Professional Documents
Culture Documents
Describing Function Analysis
Describing Function Analysis
nonlinear systems
Prof. Elio USAI eusai@diee.unica.it
University of Leicester
Department of Engineering
Plant P
process
e(t)
C
controller
u(t)
A
actuator
m(t) +
+
y(t)
z(t)
F
filter
ym(t)
+
T
trasmitter
n(t)
Controller
r: reference variable e: error signal u: control variable m: manipulated input y: controlled variable ym: measure of the output z: feedback signal d: disturbance n: measurement noise
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
+M
-M
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
Saturated actuators Relay control Gears backlash Hysteresis in magnetic materials Dead zone in electromechanical systems
-E
+U
e +E
-U
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
Saturated actuators Relay control Gears backlash Hysteresis in magnetic materials Dead zone in electromechanical systems
-U +U u (driving gear)
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
Saturated actuators Relay control Gears backlash Hysteresis in magnetic materials Dead zone in electromechanical systems
-M0 +M0 u
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
Saturated actuators Relay control Gears backlash Hysteresis in magnetic materials Dead zone in electromechanical systems
-U +U u
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
r(t)
+ _
e(t)
u(t)
m(t)
y(t)
z(t)
P(s ) =
1 s 1 + 1 .4 s + s 2
)
A(s ) = N 1 O sat ( 2 )
0.5 H (s ) = (1 + 0.1s )
C (s ) = 10
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
10
System output
-5
-10
-15
-20
10
12
14
16
18
20
time [s]
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
Magnitude (dB)
-50
-100
-150 -90 System: sys Frequency (rad/sec): 0.941 Phase (deg): -180
The NOT saturated system is not stable since its stability margin are negatives mg = -12 dB m = - 47 deg
Phase (deg)
-180
-270
-360 10
-1
10
10
10
Frequency (rad/sec)
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
z(t)
H(j)
r(t)
C(j)
r(t) + _
u(t)
f(u)
m(t)
w(t)
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
u (t ) = u0 + u (t )
r(t)=kCR + _ u(t) f(u) m(t) G (j) w(t)
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
m0
kC R
u
u0
1 kG
m0 = f (u0 ) u0 = kC R kG m0
m0 = f (u ) f (u ) = f (u 0 + u ) m0
r(t)=0
+ _
u(t)
f(u)
m(t)
G (j)
w(t)
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
If f(u) is a passive sector function absolute stability tools allow for sufficient conditions for global asymptotic stability of the variation system, i.e, the steady state is characterised by constant values of the system variables
If the origin of the variation system is not stable does the variables diverge to infinity or some periodic motion can appear?
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
Limit cycles
Limit cycle: a periodic oscillation around a constant working point
r=51() + _ u(t) m(t) w(t)
f(u)
G (j)
G (s ) =
10 (s + 1) s 2 + 1.4s + 1
6 f (u ) = sat (1, 11 , 16 ) 11
2
m
1.5
m=
1 1 u+ 10 2
0.5 m0
0 u0
5 11 5 m0 = 11 u0 =
-0.5
-1 -6
-4
-2
u
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
Limit cycles
Limit cycle: a periodic oscillation around a constant working point
r(t)=0 + _ u(t) f(u) m(t) G (j) w(t)
output w(t)
w
4
5 11 5 m0 = 11 u0 =
w0 =
0 5 10 15 20 25 30
50 = 4.545 11
time [s]
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
Limit cycles
Limit cycle can be a drawback in control systems: Instability of the equilibrium point Wear and failure in mechanical systems Loss of accuracy in regulation Parameters of the limit cicle can be used to discriminate between acceptable and dangerous oscillations oscillation frequency oscillation magnitude
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
It is based on the following assumptions There is only one single nonlinear component The nonlinear component is not dynamical and time invariant The linear component has low-pass filter properties The nonlinear characteristic is symmetric with respect to the origin
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
The system can be represented by a lumped parameters system with two main blocks: The linear part The nonlinear part
r=const.
