Sliding mode controller based on fuzzy neural network optimization for direct torque controlled PMSM

A sliding mode controller for permanent magnet synchronous machine (PMSM) is investigated in this paper, in which direct torque control (DTC) concept, variable structure control and space vector modulation (SVM) are integrated to achieve high performance. It features in very low flux and torque ripple, strong robustness and fixed switching frequency. Then, an FNN is investigated to optimize the control gain matrix of sliding mode controller. The inputs of the FNN are the switching surface and its derivative, and the output of the FNN is the estimated bound of uncertainties. The theoretical analyses for the proposed FNN sliding-mode controller are described in detail. Simulation results show that the proposed FNN sliding-mode controller provides high-performance dynamic characteristics and is robust with regard to plant parameter variations. Furthermore, comparing with the sliding-mode controller, the chattering phenomenon is much reduced by the proposed FNN sliding-mode controller.

I. Introduction

Since Depenbrock [1] and Takahashi [2] proposed direct torque control (DTC) for induction motor drives in the middle of 1980s, more than one decade has passed. The basic idea of DTC for induction motor which is to control the flux linkage and torque by selecting the voltage space vectors properly is now being adopted by the industry. In the late 1990s, DTC techniques for the permanent magnet synchronous machine (PMSM) machines have appeared [3]–[5]. Compared with traditional vector control, DTC for PMSM has the advantages of simple structure, fast torque response, good robustness etc.