Get documents about "A Spice Based Fuzzy Logic Speed Controller For Pmdc
Document Sample
A SPICE BASED FUZZY LOGIC SPEED CONTROLLER FOR PMDC MOTOR DRIVES
Ali I. Maswood A. M. Sharaf
School of Electrical and Dept. of Electrical Engineering
Electronics Engineering University of New Brunswick
Nanyang Technological University Box: 4400, Fredrickton
Nanyang Avenue, Singapore 2263. Canada.
ABSTRACT
This paper presents a novel light load or full speed
fuzzy logic rule based speed conditions. For a chopper fed dc
regulator and it's SPICE model for motor it may also occur at small
permanent magnet variable speed chopper duty cycle. Under such
(PMDC) motor ' drives. The motor circumstances, the fixed gain
speed is regulated using the speed controller loses it's effectiveness
error excursion vector Ewand a and results in current and/or speed
proportional plus integral control fluctuations.
law. The controller gain is This paper presents the
adjusted on-line using the fuzzy development of a novel. fuzzy logic
assignment table based on the error based controller for speed control
excursion vector and its rate of of PMDC. The drive utilizes a
change at any given time so the chopper dc-dc converter with a
control action is scaled on-line by combined PWM switching modulation
the error level for the control scheme as shown in Fig. 1. The PMDC
correction needed. The scaling is motor armature voltage is
based on the excursion error vector controlled by adjusting the
- 2 .2
switching functions Gl(t) and GZ(t)
= and its rate of change
Rw w w for the two mosfet chopper switches
dgw in the [e -e 1 equivalent phase E W 1 , SW2)
w w
portrait plane. The use of a fuzzy logic rule
based controller is motivated by
I. INTRODUCTION the need to tolerate any drive
For dc motor speed regulation, model parameter uncertainties, as
industries usually use PID well as the varying mechanical load
controllers with fixed or adjustable time constant. The fuzzy logic
control gain. The control is usually controller is an ideal form of a
achieved via armature voltage weighted non-linear var iab1e
regulation. However, discontinuous structure control scheme as it
armature current mode may occur at offers an effective robustness and
423
less insensitivity to any model sampling period by the movement of
parametric uncertainties and load the excursion error vector in the
variations. It does not generally (ew-eo) phase portrait. The fuzzy
require the precise knowledge of logic gain kw is assigned a minimum
the mathematical model parameters value k and an adjustable variable
0
of the controlled process. The fuzzy correction (Ak) that is
simple fuzzy logic controller shown scaled by the excursion error Rw
in Fig. 3(a) is driven primarily by and its rate of change dR as shown
w
- in assignment Table I.
the excursion error vector R0' The
controller is based on speed and The adjustable gain correction
current errors (deviations) and weighting vector iy is obtained from
the fuzzy logic assignment matrix
their time derivatives namely (Aw
m'
(Table I ) with fuzzy inputs Rw and
Awm, AIm & AIm) as depicted in the
dR each assigned three sub classes
w
error excursion plane shown in Fig.
(small, medium, large) and linear
2.
type membership weighing functions.
The chopper converter circuit
Compositional minimum-rules are
consists of two mosfet P M switched
W
-
utilized to obtain the equivalent
devices. Switch SW1 for the main
effective correction weighting y
speed regulation loop, while switch
using the centre of area (COA)
SW2 is used for the optional
defuzzification criteria applied to
supplementary current
gain correction weighing set iy.
limiting/braking loop. Only the
speed regulation loop is presented
PID Control law:
and discussed in this paper as the
a (k) = aD(k-l) + kw.AcrD(k) (1)
current limiting action can easily D
be incorporated in the control AaD(k) = (Pl.ew(k)+P2.e,(k)) (2)
level scaling for the speed
regulation loop if so desired.
