一种遗传BP网络及其在故障诊断中的应用  

A Kind of GA-BP Network and its Application in Rotating Machinery Fault Diagnosis

作　　者：潘俊辉[1] 李盼池[1] 

机构地区：[1]东北石油大学

出　　处：《微计算机信息》2012年第7期36-37,共2页Control & Automation

基　　金：基金申请人:李盼池潘俊辉;项目名称:基于量子过程神经网络的聚驱开发指标预测方法;基金颁发部门:中国博士后基金委(20090460864);基金申请人:李盼池潘俊辉;项目名称:量子神经网络模型及算法研究;基金颁发部门:黑龙江省教育厅(11551015)

摘　　要：针对旋转机械故障自动诊断问题及传统BP网络和遗传算法在进行故障诊断时所存在的缺点,提出了一种改进的遗传BP网络方法,该方法首先利用改进的遗传算法进行粗精度的学习以达到选取初值的效果,然后采用改进的BP算法完成对给定精度的网络学习,建立的网络学习收敛速度快且易于实现。仿真实验证明了该方法对旋转机械故障诊断问题具有良好的适应性。According to rotating machinery fault diagnosis problems and there existed the faults when carrying out the falut diagnosis using traditional BP network and genetic algorithm, a kind of improved genetic algorithm and BP network method was proposed in this paper. First, used the genetic algorithm to carry out the rough accuracy learning to achieve the effect of Selecting initial values, then adopted the improved BP algorithm to complete the learning for the given accuracy, the learning convergence speed of constructed networks was rapid and easily implemented. The simulation experiment proved that this method had the good adaptability for the rotating machinery fault diagnosis problems.

关 键 词：遗传算法 BP神经网络 旋转机械故障诊断 粗精度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]