The Study of FNN Control on Direct Torque Control for BLDCM

In order to improve the response speed of brushless DC motor (BLDCM) control system and simplify its structure, a novel scheme is proposed in this paper. Direct torque control (DTC) scheme and fuzzy neural network strategy are combined in the new scheme. DTC has the advantage of fast response and simple system structure without complex vector transform unit. However, its torque ripple is relatively high. In the fuzzy neural network DTC system, exact magnet flux linkage can be acquired by neural network method and the actuation duration of voltage vector is fuzzily adjusted according to both torque error and its variance ratio to reduce torque ripple. Simulation results prove that the proposed control strategy is feasible. At last, it is verified by experiment results that the fuzzy neural network strategy executes effective control on torque and current. Consequently, the experimental system operates with good steady and transient behavior