摘要
作为一种新型的神经网络,创建了具有高度柔性特性的网络结构。给出了柔性神经网络(FNN)的基本原理,并将其应用于开关磁阻电机(SRM)的建模与仿真,展示了SRM新型建模方法的主要优点。FNN在实现系统功能的同时,需要较少的神经元和迭代循环,大大降低了网络的复杂性,加速了网络的学习与实时计算速度。
A highly flexible network structure is built as a new neural network. The basic principle is given for the flexible neural network (FNN), which is applied to the modeling and simulation of switch reluctance motor (SRM). Main advantages of the new modeling method for SRM are illustrated. Less nerve cells and iterative cycles are demanded for realization of system functions by FNN, so that the network is greatly simplified and that study of network and real-time calculation speed is increased.
出处
《机车电传动》
2005年第2期23-26,39,共5页
Electric Drive for Locomotives
基金
教育部重点项目(2004104051)
北京交通大学"十五"专项科技基金(2003SM013)资助
