基于改进BP网络8/6极SRM非线性建模的研究

针对开关磁阻电机(SRM)磁化曲线高度饱和、非线性的特点,提出一种基于改进的BP神经网络建立开关磁阻电机模型的方法。该方法构造了一个将连接权值变为参数可调函数的BP神经网络。通过分析开关磁阻电机磁链与转矩特性获得神经网络的训练样本,经过训练,实现开关磁阻电机非线性建模,并在Matlab/Simulink中建立开关磁阻电机控制系统(SRD)仿真模型。仿真与实验结果的对比,证明了此建模方法可行。与传统BP神经网络建模相比,该方法节约了计算时间,具有很强的泛化能力和较高精度,有效地提高了收敛速度。

Non -Linear Modeling and Dynamic Simulation of 8/6 Poles SRM Based on Improved BP Neural Network

In view of the high saturation magnetization curve,nonlinear characteristics of SRD,a new method of modelling SRM based on improved BP neural network,that turns the connection weight into an adjustable parameter function,is presented.The network training samples are obtained by analyzing the flux and torque characteristics of SRM.Then the nonlinear model of SRM is set up after training.The simulating and experimental results show that the modeling method is feasible.And compared with the modeling method based on the traditional BP neural network,it has the advantanges of shorting computation time,strong generalization ability and higher accuracy.Besides,the method also improves the convergence rate effectively.