改进BP算法在故障诊断中的应用

针对使用BP算法的多层感知器中存在的网络学习收敛速度慢，容易陷入局部极小等问题，采用一个改进的算法，对设备故障进行了诊断和分析研究，以某柴油机供油系统为例，对其中的几种常见故障进行了诊断研究，从中可以看出，改进算法加快了收敛速度，诊断的精度也有所提高。

Application of improved neural network to fault diagnosis

A new improved algorithm (IBP) was adopted aiming at the problems o f slow convergence rate and easily falling into part minimums in network study ing of perceptions using BP algorithm. Taking the supply oil system of diesel en gine as example, the common fautls are diagnosed adopting IBP algorithms. It ca n be obtained that this method improves the precise of diagnosis and prediction, quickens the tempo of convergence.