基于VW-PCA的盾构刀盘驱动液压系统故障诊断

摘要

The weight vector is introduced into the traditional principal component analysis (PCA)so as to improve the rapidity and accuracy of fault diagnosis on hydraulic driving system of shield cutterhead;and the formula of variableweighted-principal-component-analysis (VW-PCA) method is deduced.The model of cutterhead hydraulic driving system of earth pressure balance shield with diameter of 6 450 mm is established by software AMESim;and 12 variables and 6 faults are selected and simulated respectively.And then,the weight vectors of every fault can be obtained from the simulation data.As a result,the model of VW-PCA method can be established to study the fault samples.The study results show that the contributions of every variable to different faults can be improved by variable weighting;the accuracy of system fault diagnosis by VW-PCA method can be significantly improved by more than 50％.%为提高盾构刀盘液压系统故障诊断的快速性和准确性,引入权值向量来改进传统主元分析法(PCA),并推导变量加权主元分析法(VW-PCA)公式.通过AMESim仿真软件建立φ6 450 mm土压平衡盾构刀盘驱动液压系统模型,选取12个测试变量,模拟仿真6种故障状态;运用仿真故障数据得到每一类故障的加权向量,建立变量加权主元分析模型(VW-PCA),并使用该模型对测试故障样本进行研究.结果表明:变量加权(VW)能够增强各系统变量对不同故障类型的贡献程度,变量加权主元模型(VW-PCA)能够显著提高系统诊断的正确率,最大可提高50％以上.