InternationalJournalofAutomationandPowerEngineering(IJAPE)Volume2Issue4,May2013www.ijape.

org
65
AdaptivePF(PDF)SpeedControlforServo
Drives
LaszloSzamel
*1

DepartmentofElectricPowerEngineering,BudapestUniversityofTechnologyandEconomics
EgryJzsefu.18,H1111Budapest,Hungary
*1
szamel.laszlo@vet.bme.hu

Abstract
ThispaperproposestwomodelreferenceadaptivePF(PDF)
speed control methods for servo drives. Following from the
structure (PFtype) of model reference parameter adaptive
control was developed to provide constant loop gain in
speed control loop with changing gain (moment of inertia
and/or torque factor) which makes it easier to reach
nonovershooting step response as well as fast speed
changing compensation caused by jump in load. The
algorithmevenkeepsitsstabilityatfastchanging,jumplike
loadtorque.Modelreferencesignaladaptivecontrolisused
to provide constant loop gain in speed control loop with
changingparametersexposedtoasignificantload.Theblock
diagram of the adaptive control can be seen as an extended
versionofthePFcontroller,sooneoftheadaptationfactors
(whichisthefreeparameteroftheadaptivecontrol)isgiven.
Both model reference adaptive controls drawn up can be
easily implemented because the adaptation algorithms do
not need acceleration measuring (thanks to the firstorder
model). Simulation and experimental results demonstrate
that the proposed methods are promising tools for speed
controlofelectricaldrives.
Keywords
Adaptive Control; Adjustable Speed Drives; Motion Control,
PseudoDerivative Feedback; Servo Drives; Switched Reluctance
Drives;VariableSpeedDrives
I nt r oduc t i on
Generally the controller of the speed loop in motion
control systems is of PItype. The integrator of the
controller eliminates the error caused by the step
change in load. The setting of speed controller is
difficult as the closedloop transfer functions are not
identical to step changes in the load, as well as the
reference signal. It seems to be preferable to use PF
type controller (proportional gain in a separate
feedback)insteadofthetraditionalPIone(Diep,N.V.,
and Szamel, L., 1990), (Ohm, D. Y., 1990), (Perdikaris,
G. A.andVanPatten, K.W., 1982). Phelan named this
structure Pseudoderivative feedback (PDF) control
(Phelan,R.M.,1977).
In motion control systems, there is robustness against
parameter changes and disturbance rejection of main
interest. Themodel reference adaptivecontrol has the
followingfeatures:
- It enables the compliance of the system with
varying operational conditions possible and
ensures the behavior of the controlled system
accordingtotheprescribedreferencemodel.
- It means such a special type of adaptive
systems, which results in nonlinear control
systems. This is the reason why the analytical
analysisiscompletedbytheLyapunovstability
criterionorbyhyperstabilityprinciple.
- Itsdesignandapplicationarecloselyrelatedto
theuseofcomputermethods.
- Simple implementation of the control
algorithm.
Speed Cont r ol of Ser vo Dr i ves
A model reference adaptive control is used for the
speed control. Such an adaptive control has been
succesfully elaborated by using a suitable chosen
Lyapunov function to compensate the gain of the
speed control loop (Liu Hsu et al., 2007), (Szamel,
Laszlo,2002),(Szamel,Laszlo,2004).
ModelReferenceParameterAdaptiveControl
The adaptive control of servodrives with a cascade
arrangementisthemosteffectivewhenitisappliedto
the inner loop containing the effect of variable
parametersdirectly,i.e.theinertia(
m
J )and/ortorque
factor (
m
k ). The speed control implemented by PI
controller is of cascade arrangement in fact as it
contains an inner, proportional feedback loop (PF
controller,Fig.1.),(Perdikaris,G.A.andVanPatten,K.
W., 1982). A onestorage proportional element can
describethisinnerloopneglectingthetimeconstantof
www.ijape.orgInternationalJournalofAutomationandPowerEngineering(IJAPE)Volume2Issue4,May2013
66
closed current control loop. In such a way our
adaptationalgorithmisthesimplest.

