Hybrid EAs for backup sensorless control of PMSM drives

This paper presents a robust strategy to increase the reliability of Permanent Magnet Synchronous Motor drives against encoder or resolver failures. If the position sensor fails, the "failure and recovery manager" switches the drive in sensorless mode, i.e., the control system uses the speed and position feedback given by a Sliding-Mode observer in place of the sensor. As the accuracy of the sensorless control depends on the tuning of the observer parameters according to the motor conditions, the Sliding-Mode observer is periodically tuned during the sensor-based control of the drive. A fast and robust tuning of the observer can be obtained by Hybrid Evolutionary Algorithms. This prevents untimely sensored-to-sensorless switching due to speed transients and allows better performances of the sensorless control when the position sensor fails actually. The results carried out prove that the position sensor failures do not affect the drive operation, and proposed HEA outperforms the standard search algorithms.