+ _
u(t)
N.L.
m(t)
G (j)
w(t)
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
m=M
u u +
-M
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
This is the main assumption that allows for neglecting the higher frequency harmonics that can appear when a nonlinear system is driven by a harmonic signal
G ( j ) >> G ( jn ) n = 2, 3, K
The more the low-pass filter assumption is verified the more the estimation error affecting the limit cycle parameters is small
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
This guarantees that the static term in the Fourier expansion of the output of the nonlinearity, subjected to an harmonic signal, can be neglected
r(t)=0 + _ u(t) m(t) w(t)
f(u)
G (j)
Such an assumption is usually taken for the sake of simplicity, and it can be relaxed
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
y (t ) = f (t ),
y (t ) = y (t T ), T is a real constant
a0 y (t ) = + (ak sin (k 2 k =1
2 2 a k = f (t ) cos(k T T 2 2 2 bk = f (t )sin (k T T 2
T T
2 T
t ) + bk cos(k
t )d ( 2T t )
2 T
t ))
2 T
t )dt =
f (t )cos(k
2 2
2 T
k = 0,1,2,K b0 = 0
2 T
2 T
t )dt =
f (t )sin (k
t )d ( 2T t )
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
k = G ( jk )
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
u (t ) = Ue j t
2 M k = ak + bk2 a k = arctan k bk
m(t ) = M k e jk e jk t
k =1
Taking into account the low-pass property of the linear part of the system
w(t ) = Gk e j k M k e jk e jk t G1e j1 M 1e j1 e j t
k =1
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
u(t)=-w(t)
j1
Ue
j t
= G1e M 1e e
j1
j1
j t
1 + G1e
M 1 j1 e =0 U
1 + G ( j )N (U , ) = 0
N (U , ) =
1 (b1 + ja1 ) U
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
G ( j ) =
1 N (U , )
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
=+
0
-2
Imaginary Axis
-4
+ U
(U0 , 0 ) -\1/N(U)
U=0
-6
-8
G(j)
=0
-10 -1.4 -1.2 -1
Real Axis
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
0.5
=100
-0.5
=17.7828 =0.1
Imaginary Axis
-1
U
-1.5 System: sys Real: -1.14 Imag: -2.02 Freq (rad/sec): 0.651
-2
=3.1623
-2.5
-3
=0.56234
-3.5 -8
-7
-6
-5
-4
-3
-2
-1
Real Axis
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
f(u)
N (U , ) =
1 (b1 + ja1 ) U
The evaluation of coefficients a1 and b1 can be performed by means of both analytical calculation and numerical integration, depending on the type of nonlinearity involved
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
u m
+M
2
1.5
+M
1
0.5
m(t)
T
-0.5
-1
-M
-2
-1
-M
-3 -3 -2 -1
-1.5
u(t)
-2
0.5
1.5
2.5
Time
N (U , ) =
4M U
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
The negative reciprocal of the DF is the negative real axis in backward direction
Imaginary Axis
-0.5
( ;4M/)
c
-1/N(U)
-1
-1.5
-2
A limit cycle can exist if the relative degree of G(j) is greater than two
-2.5
-3
G(j )
-3.5
-4 -1.6
-1.4
-1.2
-1
-0.8
-0.6
-0.4
-0.2
0.2
Real Axis
The oscillation frequency is the critical frequency c of the linear system and the oscillation magnitude is proportional to the relay gain M
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
1.5
M
1.5
u m
0.5
0.5
-0.5
-0.5
-1
-1
-M
-1.5
-1.5
-2 -2.5
-2
-1.5
-1
-0.5
0.5
1.5
2.5
-2
0.5
1.5
2.5
Time
4M N (U , ) = j U
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
-0.2
-0.4
Imaginary Axis
-0.6
-0.8
G(j )
-1
-1.2
-1/N(U)
-1.4
U
-1.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4
Real Axis
The oscillation frequency is lower than the critical frequency c of the linear system and the oscillations magnitude is proportional to the relay gain M and to the modulus of the transfer function at phase -/4
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
2.5 u m
1.5
+M
1.5
1
1
0.5
0.5
-
0
t
-0.5
-0.5
-1
-1
-1.5
-1.5
-M
-2 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
-2
-2.5
0.5
1.5
2.5
Time
4M j U N (U , ) = 2 4M 4 M 1 j U U 2 U
U U >
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
Imaginary Axis
The negative reciprocal of the DF is the parallel to the negative real axis and with constant negative imaginary part
-0.5
U
-1
-1/N(U)
-1.5
-2
-2.5
-3
G(j )
-3.5
-4
Real Axis
The oscillation frequency is lower than the critical frequency c of the linear system and the oscillations magnitude is proportional to the relay gain M
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
0.8
u m
+M
0.6
+M
0.4
2
0.2
-1
-0.2
-2
-0.4
-M
-3
-0.6
-M
-4 -2.5
-0.8
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
u
-1 0 0.5 1 1.5 2 2.5 3
Time
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
