基于运行大数据学习的复杂装备故障诊断技术及其典型应用  

Fault Diagnosis Technology for Complex Equipment Based on the Learning of Big Operation Data and Its Typical Application

作　　者：刘达新[1] 裘乐淼[1] 王志平[1] LIU Daxin;QIU Lemiao;WANG Zhiping(Zhejiang University, Hangzhou 310027,China)

机构地区：[1]浙江大学,浙江杭州310027

出　　处：《中兴通讯技术》2017年第4期56-59,共4页ZTE Technology Journal

基　　金：国家自然科学基金资助项目(51475417;L1524002)

摘　　要：认为通过从复杂装备运行特征大数据中挖掘出故障信息,实现运行故障的智能诊断,对保证复杂装备的安全和稳定运行具有重要意义。结合复杂装备运行大数据的特点与机器学习理论,提出了基于运行大数据学习的复杂装备故障预测诊断方法,实现了复杂装备运行特征大数据与运行故障的分层关联,基于大数据分析的复杂装备运行故障特征提取以及基于模糊反向传播(BP)神经网络的复杂装备运行故障诊断。此外,还将此技术应用到高速电梯的运行监测中,开发了高速电梯急停故障大数据分析与诊断系统,很好地验证了该方法的有效性。In this paper,we consider that it is significant to dig out the fault information from the big data of the complex equipment operation feature,and diagnose the operation fault intelligently for ensuring the safe and stable operations of complex equipment.By combining the characteristics of complex equipment’s big operation data and the machine learning theory,a fault prediction and diagnosis method for complex equipment based on the learning of big operation data is proposed.In this way,some items are realized,including1)the hierarchical correlation between the complex equipment’s big operation data and the operation trouble;2)the operation trouble feature extraction based on big data analysis;3)the operation fault diagnosis based on fuzzy back propagation(BP)neural network for complex equipment.Moreover,by applying the proposed method to the high speed elevator’s operation monitoring,the big data analysis and diagnosis system for emergency stop fault of high speed elevator is developed,and the effectiveness of the method is well verified.

关 键 词：大数据 机器学习 故障诊断 复杂装备 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP277[自动化与计算机技术—控制科学与工程]