Sensorless Direct Torque Control of PMSM Drive

for EV Application
Toshi Sharma Avik Bhattacharya
Dept. of Electrical Engineering Dept. of Electrical Engineering
Indian Institute of Technology, Roorkee Indian Institute of Technology, Roorkee
India India
[email protected]
Abstract – This paper focus on the Direct Torque Control (DTC) Consequently, several architectures are proposed for EVs viz.
of a Permanent Magnet Synchronous Motor (PMSM) drive for Battery Electric Vehicle (BEV), Hybrid Electric Vehicle
Electric Vehicle (EV) application. In order to achieve optimized (HEV), Plug-In Hybrid Electric Vehicle (PHEV) and Fuel Cell
drive efficiency, Maximum Torque Per Ampere (MTPA) control (FC) Vehicles [5]. In [6], a comparative evaluation of all the
is implemented. Moreover, to increase the mechanical robustness
types of EVs is provided. The basic architecture of any EV
of the drive and to make it cost effective sensorless speed
estimation is carried out. The technique used for sensorless speed involves an Energy Storage System (ESS) and Electric
estimation is Model Reference Adaptive Control (MRAC). The Propulsion System (EPS). ESS includes the batteries, ultra-
proposed algorithm is implemented in MATLAB/Simulink and capacitors, etc. whereas the EPS includes the electric motor
the simulation results shows the validity of proposed estimator. and power electronic converter. This paper deals with the
control of electric motor in EPS of an electric vehicle.
Keywords – PMSM, DTC, MRAC, Sensorless, Adaptive
The most commonly used motors for EV application are:
Mechanism, MTPA, Electric Vehicle
Induction Motors, Permanent Magnet motors and Switched
I. INTRODUCTION
Reluctance motors. Research in the past few decades has
The development in Internal Combustion Engine (ICE)
indicated that Permanent Magnet synchronous motor drives
based automotive industry is boon for the human race.
are becoming the strong contender of Induction Motor drives.
However, large number of automobiles across the globe has
Moreover, various advantages of PMSM over Switched
caused and continues to cause serious problems to
Reluctance motor like high efficiency, high torque density,
environment as well as to human life. Moreover, the price of
simple control, enhanced PM characteristics, comparatively
petroleum is also soaring high due to oil crisis and ever
low cost make them a good alternative for electric vehicle
increasing demand. Thus in recent decades, the concept of
application. The two high performance control i.e. the Vector
electrification of vehicles has evolved to have efficient,
Control and the Direct Torque Control can be usedfor PMSM
reliable and safe transportation with zero or reduced emission
control [7].
at affordable cost [1].
Numerous merits of DTC have attracted many researchers’
In 1976, USA published an act called “Electric and Hybrid
attention. DTC was presented by Takahashi and Noguchi [8]
Vehicle Research, Development, and Demonstration Act,
and Depenbrock [9] in 1980-90’s for Induction motor. Since
Public Law 94–413” compelled by the energy crisis and
then, it has become one of the most used control technique in
environmental degradation problems [2]. In 1990 California
industries where high dynamic performance is required. Later
Air Resources Board [CARB] adopted a zero emission
on the concept was implemented on Permanent Magnet
vehicle(ZEV) rule as part of 1990 Low Emission Vehicle
Synchronous Motors also. In the literature several types of
Program. In the ZEV rule, 2% of the automobiles produced by
DTC techniques for PMSM are discussed. They are: Classical
the large automakers should be zero emission vehicles. The
DTC (switching table based), Space Vector (SV) based DTC,
percentage was increased to 5% in 2001 and to 10% in 2003.
Model Predictive algorithm based DTC (MPDTC), Duty Ratio
The regular modifications in the ZEV mandate is forcing
Control based DTC (DDTC) [10]-[11], Fuzzy Logic based
manufacturers to rapidly ramp up the production of ZEV every
DTC [12].
year and it is expected to reach about 15% ZEV production of
In order to achieve the efficiency optimization in the constant
the total production by 2025 [3]. In 1990, an international
torque region, Maximum Torque per Ampere (MTPA) is a
membership organization “Electric Vehicle Association of
preferred control strategy of PMSM. It is similar to
Asia Pacific (EVAAP)” which is also Asian Pacific
maximizing the torque output of the machine for a particular
representative of the World Electric Vehicle Association was
amount of stator current. The maximum torque per ampere
developed. It was found with the aim of promoting the
(MTPA) control strategy is one of the methods used for high-
development and use of electric and hybrid electric vehicle in
efficiency drives [10]-[16].
Asia and Pacific region. Among the several Asian Pacific
Mechanical robustness is an important aspect in the
Region countries, India is also one of the major contributors in
demanding applications where drive is incorporating high
this sector [4].
dynamic control strategies like DTC and MTPA. Moreover,

