Various control methods of permanent magnet synchronous motor drives in electric vehicle: a technical review

Day by day use of internal combustion engines (ICE) compared to electric vehicles (EV) is deteriorating because mainly of pollution and their less fuel availability. In the present scenario, an electric vehicle plays a major role in place of an ICE vehicle. So that performance of EV can be improved by proper selection of electric motor. Initially, EV prefers induction motors for traction purposes, but complexity in controlling induction motor, permanent magnet synchronous motor (PMSM) presently used in EV by most of the electric vehicle manufacturers due to its advantages. This paper reviews on various control methods for PMSM used in EV. Various control methods are being used for EV applications. Initially, conventional direct torque control (DTC) technique being used in controlling electric motors but it has a drawback of high torque and flux ripples. Hence, intelligent controllers are predominantly using in controlling PMSM drives.


INTRODUCTION
In future days, there is a chance of depletion of petroleum products. Because the availability of petroleum products is only in particular areas of the world. For past years, we were using these petroleum products for running internal conbution (IC) engines. However, it produces harmful pollutants to the environment, which increases global warming. Hence, we should move towards the alternating source for a vehicle that replace the internal combustion engines (ICE) vehicle. Obviously, eclectic vehicles are the alternative source for the vehicle.
Pollution is not only parameter to be considered in designing electric vehicle but also very important is its driving performance [1]. So to improve the performance, choice of electric motor and corresponding technique adopted for controlling motor is also important [2]. The Figure 1 represents various parts of electric vehicles (EV).
Complete electric vehicle consists both electrical components and mechanical components. It consists of converter system, controller, electric motor, mechanical device and wheel drive. Controller, power converter and motor are electrical components of the entire EV system and rest are mechanical components. It also consists of battery system [1], [3]. Controller and electric motors are heart of the EV system. Hence, selection of electric motor plays an important role in EV system. Selecting motor should have more torque density, simple to control, more efficient and low cost.
Early day's direct current (DC) motor was better choice for electric vehicles because of their simpler control method and its characteristics are most preferable for EV motor. Presence of commutator and brushes makes the motor more complex for electric vehicle. To overcome this drawback of dc motor, induction motor  [4]. Absence of commutator and brushes makes it simple in construction [5]. Analysis become complex due its non-linear characteristics [6], [7]. To overcome induction motor drawbacks permanent magnet synchronous motor (PMSM) finds development in EV applications. PMSM has higher current density and more efficient than induction motor. Hence it is very much suitable for EV applications. Comparison is done in Table 1 on electric motors with respect to different parameters.
It is better to choose particular motor for electric vehicle depend on its requirements. It is better to go with brushless dc motor for 2-wheelers. PMSM or Induction motors are good choice for four-wheeler vehicles. But when considering to efficiency and power density PMSM is better choice for electric vehicle. By employing modern control techniques, it is possible to control PMSM drive very efficiently. Section 2 explains modelling of PMSM drive. Section 3 describes the control techniques involved in controlling PMSM drives.

PMSM MATHEMATICAL MODELLING
PMSM modelling is explained by taking rotor as reference. The following [9], [10] are expressed for PMSM with respect to rotor reference.
Where: superscript ' ' represents rotor reference frame, = -axis voltage, = -axis current, = -axis voltage, = -axis current -is number of poles, Ѱ = permanent magnet flux linkage, = mutual inductance, = -axis inductance, = -axis inductance, = stator resistance, and = rotor speed. The torque equation of a PMSM is given as: The resultant flux linkage is: In following section literature review is done on various control techniques available for PMSM drive.

CONTROL TECHNIQUES
For EV applications, controlling torque over wide range is necessary. Complexity in controlling PMSM drive is one of the drawbacks. Scalar and vector controls are the basic control methods for ac drives. Scalar control gives good steady state performance. To get high accurate and good dynamic performance vector control methods are used. Proportional integral (PI) controller is commonly applied controller rather than propotional integral derivative (PID) controller, but it leads to instability in the system. Major techniques like field-oriented control and direct torque control have their own advantages and disadvantages for controlling PMSM drive. Field oriented control (FOC) is used to improve steady state performance and to reduce the torque ripples. FOC improves the dynamic response [11], [12] and its performance can be increased. However, this method is sensitive to small changes in the parameter when temperature changes [13]. Takanashi and Noguchi [14] and Depenbrock [15]

