Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
SILVEIRA, A. W. F. V., ANDRADE, D. A., FLEURY, A., GOMES, L. C., BISSOCHI, C. A., DIAS, R. J.
Laboratrio de Acionamentos Eltricos, FEELT, Universidade Federal de Uberlndia
Av. Joo Naves de vila, N 2121, Bloco 3N, Uberlndia-MG
E-mails: gutofleury@gmail.com,darizon@ufu.br
Abstract This paper deals with the switched reluctance motor/generator drive. An alternative half-bridge electronic converter to-
pology is used to attain motor operation mode. To change from motor operational mode to generation mode the electronic converter
structure is strategically modified through a relay switch, allowing the flux of power go to the load. Simulation and experimental re-
sults obtained from a 6x4 switched reluctance machine prototype operating as a motor/generator are presented. They show the beha-
vior of the electrical machine during the transition between both operational modes. The DSP based control is successfully tested. Re-
sults presented here sustain the applicability of the switched reluctance machine as a motor/generator solution for general application.
Keywords Switched reluctance motor/generator, half-bridge converter, digital signal processor.
Resumo Este artigo trata do acionamento da mquina a relutncia varivel operando como motor/gerador. Um conversor eletr-
nico de meia ponte (half-bridge) com alteraes foi utilizado para permitir os dois modos de operao. Durante a comutao de mo-
tor para gerador, ou vice-versa, o circuito de desmagnetizao modificado com a utilizao de um rel de potncia. Isso permite que
durante a operao motora a energia armazenada nas bobinas das fases seja devolvida para o link cc na etapa de desmagnetizao,
porm, durante a operao de gerao, o caminho desviado para a carga resistiva. Resultados de simulao de uma mquina 6x4,
trifsica so apresentados e discutidos. Experimentalmente, o sistema foi montado utilizando DSP e os resultados so mostrados neste
trabalho, dando sustento a aplicabilidade da mquina a relutncia varivel como motor/gerador para aplicaes industriais diversas.
Palavras-chave Motor/gerador a relutncia varivel, conversor de meia ponte, processador digital de sinais.
1 Introduo
A crescente necessidade de otimizar processos, tor-
nando-os mais econmicos, tem levado ao desenvol-
vimento de acionamentos eletrnicos para mquinas
eltricas capazes de control-las como mo-
tor/gerador.
Na indstria automobilstica o motor de partida
e o alternador podem ser substitudos por uma nica
mquina eltrica, o que economizaria materiais e
reduziria o peso e o custo do sistema (Fahimi, 2004).
Essas vantagens tambm se aplicam industria ae-
ronutica, onde tambm so usadas mquinas eltri-
cas para partir turbinas e gerar eletricidade (Mac-
Minn, 1989). Alm de permitirem economia no pro-
cesso de fabricao, o desenvolvimento de moto-
res/geradores para aplicao automotiva permitem a
implementao de novas tecnologias que visam
reduo da poluio emitida pelos veculos, tais co-
mo (Fahimi, 2004), (Kassakian, 1996), (Miller,
1999):
Desligar o motor a combusto toda vs que
o veculo estiver parado em sinaleiros ou
congestionamentos, partindo o mesmo
quando for acelerado.
Desenvolvimento de veculos hbridos le-
ves.
Aumentar a capacidade de gerao, suprin-
do a crescente demanda por potncia eltri-
ca dos veculos mais modernos.
Visando desenvolver a tecnologia dos moto-
res/geradores algumas mquinas eltricas tm sido
exploradas atravs de experimentos. Em (Cai, 2004),
um estudo comparativo entre as principais mquinas
candidatas aplicaes como motor/gerador apre-
sentado. Neste estudo a mquina a relutncia vari-
vel (MRV) se mostrou uma forte candidata pelas
suas caractersticas construtivas, que permitem, a
mesma, operar em altas velocidades (de Andrade,
2007), com uma ampla faixa de variao de veloci-
dade e com um controle simples, quando comparada
mquina de induo. Alm disso, uma mquina
robusta e sua confeco simples e barata.
O trabalho apresentado em (Fleury, 2008), mos-
trou que a MRV, projetada para este estudo, operan-
do como gerador, apresenta a maior potncia entre-
gue a carga com a mquina operando a 1300 rpm, e
que aps atingir este pico, o valor da potncia entre-
gue a carga, diminuiu lentamente com o aumento da
velocidade. Esta caracterstica importante para
aplicaes automotivas j que o motor a combusto
opera em regime de velocidade varivel, oscilando
entre 600 rpm a 6000 rpm (1:10).
Dentro deste contexto, este artigo apresenta um
estudo de uma MRV com seis plos no estator e qua-
tro no rotor (6x4), sendo acionada por um conversor
de meia ponte acrescido de um rel para modificao
do circuito de desmagnetizao. O controle do mo-
tor/gerador a relutncia varivel (MGRV) foi simu-
lado utilizando o software MATLAB/SIMULINK
e
testado experimentalmente utilizando uma platafor-
ma DSP. Os resultados obtidos sero apresentados e
discutidos no decorrer do artigo.
4898
XVIII Congresso Brasileiro de Automtica / 12 a 16-setembro-2010, Bonito-MS
2 Motor/Gerador a Relutncia Varivel
A mquina a relutncia possui enrolamentos das
fases nas salincias do estator. A ausncia de enro-
lamento nas salincias do rotor permite que altas
velocidades sejam alcanadas (Henriques, 2003). A
figura 1 mostra uma representao de uma MRV
com um dos enrolamentos de fase presente.
