基于SVM和证据理论的多数据融合故障诊断方法.pdfVIP

学兔兔 第31卷 第8期 仪 器 仪 表 学 报 V01．31 No．8 2010年8月 Chinese Journal of Scientific Instrument Aug．2010 基于SVM和证据理论的多数据融合故障诊断方法术 姜万录 ，吴胜强 (1 燕山大学机械丁程学院 秦皇岛 066004；2 邢俞职业技术学院 邢俞 054035) 摘 要：针对D—s证据理论很难确定基本概率分配(BPA)及支持向量机(SVM)的硬判决难获得概率输出的缺陷，融合D—s证 据理论和SVM算法提出了一种多数据融合故障诊断新方法：利用“一对一”多类SVM分配了BPA，引入基于矩阵分析的融合算 法解决了证据理论存在的计算瓶颈问题。对液压泵进行了试验，首先，采集了柱塞泵松靴、缸体与配流盘磨损等故障信号，应用 小波包对采集的信号进行了预处理，提取了l2个时频特征量；最后，用所提出的基于SVM和证据理论的多数据融合新方法进 行了诊断。试验结果表明，新方法故障确诊率高，诊断有效。 关键词：故障诊断；SVM；D—S证据理论；多数据融合；小波包 中图分类号：TP306 文献标识码：A 国家标准学科分类代码：460．45 Multi—data fusion fault diagnosis method based on SVM and evidence theory Jiang Wanlu ，Wu Shengqiang ， (7 College of Mechanical Engineering，Yamhan University，Qinhuangdao 066004，China； 2 Xingtai Vocational and Technical College，Xingtai 054035，China) Abstract：Aiming at the difficulty that evidence theory can hardly determine basic probability assignment(BPA)and SVM can hardly obtain probability output，a new multi—data fusion fault diagnosis method is proposed，which is based on SVM and D—S evidence theory．In the method，BPA is assigned based on one—versus—one multi—class SVM；a fusion arithmetic based on matrix analysis is presented to solve the calculation bottleneck problem of evidence theory．The method is tested on a hydraulic pump；at first，the fault signals of slipper looseness and wearing between cylinder body and valve plate are collected；the measured signals are preprocessed using wave]et cluster；12 fault features are picked up in time domain and~equency domain；at last，the fault is diagnosed using the proposed new method．Experi