基于粗糙集神经网络的综合传动故障诊断  

Rough-set Neural Network Based Fault Diagnosis on Power-Shift-Steering-Transmission

作　　者：郑长松[1] 马彪[1] 李和言[1] 路华鹏[1] 

机构地区：[1]北京理工大学车辆传动国家重点实验室,北京100081

出　　处：《中国机械工程》2006年第S2期78-81,共4页China Mechanical Engineering

基　　金：教育部新世纪优秀人才支持计划项目(NCET-04-0200)

摘　　要：针对综合传动装置在道路试验过程中大量采集的油样,建立了油液分析数据库。采用粗糙集理论的信息约简,根据粗糙集理论能处理不精确、不完整、不一致的数据的特点,结合神经网络强大的线性逼近和模式识别的功能,建立了粗糙集神经网络模型。将该模型应用于综合传动装置的模式识别,取得了满意的诊断结果。研究结果为在不完整征兆信息下的机械故障诊断提供了新的方法和思路。An oil analysis database of the Power-Shift-Steering-Transmission(PSST)on road test was set up.By the Rough-set's information reduction and dealing with the imprecise,half-baked, inconsistent data and the neural network's linearity approximate and pattern recognition,we set up the Rough-Set Neural Network(RSNN)model.Appling this model on the pattern recognition of the Power-Shift-Steering-Transmission and the satisfactory diagnosis results were obtained.This method offers new method and idea on the machinery fault diagnosis for the half-baked information.

关 键 词：综合传动 粗糙集神经网络 信息融合 故障诊断 

分 类 号：TH132[机械工程—机械制造及自动化]