Com relao ao funcionamento da mquina, se
um plo do rotor se alinha com o plo energizado do
estator, a posio de equilbrio estvel. Assim, na
mquina a relutncia existe uma tendncia natural
de a parte mvel permanecer na posio de indutn-
cia mxima da bobina excitada. Se o rotor encontra-
se em posio de equilbrio instvel em relao a
uma determinada fase, e esta energizada, o rotor
tender a girar para a posio de equilbrio, caracte-
rizando uma operao motora. Agora, se da posio
de equilbrio estvel, o rotor forado a girar por um
agente mecnico, o torque produzido restaurador e
resulta em fora contra-eletromotriz aditiva tenso
aplicada, e a mquina gera energia eltrica. Em uma
mquina a relutncia varivel, a energia mecnica
recebida de uma mquina primria transformada
em energia eltrica forando o desalinho entre o
plo do rotor e o plo energizado do estator. Pela
figura 2 possvel observar as regies em relao
variao da indutncia, de uma das fases, em que a
MGRV opera como motor ou gerador.
Fig.1. Mquina a relutncia varivel 6/4.
Fig.2. Variao da indutncia de uma fase do MGRV.
A. Modelagem matemtica
O circuito de uma fase do MGRV pode ser equa-
cionado como:
d
dL
i
dt
di
L Ri v + + =
(1)
onde v a tenso aplicada, i a corrente da fase, R
a resistncia da fase, L a indutncia da fase e a
posio do rotor. O terceiro termo do lado direito da
igualdade a fora contra-eletromotriz e, que isola-
damente pode ser escrita como:
d
dL
i e =
(2)
onde =d/dt a velocidade angular do rotor.
O conjugado mecnico produzido pela MGRV,
desconsiderando as perdas para efeito de anlise,
pode ser expresso por (3).
( )
d
dL
i i T
2
2
1
, =
(3)
Algumas concluses podem ser feitas a partir da
equao acima: O conjugado mecnico produzido
pela mquina independente do sinal da corrente
que circula na fase, ento a corrente aplicada na fase
pode ser unidirecional. Para se obter o conjugado
necessrio o conhecimento da corrente e da variao
da indutncia em relao posio do rotor dL/d.
Para realizao da modelagem matemtica usada
no programa de simulao, o conjugado mecnico
produzido pela mquina foi calculado levando em
considerao as perdas por atrito viscoso D e mo-
mento de inrcia J, conforme apresentado pela equa-
o (4).
D
dt
d
J T T
emag
=
(4)
Considerando trs fases com indutncias e corren-
tes instantneas diferentes, o conjugado eletromag-
ntico dado por:
|
\
|
+ + =
d
dL
i
d
dL
i
d
dL
i T
c
c
b
b
a
a emag
2 2 2
2
1
(5)
A equao de velocidade do rotor (6) completa a
descrio dinmica da mquina.
dt
d
=
(6)
O modelo matemtico do MGRV, considerando
as trs fases, apresentado por (7).
1 2 3
0 0 0 0
0 0 0 0
0 0 0 0
0
0 0 0 0 1 0
0 0 0
0 0 0
0 0 0
0 0 0 0
0 0 0 0 1
a a a
b b b
c c c
m a b c
a
a a
a
b
b
b b
c
c
c c
v R i
v R i
v R i
C i r i r i r D
dL
L i
i
d
dL
i
L i
d
i
dL
L i
d
J
( ( (
( ( (
( ( (
( ( ( = +
( ( (
( ( (
( ( (
(
(
(
(
(
+ (
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(7)
onde:
d
dL
r
d
dL
r
d
dL
r
c b a
2
1
;
2
1
;
2
1
3 2 1
= = =
(8)
Designando por [V], [R], [I], [L] e [
I ] as matri-
zes na ordem em que aparecem em (6), a matriz de
estados do MGRV tem a seguinte forma:
] ][ [ ] [ ] [ ] [ ] [
1 1
I R L V L I
= (9)
4899
XVIII Congresso Brasileiro de Automtica / 12 a 16-setembro-2010, Bonito-MS
Fig. 3. Desenho esquemtico do conversor de meia ponte utilizado no acionamento do MGRV.
B. Modelagem computacional
O programa Matlab/Simulink
o
(
V
)
0.5 0.52 0.54 0.56 0.58 0.6
-40
-20
0
20
Tempo (s)
T
e
n
s
o
e
m
u
m
a
f
a
s
e
(
V
)
Fig. 6-c. Tenso gerada pela
MRV.
Fig. 6-d. Tenso em uma das fases.
0 5 10 15
0
0.05
0.1
0.15
0.2
0.25
Corrente (A)
F
lu
x
o
(
W
b
)
0 0.5 1 1.5
0
50
100
150
Tempo (s)
P
o
t
n
c
i
a
(
W
)
Fig. 6-e. Fluxo v.s. corrente. Fig. 6-f. Potncia de entrada e
potncia de sada.
0.022 0.023 0.024 0.025 0.026 0.027 0.028 0.029 0.03
0
2
4
6
8
10
12
14
16
Tempo (s)
c
o
r
r
e
n
t
e
(
A
)
Fig. 7. Variao da amplitude da corrente de uma fase em relao ao
valor do ngulo de abertura das chaves (off).
3 Resultados Experimentais
A MRV, 6x4, trifsica de 1 cv, descrita na tabela 1,
foi acoplada a um motor de induo trifsico de 4
plos, 2 cv, acionado por um inversor comercial,
para a realizao dos testes experimentais. O con-
versor de meia ponte apresentado na Fig 3 foi mon-
tado para a realizao dos testes e um DSP modelo
TMS320F2812 foi utilizado para controlar a MRV
como motor/gerador. Para obteno da posio rot-
rica, necessria para o acionamento do MGRV, fo-
ram usados sensores pticos associados a dois discos
ranhurados (Fig. 10), um para motor (S1) e um para
gerador (S2), que indicam o momento certo de mag-
netizar cada uma das fases na etapa motora e gera-
dora. A Fig. 8 mostra o diagrama de blocos da ban-
cada de experimentos e a Fig. 9 apresenta a foto da
bancada de testes.
