基于MSCO-RBFNN控制的开关磁阻电动机在工业缝纫机中的应用

Control for switched reluctance motor in the application of industrial sewing machine based on mutative scale chaos optimization radial basis function nerual network

为实现开关磁阻电动机(Switched Reluctance Motor,SRM)无位置传感器控制并减小转矩脉动,以推动SRM在工业缝纫机上的进一步应用,提出一种基于变尺度混沌优化的径向基神经网络(MSCO-RBFNN)方法,对开关磁阻电动机进行建模,以一台四相8/6极750W开关磁阻电动机为样机建立有限元模型(FEM),获得磁链、电流与转子位置关联样本数据,对MSCO-RBFNN模型进行训练和测试。Matlab/Simulink软件仿真和数字信号处理(DSP)实验结果表明,MSCO-RBFNN模型具有较好的收敛性能,所计算的转子位置与实际转子位置的误差较小,使电动机换向准确,减小了电动机的转矩脉动。

Small torque pulsation and position-sensorless control can promote application of Switched Reluctance Motor （SRM） for industrial sewing machine. The model of SRM based on Mutative Scale Chaos Optimization Radial Basis Function Nerual Network （MSCO- RBFNN） method is presented. The Finite Element Model （FEM） of a four-phase, 8/6 pole, 750W prototype of SRM is established for sample data of the flux linkage, current and rotor position, which is used to train and test the MSCORBFNN model. The simulation and experiment results demonstrate that the optimized neural network has better convergence performance, the rotor position deviation between the theoretic calculation and the experimental result is demonstrated small to enable accurate motor commutation and small torque pulsation.