978-1-5386-4769-1/18/$31.00 ©2018 IEEE

the applications, like chemical industries, electric vehicles etc. (4)
does not allow the use of any additional speed sensor. Thus, = + ∗ + ∗
sensorless PMSM drive has wide scope of research. Many
speed estimation techniques like Extended Kalman Filters III. PROPOSED CONTROL STRATEGY
(EKF), Signal Injection based, State Observer based and The block diagram for the proposed scheme is shown in Fig.1.
MRAC based speed estimation are reported in literature. In the block diagram, PMSM is fed by a two level inverter. In
Fuzzy logic based, Artificial Intelligence (AI) and Artificial the control strategy, reference torque is generated by
Neural Network (ANN) based speed are the latest contribution processing speed error through a PI controller. Speed error is
in the sensorless speed estimation techniques[17]-[21]. generated by subtracting actual speed of the motor from the
The main aim of this work is to develop a sensorless drive for commanded speed. To achieve the reference flux, Maximum
Electric Vehicle application.In the proposed work, Direct Torque per Ampere (MTPA) control strategy is used. The
Torque Control based sensorless PMSM drive is developed. torque error and the flux error are processed through a
Conventional switching table based DTC is implemented for switching pulse generator and firing signals for inverter are
drive control. Model Reference Adaptive Control is used to generated. Speed estimation is carried out using a Stator
estimate the speed of the PMSM eliminating the speed sensor. Current based MRAC eliminating speed sensor.
In order to achieve the maximum torque and optimized
efficiency, MTPA control is also incorporated. The work is
carried out in the MATLAB/Simulink environment on a test
+ −
motor with parameters shown in appendix I and results under
various operating conditions are discussed.
II. MATHEMATICAL MODELLING OF PMSM ∗+
∗ + ∆
The dynamic model of a PMSM drive is formulated in Rotor − −
Reference frame. In the Rotor Reference frame under steady
+ ∆Ψ
state, the sinusoidal quantities would become constant. The
Ψ∗ − Sector
permanent magnet in the PMSM rotor can be modelled as a Ψ Information
constant current source of magnitude . It is assumed that the
rotor flux is aligned along the d-axis, thus the direct axis rotor Motor
current is and quadrature axis rotor current is zero as there Parameters
is no flux along the q-axis. The machine equations of a PMSM
in rotor reference frame are expressed as follows [22]-[23]:
+
= + (1)
− + 0
Fig.1: Block Diagram of the Proposed Scheme
= ∗ (2)
The direct torque control scheme, MRAC for sensorless speed
where, estimation and MTPA control are explained in the sections
below.
d-axis Voltage in rotor Mutual
3.1 Conventional DTC Scheme
reference frame Inductance
q-axis Voltage in rotor d-axis The electromagnetic torque in a PMSM drive can be changed
reference frame inductance by control of the torque angle ( ). The torque angle ( ) is the
d-axis Current in rotor q-axis angle between the stator and rotor flux linkage, when the
reference frame inductances stator resistance is neglected. By using the actual voltage
q-axis Current in rotor Stator vector supplied by PWM inverter, position of the stator flux
reference frame resistance vector in stationary reference frame can be determined.
Permanent Magnet Flux Consequently, the torque angle ( ) can be changed as shown
Rotor Speed in Fig.2.
Linkage
The torque equation for an Interior Permanent Magnet The basic flux equation using machine voltage and current can
be written as [24]:
Synchronous Motor (IPMSM) can be given as:
3 = ( − ∗ ) (5)
= ∗ ∗ + − ∗ ∗ (3)
2 2
Where, P is the number of poles.
The resultant flux linkage( ) can be given as:
 +∆ TABLE I: TABLE FOR OPTIMAL VOLTAGE VECTOR SELECTION
[24]
∆ = ∗∆ I II III IV V VI
1
1 2
+∆
3
1
2 2
3
The basic block diagram of the conventional DTC scheme is
Fig.2: Control of Flux and Torque in a DTC Drive shown in Fig.3.
Neglecting the stator winding resistive drop, (5) can be re
written as: + V −

