Five-Level Inverter Fed Sensorless PMSM Drive Using Modified MPTC with Novel Reduced Search Space Algorithm

Multi-level inverters (MLI's) have gained vast interest in the field of drives to enhance the performance of the drive by minimizing the torque ripples. Developing the vector control algorithms for the MLI fed drives is one of the challenging tasks. In the era of vector control methods, the model predictive control has proven its standards in improving the performance of the drive over other methods, such as, field-oriented control, direct torque control, etc. Among the model predictive control methods, the modified model predictive torque control (MMPTC) method provides better torque and speed characteristics. However, the MMPTC method for multi-level inverter requires more computational time and it increases the burden on the controller. Therefore, a novel reduced search space algorithm is proposed based on the geometrical congruity between the two level and multi-level space vector diagrams to reduce the computational complexity of multi-level inverter switching. Further, dependency of speed sensor information in the controller makes the system vulnerable and speed sensorless drive make the system stronger. In this article, the five-level inverter fed MMPTC based sensorless PMSM drive with novel reduced search space algorithm is proposed and the performance characteristics of proposed method is compared with two-level and three-level inverters. The hardware prototype setup is developed in the laboratory and the simulation results are validated with the experimental results over a wide speed range.