DPCA与GA-SVM融合的智能台车液压系统故障诊断  

Hydraulic System Fault Diagnosis of Intelligent Trolley Based on Dynamic PCA and Genetic Algorithm Improved SVM

作　　者：陈昭明 徐泽宇[3] 赵迎[3] CHEN Zhao-ming;XU Ze-yu;ZHAO Ying(College of Mechanical Engineering,Chongqing University,Chongqing 400044,China;School of Artificial Intelligence,Chongqing School,University of Chinese Academy of Sciences,Chongqing 400714,China;Intelligent Manufacturing Technology Institute,Chongqing Institute of Green and Intelligent Technology,Chinese Academy of Sciences,Chongqing 400714,China)

机构地区：[1]重庆大学机械工程学院,重庆400044 [2]中国科学院大学重庆学院人工智能学院,重庆400714 [3]中国科学院重庆绿色智能技术研究院,智能制造技术研究所,重庆400714

出　　处：《控制工程》2020年第11期1980-1986,共7页Control Engineering of China

基　　金：自然科学基金青年基金(61605205);国家质检公益性行业科研专项(Y42Z130I10)。

摘　　要：针对智能台车液压系统故障原因复杂多样及故障诊断效率低等问题,提出动态主成分分析(DPCA)与遗传算法改进支持向量机(GA-SVM)相结合的液压系统故障诊断方法。首先,采用AMEsim软件建立液压系统仿真模型采集故障数据并进行预处理;然后采用DPCA对故障特征向量进行降维,解除特征间的相关性和缩短训练时间;再运用遗传算法对SVM进行参数优化,将抽取出来的故障特征参数样本输入优化后的SVM中进行训练,获得分类模型,从而实现故障诊断。测试结果表明该方法的效率高于常规PCA-SVM及BP神经网络,为台车设备的维修和保养提供了指导,具有良好的应用价值和前景。Due to the intelligent trolley has the problems of complex and diverse fault reasons and low efficiency of fault diagnosis,a novel diagnosis method of hydraulic system based on dynamic principal component analysis(DPCA)and genetic algorithm improved support vector machine(GA-SVM)is proposed.Firstly,a hydraulic system simulation model is established by Amesim to collect fault data and carry out pretreatment.Secondly,DPCA is used to reduce the dimension of feature vectors,relieve correlation between features and shorten training time.And then the parameters of SVM is optimized by genetic algorithm,and the samples of the extracted fault feature parameters are trained in the optimized SVM to obtain the classification model,so that the fault diagnosis can be realized.The results show that the method has higher diagnostic efficiency than that of PCA-SVM and BP neural networks,which provides guidance for the repair and maintenance of trolley,and has good prospects for future engineering application.

关 键 词：液压系统 故障诊断 动态主成分分析 遗传算法 支持向量机 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置]