= ( )
(6) Torque
or Comparator Switching
Δ = ∗Δ ∗
+ ∆ Table

where, , , and are the stator flux linkage, voltage, −

current and winding resistance respectively. Ψ∗ +

The reference torque is generated by processing the speed − ∆Ψ

Ψ Flux Sector
error though a PI controller whereas the actual torque of the Comparator Information
machine is estimated using the (5).
Torque and Flux
3 (7) Estimator
= ∗ ∗ − ∗
2 2
where, and are the stator flux linkages in the stationary
reference frame.
and are the stator currents in the stationary reference PMSM
frame.
Fig.3: Basic block diagram of DTC for PMSM
= ( − ∗ ) (8) According to the stator flux position and the output signals of
hysteresis controllers, an optimal voltage vector to be applied
= − ∗ (9) to the stator winding is selected. The optimal voltage vector is
selected such that the error of torque and stator flux in each
−1 −1 control period is minimized. The selection of optimal voltage
2 1 2 2
where, = (10) vector is implemented with reference to a look Up table
3 0 √3 −√3 mentioned in Table I.
2 2
3.2 Maximum Torque per Ampere (MTPA)
′ ′ can be either voltage or current of the machine. The torque produced in the Interior Permanent Magnet
The reference flux can be obtained from (4) and the actual Synchronous Motor (IPMSM) has two components: magnetic
flux of the machine can be calculated as: torque , produced by permanent magnets and
= + (11) reluctance torque ( ), produced by reluctance of machine.
The two torque equations are shown in (13) and (14).
= tan (12) 3
= ∗ ∗ (13)
The angle obtained in (12) gives the information about the 2 2
sector in which the stator flux is located. 3
= ∗ − ∗ ∗ (14)
In the DTC scheme, the reference stator flux ∗ and reference 2 2
torque ∗ are compared with their corresponding estimated Unlike conventional wound rotor salient pole synchronous
values. The error output is then processed through two motor, in an IPMSM q-axis inductance exceeds the d-
hysteresis comparators, namely Torque Hysteresis Comparator axis inductance ( ). Hence for the positive values of the
and Flux Hysteresis Comparator. The flux hysteresis direct axis current, the reluctance torque will be more
comparator is a two-level comparator, while the torque negative. Consequently, higher quadrature axis current is
hysteresis comparator is a three-level comparator. required to produce certain amount of torque resulting in
higher copper losses and reduced efficiency. The main idea of
MTPA is to obtain a particular and pair that causes
minimum phase currents. To implement MTPA, the
electromagnetic torque equation given in (3) has to be
differentiated w.r.t. and equated to zero. Subsequently, the
direct axis current can be given by the following equation:
′

=− − + (15) . .
2∗ − 4∗ − ′

3.3 MRAC based Speed Estimation:

Among all the techniques available for senseless speed
estimation, MRAC is widely accepted due to its simplicity and ⁄

stable operation even in regenerative operation. Also the

method does not require any extra hardware or signal injection Fig.7: Stator Current based MRAC for Speed Estimation
or huge memory like EKF or AI/ANN. The basic structure of In the proposed scheme, Stator current based MRAC is used
MRAC is shown in Fig.5. for speed estimation as shown in Fig.7. PMSM is considered
Reference Model as reference model and the stator current equations are
= ( , )
+
considered as adjustable model.The stator current equations
− can be written as:
Adjustable Model
= ( , )
=− ∗ + ∗ ∗ + (17)
Adaptation
Mechanism

Fig.5: Structure of MRAC =− ∗ − ∗ ∗ − ∗ + (18)