Imaginary Axis
The negative reciprocal of the DF is part of the negative real axis in backward direction
U
-0.5
-1/k
-1
-1.5
-2
A limit cycle can exist if the relative degree of G(j) is greater than two and gain k is sufficiently high
-2.5
-3
-3.5
G(j )
-4 -1.6
-1.4
-1.2
-1
-0.8
-0.6
-0.4
-0.2
0.2
Real Axis
The oscillation frequency is the critical frequency c of the linear system and the oscillations magnitude depends on the saturation parameters M and k
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
Dead zone
3 2
u m
1.5
k
1
0.5
-
0
+
-0.5
-1
-1
-2
-1.5
-3
-2
-3 -2 -1 0 1 2 3
u
-2.5
0.5
1.5
2.5
Time
= U sin (t )
0 2k N (U ) = k
U arcsin U
2 + 1 U > U U
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
0.5
-1/N(U)
0
-1/k
Imaginary Axis
U=+
-0.5
-1
A limit cycle can exist if the relative degree of G(j) is greater than two and gain k is sufficiently high
G(j )
-1.5
-2
-2.5 -3.5
-3
-2.5
-2
-1.5
-1
-0.5
Real Axis
The oscillation frequency is the critical frequency c of the linear system and the oscillation magnitude depends on the dead zone parameters and k
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
Dead zone
2
The nonlinear characteristics of the Dead Zone can be computed by subtracting the Saturation characteristics from a linear one
saturation ( u)
dead zone ( u)
-1
(u ) = k (u )
-2
-3 -3
-2
-1
N (U ) = k N (U )
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
N t (U ) = N1 (U ) + N 2 (U )
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
( ) =
[arcsin( ) +
2
1 2
k
- + u/U
1 k N ( ) = k ( ) < 1
-M
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
k2 k1
u/U + m -M
1 k1 N ( ) = k 2 + (k1 k 2 ) ( ) < 1
m k - + u/U
1 0 N ( ) = k (1 ( )) < 1
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
k2
- +
k1 +
u/U
1 0 N ( ) = k (1 ( )) <1 k ( ( ) ( )) < 1
-M
m +M
k
u/U -M
N ( ) = k +
4M
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
u/U
N ( ) =
4M
-M
m +M
+
-M u/U
0 N ( ) = 4 M 1 2
1 <1
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
1 0 N ( ) = k [1 (2 1)] j 4k (1 ) < 1 2
m k - + u/U
1 0 N ( ) = k [1 (2 1)] + j 4k (1 ) < 1 2
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
- +
u/U
0 N (U , ) = 4 M 4 M 1 2 j U U
1 <1
-M +M m
- +
u/U
0 N (U , ) = 4 M 4 M 2 U 1 + j U
1 <1
-M
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
1 N t (U ) = N 1 (U ) + N 1 (U ) j
N (U , ) = 4M 4M j U U
Imaginary axis
U
0 =100 -0.5 =17.7828
U
=0.1
-1
-1.5
-2
U
=3.1623
U 1 ( + j ) = 2 N (U , ) 4M 1 +
-2.5
-3 =0.56234 -3.5 -8 -7 -6 -5 -4 -3 -2 -1 0 1
Real axis
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
If the block N.L. is a pure constant k, the stability of the feedback system can be performed by means of the Nyquist criterion, which gives the number of roots with positive real roots of the Harmonic Balance Equation
1 1 + G ( j ) = 0 k
The Nyquist criterion looks at the relative position of the transfer function G(j) with respect to the point (-1/k, 0) in the complex plain. By extension the criterion can be applied to any point of the complex plain with reference to the Harmonic Balance Equation
1 + G ( j ) = 0 , C
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
0.5
G ( j ) =
=+
1 2 j ( j ) + 0.5 j + 1
-1/k
-0.5
Imaginary Axis
-1
-1.5
The closed loop is not stable with respect to the point (-1/k,0).
-
-2
-2.5
-3
-3.5
-4 -3
-2.5
Real Axis
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
YES, it can
What does it mean that a limit cycle is stable? If a perturbation of the magnitude of the periodic oscillation occurs, it tends to the original value as time passes. How can a limit cycle be considered from the Nyquist criterion point of view? It is a marginally stable condition. How can the magnitude of a limit cycle be represented? By a point in the Describing Function plot.
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
Imaginary Axis
1.5
0.5
-0.5
U
B B B
-1
-1.5
-2
G(j )
-2.5
-3 -3
Real Axis
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
0.5
Imaginary Axis
-0.5
-1
U
B
G(j )
-1.5
-2
-2.5
-3
Real Axis
The linear system G(j) is assumed to be stable in order to apply the reduced Nyquist criterion
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
+24 V
M
0V -24 V
d.c.
The position of the motor shaft is measured by means of a rotational variable resistance. The voltage on the rotational resistance drives the position of a relay that switches the motor supply voltage between +/- 24 V d.c.
