC 1 + PC
function
Disturbances with frequencies such that h S (iω )h < 1 are
The Audience is Thinking ... reduced by feedback, disturbances with frequencies such that
h S (iω )h > 1 are amplified by feedback.

Assessment of Disturbance Reduction Assessment of Disturbance Reduction

We have
Ycl(s) 1
= S ( s) = Ycl 1
Ycl(s) 1 + P ( s) C ( s) = =S
Feedback attenuates disturbances of frequencies ω such that Yol 1 + PC
h S (iω )h < 1. It amplifies disturbances of frequencies such that Geometric interpretation:
h S (iω )h > 1 Disturbances with frequen-
cies inside the circle are
1 −1
10
amplified by feedback. Dis-
0
turbances with frequencies
10
outside are reduced.
Disturbances with frequen-
cies less than ω s are re- ωs
−1
10

−2
duced by feedback.
10
−2 −1 0 1
10 10 10 10

&
c K. J. Åström August, 2001 6
 Properties of the Sensitivity function The Water Bed Effect
• Can the sensitivity be small for all frequencies? 1

– No we have S (∞) = 1!

log h S (iω )h
0
• Can we get h S (iω )h ≤ 1?
– If the Nyquist curve of L = PC is in the first and third −1

quadrant! Passive systems! −2

• Bodes integral, pk RHP poles of L(s)

Z ∞ Z ∞
−3
0 0.5 1 1.5 2 2.5 3

log h S( iω )hdω = log

1
dω ω
h1 + L( iω )h Z
0 0
X π
∞ X π
=π Re pk − lim sL( s) log h S (iω )hdω = π
lim sL(s) Re pk −
2 s→∞ 0 2 s→∞
The sensitivity can be decreased at one frequency at the cost
• The "water-bed effect". Push the curve down at one
of increase at another frequency.
frequency and it pops up at another!

Robustness Maximum Sensitivity

Effect of small process changes on T = PC/(1 + PC)
dT dP CdP 1 dP dP
= − = =S The number
T P 1 + PC 1 + PC P P
Ms = max h S (iω )h −1
−1
How much can the process is a measure of robustness. The Ms ω ms
be changes without making number 1/ Ms is the smallest dis-
the system unstable? tance from the Nyquist curve to
1+ L
the critical point -1. ωs
h C ∆ Ph < h1 + PCh
or C∆ P
h∆ Ph 1
< Reasonable values are between 1.2 and 2.
h Ph hT h

&
c K. J. Åström August, 2001 7
 Maximum Sensitivities Summary of the Sensitivity Functions
Requirement on maximum sensitivities give constraints that tell 1 L
S= , T= , Ms = max h S(iω )h, Mt = max hT (iω )h
that the Nyquist curve should avoid certain circles 1+ L 1+ L
The value 1/ Ms is the shortest distance from the Nyquist curve of
Ms = Mt = 2 Ms = Mt = 1.4 the loop transfer function L(iω ) to the critical point −1.
V log T Ycl (s)
S= =
V log P Yol (s)
How much can the process be changed without making the system
unstable?
h∆ Ph 1
<
h Ph hT h
Bode’s integral the water bed effect.
Z ∞ X π
log h S(iω )hdω = π Re pk − lim sL(s)
0 2 s→∞

5. Summary The Gangs of Four and Six

• Systems with two degrees of freedom allow a complete Response of y to load disturbance d is characterized by
separation of responses to reference signals and distur- P
bances. 1 + PC
• Design for disturbances and robustness first, then choose Response of u to measurement noise n is characterized by
feedforward F to give desired response to reference C
signals
1 + PC
• A system with error feedback is characterized by four Robustness to process variations is characterized by
transfer functions (Gang of Four)
1 PC
• The basic feedback loop with two degrees of freedom is S= , T=
1 + PC 1 + PC
characterized by six transfer functions (Gang of Six)
Responses of y and u to reference signal r is characterized by
• The effect of feedback on disturbances is given by
PC F CF
,
Yclosedloop(s) = S (s) Yopenloop(s) 1 + PC 1 + PC

&
c K. J. Åström August, 2001 8