利用人工神经网络SNC进行最优励磁控制

BP神经网络是一种多层结构的映射网络。由于它计算简单、存储量小，并具有分布并行处理特性，所以是目前应用最广的一种模型。本文设计了一种BP神经网络的监督学习控制器（SNC），在线性最优励磁控制的基础上，利用3层BP神经网络对柴油发电机的控制过程进行监督学习。通过对网络的训练，使其能达到实时控制的目的。仿真结果表明，所设计的SNC在系统运行方式较大的变化范围内，都能提供很好的控制性能。

Optimal Exciting Control Based on BP Neural Network

BP neural network is a type of many layers structure reflection network. It's calculation is simple, memory ca-pacity is small and it has specific property that distribution can be handled, so it is a model which applies the widest at present.A type of supervisory and learning controller based on BP neural network is designed. On the basis of linear optimal exciting con-trol, a three-layered BP neural network is used for supervising and learning the objective of real-time controlling, AS shownby simulation results, the controller so designed is able to provide very good controlling performance over a range of fairly greatchange in system operation mode; and it has fairly good robustness and adaptablity.