Documente Academic
Documente Profesional
Documente Cultură
Material
! Lecture slides
Course Outline
Lecture 14 Summary
Last lecture: Stable periodic solution
0.2
0.15 G (i )
r e u y 0.1
G (s) 0.05
0
0.05 1/N(A)
0.1
0.15
0.2
5 4 3 2 1 0
(s + 10)2
G (s) = with feedback u = sgn y
(s + 1)3
gives one stable and one unstable limit cycle. The left most
intersection corresponds to the stable one.
Periodic Solutions in Relay System
The relay gain N(A) is higher for small A:
Stable Unstable
periodic periodic
orbit orbit
Todays Goal
v u
G (s)
C (s)
Output
1
0
0 20 40 60 80
0.1
Input
0.1
0 20 40 60 80
Time
Actuator
e v u
K
K 1 +
Ti s
es
1
Tt
(b)
y
KTd s
K 1 +
Ti s
es
1
Tt
Choice of Tracking Time Tt
yset-point Tt = 3
Tt = 2
1
Tt = 0.1, Tt = 1
0
0 10 20 30
0.1 Tt = 3
Tt = 0.1 Tt = 2
0 Tt = 1
0.1
0 10 20 30
With very small Tt (large gain 1/Tt ), spurious errors can saturate
the output, which leads to accidental reset of the integrator. Too
large Tt gives too slow reaction (little effect).
The tracking time Tt is the design parameter
of the anti-windup.
Common choices: Tt = Ti or Tt = Ti Td .
State feedback with Observer
xm
x = Ax + B sat(v ) + K (y C x)
v = L(xm x)
State-space controller:
Idea:
Rewrite representation of control law from (a) to (b) such that:
(a) and (b) have same input-output relation
(b) behaves better when feedback loop is broken, if SB stable
Antiwindup General State-Space Controller
xc = Fxc + Gy + K (u Cxc Dy )
! "# $
=0
= (F KC )xc + (G KD)y + Ku
= F0 xc + G0 y + Ku
xc = F0 xc + G0 y + Ku
u = sat (Cxc + Dy )
State-space controller without and with anti-windup:
y xc u
G s 1 C
F KC
y xc v u
G KD s 1 C sat
K
5 Minute Exercise
yref u y
1
s Li Process
x%
L Estimator
Saturation
k(y x)
F
x y
2
0
0 10 20
Time
0.4
0.2
0
0 10 20
Time
0
0 10 20
Time
Position Control of Servo with Friction Hunting
F
Friction
xr
u 1 x
PID 1
ms v s
0.5
0
0 20 40 60 80 100
Time
0.2
0.2
0 20 40 60 80 100
Time
1
1
0 20 40 60 80 100
Time
3 Minute Exercise
What are the signals in the previous plots? What model of friction
has been used in the simulation?
Friction
N N
F F
Sticking
Boundary lubrication
Friction
Lubrication Regimes
Mixed lubrication
200
150
100
50
Friction [Nm]
-50
-100
-150
-200
Friction
Velocity
Classical Friction Models
a) b)
F F
v v
c) d)
F F
v v
&
Fc sign v (t) + Fv v (t) v (t) #= 0
c) F (t) =
max(min(Fe (t), Fs ), Fs ) v (t) = 0
Fe (t) = external applied force , Fc , Fv , Fs constants
Advanced Friction Models
! Karnopp model
! Armstrongs seven parameter model
! Dahl model
! Bristle model
! Reset integrator model
! Bliman and Sorine
! Wit-Olsson-strm
Demands on a model
! Lubrication
! Integral action (beware!)
! Dither
! Non-model based control
! Model based friction compensation
! Adaptive friction compensation
Integral Action
e(t)
e(t)
F
Friction
v x
xr
1 1
+ PID +
u
+
ms s
Ffriction
Friction
uPID 1 v 1 x
+
ms s
+
Friction
estimator
F%friction
Assumption: v measurable.
Friction estimator:
z = kuPID sgn(v )
a% = z km|v |
%
Ffriction = a% sgn(v )
Result: e = a a% 0 as t ,
since
de d a dz d
= = + km |v |
dt dt dt dt
= kuPID sgn(v ) + kmv sgn(v )
= k sgn(v )(uPID mv)
= k sgn(v )(F F )
= k(a a)
= ke
d
Remark: Careful with dt |v | at v = 0.
The Knocker
Combines Coulomb compensation and square wave dither
2.5
1.5
0.5
0.5
0 1 2 3 4 5 6 7 8 9 10
1
P-controller 1
PI-controller
0.5 vref 0.5
0 v 0
0.5 0.5
1 1
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
10 10
5 u 5
0 0
5 5
10 10
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
P-controller
1
with Coulomb estimation
0.5
0.5
1
0 1 2 3 4 5 6 7 8
10
10
0 1 2 3 4 5 6 7 8
Next Lecture
! Backlash
! Quantization