MRAS-Based Sensorless Control Scheme for Open-

End Stator Winding Six-Phase Induction Motor with

Fuzzy Logic Speed Controller: Real-Time Simulation
Saad Khadar Haitham Abu-Rub Abdellah Kouzou
Applied Automation and Industrial Department of Electrical and Computer Applied Automation and Industrial Diagnostics
Diagnostics Laboratory Engineering Laboratory
Ziane Achour University Texas A&M University, Qatar Foundation Ziane Achour University
17000 Djelfa, Algeria Doha, Qatar 17000 Djelfa, Algeria
[email protected] [email protected] [email protected]

Abstract— This paper presents a advanced field-oriented adopted for the control of the proposed drive. On other side, the
control (FOC) for an open-end stator windings six phase sensitivity to the machine parameters, and relatively slow
induction motor (SPIM-OESW) with fuzzy logic (FL) speed dynamic response are the main drawbacks of the FOC
controller. The proposed motor topology is fed by four, three- technique [4]. Also, the rotor speed is often characterized by an
2021 IEEE 30th International Symposium on Industrial Electronics (ISIE) | 978-1-7281-9023-5/21/$31.00 ©2021 IEEE | DOI: 10.1109/ISIE45552.2021.9576385

phase voltage inverters with two isolated DC sources. The FOC overshoot at startup of the machine. Recently, several works
technique requires a precise knowledge of some machine’s have shown that the FL controllers is able to replace the PI
variables such as rotor position and rotor flux which are difficult controller and reduce the response time [11], [12]. The FOC
to measure. Usually, the rotor position and motor speed are technique for SPIM drive requires a measurement of the rotor
measured by a shaft-mounted encoder, but that makes the
speed and rotor flux values. However, the presence of speed
maintenance activities more difficult, and reduces system’s
reliability. Motor flux measurement is more complicated and
sensor may substantially reduce the overall system ruggedness
requires machine’s built in sensors. Therefore, a rotor flux-based [4], [10], [13]. Therefore, many approaches have been
model reference adaptive system (RF-MRAS) is proposed in this proposed for sensorless control which mainly are based on the
paper for the estimation of both rotor flux and rotor speed. In Luenberger observers [14], the extended Kalman filter [15], the
addition, the classic integral proportional (PI) controller is model reference adaptive system (MRAS) [16], the
replaced by a FL controller to achieve better speed response at Backstepping observer [17], and the sliding mode observer
different operating points. The proposed control scheme is built [10]. These techniques provide the best performance over a
within the Simulink environment combined with the Real-Time wide range of speed. The rotor flux based MRAS is described
platform. Real time results are obtained to illustrate the as one of attractive solutions due to its good performance and
effectiveness of the proposed control under different operating ease of implementation [4], [16].
conditions. This paper proposes a sensorless control for the SPIM-
OESW topology with FOC. The main advantages of the
Keywords—Field oriented control, fuzzy logic controller, double proposed solutions are simplicity, fast transient response and
star induction motor, MRAS estimator, open-end winding. reduced overshoot. Moreover, speed estimation scheme based
on RF-MRAS estimator is investigated to reduce the number of
I. INTRODUCTION
the sensor components used. Thus, it reduces the overall system
For several decades, the multilevel inverter concepts and cost and size, decreases the maintenance requirement and
multiphase machine concepts have gained rapid developments increases the system robustness and reliability. The paper is
[1]-[3]. The double star induction machine belongs to the organized as follows. The SPIM-OESW model is presented in
multiphase machines category. Such machines have numerous Section II. The sensorless FOC control scheme based on RF-
merits such as having high flux density, reduced torque ripples, MRAS estimator is described in Section III, while the fuzzy
low common-mode voltage and fault-tolerant capability [2], logic controller is presented in Section IV. Sections V provide
[4]. These characteristics make the machine more convenient the numerical simulation results to evaluate the proposed
for high power applications [5]. In this context, the proposed control system. Finally, concluding remarks are given in
work aims to benefit from the merits of both multilevel inverter section VI.
and multiphase machine concepts. The open-end stator winding
six-phase induction motor (SPIM-OESW) topology is obtained II. MODELING OF SIX PHASE INDUCTION MOTOR
based on the opening the neutral point of the six phase
induction motor (SPIM) and supplying the stator winding ends A. SPIM modeling
by four three-phase voltage inverters with two isolated DC The six phase induction machine consists of two separate
sources. The four voltage inverters are combined with open end three-phase stator windings while the rotor winding is squirrel
stator winding without any neutral point, which combines the cage structure, as shown in Fig. 1. The two stator windings sets
advantages of multi-phase induction machines and multilevel are ( Sa1,S b1,S c1 ) and ( Sa 2,S b 2,S c 2 ) which are displaced
inverters [4], [6]. The field-oriented control (FOC) techniques apart by 30° electrical degrees. While the rotor windings sets is
for the SPIM have been widely proposed [7]-[10] so it is