=−
The above equations consist of the speed term and hence can
be used in adjustable model. For the reference model and
can be obtained directly from the PMSM three phase
current. Thus, using a PI controller as an adaption mechanism
the MRAS based speed estimation algorithm can be given as:
, = + ∗ − ∗
Fig.6: Equivalent Non Linear Feedback System
(19)
MRAC model uses outputs of two models: one independent of − −
parameter which has to be evaluated (Reference Model) and
where, , is the estimated rotor speed.
the other dependent on the parameter which has to be
IV. RESULTS
evaluated (Adjustable Model), to form an error signal.
4.1 Speed Estimation
Reference model is a Linear Time Invariant [LTI] model
Speed Command is changed from 80 rps to 40 rps.
which should be strictly positive real whereas feedback
adjustable model is a nonlinear time variant model which must In Fig.8, commanded speed , speed estimated using
satisfy Popov’s integral inequality, according to which: MRAC ( )and actual/measured speed of machine ( )
are shownin blue, red and green respectively. Fig.9 shows the
≥ 0; ∀ > 0 (16) state of charge (SoC) of battery. The stator three phase
where, and are expressed in Fig.6. currents are shown in Fig.10. In the figure, with the decrease
Generally, a PI controller is used in the adaptation mechanism in speed command, decrease in stator current frequency is
for the convergence of the error. The stability of the MRAS is observed. In Fig.11 reference quadrature and direct axis
governed by the Popov’s Hyperstability Criteria. It includes currents are shown. Fig.12 represents load torque ( ),
rigorous mathematical derivation and proves that the reference torque generated from speed error and
proportional integral based adaptation mechanism provides estimated torque or actual torque ( ) from top to bottom
quick and stable convergence of error signal to zero value. order.
 4.2 Speed Reversal
Speed Command is changed from 80 rps to -80 rps.
In Fig.13, commanded speed , speed estimated using
MRAC ( )and actual/measured speed of machine ( )
are shown in blue, red and green respectively. Fig.14 shows
the state of charge (SoC) of battery. The stator three phase
currents are shown in Fig.15. In Fig.16 reference quadrature
and direct axis currents are shown. Fig.17 represents load
torque ( ), reference torque generated from speed error
Fig.8: Speed Command of PMSM
and estimated torque or actual torque ( ) from top to
bottom order.
In Fig.18, − stator flux component’s trajectory is shown.
V. CONCLUSION
In this paper, a Direct Torque Controlled PMSM drive is
presented. In order to have low copper losses and optimized
efficiency, MTPA control is implemented. Due to high cost,
additional maintenance and limitation of drives
implementation in rugged environment, the proposed drive is
made sensorless. The speed estimation is carried out using
Fig.9: SoC of Battery stator current based Model Reference Adaptive Control.
Simulation results shows the proper functioning of a speed
sensorless, direct torque controlled PMSM drive incorporating
MTPA. The motor parameters used for simulation are given in
appendix I.