11. FUZZY RULE BASED CONTROLLER
The fuzzy logic rule based
speed regulator structure is shown
in Fig. 3(a) f o r the main speed
control loop. It is driven by the where : B,, 8, are assigned
speed error e and its derivative weightings for previous control
0
e The control law is equivalent levels and Ro is small dead zone to
w'
to a proportional plus integral avoid control hunting.
action with on-line adjustable gain
that is is scaled at every
424
Tab1e:I Ak assignment Table corresponding SPICE equivalent
circuit is shown in Fig. 3(b). The
PID control law equations(Eqns. 1
through 7) have been used to
construct the SPICE controller
model. The input signals to the
controller are the actual motor
speed(u1 and the reference
speed(uref 1. The output signal
from the controller is fed to the
series chopper switch (SW1) to
r has 9 sub class levels shown above
control it’s duty cycle. Fig. 4(a)
r = i.1, .2, .3, .91.
shows the PMDC motor block diagram
with the main speed loop and the
The weightings Pland P, are supplementary current limiting
selected using off-line digital
loop. The corresponding SPICE model
simulation to ensure the is developed and is shown in Fig.
minimization of a global error
4(b). In Fig. 4(b) the RICl values
criterion Jo of the cumulative
represent the motor armature
speed error.
circuit time constant(t 1 is found
I
L
a
m
as, t = - . R2C2 values
a Rm
where: N= Selected final sampling
represent the motor mechanical time
J
constant and is found as tm = - ’
B
Tse t t ing
count, N=
Tiample ’
IV. SAMPLE RESULTS
111. SPICE MODEL OF THE FUZZY The dynamic transient response for
CONTROLLER & THE PMDC drive motor variables are shown in
The SPICE circuit simulator is Figs. 5 through 9 for a change in
widely used in industries and has reference.
practically become a standard. If Fig. 5 shows the chopper source
the circuit is properly described, current (Is)as well as its
the results are as accurate as the effective rms value. This current
the ones from the experimental is shown for a change in the region
prototype. The model response can of the rated motor speed. Fig. 6
be obtained in time(transient1 or shows the current passing through
in frequency domain. the series mosfet switch SW1 (Isw).
Fig. 3(a) shows the fuzzy logic The duty cycle is being controlled
controller block diagram. The by the fuzzy logic speed control
loop and on-line gain adjusting
425
action. and its rate of change dRw. Switch
Fig. 7 shows the transistor gating (SW1)controls the speed regulation
signal S1(V(11,12)) generated from loop, while switch (SW2) can be
the crossing points of a triangular used as supplementary switch for
carrier switching waveform (V(55)) limiting the motor armature current
and the control reference waveform and provide motor braking.
obtained from the fuzzy controller
(V(53)). Fig. 8 depicts the motor VI. REFERENCES
armature current (Im)and its rms [ll T.S.Low , M.A Jabbar , M. A.
value. Rahaman, "Permanent-Magnet Motors
Fig. 9 illustrates the motor for Brushless Operation", IEEE
speed (wm), shown only near its Transcations on Industry
rated value. This is due to the Applications, 1990,pp. 124-129.
simulation package dynamic memory [21 B. Puthal, "Novel Closed Loop
limitation in the PC environment. Control Scheme fot Thyristor-fed DC
Full speed build up from standstill Motor",J. 1nst.Engg (India) Electr.
characteristic could not be shown Engg. Div., 1978, pp 333-338.
entirely. [1
3 Y. Y.Hsu, and W.C.Chan,
" Opt imal Variable Structure
V . CONCLUS IONS Controller for DC Motor Speed
The paper presents a novel Control", IEE Proc.,Vol.131, 1984,
fuzzy logic based speed regulator pp. 233-237.
with an on-line adjustable gain for [ 1 A. M. Sharaf, A. Ghosh, "Speed
4
PMDC motors. The speed control is and torque regulation of permanent
achieved via the armature voltage magnet DC motors using rule based
regulation using Type A chopper Fuzzy logic", proceedings of
drive circuit. Two MOSFET switches, Intelligent vehicles symposium,
SW1, SW2, using PWM switching Tokyo, July 14-16, 1993.
strategy can be used. An effective [51 S . Weerasoorya, and
proportional plus integral control M.A.El-Sharkawi,"Identification and
law with the on-line fuzzy gain Control of DC Motor Using
adjustment is utilized to ensure Back-Proportional Neural Networks'"
effective speed reference tracking. [ l J.