s
A
s Y
i
i i
t
=

) (
) ( , e
(1)
where
m
m
i
J
k
A = .
Arrangement of control circuit can be seen in the
followingfigure:
i
l

e
r
e
ri
i e
s
A
i

s
i
K

p
K
p
K

FIG.1BLOCKSCHEMEOFPARAMETERADAPTIVEPFSPEED
CONTROL
where
e isthespeed,

r
e isthespeedreferencesignal,
i isthecurrentofthemotor,
l
i is the current equivalent to the load
torque.
Sectiondeterminedbythetransferfunction ) (
) ( ,
s Y
t
i i e

isfedbackbyaproportionalmemberofgain
P
K .The
task is to change the gain
P
K in such a way that the
product
P i
K A should remain constant despite the
changingofparameter
i
A .
Transferfactoroftheinnerclosedloopisgivenbythe
reciprocal (
P
K / 1 ) of feedback member that is not
constantbecauseofthechangeininertiaand/ortorque
factor. Consequently loop gain of the outer speed
control loop would change as well. In order to get a
onestorage element with unity transfer factor, we
have to insert a member with gain
P
K between the
integratorofthePFcontrollerandthereferencesignal
of the inner loop. Firstorder reference model with
time constant
m
T gets sum of the input signals of
abovemember(
ri
e )andthesignal
lm
e compensating
theloadeffectforthemodel.Sodynamicsofreference
model can be described by the following differential
equation:

lm ri m m m
T e e e e + = + . (2)
Dividing (2) with
m
T and applying designation
m m
T q / 1 = ,thefollowingequationcanbeobtained.
) (
lm ri m m m m
q q e e e e + = + (3)
The differential equation of the firstorder controlled
sectionisasfollows:

l i ri i p i p
i A A K A K = + e e e ) ( ) ( . (4)
Factor
p
K can be described as the sum of
p0
K
determined by mean
i
A and
p
K A accomplished by
theadaptationalgorithm.So:
q q A K K A K
i p p0 i p
A + = A + = ) ( , (5)
where
P0
K ,and q areconstant.
It is assumed that the change of
i
A is slow from the
viewpointofadaptationandthereforetheeffectofthis
changecanbeneglected.
Substituting(5)into(4)weget:

l i ri
i A q q q q A + = A + + e e e ) ( ) ( . (6)
By using (3) and (6) and substituting expression of
modelerror e e c =
m
thedynamicequationwillbe:

l i lm m ri m
i A q x x q + + + = e e e c c , (7)
where
m
q q q x A + = ) ( .
Dynamic of model error should be asymptotically
stable to follow the system with proposed model. For
determinationof q A thefollowingLyapunovfunction
shouldbecomposed:
( )
2 2

2
1
x V | c + = , (8)
where | isapositivenumber.
WhenchoosingtheLyapunovfunctionbothpurposes,
i.e. the termination of the model error ( e e c =
m
)
andloopgaindeviationhavebeentakenaccount.
ThetimederivativeoftheLyapunovfunctionis:
x x V

| c c + = . (9)
Substituting(7)into(9)thefollowingequationisvalid:
InternationalJournalofAutomationandPowerEngineering(IJAPE)Volume2Issue4,May2013www.ijape.org
67

. ) (

2
x x i A q
x x q V
l i lm m
ri m

| e c
e c e c c
+ +
+ + =
(10)
If

0 = + x x x x
ri
| e c e c
, (11)
thatis

| e e c / ) ( =
ri
x
(12)
and

0 ) ( < +
l i lm m
i A q e c
, (13)
then

2
c
m
q V <

. (14)
Theaboveequationisanegativedefinitefunctionthat
showstheasymptoticstabilityoftheerrordynamic(7).
By using (5), (7) and (12) the following adaptation
algorithmistrue:

) ( e e c = A
ri p
K

, (15)
where may be an arbitrary positive number. The
inequality (13) shows how we have to change the
signal
lm
e representingtheloadofmodel.
If

m i l lm
T A i
max
, 0 then < > e c , (16)
respectivelyif

m i l lm
T A i
max
, 0 then > < e c .
ModelReferenceSignalAdaptiveControl
The controlled loop has been approximated by an
integral element. Time constant of the closed current
control loop has been neglected. The control consists
of a Pelement with the gain
p
K . Input of Pelement
contains not only the control error signal but an
adaptation signal (g) as well. Applying the signal
adaptation control, a P type controller with
p
K gain
can ensure zero speed error as the adaptation signal
can produce a current reference signal to compensate
theloadingcurrentatzerospeederror.

