Closed-loop transfer functions

Closed-Loop Transfer Functions
1.
Introduction
2.
Stirred tank heating system
3.
Closed-loop block diagrams
4.
Closed-loop transfer functions
5.
Simulink example
Introduction

Block diagrams
» Convenient tool to represent closed-loop systems
» Also used to represent control systems in Simulink

Closed-loop transfer functions
» Transfer function between any two signals in a
closed-loop system
» Usually involve setpoint or disturbance as the
closed-loop input and the controlled output as the
closed-loop output
» Conveniently derived from block diagram
» Can be derived automatically in Simulink
» Used to analyze closed-loop stability and compute
closed-loop responses
Stirred Tank Blending System

Control objective
» Drive outlet composition (x) to
setpoint (xsp) by manipulating pure
stream flow rate (w2) despite
disturbances in flow rate (w1) and
composition (x1) of other feed
stream

Control system
» Measure x with composition
analyzer (AT)
» Perform calculation with
composition controller (AC)
» Convert controller output to
pneumatic signal with currentpressure converter (I/P) to drive
valve
Blending Process Model

Mass balances for constant volume
0  w1  w2  w

d ( Vx )
 w1 x1  w2 x2  wx 
dt

w  w1  w2
dx w1 ( x1  x)  w2 ( x2  x)

 f ( x, x1 , w2 )
dt
V
Linearized model
dx'  w x '  w1 x1'  (1  x ) w2'

dt
V

Transfer function model
w1 w
(1  x ) w '
K1
K2
'
'
X ( s) 
X 1 ( s) 
W2 ( s) 
X 1 ( s) 
W2' ( s)
V
V
s  1
s  1
s 1
s 1
w
w
'
Control System Components

Composition analyzer – assume first-order
dynamics

Controller – assume PI controller

I/P converter – assume negligible dynamics
Control System Components cont.


Control valve – assume first-order
dynamics
Entire blending system
Closed-Loop Block Diagrams






Gp(s) – process transfer
function
Gd(s) – disturbance transfer
function
Gv(s) – valve transfer function
Gc(s) – controller transfer
function
Gm(s) – measurement transfer
function
Km – measurement gain







Y(s) – controlled output
U(s) – manipulated input
D(s) – disturbance input
P(s) – controller output
E(s) – error signal
Ysp(s) – setpoint
Ym(s) – measurement
Transfer Function for Setpoint Changes
Y  Yu  Yd  Yu  G pU  G p Gv P  G p Gv Gc E
~
E  Ysp  Ym  K mYsp  GmY
Y  G p Gv Gc E  G p Gv Gc K mYsp  GmY 
K mGc Gv G p
Y

Ysp 1  Gc Gv G p Gm
Transfer Function for Disturbance Changes
Y  Yu  Yd  G pU  Gd D  G p Gv P  Gd D  G p Gv Gc E  Gd D
~
E  Ysp  Ym  Ym  GmY
Y  G p Gv Gc E  G p Gv Gc  GmY   Gd D
Gd
Y

D 1  Gc Gv G p Gm
Simultaneous Changes

Principle of superposition
K mGcGvG p
Gd
Y
Ysp 
D
1  GcGvG pGm
1  GcGvG pGm

Open-loop transfer function
» Obtained by multiplying all transfer functions
in feedback loop
GOL  Gc Gv G p Gm
Y
K mGc Gv G p
1  GOL
Gd
Ysp 
D
1  GOL
General Method

Closed-loop transfer function
Pf
Z

Zi 1  P e
»
»
»
»

Z = any variable in feedback system
Zi = any input variable in feedback system Z and Zi
Pf = product of all transfer functions between Z and Zi
Pe = product of all transfer functions in feedback loop
Setpoint change
P f  K mGcGvG p

P e  GcGvG p Gm  GOL
Disturbance change
P f  Gd
P e  GcGvG p Gm  GOL
Closed-Loop Transfer Function Example
Simulink Example
0.05
Inlet flow
0
6.37
50
PID
Setpoint
Km
Add
PID Controller
0.12
0.0103
Kip
Kv
y
5s+1
Add 1
Tank
Level
50
Km1
>> gp=tf([6.37],[5 1]);
>> gcl=gp/(1+gc*kv*gp*km)
>> kv=0.0103;
>> kip=0.12;
>> km=50;
>> gc=tf([2.5 5],[0.5 0]);
Disturbance transfer function:
15.93 s^2 + 3.185 s
----------------------------------12.5 s^3 + 46.01 s^2 + 90.72 s + 16.4