978-1-7281-9023-5/21/$31.00 ©2021 IEEE

Authorized licensed use limited to: Consortium - Algeria (CERIST). Downloaded on December 29,2022 at 21:09:40 UTC from IEEE Xplore. Restrictions apply.
( ra, rb, rc ). The simplified model of the machine can be For the stator winding 2:
described as next [18]. Vsa 2 = Va210 − Va220
The stator and rotor voltage can be obtained as:  2 2 (4)
Vsb 2 = Vb10 − Vb 20
[Vsi1 ] = [ Rs1 ][isi1 ] + p [ϕ si1 ] 
 (1)
2 2
Vsc 2 = Vc10 − Vc 20
[Vsi 2 ] = [ Rs 2 ][isi 2 ] + p [ϕ si 2 ]
 where, Vsa1 ,Vsb1 ,Vsc1 and Vsa 2 , Vsb 2 , Vsc 2 are the three-phase
[Vri ] = [ 0] = [ Rr ][iri ] + p [ϕ ri ]
voltages of stator 1 and stator 2, respectively.
where, i = a, b, c , [Vsi1 ] , [Vsi 2 ] , and [Vri ] are the matrixes
of the stator and rotor voltages components respectively. [ Rs1 ] ,
-Stator 1-
[ Rs 2 ] and [ Rr ] are the matrixes of the stator and rotor
resistances respectively. [isi1 ] , [isi 2 ] and [iri ] are the matrixes Vdc1

of the stator and rotor currents components respectively. [ϕsi1 ] ,

[ϕsi 2 ] , and [ϕri ] are the stator and rotor flux components
respectively. 30 o
The stator and rotor flux equations can be expressed as:
[ϕs1 ] = [ Ls1 ][isi1 ] +  Ls1,s 2  [isi 2 ] +  Ls1,r  [iri ]

 T
(2)
[ϕs 2 ] = [ Ls 2 ][isi 2 ] +  Ls1,s 2  [isi1 ] +  Ls 2,r  [iri ]
Vdc2

 T T
[ϕri ] = [ Lr ][iri ] +  Ls1,r  [isi1 ] +  Ls 2,r  [isi 2 ] -Stator 2-

where, [ Ls1 ] , [ Ls 2 ] , and [ Lr ] are the matrixes of the stator

Fig. 2. Six phase induction motor configuration under open-end stator
and rotor inductances respectively.  Ls1,s 2  is the mutual winding

inductance matrix between star 1 and star 2.  Ls 2,s1  is the III. SENSORLESS FIELD ORIENTED CONTROL
mutual inductance matrix between star 2 and star 1.
A. Field oriented control
The FOC technique assures decoupling the torque and rotor
flux in order to make the control similar to a separately excited
DC machine [7]-[10], [20]. Accordingly, the electromagnetic
torque is proportional to the product of the armature current.
The rotor direct flux converges to its reference while the rotor
quadrature flux equal zero, thus:

ϕ rd = ϕr
 (5)
ϕ rq = 0
By applying this principle (5) and to simplify the motor
Fig. 1. Representation of the SPIM windings. model and to obtain constant coefficients for the  L s 2, s1  and
 
B. Open-end stator winding six phase induction motor  Ls1,r  , it is necessary to apply Park transformation in
The power circuit of SPIM-OESW is shown in Fig 2 [2]. synchronous reference frame. Such approach helps removing
The SFIM-OESW is achieved by disconnecting the neutral the time varying coefficients. The Park transformation matrix is
point of a SFIM and four three-phase voltage inverters with defined as [19]:
two isolated DC sources, which are associated with these open
ends. The output voltages can be obtained as follows:  2π 4π 
For the stator winding 1:  cos(θ ) cos(θ − 3 ) cos(θ − 3 ) 
  (6)
2 2π 4π 
Vsa1 = Va110 − Va120 [T ] =  sin(θ ) sin(θ − ) sin(θ − ) 
 3 3 3
1 1 (3)  
Vsb1 = Vb10 − Vb 20  0.5 0.5 0.5 
  