6 Zhang,and T. H. Barton ,
The fuzzy logic controller action "Robustness Enhancement of DC
is based on the excursion vector Drives With a Smooth Optimal
location in the (eU-eU) plane. Sliding Mode Control", IEEE
The duty cycle ratio aD is Trans.on
adjusted on-line using the speed Ind.APPL.,Vol.27,No.4,July/August
error excursion vector magnitude Ro 1991, PP. 686-693.
e
Motor armature tu
Current sensor
Sourcr -
IF u z z q ControI I er1 Sensing
Fig. 1. Complete circuit diagram of the Permanent magnet dc (PMDC) Fig. 2, eW -e O
phase portrait.
motor chopper drive circuit.
=
Fig. 3(a), PMDC motor drive block diagram with main speed regulation
loop and supplementary current limiting loop,
XROlO
RD
n
0
T
0
R
R02
z
F
U
T
OPRMP 7 XOPAMPf2
Fig. 3(b). Equivalent PMDC SPICE model
Sprrd
Sprrd
Acquisition 8 qU= Cr - &,> Fuzz9 C 8 i n Adaustment
Rf r r r n c r TrrnsCormrtion Ruler
Ku
Rrrmturr
Currrnt
I n t e9r.t or
Lau
output
SUnnER
Fig. 4(a), Fuzzy logic controller block diagrame.
Fig. 4(b), Equivalent controller SPICE model.
427
.......7 12Vt ......... .....
+ ---.+. . . . +. . . . +-.- .....+
.... . . . . . . . .----t +
.. ........
35 mi ;
34 Bt j
3, on+ i
33 5 R i i
31
33 i_:
138 8ms
a 111111 .
139 Gnl
Iti 1~~~~1'111I'i
. . . . . . . .+ . . . . + . . . . . . . ... . .
-4 . . . . +. . . . .. . . . . . . . .+. . . .t. . .
139 h 3
"(lllflll
Il~~
139 *3
TI*
139 -3 139 h 3 1'
....
...
;
-,?Vi
. . .......... ........ ........ ........ ......... ........
....
. . . . .+
b VI531139VI551
s 0.5
t
139 ms
+
v111,121 139 cs
rlr
+
139 -3
.+
139 m, .
10
+
.
o,
4
Fig 7, PUH sultching function for the mosfet witch generated froa
Fig. 5 . Chopper source current and its rms value. the crossing points of a triangular carrier waveform and the
waveform obtained from the fuzzy controller
. . . . . . . . . + . . . . .... .... . . . . + . . . .
m+ . . . .+. . . . . .. . . .+. . . . +. . . . .+ . . . . . .. . . . +
51 . . . . . . . . .... . . . . . . . . . . . . . . . .
. . . . +. . . .+.... .+. . . .+. . . . +. . . . .+. . . . ~
400 + ?
51 b a t
,004 i
51 ,I,+ t
-00+ j
-mi i
51 C M i i
51 5Sai
-4on; T
51 mi
....
-m+. . . . ... .... .
. . . +. . . . +. .......+..... . . . . +
+ ....
.
.
m
138 b , 13
11rtet1 . m,~llrtclll
139 4m3 139 b s 139 h a 140 Gn3
51 5,a; . . . .4. . . . . + . . . ..+ . . . . +. . . . .+ .... .+. . . .
. . . . . . . . . . . . . . . . . . . . .... . . . .
. .
Tlrr
156 6nn.i 139 01s 139 -3 139 4 3 139 MI 139 Cms 140 On3
11111 n,llll")l
TI*
Fig. 6 , Current passing through the serlee mosfet switch and its rms
Fig. 0 . Motor armature current and its rms value.^
value.
. . . . . . . . +. . . . . . . . . . . . . . . . +................
........ . . . . . . . . ........
/
35 $ V I
35 b"+ ........ . . . . . . . . ........ .......
. . . . . . . .+. . . . . . . . . . . . . . . . +. . . . . . . ..+
+
Fig. 9. Motor speed show over 95% of its rated value.
428
Related docs