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
+ 24 < d vr (t ) = 24 d
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
kt (s ) W (s ) = = Vs (s ) (sLa + Ra )(sJ + B ) + k t k e
kt ( j ) = = ( j ) + j( j ) W ( j ) = Vr ( j ) j (( jLa + Ra )( jJ + B ) + k t k e ) =-
2 (Ra J + La B )2 + (Ra B + k t k e 2 La J )
j
k t ( R a J + La B )
2 (Ra J + La B )2 + (Ra B + k t k e 2 La J )
k t R a B + k t k e 2 La J
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
R=0.4; % rotor resistance L=0.001; % rotor inductance ke=0.3; % voltage feedback constant kt=0.3; % torque constant
Imaginary Axis
U
-2 -4
-6
Jm=0.01; % motor inertia Jl=0.09 % load inertia Bm=0.05; % motor friction coefficient Bl=0.05; %load friction coefficient
-8
-10
-12
-14
-16
J=Jm+Jl; B=Bm+Bl;
-18
-20 -2
Real Axis
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
R=0.4; % rotor resistance L=0.001; % rotor inductance ke=0.3; % voltage feedback constant
Imaginary Axis
kt=0.3; % torque constant Jm=0.01; % motor inertia Jl=0.09 % load inertia Bm=0.05; % motor friction coefficient Bl=0.05; %load friction coefficient J=Jm+Jl; B=Bm+Bl;
U
-1
-2
-3
-4
Real Axis
(W
= cr = p ( j ))= 0
Ra B + k t k e = 36.056 rad/s La J
U=-
4M
(W p ( j cr )) = 0.1759 rad
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
R=0.4; % rotor resistance L=0.001; % rotor inductance ke=0.3; % voltage feedback constant
Imaginary Axis
kt=0.3; % torque constant Jm=0.01; % motor inertia Jl=0.09 % load inertia Bm=0.05; % motor friction coefficient Bl=0.05; %load friction coefficient J=Jm+Jl; B=Bm+Bl;
U
-1
-2
-3
-4
Real Axis
(W
= cr = p ( j ))= 0
Ra B + k t k e = 36.056 rad/s La J
U=-
4M
(W p ( j cr )) = 0.1759 rad
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
60
40
20
-20
-40
-60
-80
10
Time
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
2.5
1.5
0.5
10
Time
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
1.75
Angle [rad]
1.7
1.65
1.6
1.55
2U=8M|G(jc)|/
1.5
1.45
1.4
8.4
8.45
8.5
8.55
8.6
8.65
8.7
8.75
8.8
Time [s]
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
ang
V w Cr
Scope
i
DC Motor
kt W (s ) = (sLa + Ra )(sJ + B ) + k t k e
The linear plant is characterised by an all-pole transfer function with relative degree two, therefore there is no contact point between the Nyquist plot and the real negative axis, but the origin at =+ (corresponding to U=0 in the DF plot)
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
U
-0.2
U=0 =+
=0
-0.4
Imaginary Axis
-0.6
The Nyquist plot of the linear system is tangent to the negative reciprocal of the DF at the origin, i.e., U=0 and =+ Maybe a sliding mode behaviour is established asymptotically!?