Os testes experimentais foram efetuados com a
mquina primria (MIT) girando no mesmo sentido
do MGRV, operando como motor. Esta situao
semelhante a que ocorre quando o motor/gerador de
um automvel parte o motor a combusto e continua
sendo acionado como motor momentos antes de seu
modo de operao ser comutado para gerador.
Fig. 8. Diagrama de blocos do sistema montado para experimentos.
Fig. 9. Foto da bancada de testes montada no laboratrio.
4901
XVIII Congresso Brasileiro de Automtica / 12 a 16-setembro-2010, Bonito-MS
Os testes foram realizados com o MGRV ope-
rando a 1350 rpm. A comutao de motor para ge-
rador ou vice-versa controlada, pelo usurio que
utiliza computador atravs do software Code Com-
poser
d
dL
i
dt
di
L Ri v + + =
(1)
onde v a tenso aplicada, i a corrente da fase, R
a resistncia da fase, L a indutncia da fase e
a posio do rotor. O terceiro termo do lado di-
reito da igualdade a fora contra-eletromotriz e,
que isoladamente pode ser escrita como:
d
dL
i e =
(2)
onde =d/dt a velocidade angular do rotor.
O conjugado mecnico produzido pela GRV,
desconsiderando as perdas para efeito de anlise,
pode ser expresso por (3).
( )
d
dL
i i T
2
2
1
, =
(3)
Algumas concluses podem ser feitas a partir da
equao acima: O conjugado mecnico produzido
pela mquina independente do sinal da corrente
que circula na fase, ento a corrente aplicada na
fase pode ser unidirecional. O sinal do conjugado
determinado pela fase da corrente em relao
variao da indutncia dL/d.
Para realizao da modelagem matemtica usa-
da no programa de simulao, o conjugado mec-
nico produzido pela mquina foi calculado levando
em considerao as perdas por atrito viscoso D e
momento de inrcia J, conforme apresentado pela
equao (4).
D
dt
d
J T T
emag
=
(4)
Considerando trs fases com indutncias e cor-
rentes instantneas diferentes, o conjugado eletro-
magntico dado por:
|
\
|
+ + =
d
dL
i
d
dL
i
d
dL
i T
c
c
b
b
a
a emag
2 2 2
2
1
(5)
A equao de velocidade do rotor (6) completa a
descrio dinmica da mquina.
dt
d
=
(6)
O modelo matemtico do GRV, considerando as
trs fases, apresentado por (7).
2615
XVIII Congresso Brasileiro de Automtica / 12 a 16-setembro-2010, Bonito-MS
1 2 3
0 0 0 0
0 0 0 0
0 0 0 0
0
0 0 0 0 1 0
0 0 0
0 0 0
0 0 0
0 0 0 0
0 0 0 0 1
a a a
b b b
c c c
m a b c
a
a a
a
b
b
b b
c
c
c c
v R i
v R i
v R i
C i r i r i r D
dL
L i
i
d
dL
i
L i
d
i
dL
L i
d
J
( ( (
( ( (
( ( (
( ( ( = +
( ( (
( ( (
( ( (
(
(
(
(
(
+ (
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(7)
onde:
d
dL
r
d
dL
r
d
dL
r
c b a
2
1
;
2
1
;
2
1
3 2 1
= = =
(8)
Designando por [V], [R], [I], [L] e [
I ] as ma-
trizes na ordem em que aparecem em (6), a matriz
de estados do GRV tem a seguinte forma:
] ][ [ ] [ ] [ ] [ ] [
1 1
I R L V L I
=
(9)
2.2 Modelagem Computacional
O programa Matlab/Simulink
o
n
a
f
a
s
e
(
V
)
Fig. 10. Tenso em uma das fases.
1 2 3 4 5 6 7 8
0
10
20
30
40
50
60
70
Tempo (s)
T
e
n
s
o
(
V
)
Fig. 11. Curva de tenso na carga durante o teste de transitrio de
carga.
2617
XVIII Congresso Brasileiro de Automtica / 12 a 16-setembro-2010, Bonito-MS
4 Resultados Experimentais
Fig. 12. Foto da bancada de testes montada no laboratrio.
Com intuito de validar experimentalmente a estra-
tgia de controle apresentada, o GRV descrito an-
teriormente foi acoplado a um motor de induo
trifsico de 2 cv, 4 plos, acionado por um conver-
sor de freqncia comercial. O conversor HB foi
construdo para acionar o GRV e a estratgia de
controle foi programada para ser executada em um
DSP TMS320F2812, utilizado no sistema. A posi-
o do rotor, necessria para a aplicao dos sinais
de gatilho durante o perodo de magnetizao das
fases, foi obtida utilizando sensores pticos associ-
ados a um disco, que representa o exato instante
em que cada fase deve ser magnetizada. A figura
12 apresenta uma foto da bancada de testes monta-
da no laboratrio.
O controlador PI, utilizado na estratgia de-
senvolvida, foi sintonizado com facilidade por ten-
tativa e erro, e os coeficientes empregados foram:
K
p
= 2 e K
i
= 0.8.
Um teste de transitrio de carga foi realizado,
onde a resistncia de carga foi reduzida de 20
para 15 , conforme foi realizado em simulao. A
referncia de tenso na carga foi configurada como
sendo de 42 V. Neste teste o GRV operou a 1350
rpm e pela figura 13 pode ser observado o sinal de
corrente, que aumenta de amplitude aps a reduo
da carga, e o sinal de tenso, que apresenta uma
rpida depresso em seu valor logo aps o transit-
rio, voltando ao valor de referncia instantes de-
pois.
