基于神经网络的电站仿真器建模方法及应用

Modelling Method for Power Plant Simulator Based on Neural Network

本文将神经网络理论应用到电站系统仿真实践中，有效地解决了电站系统仿真研究常遇到的机理复杂的非线性环节的建模问题。并以制粉系统为例，提出了作用函数Φ（ｘ）、学习效率系数α的选取方法以及权值Ｗｉ，ｊ的初值设定及要求，给出了提高逼近精度的方法。

In this paper, the neural network theory is applied in practice of power plant systemsimulation. It effectively solves modelling problems of nonlinear ring joint. the principle of which is complex.Furthermore, this paper gives powder--robing system as an example.offers methods to select contribution function 900 as well as efficiency coefficient a,to the original value of weighting value W, andits demands. Meanwhile. it offers the method which can increase attaching accuracy.