电梯故障诊断的RBF神经网络方法研究  

Research on elevator fault diagnosis using RBF neural network

作　　者：邱春玲[1] 张广明[1] 钱夏夷[2] 

机构地区：[1]南京工业大学自动化学院,南京210009 [2]江苏省特种设备安全监督检验研究院,南京210009

出　　处：《机械设计与制造》2009年第9期175-177,共3页Machinery Design & Manufacture

基　　金：江苏省社会发展计划项目(BS2007083);江苏省工业装备数字制造及控制技术重点实验室资助(BM2007201)

摘　　要：以电梯曳引机为研究对象,在阐述其振动故障的基础上,针对涡轮蜗杆减速器故障,给出了故障特征现象和故障代码。对已发生的蜗杆不平衡故障,分别采用RBF神经网络和BP神经网络,利用Matlab工具建模,给出了故障诊断结果,并进行结果分析比较。多次数据离线检验和实时在线监测结果均表明采用RBF神经网络方法能够及时、有效地检测出曳引机工作过程中的故障,并能够满足曳引机运行的实时性和鲁棒性等要求。Researching for elevator tractive motor,based on vibration fault description,in accordance with turbine worm retarder fault,the fault characters and fault code are given. Aimed at worm unbalance fault been happened,RBF and BP neural network are adopted,using the tool of Matlab in modeling and giving the fault diagnosis results ,finally,conclution anlysis and compare are given. The RBF neural network algorithm realized was verified with a great deal of historical test-data and also validated in the practical ground tests of the motor. The results show that the algorithm not only can detect the fault of the motor in time and efficiently,but also can meet the real-time and robustness requirement.

关 键 词：电梯 曳引机 故障诊断 RBF网络 BP网路 

分 类 号：TH16[机械工程—机械制造及自动化] V434[航空宇航科学与技术—航空宇航推进理论与工程]