T
T
1 > 1 > 1 > 1 >
1 > 1 > 1 > 1 >
2 > 2 > 2 > 2 >
2 > 2 > 2 > 2 >
1) Ch 1: 2 V 500 ms
2) Ch 2: 20 V 500 ms
Fig. 13. Corrente em uma fase (escala: 1 V - 4 A) e tenso na
carga para transitrio de resistncia de carga.
TT
1 > 1 > 1 > 1 >
1 > 1 > 1 > 1 >
2 > 2 > 2 > 2 >
2 > 2 > 2 > 2 >
1) Ch 1: 10 V 500 us
2) Ch 2: 500 mV 500 us
Fig. 14. Curva de corrente (escala: 1 V - 4 A) e sinal de gatilho.
O sinal de corrente em uma das fases e o respec-
tivo sinal de gatilho aplicado chave superior do
conversor referente a esta fase, podem ser observa-
dos pela figura 14, j a figura 15, mostra o sinal de
tenso e gatilho de uma das fases para a mesma
situao.
Foi realizado um teste de transitrio de refern-
cia para o controlador. Inicialmente a referncia
foi configurada com 42 V e depois modificada ins-
tantaneamente para 30 V. A figura 15 mostra o
resultado deste teste, onde possvel verificar o
comportamento da corrente, que reduz sua ampli-
tude devido ao transitrio, e o sinal de tenso na
carga, que converge para o valor de referncia a-
plicado ao controlador.
Um outro teste foi realizado, agora com transi-
trio de velocidade. O GRV controlado em malha
fechada com referencia de 42 V foi acelerado de
800 rpm para 1800 rpm. A figura 16 apresenta o
resultado, por onde pode-se observar o sinal de
tenso, que permaneceu com valor em torno da
referncia, e o sinal de corrente, que reduz de am-
plitude na medida que a mquina ganha velocida-
de. Note que a amplitude da oscilao presente no
sinal de tenso na carga reduz com o aumento da
velocidade, o que caracterstico deste tipo de m-
quina.
T
T
1 > 1 > 1 > 1 >
1 > 1 > 1 > 1 >
2 > 2 > 2 > 2 >
2 > 2 > 2 > 2 >
1) Ch 1: 10 V 1 ms
2) Ch 2: 20 V 1 ms
Fig. 15. Tenso em uma fase e sinal de gatilho.
T
T
1 > 1 > 1 > 1 >
1 > 1 > 1 > 1 >
2 > 2 > 2 > 2 >
2 > 2 > 2 > 2 >
1) Ch 1: 2 V 500 ms
2) Ch 2: 20 V 500 ms
Fig. 16. Corrente em uma fase (escala: 1 V - 4 A) e tenso na
carga para transitrio de referncia.
T
T
1 > 1 > 1 > 1 >
1 > 1 > 1 > 1 >
2 > 2 > 2 > 2 >
2 > 2 > 2 > 2 >
1) Ch 1: 2 V 5 s
2) Ch 2: 10 V 5 s
Fig. 17. Curva de tenso na carga e corrente na
fase (escala:1 V - 4 A).
2618
XVIII Congresso Brasileiro de Automtica / 12 a 16-setembro-2010, Bonito-MS
Visando verificar o perfil da eficincia do GRV
operando com a estratgia de controle apresentada,
que visa manter constante a tenso na carga, testes
foram realizados com a mquina operando em di-
ferentes velocidades. A figura 18 mostra que a
potncia gerada, ou seja, a potncia eltrica de
sada (potncia dissipada na carga) diminuda da
potncia eltrica de entrada no conversor HB, au-
menta com o aumento da velocidade.
800 1000 1200 1400 1600 1800 2000 2200
85
90
95
100
105
110
Velocidade (rpm)
P
o
t
n
c
i
a
G
e
r
a
d
a
Fig. 18. Curva de potncia gerada pelo GRV em funo da veloci-
dade.
5 Concluses
O funcionamento de uma mquina a relutncia
varivel 6x4 operando como gerador controlado
em malha fechada de tenso na carga foi estudado
neste trabalho. A estratgia de controle apresenta-
da baseada na utilizao de um controlador PI
para variar a mdia da tenso aplicada nas fases
durante o perodo de magnetizao, utilizando
sinal PWM. Esta estratgia foi testada atravs de
simulao do prottipo montado e experimental-
mente utilizando um DSP de ponto fixo. Os resul-
tados demonstraram que a tcnica de controle mos-
trada no artigo controla bem o GRV, permitindo
oper-lo em malha fechada de tenso na carga em
situaes de transitrio de resistncia da carga,
referncia de tenso e de velocidade. Este tipo de
controle interessante para aplicaes embarcadas,
como por exemplo, automotiva, aeronutica e em-
barcaes.
Agradecimentos
Agradecimentos FAPEMIG e ao CNPq pelo
apoio financeiro e Universidade Federal de Uber-
lndia pela estrutura disponibilizada.
Referncias Bibliogrficas
Fahimi, B., et. al. A switched reluctance machine-
based starter/alternator for more electric cars
Trans. Energ. Conver., vol. 19, n1., maro 2004.
Miller, J. M., Rajarathnam, A. V. and Ehsani, M.
Current status and future trends in more electric
car power systems, Proc. IEEE Veh. Technol.
Conf., Houston, TX, Maio 1999.
Cai, W. Comparison and review of electric machines
for integrated starter alternator applications, IAS-
IEEE, 2004.