1227
(DTC) in place of FOC drawbacks. DTC has a capability to control torque and flux directly by increasing efficiency of PSMS drive. Simplicity of DTC has its one of major advantage. The conventional DTC produces more torque and flux ripples and parameters like torque and flux are measurable efficiently [16]- [18]. The main limitation of DTC PMSM is high torque and flux ripple. It explains new technique called sliding mode controller (SMC) to increase the drive performance [19]- [22]. Chattering phenomenon is major drawback of SMC technique. A combination of two controller's namely fuzzy logic controller (FLC) and sliding mode controller are used to maximize the performance of the system [23], [24]. But the process of using hybrid technique is complex compared to individual technique. Viswanathan and Thathan [25] gives intelligent controllers are better compared to conventional controllers.
Liu et al. [26], and Sun et al. [27] explains fuzzy logic control based PMSM drive. Viswanathan and Thathan [28] DTC which is based on neural network controller for switched reluctance motor (SRM) to reduce torque ripples. Niu et al. [29] uses predefined switching table to minimize the torque and flux ripples by using online optimization method. Pushparajesh et al. [30] uses look up table-based vector control for switched reluctance motor to improve the torque response. Navardi et al. [31] combines conventional DTC and finite predictive control methods to improve performance of PMSM. Fu et al. [32] paper analyses the sliding mode controller in place of PI and traditional DTC to improve the stability of the system. But SMC algorithm has drawback of chattering phenomenon. Sharma, and Bhattacharva [33] explains model reference adaptive control (MRAC) for improving efficiency of PMSM using DTC method.
Vafaie et al. [34] proposes new predictive DTC method, where it improves both steady state and transient state responses of PMSM. Sekour et al. [35] paper proposes PI-fuzzy resistance estimator to remove the error caused by the change in the resistance which happen in low speed of DTC. Shakunthala et al. [36] explains how flux and torque ripples can be lowered in PMSM by using combination of PI and FLC. Rahideh et al. [37] improves performance of PMSM drive compared to conventional DTC by using genetic algorithm (GA)-FLC method. Here propotional intergral ontroller is also used and it is tuned using GA. Nicola et al. [38] explains new strategy called combination of PI and iterative learning control (PI-ILC) for repetitive task in control system. This control technique is used to reduce the speed signal ripples if this motor has repetitive tasks. Jlassi and Cardoso [39] explains model predictive torque (MPT) control for PMSM drives. It has high complexity in calculation process, which reduces the performance of the system, because complexity increases torque ripples. Nicola et al. [40] proposes multiple neural networks used to improve the performance of PMSM. Depending on estimated value of load torque corresponding ANN is selected. It increases the complexity of the network.
Cheng et al. [41] replaces PI controller and conventional DTC (CDTC) controller by model predictive controller (MPC) to get required voltage vector in order to reduce the ripples in the torque, still it suffers from varying switching frequency. It marginally decreases torque ripples compared to conventional MPC. Kakouche et al. [42] replaces combination of PI and DTC by two fuzzy controllers. It gives good performance compared to PI and DTC. As it uses two fuzzy sets, it increases the complexity of the system. This paper does not talk about maximum torque range, as it is applicable for EV. Li et al. [43] proposes new predictive control technique to reduce the cost function and to reduce the computational burden. Wu et al. [44] proposes DTC based on random zero vectors space vector pulse width modulation (SVPWM) to reduce torque ripples. But it has slow torque responses. Marulasiddappa, and Pushparajesh [45] reviews on different control algorithms for induction motor-based EV. The technique used Zhang et al. [46] gives slow torque response. This Table 2 compares four different controllers namely CDTC, SMC-DTC, MPC-DTC and Intelligent controllers. Torque ripples can be minimized by using intelligent controllers in PMSM drives. And also torque response is very high for intelligent controller in PMSM drives.

CONCLUSION
This review paper is done in over viewing the different control methods for PMSM drive, which are used in electric vehicle. Electric vehicles are used over wide range of torque and speed. It is observed that PMSM drive is more preferable than induction motor drive due to various advantages. Various control methods have been employed in controlling over wide range of torque for PMSM. It has been observed that conventional DTC produce more torque ripples compared to other control methods. Due to high ripples performance of PMSM drive will reduce. Hence in order to improve performance of PMSM intelligent controllers are introduced. Comparison is made for different electric motors used in EV. Advantages and drawbacks of each control algorithms are briefed. Control techniques are compared on various performance parameters.