g i
l
e
r
i e




p
K
s
A
i

FIG.2INITIALBLOCKSCHEMEOFSIGNALADAPTIVESPEED
CONTROL
Regarding the block diagram of control loop,
following differential equation is valid for the closed
loop:
( ) .
l i r p i p i
i A g K A K A + = + e e e (17)
The feature of the closed speed control loop has been
takenintoconsiderationbyaparallelcontrolmodelto
beexpressedbyafirstorderproportionalelement.The
differentialequationofthefirstordersystemis:

r m m m m
q q e e e = + , (18)
where index m refers to the model and
m
q is the
reciprocalofmodeltimeconstant.
Using (17) and (18) and introducing the expression
e e c =
m
formodelerror,thedynamicequationfor
themodelerrorisasfollows:
( )( ) ( ) g K i A K A q q
p l i r p i m m
+ = + e e c c . (19)
The adaptation signal ) (t g can be written in the
followingform:
( ) ) ( (t) ) (
2 1
t g g t g
r
+ = e e . (20)
Substituting(20)for(19):
( )
2 1
b b q
r m
+ + = e e c c , (21)
where
( ) ) ( 1
1 1
t g K A q b
p i m
+ = ,
( ) ) (
2 2
t g K i A b
p l i
= .
Let us compose the following Lyapunov function to
producethesignal ) (
1
t g and ) (
2
t g :
( )
2
2 2
2
1 1
2
2
1
2
1
b b V | | c + + = , (22)
where
1
| and
2
| arepositiveconstants.
TimederivationoftheLyapunovfunctionis:

2 2 2 1 1 1
b b b b V

| | c c + + = . (23)
www.ijape.orgInternationalJournalofAutomationandPowerEngineering(IJAPE)Volume2Issue4,May2013
68
Substituting(21)for(23):

( )
.

2 2 2 1 1 1
2 1
2
b b b b
b b q V
r m

| |
c c e e c
+ +
+ + =
(24)
If

( )
1 1
/ | c e e =
r
b

(25)
and

2 2
/ | c = b

, (26)
then

2
c
m
q V =

, (27)
anditensuresasymptoticalstabilityofthemodelerror.
On the basis of (21), (25), (26) and by assuming that
variation of
i
A can be neglected compared to the
speed of adaptation, the following adaptation
algorithmisvalid:
( ) ) (
1 1
e e c =
r
t g , (28)
) (
2 2
c = t g , (29)
where
1
and
2
are positive constants, the free
parameters of adaptation. Taking relations (20), (28),
(29) into consideration, the following equation comes
true:
( ) ( )
} }
+ = dt dt t g
r r
) (
2 1
c e e c e e . (30)
A block diagram of the control circuit introducing
adaptation signal ) (t g and assuming that ) (
1
t g is
constantcanbeseeninFig.3.Thestructureofcontrol
containstwoparts.Inthefirstpartthereferencesignal
is led through a first order system and a PI controller
with variable gain and integration time. The second
oneisadifferentiatingfilterwhichtakeseffectonlyon
changing of reference signal. The gain and
differentiation time are also changing. The adaptation
gain factor
2
gives the reciprocal of integrating time
constant of controller type PI, assuming ) (
1
t g = 0. To
fulfill the constant integrating time constant, it is
preferabletosubstitute
2
by ( ) ) ( 1
1 2
t g + .Insucha
way the neglect of time constant of current control
loopcanbecompensated.


il
er c i e






s
A
i

s
p 2
K

( )
1 p
g 1 K +
( )
1 p m
g 1 K T + s
m
T 1
1
s +

FIG.3BLOCKSCHEMEOFTHESIGNALADAPTIVESPEED
CONTROLWITHADAPTATIONSIGNAL
The approaching block diagram of the adaptive
control can be seen as extended version of the PI
controller,so
2
,oneoftheadaptationfactors(which
isthefreeparameteroftheadaptivecontrol)isgiven.
Contraction of modelfiltered reference signal and the
PI controller can be transformed into a socalled PF
controller when integration time of PI controller
equalstothetimeconstantofthemodel(Fig.4.).
i
l