1 1
 
Vsc1 = Vc10 − Vc 20

Authorized licensed use limited to: Consortium - Algeria (CERIST). Downloaded on December 29,2022 at 21:09:40 UTC from IEEE Xplore. Restrictions apply.
 The final expressions of the V sd and Vsq could be given as:
 L  Reference Model
disd 1
Vsd 1 = Rs1isd 1 − Ls1 − ωs Ls1  r ϕr .ωg + Ls1isq1 
 dt  Rr  (7) +

V = R i − L di sd 2
 L  -
 sd 2 s 2 sd 2 s2 − ωs Ls 2  r ϕr .ωg + Ls1isq 2 
 dt  Rr 
 disd 1 Adaptive Model
Vsq1 = Rs1isq1 − Ls1 dt + ωs Ls1 (ϕ r + Ls1isd 1 ) (8)

V = R i − L disq 2 + ω L (ϕ + L i )
 sq 2 s 2 sq 2 s2
dt
s s2 r s 2 sd 2 Adaptation
Mechanism
where, ωr the rotor speed, ωs the synchronous speed, the σ is
Blondel’s coefficient. The expressions of the electromagnetic Fig. 3. Structure of RF-MRAS estimator
torque and the slip angular frequency are given by:
np Lm IV. FUZZY LOGIC SPEED CONTROLLER
Tem = ϕr (isq1 + isq2 ) (9)
Lm + Lr The block diagram of the principle of fuzzy logic controller
is shown in Fig. 4, [5], which is applied to the speed control
where, p is number of pole-pairs loop for the purpose of optimizing the control performance of
the SPIM-OESW topology. The FL controller is composed of
B. Rotor flux Based Model reference adaptive system concept three blocks: fuzzification, rule bases, and defuzzification.
The block diagram of RF-MRAS based speed estimation is
shown in Fig 3 [4, 16, 19, 21]. This estimator consists of three
models, which can be described as:
Reference model:

 d ϕ rα -v Lr  σ Ls disα 1 Lm Ls1 disα 2 

 =  vsα 1 - Rs1isα 1 - - - 
 dt Lm  dt Lr dt  (10)
 Fig. 4. Structure of the FL controller.
 d ϕ r β -v = Lr  v - R i - σ Ls dis β 1 - Lm Ls1 - dis β 2 
 dt  s β 1 s1 s β 1 
 Lm  dt Lr dt 
where, e is the error, can be expressed as:
Adjustable model: e(k ) = ω& ( k ) - ω ( k -1) (14)

 d ϕˆ rα - i  Rr Lm R ϕˆ  The error derivative (de) of the variable to be controlled can

 = (isα 1 + isα 2 ) - r rα -i - ωϕˆ r β - i  be obtained as:
 dt  Lr Lr  (11)
 e ( k ) - e ( k -1)
 d ϕˆ r β - i =  Rr Lm (i + i ) - R r ϕˆ r β -i - ωϕˆ  de = (15)
 dt  sβ 1 sβ 2 rα - i  T
  Lr Lr 
The output of the FLC is given by:
Adaptation mechanism
Tem∗ ( k ) = Tem ( k -1) + ∆ u ( k ) (16)
The error signal in stationary reference frame between these
two models for the corrector is calculated by: The membership function of input variables e( k ) , de ( k )
ε = ϕˆ rα-iϕ rβ-v - ϕˆ rβ-iϕ rα-v (12) and output variables ( ∆u ( k ) ) as shown in Fig. 5. The Table I
presents the linguistic of input variables and the output
The expressions for estimated speed signal using a PI variable.
controller can be given as:

ωˆ = ε K I ω + K Pω  ε dt (13)

where, KIω and KPω are the proportional and integral positive
coefficients.

