Sei sulla pagina 1di 6

Modeling of Line-Starting of Reluctance

Synchronous Motors Considering Magnetic


Saturation, with Experimental Validation
Sorin Enache, Aurel Campeanu, Ion Vlad, Monica-Adela Enache and Mircea Dobriceanu
University of Craiova, Faculty of Electrical Engineering
senache@em.ucv.ro
Abstract-This paper presents aspects regarding the saturation
influence upon the operation of the reluctance synchronous
motors in case of some representative dynamic regimes. In this
purpose, there have been detailed the mathematical model and
the Matlab model conceived with its help. There are also
presented the simulations obtained for a few concrete situations.
The paper ends with experimental results and conclusions.

I.

saturated synchronous motors are detailed in the papers [4],


[9], based on some large and present references.
Further on, there is tackled, starting from the paper [2], a
modern manner of including the effects of the saturation in
the computations, in the two-axes theory (fig. 1), by using the
equations with representative phasors, written with the
components of the current (id, iq, iE, iD, iQ).

INTRODUCTION

The reluctance synchronous machine is more and more


used in present drives of high performance rated at low and
middle power, fact reflected in many reference papers [1],
[10]. In these conditions the pre-determination of the dynamic
regimes, together with the steady state ones, by modelling
and simulation, becomes a compulsory stage of the design
process.
The mathematical models of this machine are established
with shades of difference because there must be considered,
in connection, the effects of the magnetic saturation when
considering the fundamental harmonic of the magnetic field
and the magnetic asymmetry.
Recent researches [2], [8] show that it is possible to accept
for synchronous machines, in a relative large range of power
and speed, the same level of saturation in the d and q axes;
this way, it is possible to use exclusively the theory of the
circuits.
In what follows, there will be presented, starting from the
general case of a synchronous machine, a mathematical
model which takes account of the magnetic saturation for
reluctance synchronous machines.

Fig. 1. Equivalent circuit of the synchronous machine in the two-axes theory.

In these conditions the voltage equations, particularized for


a synchronous motor, can be written in the following matrix
form:
d
U = A
X +BX
(1)
dt
where:

U = ud

uq

X = id

iq

uE

0 0

II. MATHEMATICAL MODEL WITH SATURATION


The present command methods of the synchronous
machine, used in automatic systems, which need execution
precision and high speed of response, inevitably involve the
estimation of the magnetic stresses.
In this respect, the specialty literature has been trying to
catch the saturation by computation, for the analysis of the
steady sates and dynamic regimes.
Thus, in [3] there are reviewed the efforts of considering
the saturation in the equations of the synchronous machine.
At the same time, some aspects regarding the command of the

978-1-4799-5183-3/14/$31.00 '

2014 IEEE

324

A=

iE

iD

iQ

(2)

(3)

Ls + Lmdt

Lmdt

Lmdt

Ls + Lmq

Lmq

Lmdt

LE + Lmdt

Lmdt

Lmdt

Lmdt

LD + Lmdt

Lmq

LQ + Lmq
(4)

( Ls + Lmq )

Rs

( Ls + Lmd )
B=

Lmq

Rs

Lmd

Lmd

RE

RD

RQ

Ls + Lmdt
0
A=
Lmdt
0

J d
p dt

(6)

0
Lmq
(15)
0
LQ + Lmq
Lmq
0
0

(16)

RQ

The motion equation written in the form (6) is added to the


equation (1), where:

where:

3
m = p[( Lmd Lmq )id iq + Lmd (i E + i D )iq Lmq id iQ ] (7)
2

Lmdt
0
LD + Lmdt
0

0
( Ls + Lmq )
Rs
( Ls + Lmd )
Lmd
Rs
B=
0
0
RD
0
0
0

(5)
The motion equation written in the following form is added
to the equation (1):
m mr =

0
Ls + Lmq
0
Lmq

m=

3
p[( Lmd Lmq )id iq + Lmd iD iq Lmq id iQ ]
2

(17)

with ud and uq given by (8).