e
r
i e
s
A
i

( )
s
1 p 2
g 1 K +

( )
1 p
g 1 K +
( )
m
1 p m
T 1
g 1 K T
s
s
+
+

FIG.4.BLOCKSCHEMEOFTHESIGNALADAPTIVEPFSPEED
CONTROL(EXTENDEDVERSION)
The basic structure of the signal adaptive speed
controller is also PF type which on the one hand
provides nonovershooting step response with its
structure. Moreover, it ensures fast compensation of
speedvariationcausedbyajumpinmotorload.
Thetransferfunctionbetween e and
r
e isasfollows:
) 1 (
1 1
1
1
1
1
1 2
2
2
2
2
g K A
s s
T
s sT
sT
p i
m
m
m r
+
+ +
+ +
+
=

e
e
. (31)
Assumingthattheadaptationsignal ) (
1
t g isconstant
(at the end of the adaptation), the following equation
isvalid:

( )
P i
K A
g
2
1
1

= +
. (32)
Substituting(32)into(31):
InternationalJournalofAutomationandPowerEngineering(IJAPE)Volume2Issue4,May2013www.ijape.org
69

2
2
2
2
2
2
1 1
1
1
1
1

e
e
s s
T
s sT
sT
m
m
m r
+ +
+ +
+
=
. (33)
Choosing
2
1

=
m
T ,thetransferfunctionisasfollows:

m r
sT +
=
1
1
e
e
. (34)
Sothesystemfollowsthemodelwithouttimedelay.
This control has been tested by a simulation program
developed in our Department. Firstly the adaptation
has been examined without current limitation and
load in order to take into consideration only the non
linearity of the motor and the adaptation. In the
interestoftheadaptationstability,thespeedofchange
of the adaptation signal
1
g has to be limited. The
signal
1
g can results in the significant oscillations
without limitations as the change of the signal is
possibleindiscretetimes.
The current limitation can result in further problems.
This limitation hinders the tracking of the model, so
theeffectoftheabovesignal
1
g willbetoolargeorit
can change in the reverse direction. To eliminate the
above problem, signal
1
g is not to be changed in the
periodofthecurrentlimitation.
Dr i ve Syst em
Block scheme of the examined drive system with
switched reluctance motor (SRM) (Borka, Jozsef et al,
1993),(BoseB.K.andMiller,T.G.E.,1985)isshownin
Fig. 5. This work is a part of the investigation of an
HTS (high temperature superconductor) energy
storage flywheel system (Vajda, Istvan et al., 2003).
The main advantages of this motor type are the
following:
- noneedforanextragenerator;
- relatively big torque can be achieved, it is easier
topasscriticalspeeds;
- noironlossesatstandby.
FIG.5BLOCKSCHEMEOFDRIVESYSTEM
It is followed from the operational principle of SRM
that its phase windings are to be excited at a well
determined angle of the rotor position in an
appropriateorder.ThisiswhyaRotorPositionSensor
istobemountedontheshaftofthemotor.Inourcase
the position sensor is a resolver. It can be calculated
fromtheteethnumbers(6/8)thatthephaseswitchings
havetofolloweachotherby15degree.Theresolveris
supplied by an oscillator circuit, whose signals are
evaluatedbyaPositionDecoder.
The PositionDecoder hastwo outputs: the Angleand
Speed signals. Based on the two signals, the Angle
Controller composes the Control signals for the phase
switchingtransistors.
Thesupplyunitconsistsofthreemainblocks,namely
the RECTIFIER, the FILTER and the INVERTER. The
inverter is a pulsed width modulated (PWM) one,
marked by QP in the figure and it contains a oneone
switching transistor per phase as well as a brake
chopper, which excludes from showing in the figure.
The common point of phase windings is supplied by
the PWM inverter. It is of autonomous operation and
has an inner current control loop. The other ends of
phase windings are connected to the phase switching
transistors.
Fundamentally, SRM drives have two control loops,
that is, the outer one is the speed loop, Speed
Controller and the inner one is the current loop,
CurrentController.
Neglecting the saturation, the motor torque is
proportional to the square of current, which means
thatthecurrentreferencesignalcanbecomposedfrom
RECTIFIER
Mains
FILTER INVERTER
Reluctance
Motor
Switched
Rotor
Position
Sensor
signal
Torque limit
Speed signal
Angle
Position
Decoder
Current
Controller Controller
Speed
Control
signals
signal
Current signal
+
_
Angle
Controller
Speed reference
Current reference
QP
www.ijape.orgInternationalJournalofAutomationandPowerEngineering(IJAPE)Volume2Issue4,May2013
70
the torque reference signal, produced by the speed
controller by the help of a squareroot function after
composingitsabsolutevalue.Thesaturationofmotor,
depending greatly on the motor construction,
complicates the transient analysis of the system
further. In consequence of the saturation, the square
relation between the current and the torque will be
valid approximately only. Furthermore, the inner
voltage of motor at high speed causes an additional
problem, since the current control of PWM cannot
produce constant current, given by the command
signal. One solution to solve the task is the
compensation in the function of speed and torque,
usingcalculatedvaluesfromalookuptable,storedin
thememory.Theothersolutionistheapplicationofan
adaptivecontrol.
The output signal of theSpeed Controller serves for a
CurrentreferencesignaloftheCurrentController.The
hardware and software tools together fulfil the two
loop control. The Current Controller produces the
control signal for the PWM inverter, and receives the
Current signal from the PWM inverter at the same
time. Based on the current reference signal, the
Current Controller controls the PWM inverter of fix
frequencybyinstallingananalogcontroller.
Forthecontrolofthesumofphasecurrents(Fig.5.)it
issuitableasimplerfourtransistorinverterandisnot
necessary a sixtransistor one as in the case of control
of phase currents independently from each other. But
the detriment of the previous solution is that the
torque pulsation can be decreasedina smaller degree
bychangingtheturnonandturnoffangles.
Namely, in the case of the constant current reference
signal,thecurrentincreaseislimitedbytheswitched
off, but conducting phase current as the regulator
controls the sum of two phase currents. The increase
ofthephasecurrentatstartingoftheconductingstate
can be forced by the modification of the current
referencesignal(Szamel,Laszlo,2001):