(a)

Authorized licensed use limited to: Consortium - Algeria (CERIST). Downloaded on December 29,2022 at 21:09:40 UTC from IEEE Xplore. Restrictions apply.
 (b)
Fig. 5. The membership function: (a) input variables ( e , de ), and (b) output
variables ( ∆u )

TABLE I. RULES BASE FOR FL CONTROLLER

Fig. 6. Responses of speeds, estimation error, electromagnetic torque under
e different values of load torque with constant speed (25rad/s/div)
de NB NM NS ZE PS PM PB
Fig. 7 presents the rotor flux components and stator current
NB NB NB NB NB NM NS ZE
components in d − q reference frame, the isd component as an
NM NB NB NB NM NS ZE PS
NS NB NB NB NM NS ZE PS
image for the ϕ r and the isq component as an image for the
ZE NB NM NS ZE PS PM PB Tem . The isa1 , isb1 , isc1 and isa 2 isb 2 isc 2 are sinusoidal and
PS NM NS ZE PS PM PB PB shifted by 30° electrical as shown in Fig 8. The rotor position
and estimated position are shown in Fig. 9, the rotor position
PM NS ZE PS PM PB PB PB
varies between 0 and 2 π rad. It is clear that it presents a
PB ZE PS PM PB PB PB PB periodic function under different values of applied load torque.

V. REAL TIME SIMULATION RESULTS

To verify the effectiveness of the proposed sensorless
algorithm, the system has been investigated under different
operation conditions. To illustrate the performances of the
proposed control, the classical PI speed controller has been
replaced by a FL speed controller in all tests. Extensive real-
time simulations were conducted to demonstrate the proposed
control scheme using OPAL-RT 5600 [2]. The simulation is Fig. 7. Responses of d-q stator currents and d-q rotor flux components under
different values of load torque with constant speed (25rad/s/div)
carried out under the following conditions:

A. Load torque variation

The first test is performed to verify the performance of the
proposed control algorithm for the studied SPIM-OEW
topology with constant reference speed (100 rad/s/div) under
different values of load torque (5 and 18 N. m) at time (1.2 and
1.3 s). The speed, the estimation error, the electromagnetic
torque, and applied load torque are presented in Fig. 6. It can
be observed that the rotor speed and estimated speed can track
the desired reference speed without overshoot. Also, despite
the presence of a load torque, the RF-MRAS estimator with Fig. 8. Responses of three phase stator currents of stator 1 and stator 2 under
FL controller tracks the real value of the speed with a very different values of load torque with constant speed (25rad/s/div)
small estimation error compared to the results presented in
[13]-[17]. The electromagnetic torque of SPIM-OEW topology
presents a good high dynamic and confirms the performances
of the FL controller in terms of reduce overshoot. Also, the
values of the Tem follow the applied load torque with high
accuracy.

Fig. 9. Responses of rotor position and estimated position under different

values of load torque with constant speed (25rad/s/div)

Authorized licensed use limited to: Consortium - Algeria (CERIST). Downloaded on December 29,2022 at 21:09:40 UTC from IEEE Xplore. Restrictions apply.
B. Speed Reversal Response C. Low speed operation
The second test study is conducted to verify the speed control The third test study is carried out to verify the performance
performance of the proposed control algorithm. The direction of the proposed control algorithm for low reference speed (10
of rotor rotation is reversed from 100 rad/s to -100 rad/s at t = rad/s) under different values of load torque (5 and 10 N. m) at
1.3s under load torque constant (which was applied at 1.1s). times 1.2s and 1.3 s. The speed, estimation error,
Fig. 10 presents the rotor, estimated, and reference speeds. The electromagnetic torque, d-q rotor flux components, and rotor
estimated and rotor speed track to the reference speed without position are presented in Fig. 14 to Fig. 17, respectively. A
overshoot and with small estimation error. This indicates that reversal speed test is afterwards generated, and a change in the
the proposed RF-MRAS estimator with FL controller has good reference speed from 10 to -10 rad/s is forced. The change is
performance. The d − q rotor flux components have good imposed at t=0.3 s and the results are shown in Fig. 18 to Fig.
19, respectively. The obtained results show the effectiveness
reference tracking as shown in Fig 11. The ϕ rq is maintained
of the proposed sensorless control for the SPIM-OEW
at almost zero value. The electromagnetic torque and the rotor topology with RF-MRAS estimator and FL controller. The
position signals are shown in Fig. 12 and Fig. 13, respectively. decoupling between the torque and the flux is guaranteed, and
the good reference speed tracking is ensured even at under low
speed reversal.

