Abstract:
A Pulsewidth Modulated Voltage-Fed Inverter Vector-Controlled Permanent Magnet Synchronous Motor (PMSM) Drive based on Hysteresis Current Control (HCC) is presented in this work. A detailed conceptual dq modelling of the PMSM was undertaken in the rotor reference frame for open loop studies, thereby setting the pace for the variable speed drive (VSD) of the PMSM which, inherently, is not capable of variable speed operation. Subsequently, vector control by Field Orientation Control (FOC), is used to decouple the flux and torque producing stator current components of the PMSM thereby permitting independent and precise control of flux and torque as obtainable in separately excited dc machines. A complete closed loop control system employing an outer PI speed controller and an inner hysteresis current controller was implemented to realize this speed-controlled drive. Since torque can be made proportional to current either in the stationary or rotor reference frames and effective control of current gives effective control of torque, speed and position, the HCC strategy is aimed at ensuring that the actual motor phase currents track their respective sinusoidal references. The HCC algorithm was developed and employed for the logical firing of the power semiconductor switches of the inverter. The control algorithm was optimised to obtain fast speed response, while maintaining effective current and torque tracking for all practical speed inputs namely Constant, Step and RAMP reference speed inputs. The optimal control variables were identified with emphasis on effective current and torque tracking. Four quadrant operation of the PMSM was also implemented as obtains in numerous applications in industry where controlled starts and stops are required in both forward and reverse directions. Compared to the standard AC6 of MATLAB Simpower systems, the developed model achieved rise time and settling time of 0.0108 seconds and 0.0143 seconds respectively while the corresponding values for AC6 model are 0.1944 seconds and 0.1984 seconds respectively. This, clearly, shows that the developed model has an enhanced speed response.