j
j
j
j
j r
i C C u i + =

= =
) 1 (
3
1
3
1
. (35)
where:
r
i isthecurrentreferencesignal,
j
i isthecurrentsignalofphase j ,
u istheoutputofthespeedcontroller,
j
C isthecontrolsignalofphase j (0or1).
The supplement of the first member of Eq. 35 makes
the overlap of the phase conduction possible, while
theeffectofsecondmemberistoincreasethereference
signal with the current of the switchedoff, but not
currentfreephase.
The ripple free operation can be realize only with a
current waveform depending on theangle, speed and
torque(Szamel,Laszlo,2001),(Vujici,V.P.,2012).The
proposed ripple reduced method changes only the
turnonandtheturnoffangleinfunctionofthespeed
andcurrentreference.Theoptimumturnonandturn
off angles of the SRM drive has been determined by
computersimulationbasedonthemeasuredresultsof
the analysed drive. The optimum solution has been
fulfilledbyfourcyclesembeddedintoeachother.Two
outer cycles give the current and speed reference
signals,whiletwoinneronesprovidetheturnonand
turnoff angles. By this oneone optimum angle pair
canbedeterminedtoalloperatingpoints.
It can be considered an interesting result that the
criteria of the minimum torque pulsation does not
provide an optimum solution in all cases. The torque
pulsationwillbeminimuminthespeedcurrentplane
onlyinthatcaseifthetorqueofthemotorisrelatively
small.Forthisreasonagoodresultcanbeachievedin
such a way if the relative, i.e. compared to the torque
ofmotor,torquepulsationisminimised.
The angle control of the drive determines the actual
turnon and turnoff angles with a twovariable
interpolationfromtheresultsstoredinalookuptable
andcalculatedbytheabovemethod.
Si mul at i on Resul t s
InFig.6andFig.7twoofmanyexecutedsimulations
are shown. Fig. 6 shows the runup with model
reference parameter adaptive control (15), (16), while
Fig.7withmodelreferencesignaladaptivecontrol(30)
and in both cases with turnon and turnoff angles
depending on the speed and current referenceas well
aswithcurrentreferencecompensation(35).
InternationalJournalofAutomationandPowerEngineering(IJAPE)Volume2Issue4,May2013www.ijape.org
71
0 100 200 300 400 500 600 700 800
0
5
10
15
20
25
30
35
40
Speed [rad/s]
Time [ms]