Fig. 10. Responses of speeds, and sped estimation error with the change in
reference speed under load torque constant (5N.m /div)
Fig. 14. Responses of speeds, and speed estimation error under different
values of load torque with constant speed (2.5rad/s/div)

Fig. 11. Responses of d-q rotor flux components with the change in reference
speed under load torque constant (5N.m /div)

Fig. 15. Responses of electromagnetic torque, and q-stator current under

different values of load torque with constant speed (2.5rad/s/div)

Fig. 12. Responses of electromagnetic torque, load torque and q stator

current component with the change in reference speed under load torque
constant (5N.m /div)
Fig. 16. Responses of d-q rotor flux components under different values of load
torque with constant speed (2.5rad/s/div)

Fig. 13. Responses of rotor position, and estimated position with the change in
reference speed under load torque constant (5N.m /div)
Fig. 17. Responses of rotor position, and estimated position under different
values of load torque with constant speed (2.5rad/s/div)

Authorized licensed use limited to: Consortium - Algeria (CERIST). Downloaded on December 29,2022 at 21:09:40 UTC from IEEE Xplore. Restrictions apply.
 [8] R. Bojoi, M. Lazarri, F. Profumo and A. Tenconni, “Digital field-
oriented control for dual three-phase induction motor drive”, IEEE
Trans. Industrial Application, vol. 39, no. 3, pp. 752–760, 2003
[9] R.O.C. Lyra and T.A.Lipo, “Torque density improvement in a six-phase
induction motor with third harmonic current injection”, IEEE Industrial
Application Society Annual Meeting, Chicago II, pp. 1779- 1786, 2001
[10] S. Khadar, A. Kouzou, M. M. Rezzaoui and A. Hafaifa, “ Sensor-less
Control Technique of Open-End Winding Five Phase Induction Motor
under Partial Stator Winding Short-Circuit,” Periodica Polytechnica
Fig. 18. Responses of speeds, and speed estimation error with the change in Electrical Engineering and Computer Science, vol. 64, no. 1, pp. 2-19,
reference speed under load torque constant (5N.m /div) 2019.
[11] A.G. Aissaoui, A. Tahour, N. Essounbouli, F. Nollet, M. Abid, M.I.
Chergui, “A Fuzzy-PI Control to Extract an Optimal Power from Wind
Turbine”, Energy Conv. and Management, vol. 65, pp. 688 - 696, 2013.
[12] F. J. Lin, I. F. Sun, K. J. Yang, and J. K. Chang, “Recurrent fuzzy neural
cerebellar model articulation network fault-tolerant control of six-phase
permanent magnet synchronousmotor position servo drive,” IEEE Trans.
Fuzzy Syst.,vol. 24, no. 1, pp. 153-167, 2016.
[13] S. Bolognani, R. Oboe, M. Zigliotto, “Sensorless full-digital PMSM
drive with EKF Estimation of speed and rotor position,” IEEE Trans.
Industrial electronics, vol. 46, no. 1, pp. 184-191, 1999.
Fig. 19. Responses of three phase stator currents of stator 1 and stator 2 with [14] S. Khadar, A. Kouzou and A.Hafaifa, “ Sensorless Direct Torque
the change in reference speed under load torque constant (5N.m /div) Control of Induction Motor with an Open-End Stator Winding Using an
Adaptive Luenberger Observer,” 15th Inter. Multi-Conference on
VI. CONCLUSION Systems, Signals & Devices (SSD), 19-22 March 2018, Yassmine
Hammamet, Tunisia. https://doi.org/10.1109/SSD.2018.8570691.
In this paper, sensorless speed FOC technique based on RF- [15] Y. Zhonggang, L. Guoyin, Z. Yanqing, L. Jing, S. Xiangdong, Z. Yanru,
MRAS approach fuzzy logic speed controller is propped for “A speed and flux observer of induction motor based on extended
SPIM-OESW supplied by four voltage inverters, which kalman filter and markov chain,” IEEE Trans. Power Electron, vol. 32,
regroups the advantages of multilevel inverters and multiphase no. 9, pp. 7096-7117, 2017.