Chdot, L. Integrated Starter Generator: The Need for
an Optimal Dsing and Control Approach.
Application to a Permanent Magnet Machine,
Transactions on Industry Applications, IEEE, Vol.
43, No. 2, 2007.
Perreault, D. J. Automotive Power Generation and
Control Transactions on Power Electronics, Vol.
19, N. 3, 2004.
Vries, A., at al. A Switched Reluctance Machine for a
Starter-Alternator System IEMDC-IEEE 2001.
Torrey, D. A. Switched reluctance generator and their
control, IEEE Trans. Ind. Electron., vol. 49, pp. 3-
14, Fev. 2002.
Schofield, N., Long, S. A. Generator Operation of a
Switched Reluctance Starter/generator at Extended
Speeds IEEE 2005.
Radun, V., at al. Two-channel switched reluctance
starter/generator results Transactions on Industry
Applications, IEEE, Vol. 34, No. 5, 1998.
Fleury,A., Silva; F. S., Arajo, W. R. H., Andrade, D.
A., Silveira, A. W. F. V. Reduced Switch count
converter for switched reluctance generators
Eletrnica de Potncia, Vol. 13, No. 3, August
2008.
2619
XVIII Congresso Brasileiro de Automtica / 12 a 16-setembro-2010, Bonito-MS
Modeling, Simulation and a Comparative Study Between a Single and a Three-
Phase Switched Reluctance Machine
Abstract. The comparative study of electric machines has
been in vogue due to the growing demand for electromechanical
converters with maximum possible efficiency. In this scene, the
switched reluctance machines have proven to be competitive.
Comparative studies between these machines and the already
established induction machines can be easily found in the
scientific literature, but studies on various configurations of the
switched reluctance machines are not as widespread. This paper
shows the modeling, simulation and presents a comparative
study of two Switched Reluctance Machines to a single phase
(6x6) and three phase (6x4). Aspects of construction, drive and
efficiency are discussed in order to find advantages and
disadvantages to each of these machines.
Key words
Single phase Switched Reluctance Machine, Three phase
Switched Reluctance Machine, comparison of efficiency.
1. Introduction
The interest in Switched Reluctance Machines (SRM)
has getting a competitive market space. The industries are
still in majority, induction machines; some synchronous
machines that require a more rigorous maintenance due
to the presence of brushes and rings; and fewer appear
permanent magnet machines, losing competitiveness
because of the high cost of magnets.
Several years ago, the obstacle to the interest in SRM
was the high cost of power electronics, nowadays is not a
problem due to the decrease in cost of microprocessors
and semiconductor switches [1].
Because there are no windings, brushes and magnets on
the rotor, the MRV in addition to having a simple
structure and be more robust, have lower cost of
manufacturing compared to other existing
machines[1].Windings concentrated only in the stator,
phases considered magnetically independent of one
another, high torque per amp, high power density and
high effiency are other advantages of this machines.
These advantages are leading more and more researchers
to study on their application as electric motor cars, small
domestic appliances, pumps, fans and others [2] [3] [4].
But there are also unfavorable characteristics for use as
vibration and acoustic noise, yet several studies are
already being taken to reduce these problems [5].
One study, not very rencent,[7] compares SRM with the
indcuction machines.
The objective proposed in this paper is the comparasion
of two SRM: a Single Phase Reluctance Motor
(SPSRW) and a Three Phase Reluctance Motor
(TPSRW). The comparative study was carried out
through simulations and experimental results.
2. Structure of the machines
A Switched Reluctance Machine (SRM) is composed of a
laminated structure of double salience, simple, in which
the coils are restricted only to the stator teeth. Moreover,
in the case of this work, each coil of a pair of teeth
opposite of the stator are a phase, as shown in Fig.1
This figure shows a 6x4 SRW, in other words, a SRM
with six stator poles and four poles in the rotor. Because
each pair of poles in the stator form only one phase, this
is a Three-Phase Switched Reluctance Machine to
(TPSRM).
It also can be seen in Fig.1 the connection in series of
coils of each pair of teeth opposite to form of mentioned
phase. Thus, the current that runs through these coils is
the same. Fig.1 shows the coils of only one phase, phase
A, but this configuration is repeated for the other two
phases. In this case, phase A is in its position of complete
alignment, this position was chosen to be the reference in
this work, ie, everytime the rotor is aligned with the
stator at a certain phase, it is said that the rotor is in zero
degree of that phase.
Fig.1 Cross Section of a TPSRM, showing the winding of the
phase A.
The Fig.2 shows a 6x6 SRM. As can be seen, what
distinguishes the two machines studied is only the
number of teeth on the rotor and its drives. Again, each
pair of teeth opposite in the stator was connected in
series, but the energization of all coils will be held at the
same time, no delay between pulses, hence this
configuration of SRM , where the number of teeth rotor
is equal to the number of stator teeth, is called a Single-
phase Switched Reluctance Machine (SPSRM). The
polarity of the windings of the other teeth are also
represented in Fig.2.
Fig.2 Cross Section of a SPSRM, showing a part winding of
the phase A.
3.Principles of Drive
For that each phase could be energized in the correct
moment is necessary to know, every moment, the rotor
position. For this, a positioning disc with optical sensors
was placed on the machine shaft as shown in Fig.3.
Fig.3 - Test bench showing SRM.
A-Three Phase Switched Reluctance Machine
To operate the TPSRM, was used a half-bridge converter.
This converter was chosen to be the most applied to drive
these machines [9].
Fig.4 shows a three phase half-bridge converter. In this
converter, the gate of each switch will be controlled by
positioning sensor, insomuch that each phase will be
energized when are completely disaligned (Fig. 5), at
this point the switch S
1
and S
2
will be closed and the
current coming from the source will flow through the coil
Phase A.