0 100 200 300 400 500 600 700 800

-2
0
2
4
6
8
10
12
14
16
Torque [Nm]
Time [ms]

0 100 200 300 400 500 600 700 800
0
2
4
6
8
10
12
14
16
18
Current [A]
Time [ms]

0 100 200 300 400 500 600 700 800
0
2
4
6
8
10
12
Phase current [A]
Time [ms]

FIG.6SIMULATIONRESULTSWITHMODELREFERENCE
PARAMETERADAPTIVESPEEDCONTROL
0 100 200 300 400 500 600 700 800
0
5
10
15
20
25
30
35
40
45
Speed [rad/s]
Time [ms]

0 100 200 300 400 500 600 700 800
-2
0
2
4
6
8
10
12
14
16
Torque [Nm]
Time [ms]

0 100 200 300 400 500 600 700 800
0
2
4
6
8
10
12
14
16
18
Current [A]
Time [ms]

0 100 200 300 400 500 600 700 800
0
2
4
6
8
10
12
Phase current [A]
Time [ms]

FIG.7SIMULATIONRESULTSWITHMODELREFERENCE
SIGNALADAPTIVESPEEDCONTROL
www.ijape.orgInternationalJournalofAutomationandPowerEngineering(IJAPE)Volume2Issue4,May2013
72
According to the simulation investigations
convergence of the model reference adaptive controls
at switched reluctance drives with significant torque
ripplescanbeensuredwiththenextconditions:
- Input of the firstorder reference model only
determinesitsoutputwhenthefilteredcurrent
reference signal is lower with a given AI value
(inthesimulationithasbeensetto1A)thanits
limitotherwiseoutputisidenticaltothespeed
feedback signal. So the model works like a
firstorder proportional lag element only in
operationwithoutcurrentlimitations.
- Adaptation is executed when two conditions
aretrueatthesametime:
- Filtered current reference signal is lower with
a given AI value (in the simulation it has been
setto1A)thancurrentlimit.
- Absolute value of the speed error signal is
higherthanagivenAnvalue(inthesimulation
ithasbeensetto20revolution/min).
- The appropriate selection of the adaptation
factoralsohasimportanteffectonthesufficient
convergence.
Because of the restrictions described in point 2.
adaptation practically works only in a relatively
narrow speed error track (adaptation range) which is
approximately20100revolution/minabsolutevalueof
speed error. The drawback of this limitation is the
relatively short time for the algorithm to operate. At
the same time the convergence of the algorithm is
extremelyfast,whichsignificantlyreducestheeffectof
this drawback. Two more important advantages
emergewhenadaptationworksonlywithsmallspeed
errors. First of all the controller at changing drive
parameters adapts to parameters around the value
specified by speed reference signal which also assists
to speed theadaptation. The other significant positive
effect is the disappearance of the problem coming
from nonlinear systems that the response of the
systemcanevendifferinitscharacterwhenthevalue,
amplitudeofthereferencesignalischanged.
Resul t s
The tests were completed by the described drive
system.Thetestresultshavesupportedourtheoretical
investigations. The oscillograms in the following
figures illustrate some typical starting curves and
waveforms.TheloadingmachinewasaDCmotor.Its
inertiaisaboutatripleofthatofSRM.
Fig.8andFig.9showthespeedandcurrentcurvesin
the course of starting. The upper curve is the speed,
and the lower one is the current flowing in the
common point of stator windings. It is related to the
noloadoperationmode.
The experiences show that the model reference
parameter adaptive and signal adaptive control
suggestedinthispaperworkwithoutovershooting.