machines. On the other side, the obtained results show that the [16] A. Pal, S. Das and A. K. Chattopadhyay,“An Improved Rotor Flux
proposed sensorless control scheme with fuzzy logic speed Space Vector Based MRAS for Field-Oriented Control of Induction
Motor Drives,” IEEE Trans. Power Electronics, vol. 33, no. 6, pp. 5131-
controller has attractive advantages such as good dynamic 5141, 2018
performances with small overshoot and low steady state error
[17] R. Trabelsi, A, Khedher, F. M. Mimouni, F. M'Sahli, “An Adaptive
even at low speed operation. The proposed solution is suitable Backstepping Observer for on-line rotor resistance adaptation”, Inter.
for applications requiring high power and high tracking Journal of Sciences and Techniques of Automatic Control and
accuracy in the presence of load torque variations. Computer Engineering, vol. 4, no.1, pp. 1246-1267, 2010.
[18] A. S. Abdel-Khalik, M. S. Abdel-Majeed, S. Ahmed, “Effect of Winding
Configuration on Six-Phase Induction Machine Parameters and
REFERENCES Performance,”IEEE Access, vol. 8, 2020, pp. 223009- 223020
[1] M. Trabelsi, H. Vahedi, H. Abu-Rub, “Review on Single-DC-Source [19] S. Khadar, A. Kouzou, “Stator Resistance Compensator Based on Model
Multilevel Inverters: Topologies, Challenges, Industrial Applications, Reference Adaptive System Scheme for Sensor-Less Direct Torque
and Recommendations,” IEEE Open Journal of the Industrial Electronics Control of an Open End-Winding Induction Motor with First Coil
Society., vol. 2, 2021, pp. 112–127. Faults,” Advances in Modelling and Analysis C, vol. 74, no. 2-4, pp. 80-
[2] S. Khadar, A. Kouzou, Shady S. Refaat, H. Abu-Rub, “Real-Time 86, 2019.
Digital Simulation for Sensorless Control Scheme based on Reduced- [20] H. Abu-Rub, A. Iqbal. J. Guzinski, “High Performance Control of AC
Order Sliding Mode Observer for Dual Star Induction Motor,” IEEE Drives with MATLAB/Simulink Models,” Book, 15 March 2012.
IECON 2020, 18-21 Oct. 2020, Singapore, pp. 4869- 4874. https://doi.org/10.1002/9781119969242.
[3] L. De Camillis, M. Matuonto, A. Monti, A. Vignati, “Optimizing current [21] S. Khadar, H. Abu-Rub and A. Kouzou, “Sensorless Field-Oriented
control performance in double winding asynchronous motors in large Control for Open-End Winding Five-Phase Induction Motor With
power inverter drives,” IEEE Trans. on Power Electronics, vol. 16, no. 5, Parameters Estimation,” IEEE Open J.Industrial Electro.Society., vol. 2,
pp. 676 – 685, 2001. pp. 266-279, 2021. https://doi.org/10.1109/OJIES.2021.3072232.
[4] S. Khadar, A. Kouzou, A. Hafaifa, A. Iqbal, “Investigation on SVM-
Backstepping sensorless control of five-phase open-end winding
induction motor based on model reference adaptive system and
parameter estimation,” Engineering Science and Technology, an
International Journal, vol. 22, no. 4, pp. 1013-1026, 2019.
[5] R. Sadouni, and A. Meroufel, “Indirect Rotor Field-oriented Control of a
Dual Star Induction Machine (DSIM) Using a Fuzzy Controller”, In: Act
a Polytechnic a Hungarica, vol. 9, no. 4, pp. 177-192, 2012.
[6] S. Khadar, A. Kouzou and H. Benguesmia, “Remedial Robust Control of
Five-Phase Fault-Tolerant Induction Motor with Open-End Winding
using Reduced-Order Transformation Matrices”,Modelling,
Measurement and Control A, vol. 92, no.2-4, pp. 16-23, 2019.
[7] A.R. Munoz and T.A. Lipo, “Dual stator winding induction machine
drive”, IEEE Trans. Industrial Application, vol. 36, no. 5, pp. 1369-
1379, 2000.

Authorized licensed use limited to: Consortium - Algeria (CERIST). Downloaded on December 29,2022 at 21:09:40 UTC from IEEE Xplore. Restrictions apply.