Direct-axis Dead-time Effect Compensation Strategy Based on Adaptive Linear Neuron Method for PMSM Drives

Abstract

The dead-time effect of inverter will deteriorate motor control performance. This paper presents a compensation strategy based on adaptive linear learning to improve the dead-time problem of PMSM drives. The essence of the compensation strategy is to suppress the sixth harmonic current in the synchronous reference frame. An adaptive harmonic current decomposer, self-tuned by the recursive least square algorithm, is applied to extract the sixth harmonic current in the synchronous reference frame, then the direct-axis voltage is compensated by PI controller. This method greatly suppresses the fifth and seventh harmonics in the phase current. In addition, the proposed method is easy to implement, and the effectiveness of the compensation method is verified by experiments.