-0.8
-1
-1.2
-1
Real Axis
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Time
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
Speed [rad/s]
By reducing the integration step of the simulation it is apparent that the stabilisation is asymptotic
Time [s]
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
By reducing the integration step of the simulation it is apparent that the stabilisation is asymptotic
50
40
30
20
10
The rotor current tends to a constant value, i.e. an asymptotic (2nd order) sliding sliding mode appears
-10
-20
-30
-40 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Time [s]
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
kt k d (t ) B B 0 = (t ) + t ir (t ) J (t ) J J dt J = k + 1 v r (t ) k dir (t ) R R 1 = e (t ) r ir (t ) + v r (t ) e r ir (t ) L r Lr dt Lr Lr Lr Lr kt B B k t (t ) s (t ) = (t ) + ir (t ) + 2 (t ) = 2 + J J J J ir (t )
The system transfer function has a zero in 2, and if the system output is steared to zero in a finite time, than the system behaves as a first order system with time constant =0.5
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
ang
24
V w
r=2
Relay
Manual Switch
Gain
Cr
Scope
i
2 Gain1
Take care that because of the feedback, the new closed loop gain (assuming the rotor velocity as the output) will be 1/2
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
2.5
It is apparent that the shaft speed tends to the steady state value as a first order system with = 0.5 s
Speed [rad/s]
1.5
0.5
2.5
0
0.5
1.5
2.5
3.5
Time [s]
2
Apart from a very short transient, the state trajectory tends to the steady-state values sliding on a linear manifold of the state space
1.5
0.5
0.5
1.5
2.5
3.5
Speed [rad/s]
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
U
0
Imaginary Axis
The Nyquist plot of the linear system cross the negative reciprocal of the DF at the origin, i.e., U=0 and =+ A sliding mode behaviour is established in a finite time
-1
-2
-3
-4 -2
-1
Real Axis
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
Sliding modes are characterised by infinite frequency of proper ideal switching devices Most sliding mode controllers use a sign function in the controller, i.e., an ideal relay, which can be approximately represented by its Describing Function By the example it is apparent that a sliding mode behaviour is established in a finite time if the Nyquist plot of cross the inverse negative of the describing function at the point (U=0,=+) It can also be derived that a sliding mode behaviour is established asymptotically if the Nyquist plot of is tangent to the inverse negative of the describing function at the point (U=0,=+) TAKE CARE: the Describing function is an approximate tool
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
Sign
ang
V w
Scope
Cr i
2 Gain1
It is apparent that an ideal sliding mode cannot appear because of the switching delay
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
The limit cycle is apparent, with a period Tlc0.6 ms The magnitude is very small because of the low-pass filter property of the motor transfer function
20 10 0 -10 -20 0.4 0.4001 0.4002 0.4003 0.4004 0.4005 0.4006 0.4007 0.4008 0.4009 0.401
x 10
-5
Speed [rad/s]
0.5
-0.5
-1 0.4
0.4001
0.4002
0.4003
0.4004
0.4005
0.4006
0.4007
0.4008
0.4009
0.401
Time [s]
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
-1/N(U) U=1
0
Imaginary Axis
=+
-1
=0
-2
G(j)
-3
-4 -2
-1
Real Axis
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
Analysis of the characteristics of a chattering behaviour due to unmodelled dynamics of sensors and/or actuators
What is chattering? It appears as oscillations of the system variables, whose magnitude is related to the influence of the neglected dynamics on the system bandwidth
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
Scope2 Sign
Scope3
ang V w Cr
Scope
i
2 Gain1
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
-1/N(U)
0
Imaginary Axis
-1
-2
-3
G(j)
-4 -2
-1
-1/N(U)
System: untitled1 Real: -0.0507 Imag: -0.0328 Freq (rad/sec): 662
Real Axis
Nyquist plot parameters lc=662 rad/s Ulm=1.1833 rad/s2 Uls represents the magnitude of the oscillation of the sliding variable
G(j )
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
1.5
]
0.5 0 -0.5 -1 -1.5 -2 2.9
Differences are due to: Numerical solution of the simulation Approximation of the DF approach
2.91
2.92
2.93
2.94
2.95
2.96
2.97
2.98
2.99
Time [s]
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
The rotor velocity tends to its steady-state value as a first order system with time constant =0.5s Rotor wound current presents large variations An approximate sliding mode is established
2.5
Velocity Current
1.5
0.5
0.5
1.5
2.5
3.5
Time [s]
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
kt B ir = 2 + J J
& = 2 + s
& = 2 + r + 2 + r 2 e t + t v r L J J Lr J Lr JLr r B R B R k k k
Input-output dynamics
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
& = 2 +
1
(s ) =
1 (s ) s+2
(t ) = 0 + U sin ( lc t )
1 j ls + 2 0.0015
1 lc t + arg sin (t ) = 0 + U j + 2 j lc + 2 lc
3.088
3.086
Speed [rad/s]
Nyquist plot parameters lc=662 rad/s Uls=1.1833 rad/s2 ls=0.0018 rad/s Simulation results Tls=9.7 ms lc=646 rad/s Uls=1.8544 rad/s2 ls=0.0029 rad/s
3.084
3.082
3.08
3.078
3.076
3.074
3.072
3.07 2.9
2.91
2.92
2.93
2.94
2.95
2.96
2.97
2.98
2.99
Time [s]
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008
The Describing Function approach to the analysis of the chattering phenomenon in sliding mode control systems can be used to have an estimate of the chattering parameters, i.e., frequency and magnitude The estimates are affected by an error which depends on the low-pass properties of the linear part of the plant The sliding variable must be considered as the output of the nonlinear feedback system The actual system output behaviour can be estimated by considering the reduced order dynamics In the presence of a constant reference value, the nonlinear function can be not symmetric, therefore an equivalent gain of the nonlinearity has to be considered to estimate the constant mean steady-state value of the output
Describing Function analysis of nonlinear systems Prof Elio USAI March 2008