FIG.8OSCILLOGRAMOFSPEEDANDCURRENTWITHMODEL
REFERENCEPARAMETERADAPTIVESPEEDCONTROL

FIG.9OSCILLOGRAMOFSPEEDANDCURRENTWITHMODEL
REFERENCESIGNALADAPTIVESPEEDCONTROL
Conc l usi on
To provide constant loop gain in speed control loop
with changing parameters (moment of inertia and/or
torque factor), parameter and signal adaptive model
referenceadaptivecontrolweredeveloped.
Following from the structure (PFtype) of model
InternationalJournalofAutomationandPowerEngineering(IJAPE)Volume2Issue4,May2013www.ijape.org
73
reference parameter adaptive control was developed
to provide constant loop gain in speed control loop
with changing gain (moment of inertia and/or torque
factor)whichmakesiteasiertoreachnonovershooting
step response as well as fast speed changing
compensation caused by jump in load. The algorithm
evenkeepsitsstabilityatfastchanging,jumplikeload
torque.
Model reference signal adaptive control is used to
provideconstantloopgaininspeedcontrolloopwith
changingparameters(momentofinertiaand/ortorque
factor) exposed to a significant load. The approaching
block diagram of the adaptive control can be seen as
anextendedversionofthePFcontroller,sooneofthe
adaptation factors (which is the free parameter of the
adaptivecontrol)isgiven.
The adaptive controls suggested in this paper work
without overshooting. Though these methods require
a longer calculation period, it is less sensitive to the
variations of parameters. Both model reference
adaptivecontrolsdrawnupcanbeeasilyimplemented,
because the adaptation algorithms do not need
acceleration measuring (thanks to the firstorder
model). Simulation and experimental results
demonstratethattheproposedmethodsarepromising
toolstospeedcontrolofelectricaldrives.
REFERENCES
Borka, Jozsef et al., Control aspects of switched reluctance
motordrives,ISIE93,IEEEInternationalSymposiumon
IndustrialElectronics,Budapest(Hungary),296300,Juni
13,1993.
Bose B. K. and Miller, T.G.E. Microcomputer Control of
Switched Reluctance Motor, IEEE/IAS Paper presented
attheannualmeeting,542547,1985.
Diep, N.V., and Szamel, L. Uptodate Control Strategy in
the Regulators of Robot Drives, PEMC90, Budapest,
Hungary,811815,October13,1990.
LiuHsuetal.,Lyapunov/PassivityBasedAdaptiveControl
of Relative Degree Two MIMO Systems With an
Application to Visual Servoing, IEEE Transactions on
AutomaticControl,vol.52,364371,Feb.2007.
Ohm,D.Y.APDFFControllerforTrackingandRegulation
in Motion Control, Proceedings of 18th PCIM
Conference, Intelligent Motion, Philadelphia, 2636,
October2126,1990.
Perdikaris,G.A.andVanPatten,K.W.ComputerSchemes
for Modeling, Tuning and Control of DC Motor Drive
Systems,PCIProc.,8396,Mar.1982.
Phelan, R. M. Automatic Control Systems, Cornell
UniversityPress,NewYork,1977.
Szamel, Laszlo. Ripple reduced control of switched
reluctance motor drives, EDPE 2001, International
Conference on Electrical Drives and Power Electronics,
Podbanske(Slovakia),4853,October35,2001.
Szamel, Laszlo. Model Reference Adaptive Control of
Ripple Reduced SRM Drives, Periodica Polytechnica
ElectricalEngineeringvol.46.,No.34,163174,2002.
Szamel,Laszlo.InvestigationofModelReferenceParameter
AdaptiveSRMDrives,EPEPEMC2004,11thInt.Power
ElectronicsandMotionControlConf.,Riga(Latvia),Full
PaperNo.A95117,September24,2004.
Vajda, Istvan et al., Investigation of Joint Operation of a
Superconducting Kinetic Energy Storage (Flywheel) and
Solar Cells, IEEE Transactions on Applied
Superconductivity,vol.13,No.2,21692172,June2003.
Vujici, V.P. Minimization of Torque Ripple and Copper
LossesinSwitchedReluctanceDrive,IEEETransactions
onPowerElectronics,Vol.27,388399,2012.

LaszloSzamelwasborninBudapest,in
1964. He received the M.Sc., dr. univ.
and Ph.D. degrees in electrical
engineering from Technical University
of Budapest in 1988, 1995 and 2005,
respectively.
Until1997heworkedforComputerand
Automation Research Institute of
Hungarian Academy of Science, since then he is with
Budapest University of Technology and Economics. His
educational and research fields are electrical drive control,
servo and robot drives, digital systems, switched reluctance
motor drives, highaccuracy position and speed
measurementsofelectricaldrives.
Dr. Laszlo Szamel is associate professor, a member of the
Hungarian Electrotechnical Association and several
Committees of Hungarian Academy of Science.