Due to the high volume of mainstream vehicles with internal combustion engines (ICEs), environmental pollution and energy shortages are now a concern. Electric vehicles (EV) can restrict the energy source and the ideal method to save resources and provide zero-emission vehicles.

The EV's critical feature is an electric motor as illustrated in Figure 1, so its selection is crucial. In recent decades, many types of electric motors were evaluated and analyzed for EVs. Currently, switched reluctance machines (SRMs) are attracting attention in the research community because of their various advantages. These operate with overall efficiency compared to an induction motor having the same rating, as their windage and friction losses are identical.

This article introduces the SRM and its operating principle along with its pros and cons. Finally, the current control methods for SRM will also be discussed.

Figure 1: Internal structure of an EV depicting electric motors.

Switched reluctance motor and its operating principle

A switched reluctance motor produces torque by changing its magnetic reluctance as shown in Figure 2. Its stator has salient poles and includes windings identical to a brushless DC motor, but the rotor is made of steel cut into salient poles without magnets or windings. The power is supplied to its stator windings instead of the rotor, unlike the standard brushed DC motors. SRM works by alternating currents in the stator when the magnetic field developed by stator and rotor changes. To prevent a condition where both rotor and stator poles align up together and no torque is produced, switched reluctance motors have fewer rotor poles than the stator.

Figure 2: Structure of a switched reluctance motor .

The magnetic circuit developed between rotor and stator has high reluctance when they both are out of alignment. At this time, the stator pole pairs get energized, and the rotor tries to get in line with the powered stator poles, which decreases the magnetic reluctance. This ability of rotor to reach the minimum point of reluctance produces a torque, known as reluctance torque. Excitation of the stator poles must be accurately timed to make sure that it happens only when the rotor is trying to be aligned with the excited pole. For this purpose, SRM may need positive feedback from Hall effect sensors or encoders to control the excitation of stator based on an accurate rotor position.

Advantages and disadvantages

A switched reluctance motor offers many benefits against other types of electric motors because of its control flexibility, simple structure, lower cost and high efficiency. The lack of winding or permanent magnet on the rotor means that an SRM is appropriate for extremely high speed applications, and can withstand high temperatures. Furthermore, it results in a rugged and simple structure and low manufacturing cost. At the same time, if a fault occurs in any one winding or phase, the motor can still work but at a reduced load. This is particularly important for fans and pumps, and also in the electric vehicle world. However, SRM has non-linear characteristics because of magnetic saturation, which makes it complicated to accurately control its torque.

For medium and small size EVs, different types of permanent magnet motors are currently in use, but they are becoming expensive with time and will be very costly if used in heavy duty vehicles. Thus, switched reluctance motors being cost-effective and rugged are high tech for the electric vehicles. The usage of electric vehicles will not be limited in harsh environments, as SRMs can work under high temperatures. The most crucial disadvantages of SRMs are high noise, vibrations and torque ripple. It requires advanced control methods when compared with other AC and DC motor drives. Some of the current control methods are discussed in the next section.

Control methods

Switched reluctance motors suffer from high torque ripple because of discrete torque production and independent phases. Thus, it is its major drawback and may restrict its use in EVs. However, many control strategies are currently being used to solve this problem. Broadly speaking, there are two main torque control methods for SRM: direct and indirect torque control.

Direct torque control strategy includes a simple control system with a hysteresis controller. The non-linear characteristics of SRM are taken into account to compensate for the output torque ripple based on current and rotor position. Its main benefits are simplicity, high performance, and fast torque response. In many advanced strategies, these characteristics are utilized to fluctuate the reference current depending upon rotor position and reference torque. This type of torque control is also subdivided into advanced direct instantaneous torque control (ADITC) and direct instantaneous torque control (DITC).

In DITC, the average torque may be controlled within a certain bandwidth using estimation of the instantaneous torque. However, it is not possible to control the instantaneous torque and performance is dependent upon sampling time. In the ADITC method, the different phase currents in a single sample time are controlled by regulating the average phase voltage. Compared with DITC, it can get a smaller torque ripple by increasing the sampling time. But, if the sampling time is the same, ADITC switching frequency will be almost double, and thus there will be more electromagnetic emissions and switching losses than DITC.

There are three types of indirect torque control: nonlinear, cosine and linear logical torque sharing function. The nonlinear TSF method has given promising performance in comparison to the other two. The torque is controlled by directly controlling the phase currents, which in turn control the torque. The torque reference is converted into a corresponding phase current reference. However, this conversion is complicated as the torque is dependent upon the rotor position. The relationship between rotor position, torque and current is nonlinear, and thus it cannot be expressed as an analytical expression. However, it can be formulated using artificial neural networks or a lookup table.