Forging a New Path for Electric Vehicles with Sensorless Motor Drives

This a retrieved form of the original article from the series, ‘A Spotlight on Undergraduate Research’. Edited by Niharika Gunturu.

The author likes his coffee (Kalman) filtered, and plays his AC/DC records on Synchronous Motors

Sameer Kulkarni, a 5th-year Dual Degree student from the Department of Electrical Engineering, gives us a glimpse into his Dual Degree Project. It involves the development and hardware implementation of a sensor-less motor drive for a Synchronous Reluctance motor – a machine that might soon become the most preferred motor for Electrical Vehicles in the Indian context.

1 A quick intro to electric motors

Electric Motors are devices that convert electrical energy into mechanical energy. They have been at the heart of modern industry ever since they were introduced and are still the driving factor behind a lot of innovation.

An electric motor works on the principle of Lorentz Force. Essentially, when a conductor is placed in a magnetic field, it experiences a force perpendicular to the flow of current as well as the magnetic field. Motors make use of this fact in order to produce a rotational motion. They consist of a stator which is the fixed component and a rotor which is the rotating component. Since the electric motor is primarily an electromagnetic device, the rotor and stator are both made of magnetic materials – such asiron. Further, in most motors, the stator as well as the rotor, both produce a magnetic field – either by using magnets or using a coil – and the interaction of these two fields results in a rotational motion.

Depending on how the field is produced and which components produce it, electric motors can be broadly classified into 3 types.

The DC motor – In this type of motor, a static magnetic field is produced in the stator windings and the direction of current flowing in the rotor winding is reversed using a mechanical commutator. Needless to say, the supply to this motor is DC.

Figure 1: DC motor working principle

The Induction motor – In this type of motor, a rotating magnetic field is produced in the stator and this rotating magnetic field induces a current in the rotor windings. Needless to say, induction motors require an AC supply to produce the rotating magnetic field.

Figure 2: Induction motor working principle

The Synchronous motor – These motors are sometimes referred to as a hybrid between the induction motor and the DC motor. They have an AC supply for the stator winding and a DC supply for the rotor windings. As a result, the rotor rotates at the same frequency as the rotating magnetic field set up in the motor – hence the name synchronous.

Figure 3: Synchronous motor construction

1.1 Synchronous Reluctance Motor

In my DDP, I work on what is called a Synchronous Reluctance Motor(SyRM). It’s largely like the synchronous motor except for the fact that there are no windings in the rotor and the magnetic field is only produced in the stator windings. This motor uses the principle of Magnetic Reluctance. Reluctance is to magnetic circuits what resistance is to electric circuits. In the SyRM, the rotor isn’t radially symmetric and hence for a given orientation of the magnetic field, the rotor always tries to align itself such that its reluctance is minimum.

Figure 4: Simple reluctance motor

These motors have been around since the 1920s[1] and have been proven to be more efficient than the most popular motors around – the induction motors[2]. They also have a better power density compared to most motors and are only bettered by the PMSM(Permanent Magnet Synchronous Motor) in that regard. Unfortunately, these motors have not seen much commercial use. This was mainly due to the difficulty involved in manufacturing the rotor, and due to the inherent ripple present in the torque response of this motor. They also invariably require a variable frequency drive(VFD) for their operation. However, with current technology, these problems no longer persist and hence SyRMs are being looked at as the motor of choice for Electric Vehicle applications.

2 Motor drives

With the introduction of motors, it also became necessary to effectively control their output characteristics, namely their speed and torque. For the early motors like DC motors and induction motors, the control method usually involved mechanical gearboxes or voltage control using resistors. Although these methods got the job done, they were slow and extremely inefficient. Further, the degree of control they offered was limited. This made them unsuitable for applications requiring high precision and also led to significant resistive losses whenever resistors were involved. This drawback of conventional control systems paved the way for the field of motor drives. With the introduction of semiconductor switching devices like IGBTs and MOSFETs, mechanical switching and resistive voltage dividers were almost entirely replaced by electrical switching, and the use of microcontrollers allowed for relatively complex closed loop control methods to be implemented easily. This system, involving a power electronic switching circuitry, a sensing system and a microcontroller in conjunction with the motor itself is essentially what a modern ‘motor drive’ is.

Figure 5: A basic motor drive block diagram

2.1 Sensorless controller

For effective closed loop control of a motor, it is necessary to accurately know the rotating speed and the currents flowing into the motor. In case of a SyRM, it is also necessary to know the exact orientation of the rotor(remember, the rotor is not radially symmetric). Although one can use rotary encoders for sensing the rotor position and speed, a failure in the sensor would lead to a malfunction of the entire drive system. This is actually a more dangerous situation than the motor coming to a standstill. To begin with, detecting a malfunction would be difficult. Further, depending on where this motor is being used, this malfunction could lead to adverse effects (e.g. wasted raw material due to incorrectly manufactured parts). And finally, the motor itself could break down due to operation beyond rated quantities. Thus, the control loop performance was closely tied to the rotary encoder accuracy and integrity. Of course, this also meant that one had to shell out more money to get better precision.

To tackle these issues, a rotational sensorless system was proposed. Since a microcontroller is capable of quick computations, part of its computational power was now directed to the estimation of the rotor angle and speed using the dynamic equations of the motor. As a result, it was now possible to completely forego the rotary encoder and have the drive performance depend almost entirely on the control algorithm itself. Additionally, this also reduced the cost of the system and made it more robust.

Various techniques have been proposed to estimate the rotor angle in the SyRM with the popular ones being flux angle estimation and signal injection. However, for my DDP, I am using a somewhat less frequently used approach – the Extended Kalman Filter(EKF). It is computationally more expensive compared to the former two approaches but it takes away the need to explicitly filter the remaining variables.

The Kalman Filter

Estimation of a variable from its equations is more of a mathematical problem. Essentially, you are given a set of noisy readings and you are required to estimate the true value of a variable as accurately as possible from those readings. Of course, it is assumed that you know how the variable being estimated is related to the readings you are taking. While there exist various techniques to solve this problem – some being system specific like the ones mentioned in the earlier paragraph, and some being applicable for any general system – one technique which has proven to be particularly effective for a wide range of linear systems is the Kalman filter.

Figure 6: Kalman Filter working principle

A Kalman filter consists of two steps – prediction and correction. In the prediction step, the dynamic equations of the system are used to predict the next state of the system. In the correction step, the measurements obtained from the system are used to update the predicted system state. This process repeats at each time step and in most cases, it narrows down to the true value of the variable very quickly. While the Kalman Filter is guaranteed to work on any linear system, the SyRM, unfortunately is non-linear. Therefore, an EKF, which is basically an add-on that adapts the Kalman Filter for a non-linear systems, is used for the SyRM rotor angle estimation.

The Project

Now that you know what motors and motor drives are all about, and why sensorless control is necessary, it should be easy to understand what my DDP is about with just the title. So, brace yourselves.

My DDP is titled – Design Of An Extended Kalman Filter Based Sensorless Synchronous Reluctance Motor Drive.

Since motor drives are very well researched, it is almost impossible to have any breakthrough in this field. However, minor tweaks to existing works are always possible. In my case, I am using the EKF to not only estimate the rotor position and speed, but also the load torque on the motor. If implemented successfully on the motor, this drive could be used as an example to demonstrate the capabilities of the SyRM – which has for long been neglected – and provide a viable alternative to the Indian EV industry when it comes to motor drives.