PWM-Based Predictive Direct Torque Control of Switched Reluctance Machine for Accurate Torque Tracking With Minimization of Phase RMS Currents

This paper proposes a novel PWM-based optimal predictive torque controller for Switched Reluctance Machines (SRMs). The highly non-linear characteristics of flux-linkage and phase torque present challenges in achieving real-time optimal torque control in SRMs. In this work, a cost function, encompassing the instantaneous torque error and the RMS values of phase currents is formulated to be minimized. An analytical expression for the optimal duty ratio towards this objective is derived resulting in improved computational efficiency. The performance of the proposed torque controller is compared with two state-of-the-art finite control set predictive torque controllers through extensive simulations and experiments. The tests are conducted at various operating torque and speed levels on a 4-phase, 8/6 pole SRM. Results demonstrate that the proposed controller delivers better torque tracing performance in terms of lower average and RMS torque errors than the existing techniques. Also the proposed method results in higher torque per ampere, lower sound pressure levels (SPL emission), and lower computational time. The proposed controller tracks the torque accurately even in the presence of −20% to +20% modeling errors.

I. Introduction

Switched reluctance machine (SRM) is a promising permanent-magnet-less alternative for emerging applications like electric vehicles [1], [2]. SRM is known for its rugged structure with no rotor windings [3]. Simple magnetic structure and concentrated windings make this machine easy and cost-effective to manufacture [4], [5]. Inherent fault tolerance [6] and robust rotor made SRM popular in high-speed applications such as gas turbine generation [7] and aerospace [8].