Sensorless Predictive Control of SPMSM-Driven Light EV Drive Using Modified Speed Adaptive Super Twisting Sliding Mode Observer With MAF-PLL

Abstract

Sensor reduction is a key sought after requirement for a surface mounted permanent magnet synchronous motor (SPMSM) electric vehicle (EV) drive to improve reliability and to reduce its cost. However, achieving a smooth and successful wide speed range using sensorless control is still a challenge. This article presents a modified higher order sliding mode observer (SMO)-based position sensorless predictive current control for wide range speed operation of an SPMSM-driven light EV. For reducing chattering and singularity phenomenon in conventional SMO, a modified speed adaptive super twisting SMO with moving average filter phase-locked loop (PLL) is deployed to accurately estimate the position and speed of SPMSM. Moreover, optimal switching pulses are provided to a voltage source inverter using time delay-compensated finite control set-model predictive current controller, eliminating the need of proportional integral controllers and modulators. Simulated performance of the control scheme is presented in MATLAB Simulink and its hardware implementation is executed using an SPMSM EV drive laboratory test setup. A comprehensive study of different PLL-based SMOs is carried out and the effectiveness of proposed system against conventional and other methods is demonstrated experimentally. It is observed that this method allows seamless sensorless operation at 3.95% of rated speed.