On the othe hand:


u d = 2U cos(t B ) i u q = 2U sin(t B ) , (8)

where B = dt + 0 [2].
The speed of the magnetization field is given by the
relation:

d
+ ,
dt

(9)

III. MATLAB MODEL WITH SATURATION


With the help of the previous equations there has been
obtained the variant of the Matlab [7], [12] model for the
RSM when considering the saturation.
In this purpose, there has been used a s-function file,
detailed in what follows.
function [sys, x0] =
msrvs_sat(t,x,u,flag,cd,lambdad,cdt,lambdadt,cq,lambdaq)
% Input (u)

where:
d
1
=
dt m

d mq

d md

dt cos dt sin ,

(10)

respectively:

mq

sin =
, cos = md ,
m
m

(11)

with:
2
2
m = md
+ mq
.

(12)

RSM has not got excitation winding.


In these conditions the voltage equations, particularized for
a RSM, can be written in matrix form (1), where:

U = ud
X = id

uq
iq

0 0

iD

iQ

(13)
(14)

% 1 usd
% 2 usq
% 3 omg (rotor speed)
%
%

Output (x)
1 id
2 iq
3 iD
4 iQ
5 me

if (nargin==0), sys=[4,0,5,3,0,0]; x0 = zeros(4,1); return, end


if (abs(flag)==1)+(abs(flag)==3) >= 1
Rs=3.77;
% stator resistance
RD=1.5;
% D axis resistance
RQ=4.5;
% Q axis resistance
Lssigma=0.0081;
% Ls_sigma inductance
LDsigma=0.0059;
% LD_sigma inductance
LQsigma=0.0067;
% LQ_sigma inductance
p=2;
% number of the poles pairs
Lmd_min=0.199; Lmd_max=0.45;
Lmdt_min=0.052; Lmdt_max=0.45;
Lmq_min=0.071; Lmq_max=0.078;
imdc=abs(x(1)+x(3));
imdc=imdc*(imdc>=1e-6)+1e-6*(imdc<1e-6);
imqc=abs(x(2));
imqc=imqc*(imqc>=1e-6)+1e-6*(imqc<1e-6);
Lmd=cd(1)*exp(-imdc*lambdad(1))...
+cd(2)*exp(-imdc*lambdad(2))...

325

else sys=[];

+cd(3)*exp(-imdc*lambdad(3))...
+cd(4)*exp(-imdc*lambdad(4))...
+cd(5)*exp(-imdc*lambdad(5))...
+cd(6)*exp(-imdc*lambdad(6))...
+cd(7)*exp(-imdc*lambdad(7))...
+cd(8)*exp(-imdc*lambdad(8));
Lmd=Lmd*((Lmd>Lmd_min)*(Lmd<Lmd_max))+
Lmd_min*(Lmd<Lmd_min)+Lmd_max*(Lmd>Lmd_max);
Lmdt=cdt(1)*exp(-imdc*lambdadt(1))...
+cdt(2)*exp(-imdc*lambdadt(2))...
+cdt(3)*exp(-imdc*lambdadt(3))...
+cdt(4)*exp(-imdc*lambdadt(4))...
+cdt(5)*exp(-imdc*lambdadt(5))...
+cdt(6)*exp(-imdc*lambdadt(6))...
+cdt(7)*exp(-imdc*lambdadt(7))...
+cdt(8)*exp(-imdc*lambdadt(8))...
+cdt(9)*exp(-imdc*lambdadt(9))...
+cdt(10)*exp(-imdc*lambdadt(10));
Lmdt=Lmdt*((Lmdt>Lmdt_min)*(Lmdt<Lmdt_ma
x))+Lmdt_min*(Lmdt<Lmdt_min)+Lmdt_max*(Lmdt>Lmdt
_max);
Lmq=cq(1)*exp(-imqc*lambdaq(1))...
+cq(2)*exp(-imqc*lambdaq(2))...
+cq(3)*exp(-imqc*lambdaq(3))...
+cq(4)*exp(-imqc*lambdaq(4))...
+cq(5)*exp(-imqc*lambdaq(5))...
+cq(6)*exp(-imqc*lambdaq(6))...
+cq(7)*exp(-imqc*lambdaq(7))...
+cq(8)*exp(-imqc*lambdaq(8));
Lmq=Lmq*((Lmq>Lmq_min)*(Lmq<Lmq_max))+
Lmq_min*(Lmq<Lmq_min)+Lmq_max*(Lmq>Lmq_max);
end