This energizing will have a duration of 30 degrees, ie
when is missing 15 degrees to the complete alignment
this phase will be turned off, turning off the switches S
1
and S
2
(Fig. 6). At this point, the energy that was stored
in the coil of phase A will be returned to the source,
establishing a freewheel with the diodes d
1
, d
2
and
source.
The name Switched Reluctance Machine is due to the
fact that his machine is always varying the reluctance. In
the case of the 6x4 configuration the profile of reluctance
is 90 degrees. As is known, the reluctance is inversely
proportional to inductance. Figure 7 shows the
inductance profile of a TPSRM
Fig.4 Three Phase Half-Bridge converter.
Fig.5 Phase A completely disaligned in a TPSRM.
Fig.6 Phase A missing 15 graus to aligned in a TPSRM.
Fig.7 Inductance profile of a TPSRM
B- Single Phase Switched Reluctance Machine
A Single Phase Switched Reluctance Machine (SPSRM)
is a machine where, typically, the number of teeth on the
rotor and stator are equal. Thus, there is no discrepancy
between the profile of inductance of each stator tooth.
Thus, these machines are seen as having only one phase.
It is common to find projects with 2x2, 4x4, 6x6 and 8x8
poles on the stator and rotor, respectively [8]. Here, we
will analyze the machine with 6x6 configuration.
To operate this machine, again used a half-bridge
converter, however, in this case, it is necessary just a
phase, as shown in Fig.8
Fig.8 Single-phase half-bridge converter.
Thus, when the rotor is in complete disalignment, 30
degrees of alignment position, the machine will be
energized (Fig. 9). At this point, switches S
1
and S
2
will
be closed and the current coming from the source will
flow through the coil of phase A.
This energizing will have a period of 15, ie, when is
missing 15 for the complete alignment the phase will be
turned off , turning off the switches S
1
and S
2
(Fig.10). At
this point, the energy that was stored in the coil of phase
A will be returned to the source. Establishing then a
freewheel with diodes d
1
, d2 and source.
Fig.9 Phase A completely disaligned in a SPSRM.
In the case of the 6x6 configuration the profile of
reluctance have 60. The Fig.11 shows the inductance
profile of a SPSRM.
Fig.10 Phase A missing 5 to aligned in a SPSRM.
4. Mathematical Model
In an inductor the flux linkage by the coil () is
proportional to current (i) that runs through the coil and
its inductance (L), have:
e Ri v (3)
Where:
t
Ri v
w
w
O
(4)
Fig.11 Inductance profile of a SPSRM
In this way, the voltage at the terminal of each phase
voltage have resistive in nature, due to resistance of the
wires; and inductive, due to the coils of each tooth, so:
e Ri v (3)
Where:
t
Ri v
w
w
O
(4)
As the flux linkage by the coils is proportional to the
inductance and current the equation solution (4) involves
a partial derivative, where firstly L is considered constant
and i variable, and then i is consired constant and L
variable. As L is variable in relation to the rotor position
and the time, replacing equation (1) in equation (4) and
solving, have:
dt
d L
i
t
i
L Ri v
T
T w
w
w
w
(5)
Conceptually, as the derivative of rotor angular position
(T ) in relation to the time is the angular speed ():
T
Z
w
w
w
w
L
i
t
i
L Ri v (6)
Equation (6) describes the SRM electrically, but for a
complete modeling it is necessary other equation that
glimpse the machine mechanically.For this just do the
equalization the powers so that they are balanced. Thus
the energy entering the system will generate a
electromechanical torque (
emg
T ). In result of this input
will be generated a mechanic torque (
mec
T )on the system
output. Moreover, should be considered that this
machinhe should be able to win the rotational inertia of
the speed variation (
dt
d
J
Y
), and win the dynamic
friction of the bearings (
Y
D ) [9].Thus:
dt
d
J D T T
mec emg
Y
Y
(7)
Where:
D
is the coeficient of friction;
J is the moment of inertia.
The electromechanical torque can be shown
mathematically as :
T d
dL
i T
emg
2
2
1
(8)
Therefore the equation (6) and the equation (7),
together,describe in a complete ( electrically and
mechanically) a single-phase SRM, so this work will
consist on the comparison between a single-phase
machine and a three-phase machine which should be
represented in a matrix way so that all phases will be
considered in the model then for the three-phase machine
have:
w
w
w
w
w
w
T
Y
T
T
T
T
Y
3
2
1
3
3 3
2
2 2
1
1 1
3
2
1
3 3 2 2 1 1
3
2
1
3
2
1
1 0 0 0 0
0 0 0 0
0 0 0
0 0 0
0 0 0
0 1 0 0 0
0
0 0 0 0
0 0 0 0
0 0 0 0
0
I
I
I
J
L
i L
L
i L
L
i L
i
i
i
D r i r i r i
R
R
R
T
v
v
v
mec
(9)
Where:
1
r =
T w
w
1
1
L
;
2
r =
T w
w
2
2
L
;
3
r =
T w
w
3
3
L
1
I
+ =
(4)
Fig.11 Inductance profile of a SPSRM
In this way, the voltage at the terminal of each phase
voltage have resistive in nature, due to resistance of the
wires; and inductive, due to the coils of each tooth, so:
e Ri v + = (3)
Where:
t
Ri v
+ =
(4)
As the flux linkage by the coils is proportional to the
inductance and current the equation solution (4) involves
a partial derivative, where firstly L is considered constant
and i variable, and then i is consired constant and L
variable. As L is variable in relation to the rotor position
and the time, replacing equation (1) in equation (4) and
solving, have:
dt
d L
i
t
i
L Ri v
+ = (5)
Conceptually, as the derivative of rotor angular position
( ) in relation to the time is the angular speed ():
+ =
L
i
t
i
L Ri v (6)
Equation (6) describes the SRM electrically, but for a
complete modeling it is necessary other equation that
glimpse the machine mechanically.For this just do the
equalization the powers so that they are balanced. Thus
the energy entering the system will generate a
electromechanical torque (
emg
T ). In result of this input
will be generated a mechanic torque (
mec
T )on the system
output. Moreover, should be considered that this
machinhe should be able to win the rotational inertia of
the speed variation (
dt
d
J
), and win the dynamic
friction of the bearings (
D ) [9].Thus:
dt
d
J D T T
mec emg
+ + = (7)
Where:
D
is the coeficient of friction;
J is the moment of inertia.
