High-grade torque control of switched reluctance motor based on neural-fuzzy network
基于模糊神经网络开关磁阻电动机高性能转矩控制

The primary disadvantage of an SRM was the higher torque ripple which was due to the highly nonlinear and discrete nature of torque production mechanism. Based on the experimental data of static torque characteristic, a fuzzy-neural network(FNN) was applied to the learning of its inverse model off line. Then according to the predefined torque distribution function (TDF), optimal current profile was real-time gained by the FNN on line, which resulted in a linear, decoupled, low ripple control of torque. The effectiveness of the proposed method was demonstrated by computer simulation results.