end
Before running this model, it is necessary to know the
values of the inductances Lmd, Lmdt and Lmq. In this purpose,
there have been conceived another three Matlab programs
which, starting from the data experimentally obtained,
approximate the curves corresponding to the three
inductances by polynomial developments.
For exemplification, further on there is presented the
content of the file corresponding to the inductance Lmdt.
% LMDT Establishes the coefficients and the exponents of
the development with ten exponentials for Lmdt.
%
% y =
c(1)*exp(-lambda(1)*t) + ... + c(10)*exp(lambda(10)*t)
%
clf
echo off
warning off MATLAB:fmins:ObsoleteFunction
% Experimental points for Lmdt. Column 1 = imd, column 2
= Lmdt
global datadt
load datadt;
t = datadt(:,1);
y = datadt(:,2);
axis([0 max(t) 0 max(y)])
hold on
grid on
plot(t,y,'ko','EraseMode','none')
title('Lmdt experimental')
xlabel('imd [A]')
ylabel('Lmdt [H]')

if abs(flag)==1 % (States derivative)


mR=[Rs -p*u(3)*(Lssigma+Lmq) 0 -p*u(3)*Lmq;...
p*u(3)*(Lssigma+Lmd) Rs p*u(3)*Lmd 0;...
0 0 RD 0;...
0 0 0 RQ];

global Plothandle
Plothandle = plot(t,y,'r-','EraseMode','xor');

mL=[Lssigma+Lmdt 0 Lmdt 0;...


0 Lssigma+Lmq 0 Lmq;...
Lmdt 0 LDsigma+Lmdt 0;...
0 Lmq 0 LQsigma+Lmq];
invmL=inv(mL);
uu=[u(1);u(2);0;0];
sys=-invmL*mR*x+invmL*uu;
return
elseif abs(flag)==3 % (Iesiri isd, isq, irD, irQ, me)
%
sys=[x(1);x(2);x(3);x(4);1.5*p*(Lmd*(x(1)+x(3))*x(2)Lmq*x(2)*x(1))];
sys=[x(1);x(2);x(3);x(4);1.5 * p * ((Lmd - Lmq) * x(1) * x(2)
+ Lmd * x(3) * x(2) - Lmq * x(1) * x(4))];
elseif abs(flag)==0
sys=[4,0,5,3,0,0]; x0 = zeros(4,1);

lam = [4.4342 4.434 238.3 4.434 -0.18623 -0.19273 0.071536 -0.20015 0.0024174 -0.0034707]';
trace = 0;
tol = .00001;
lambda = fmins('lmdterr',lam,[trace tol]);
grid on
cdt=lmdtc(lambda);
lambdadt=lambda;
hold off
s = ['lambda = fmins(''lmdterr'',[1 1 1 1 1 1 1 1 1 1]'',[0
.0001]);'];
uicontrol('Units','normal','Position',[.2 .01 .1 .06],
'String','Again','callback',s)
% Lmdt_max=max(y);
% Lmdt_min=min(y);
clear datadt

326

clear Plothandle
clear lam
clear lambda
clear s
clear t
clear tol
clear trace
clear y
pack
clc;
disp('End')
IV. SIMULATIONS

Fig. 4. Dependence Lmdt=f(imd).

The Matlab model detailed before has been used to


emphasize the influences of some parameters upon the
dynamic characteristics of the motors studied.
In this purpose there have been performed a lot of
simulations for a representative dynamic regime:
asynchronous starting.
The motor studied here has the following parameters:
Rs=3,77 ; RD=1,5 ; RQ=4,5 ; Ls=0,0081 H; LD=0,0059
H; LQ=0,0067 H; p=2; J=0,005 kg m2. The cross section of
the rotor is presented in the Fig. 2.
Fig. 5. Variation Lmq=f(imq).

The following graphic representations have been obtained


by performing the simulations.

Fig. 2. The cross section of the rotor.

The graphic dependences of the inductances Lmd, Lmdt and


Lmq, experimentally established [5], in function of the
magnetization current are given in figures 3, 4 and 5.

a)Real parameters (RD=1,5 ; RQ=4,5 ; Ls=0,0081 H; J=0,005 kg m2) for


Mr=0.