The electromechanical torque can be shown
mathematically as :
d
dL
i T
emg
2
2
1
= (8)
Therefore the equation (6) and the equation (7),
together,describe in a complete ( electrically and
mechanically) a single-phase SRM, so this work will
consist on the comparison between a single-phase
machine and a three-phase machine which should be
represented in a matrix way so that all phases will be
considered in the model then for the three-phase machine
have:
&
&
&
&
&
3
2
1
3
3 3
2
2 2
1
1 1
3
2
1
3 3 2 2 1 1
3
2
1
3
2
1
1 0 0 0 0
0 0 0 0
0 0 0
0 0 0
0 0 0
0 1 0 0 0
0
0 0 0 0
0 0 0 0
0 0 0 0
0
I
I
I
J
L
i L
L
i L
L
i L
i
i
i
D r i r i r i
R
R
R
T
v
v
v
mec
(9)
Where:
1
r =
1
1
L
;
2
r =
2
2
L
;
3
r =
3
3
L
1
I
&
is the derivative of the current of the phase 1 in time;
2
I
&
is the derivative of the current of the phase 2 in time;
3
I
&
is the derivative of the current of the phase 3 in time;
& is the angular speed variation in time;
&
is the variation of rotor position in time.
5. Simulations
The simulation tools have facilitated the development of
projects more accurate and cheaper in electric machines.
Parameters that were previously impossible to consider,
are now easily predicted in the design of these machines
by simulation.
The simulations presented in this work were performed
using MATLAB SIMULINK.
Fig.12 Three-phase half-bridge converter.
Fig 12. Three-phase half-bridge converter.
Thus, each machine was fought degree in degree, and its
inductance was measured. The result can be seen in Fig.7
and Fig.11.
A.Simulation of a TPSRM
In the simulation of a MRVT were used some tools ready
of SIMULINK. The converter shown as electrical
diagram in Figure 4 can be seen in Fig.12, but in
SIMULINK.
In this converter was used controlled a current source by
a signal in the simulation of the coils.
In addition to the converter is another block that
simulates the mechanical and electrical behavior of the
machine. This block can be seen in Figure 13. In this
block were set as input, the voltage of the three phases
and the load on the motor shaft. The load may have a
quadratic behavior with respect to angular speed (),
characteristic of typical loads such as fans for example or
be a constant load.
The control signal from the current source of the
converter is the phase current which is the output from
block Fig.13.
The Fig.13 is simply an S-function that solves the
equation of state (9).
Fig.13 - Resolution of equation (9) in SIMULINK.
Input parameters of the Fig.13 are taken from the simulation of
half-bridge converter in Fig.12.
B. Simulation of a SPSRM
The simulation of SPSRM, of course, was conducted the
same way as was done for TPSRM.
Thus, Fig.14 shows the half-bridge converter which is
electrically shown in Fig.8, but in the SIMULINK.
Again, the blocks of Fig 14 and Fig.15 form a closed
system because the output of Fig 15 (currents) are the
input in Fig.14. And the output of Fig 14 (a voltage Va)
is input in Fig 15.
Fig.15 is merely the solution of equations (7) and (8) of
the mathematical model.
Fig.14 - Single-phase half-bridge converter.
Fig.15 - Resolution of equation (7) and (8) in SIMULINK.
5. Results of Simulations
The simulations were done with machines fed with 311
V with rated load on their axes, with the aim of analyzing
parameters such as wave form current on the phase,
variations speed in machine and efficiency.
A. Waveforms current on the phase
The waveforms of the currents of the phases in the two
machines can be seen in Fig.16 and Fig.17.
Fig. 16 shows the waveform of the current arising from
the three phases of TPSRM.
Fig. 17 shows the waveform of current in a SPSRM.
These figures make it clear that the three-phase machine
has a behavior more constant, because the phase are
powered separately and there is always an active phase.
In the SPSRM exists an interval where the machine is
totally turned off, this would provide a more fickle
behavior.
Fig.16 - Waveform of current in a TPSRM
Fig.17 - Waveform of current in a SPRM.
B. Speed of the machines
The SRM are known to have problems in their oscillation
parameters, including in the generated torque and in the
speed.
The TPSRM showed a variation of speed sharp less than
presented on the SPSRM. The range was of 127 rad / s to
119 rad / s with speed around 1200 rpm desired for this
machine.
However, SPSRM presented a disastrous change of
speed, getting between 40 rad / s and 200 rad / s. Since
then this characteristic a huge disadvantage for this type
of machine (6x6 MRVM 1 hp) when is driven in the way
it was done in this work.
C. Effiency of the machines
Both machines showed good results regarding the
efficiency item. TPSRM showed 91.5% and 81%
SPSRM.
D. Converters
The single-phase converter of the SPSRM is simpler and
cheaper than the converter of the MRVT by being
composed of only one arm, and thus need only one
switch to be operated.
