基于改进BP算法的水电机组轴瓦温度预测

应用人工神经网络BP算法，对水电机组轴瓦温度与影响瓦温变化的主要因素之间的映射关系进行表达，即建立水电机组轴瓦温度预测模型，并对轴瓦温度及其变化趋势作出预测；同时对BP算法进行改进，引入误差分布函数，动量项因子和组合转移函数，在一定程度上克服原有算法的局部最小问题，获得全局最小解，而且加快了网络的收敛速度。

The application of improved BP neural network to temperature modeling and forecasting in shaft bushing of hydroelectric unit

The objective of this study is to model and forecast the shaft bushing temperature of hydroelectric unit by back propagation (BP) neural network. There is the established mapping between before mentioned temperature and primary factors that affect the temperature. It is difficult to describe this mapping by conventional analytic means. So neural network is effective. The BP neural network is improved by integrating an error distribution function (EDF), a momentum factor and combined transfer functions so as to overcome local minimum problems and to find the global minimum solution and greatly accelerate the convergence speed.