Fig. 3. Variation Lmd=f(imd).

b)RD=3

327

c)RQ=2,25

b)RD=3

d)Ls=0,012 H

c)RQ=2,25

e)J=0,05 kg m2

d)Ls=0,012 H

Fig. 6. Characteristics iA=f(t) for different values of the RSM parameters.

a) Real parameters (RD=1,5 ; RQ=4,5 ; Ls=0,0081 H; J=0,005 kg m2) for


Mr=0.

e)J=0,05 kg m2
Fig. 7. Characteristics =f(t) for different values of the RSM parameters.

328

V. EXPERIMENTAL DETERMINATIONS
For monitoring the dynamic regimes presented before,
there has been conceived a data acquisition system detailed in
[6], achieved with a board KPCI [11].
A global view of this circuit is presented in figure 8.
By monitoring an asynchronous starting on no-load the
representations from figures 9 and 10 have been obtained.

can be concluded that the experimental determinations


confirm the validity of the simulations performed in dynamic
regime.
VI. CONCLUSIONS
The analysis of the graphics presented before has led to the
following conclusions:
- the increase of the resistance RD has a non-stabilizing
effect, materialized, in the case of its increase, in an increase
of the transient regime duration;
- a low value of the resistance RQ, even in the case of a null
resistant torque and a low inertia moment, can lead to an
unstable operation; in the case analyzed here, its decrease
leads to the increase of the transient regime duration (fact
which is visible especially on the speed curve);
- the increase of the Ls has a weak stabilizing effect;
- the increase of the inertia moment makes the
synchronization to be reached after a great number of
oscillations of the current.
ACKNOWLEDGMENT
This work was partially supported by the grant number
30C/27.01.2014, awarded in the internal grant competition of
the University of Craiova.

Fig. 8. General view of the computerized micro-bench.

REFERENCES
[1]

Fig.9. Experimental variation of the current.

Fig. 10. Experimental variation of the speed.

By comparing these graphic results to the results obtained


by simulation (fig. 9 to 6.a, fig. 10 to fig. 7.a, respectively), it

I. Boldea, Reluctance Synchronous Machines and Drives. Clarendon


Press, Oxford, 1996.
[2] A. Campeanu, I. Cauil, I. Vlad, S. Enache, Modelling and simulation
of the alternating current electrical machines, Publishing House of the
Romanian Academy, Bucharest, 2012, ISBN 978-973-27-2233-6.
[3] A. Campeanu, Nonliniar dynamic models of saturated salient pole
synchronous machine, Revue Roum. Sci. Tech. Electrotechn. et
Energ., 47, 3, 2002, pp. 307-317.
[4] A. M. El-Serafi, Factors Affecting the Accuracy of Determining the
Saturated Synchronous Machines Cross-Magnetizing Parameters by the
Finite-Element Method, Proc. of IEEE ICEM 2008, Villamoura,
Portugal.
[5] Enache, M.A., Enache, S., Dobriceanu, M., Considerations regarding
dynamic regimes of reluctance synchronous motors, 19-th
International Symposium on Power Electronics, Electrical Drives,
Automation and Motion SPEEDAM 2008, Ischia (Italy), 11-13 June,
pp. 284-289.
[6] S. Enache, A. Campeanu, I. Vlad, M.A. Enache, Method and device
of synchronization for determination of squirrel cage asynchronous
motors parameters, 8th International Symposium on Advanced Topics
in Electrical Engineering (ATEE), May 23-25, 2013, Bucharest,
Romania.
[7] M. Ghinea, V. Fireteanu, MATLAB. Numerical computation graphics
- applications, Teora Publishing House, 1995.
[8] E. Levi, Saturation modeling in D-Q axis models of salient pole
synchronous machines, IEEE Trans, on Energy Conversion, 14, 1, 1999,
pp. 44-50.
[9] E. Schlemmer, G. Hofner, Finite-Element Investigations of the
Transient Behaviour of Salient pole Synchronous Machine with
Different Dumper and Connections, Proc. of IEEE ICEM 2008,
Villamoura, Portugal.
[10] A. Vagati, Handbook of Electric Motors - Synchronous Reluctance
Motors. Marcel DEKKER Inc. (USA), Vol.1 ISBN:0-8247-4105-6,
2004.
[11] ***, KPCI 3101-3104 Series PCI Bus Data Acquisition Board Users Manual. Keithley Metrabyte, 2007.
[12] ***, Matlab User's Guide, The MathWorks, Inc., Natick, Massachusetts,
2008.

329

Potrebbero piacerti anche