6. Experimental Results
To perform comparative tests between the two
configurations of SRM was mounted a bench containing
two SRM, a three-phase induction machine (MIT) and a
Half-bridge converter, as shown in fig.3. A MIT was
used as a load. For this, a direct current was applied in
phases in order to generate a fixed magnetic field and so
when the SRM are operated and make the MIT turning, a
power appears making that this works as a load.
In the first test the TPSRM was fed with a voltage of 200
V on the dc bus and the current values was incremented
one to one starting from zero (empty machine) and go up
to 7 A (Maximum load by MIT).
The graph in Fig. 18 shows the decrease of speed and
increase of the power when the load on the shaft of the
TPSRM increase.
0 1 2 3 4 5 6 7
400
600
800
1000
1200
1400
1600
1800
2000
2200
2400
Current in ampre (Load)
P
o
w
e
r
(
W
)
a
n
d
s
p
e
e
d
(
r
p
m
)
Fig.18 Power (- - - -) and Speed (_____) in a TPSRM.
Then the same test was repeated for SPSRM, the result is
shown in Fig.19.
0 1 2 3 4 5 6 7
1000
1200
1400
1600
1800
2000
2200
2400
Current in ampre (Load)
P
o
w
e
r
(
W
)
a
n
d
s
p
e
e
d
(
r
p
m
)
Fig.19 Power (- - - -) and Speed (_____) in a SPSRM.
These two figures (Fig. 18 and Fig.19) show that TPSRM
when fed with constant voltage of 200 V always keep a
faster speed than SPSRM, despite its derivative of speed
in relation to the load to be greater. Moreover, the input
power required by SPSRM is always greater than the
TPSRM.
However, the analysis of input power was compromised
in this test, because despite the magnetic power generated
at MIT have relation to the current (x-axis of figures) it is
also related to the speed of the machines which, as has
been shown is different. So this first test reveals only the
behavior of the machines for their speed when there is a
increase in the load and voltage control in the dc bus.
Thus, takes necessary a second test in which speed in
SRM were kept constant at 1200 rpm (nominal speed)
and the current MIT (load) was being added. The result
is shown in Fig. 20.
0 1 2 3 4 5 6 7
0
500
1000
1500
2000
2500
Current in ampre (Load)
P
o
w
e
r
(
W
)
Fig.20 Input Power in a SPSRM (- - - -) and Input Power in a
TPSRM (_____).
In Fig.20 shows that when subjected to the same load, the
SPSRM requires a power input more than four times the
input power required by a TPSRM. When these are
operated the way it was done in this work.
8. Conclusion
Both machines have very similar aspects of construction,
so there is no way to elect between the two that is better
in this item.
TPSRM machine proved to be a more constant, with
fewer variations in speed. This did not occur with
SPSRM which was very swinging.
The converter SPSRM is simpler and inexpensive,
requiring only two switch to operate it, but these switch
must support peak currents greater than those that occur
in TPSRM. In TPSRM is necessary to use six switchs.
In the simulations, the TPSRM showed a yield of around
91.5% and SPSRM showed a yield around 81%. The
simulations were performed using nominal values of
input voltage and load.
The experiments were performed at various points of
operation where the increase in load forced the machine
to slow down, increasing its power. This test showed that
with the same dc bus voltage the TPSRM work with
greater speed and requiring less power input.
Furthermore, when subjected to the same load, the
TPSRM needed less input power (4 times less).
So for all these factors it is concluded that the TPSRM
when operated in the same way as was done here, and far
more advantageous than SPSRM.
Acknowledgement
The authors thank the Federal University of Uberlandia
and the PUC-GO, for multiple collaborations. And to
CAPES for the scholarships.
References
[1] K.A, Joseph, Opportunities for switched reluctance
motor-drives, Pulp and Paper Industry Technical
Conference, 1999. Conference Record of 1999 Annual,
1999,pp.42-47.
[2] H. Chen, Y. Guo, Green methodologies and technologies
of switched reluctance motor drive., Proceedings of the 3
World Congress on Intelligent Control and Automation,
vol 5,2000, pp.3717-3720.
[3] L. Chang, Switched Reluctance Motors: Small Motors of
the Next Generations for Automobiles?, Vehicular
Technology Conference ,vol 5, 2003, pp.3316-3320.
[4] H. Chen, Implementation of a Three-Phase Switched
Reluctance Generator System for Wind Power
Applications, Electromagnetic Launch Technology,2008
14
th
Symposium on, 2008, pp.1-6.K.
[5] C.Yookpakdee, N.H. Fuenqwarodsakul, Variable Speed
Switched Reluctance Drive for a Low Cost Applications,
Electrical Engineering/Electronic, Computer,
Telecommunications and Information Technology,2009,
ECTI-CON,2009.6
th
International Conference on, vol 1,
2009, pp.262-265.
[6] S. Jose, E.S. Antonio, C. Maria Rosario, Design of a
system for analysis and monitoring of vibrations in Linear
Switched Reluctance machines, MELECON 2010-2010
15
th
IEEE Mediterranean Electrotechnica Conference,
2010, pp.768-773.
[7] M.R. Harris, T.J.E. Miller, Comparison of Design and
Performance Parameters in Switched Reluctance and
Induction Motors, Electrical Machines and Drives, 1989.
Fourth International Conference on, 1989, pp.303-307.Y
[8] bKRISHNAN, Ramu, Switched Reluctance Motors
Drives. CRC Press, 2001.
[9] S V. F, Augusto, Modelagem, Construo, Testes e
Analise de Desempenho de um Gerador Relutncia
Chaveado, Tese de Doutorado, Universidade Federal de
Uberlndia, Abril,2008.