Fig.10: Three Phase Stator Current

Fig.13: Speed Reversal for PMSM

Fig.11: d-q Components of Stator Current Fig.14: SoC of Battery

Fig.12:Torque Response of PMSM drive

Fig.15: Three Phase Stator Current
 [5] C. C. Chan; “The State of the Art of Electric and Hybrid Vehicles,”
Invited Paper, Proceedings of the IEEE, Vol.90, No. 2, February 2002
[6] Chan, C.C.; , "The State of the Art of Electric, Hybrid, and Fuel Cell
Vehicles," Proceedings of the IEEE , vol.95, no.4, pp.704-718, April
2007
[7] M. S. Merzoug, F. Naceri, "Comparison of Field-Oriented Control and
Direct Torque Control for Permanent Magnet Synchronous Motor
(PMSM)," WASET, International Journal of Electrical and Computer
Engineering, Vol:2, No:9, 2008
[8] Takahashi, I., Noguchi, T., "A new quick-response and high efficiency
control strategy of induction motor," IEEE Trans. Ind. Appl., 1986,22,
(5), pp. 820–827
Fig.16: d-q Components of Stator Current [9] Depenbrok, M., "Direct self-control (DSC) of inverter-fed induction
machine," IEEE Trans. Power Electron., 1988, 3, (4), pp. 420–429
[10] L. Zhong, M. F. Rahman, W. Y. Hu, K. W. Lim, "Analysis of Direct
Torque Control in Permanent Magnet Synchronous Motor Drives,"
IEEE Trans. on Power Electronics, VOL. 12 No. 3, May 1997
[11] Feng Niu, Bingsen Wang, Andrew S. Babel, Kui Li, Elias G. Strangas,
"Comparative Evaluation of Direct Torque Control Strategies for
Permanent Magnet Synchronous Machines," IEEE Trans. on Power
Electronics, VOL. 31, No. 2, February 2016
[12] O. Ouledalia , A .Meroufelb , P. Wira ,S.Bentouba, "Direct torque
fuzzy control of PMSM based on SVM," Elsevier, Energy Procedia, 74
(2015) 1314 – 1322
[13] Mohamed Taha Elsayed, Osama Ahmed Mahgoub, Sherif Ahmed
Zaid, "Simulation study of Conventional Control vs. MTPA Based for
PMSM Control," Proceedings of the 14th International Middle East
Fig.17:Torque Response of PMSM drive for speed Reversal Power Systems Conference (MEPCON’10), Cairo University, Egypt,
December 19-21, 2010
[14] Shoudao Huang, Ziqiang Chen, Keyuan Huang, Jian Gao, "Maximum
torque per ampere and flux-weakening control for PMSM based on
curve fitting, IEEE Conf. on Vehicle Power and Propulsion," Sept
2010
[15] Atsushi Shinohara, Yukinori Inoue, Shigeo Morimoto, Masayuki
Sanada, "Maximum Torque Per Ampere Control in Stator Flux
Linkage Synchronous Frame for DTC-Based PMSM Drives Without
Using q-Axis Inductance," IEEE Transactions on Industry
Applications, Vol. 53, No. 4, July/August 2017
[16] Abbas Najjar-Khodabakhsh, Jafar Soltani, "MTPA control of
mechanical sensorless IPMSM based on adaptive nonlinear control,"
ISA Transactions 61 (2016) 348–356
[17] Suman Maiti, Chandan Chakraborty, Sabyasachi Sengupta,
"Simulation studies on model reference adaptive controller based
Fig.18: − Component of Stator Flux speed estimation technique for the vector controlled permanent magnet
Appendix I: synchronous motor drive," Simulation Modelling Practice and Theory
17 (2009) 585–596
Motor Parameters [18] Suman Maiti, Chandan Chakraborty, "A new instantaneous reactive
Pole Pair 4 power based MRAS for sensorless induction motor drive," Simulation
Stator Resistance ( ) 0.62 Ω Modelling Practice and Theory 18 (2010) 1314–1326
Direct Axis Inductance ( ) 0.001971 H [19] Osama M. Arafa, Ghada A. Abdel Aziz, Mohamed I. Abu El-Sebah,
Quadrature Axis Inductance 0.002179 H Ahmed Aly Mansour, "Observer-based sensorless speed control of
PMSM: A focus on drive’s startup," Journal of Electrical Systems and
Flux Linkage established by magnets 0.08627 Wb
Information Technology 3 (2016) 181–209
Co-efficient of Inertia ( ) 0.0003617 Kg.m2 [20] E.G. Shehata, "Speed sensorless torque control of an IPMSM drive
Viscous Co-efficient 5e-04 with online stator resistance estimation using reduced order EKF,"
Rated Torque 6 N-m. Electrical Power and Energy Systems 47 (2013) 378–386
Rated Speed 80 rps [21] Jemaa Brahmi, Lotfi Krichen, Abderrazak Ouali, "A comparative
Acknowledgement: study between three sensorless control strategies for PMSG in wind
The work is funded by theCouncil of Scientific & Industrial energy conversion system," Applied Energy 86 (2009) 1565–1573
[22] R. Krishnan, Electric Motor Drives: Modelling, Analysis and Control.
Research(CSIR) – GoI. [23] P. Pillay, R. Krishnan, "Modelling simulation and analysis of
Reference: permanent magnet synchronous motor drives. Part-I: The permanent
magnet synchronous motor drive," IEEE Trans. Ind. Applicat. (1989)
[1] M Ehsani, YiminGao, S. E. Gay, Ali Emadi, “Modern Electric Hybrid
265–273.
Electric and Fuel Cell Vehicles – Fundamental, Theory and Design”,
[24] B.K. Bose, Power Electronics and Variable Frequency Drives
CRC Press
Technology and Applications, IEEE press, New York, 1996.
[2] http://www.ucsusa.org/clean-vehicles/california-and-western-
states/californias-zevprogram#.V_YYK-Z97IV
[3] http://www.zevfacts.com/zev-states.html
[4] http://www.evaap.org/info/Psgubun-1_info.html