Experimental comparison of MRAS and Sliding Mode Observer for PMSM sensorless control

This paper presents the comparison between sensored field oriented control, model reference adaptive system, and sliding mode observer sensorless techniques for isotropic permanent magnet synchronous machines considering performance and microprocessor resources usage. These three algorithms were implemented on the TMS320F28069M microcontroller using MATLAB Simulink<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">TM</sup> Embedded Coder for Texas Instruments for the 2MTR-DYNO motor kit. After recalling the theoretical aspects of these techniques, the experimental testbench is presented. The tests compare these strategies in terms of torque-speed operating region, average estimated speed error, and microcontroller computation time and memory allocation. As a conclusion, MRAS and SMO had a similar performance in terms of torque-speed characteristic, speed precision and memory allocation. When compared with SMO, MRAS showed a superior torque capacity at lower speeds and a more straightforward tuning procedure due to its lower number of tuning parameters. On the other hand, SMO required a lower computational time than MRAS.