Self-learning MTPA control of interior permanent magnet synchronous machine drives based on virtual signal injection

Abstract

This paper describes a novel self-learning maximum torque per ampere (MTPA) control scheme for interior permanent magnet synchronous machine (IPMSM) drives to achieve fast dynamic response in tracking the MTPA points without accurate prior knowledge of machine parameters. The proposed self-learning control scheme (SLC) generates the optimal d-axis current command for MTPA operation after training. Virtual signal injection control (VSIC), which has been recently developed as a novel parameter-independent MTPA points tracking scheme, is utilized to train the SLC and compensate the error of the SLC during its operation. In this way, the proposed SLC can achieve the MTPA operation accurately with fast response and the online training of the SLC will not affect MTPA operation of IPMSM drives. The proposed control scheme is verified by simulations under various operation conditions on a prototype IPMSM drive system.