基于神经网络的水轮发电机组的建模分析

水轮发电机组是一个具有非线性的复杂受控对象。本文分别采用BP和RBF神经网络模型对水轮发电机组进行动态建模，经MATIAB仿真实验，结果表明用神经网络可方便的建立非线性系统的模型。通过分析比较两种网络动态建模方法，可知采用RBF网络进行建模相对采用BP网络具有明显的优点，RBF所用的学习时间和所用到的神经元个数大大减少，在某种程度上克服了BP网络的训练时间长、训练不完全和容易到达局部极小的缺陷。

MODELING AND ANALYSIS FOR HYDROELECTRIC GENERATING UNIT BASED ON NEURAL NETWORKS

Hydroelectric generating unit is a complex nonlinear controlled object.BP neural networks and RBF neural networks are separately used to simulate the dynamic model of hydroelectric generating unit in this paper.The simulating results by MATLAB show that neural networks can conveniently obtain the dynamic model of nonlinear system.After comparing the dynamic modeling performance of two neural networks,less nerve cells and less learning time are needed in RBF networks.