基于Bagging算法的盾构机故障诊断方法

Faults Diagnoses Method of Shield Machine Based on Bagging Algorithm

针对盾构机故障诊断过程中,各类故障数据集存在类别不平衡的问题,采用基于SMOTE算法和Bagging算法的故障诊断分类器对盾构机数据进行故障诊断。仿真结果表明:采用SMOTE人工合成算法、Bagging算法和投票机制组成的分类器G-mean值高于Bagging_SMOTE_CART,Bagging_SMOTE_BPNN和Bagging_SMOTE_KNN算法。

Shield machine is a kind of huge underground excavation equipment,which works in complex and variable environment commonly.It’s not rare that some faults occur in shield machine.When some faults occur,it may slow down the construction,and even lead to collapse of surface.So,it’s of significance to discover,classify and handle timely when faults occur.on the fault diagnosis of shield machine,the diagnoses data get from shield machine is always imbalanced,which makes the traditional classification algorithms difficult to achieve desired results.To solve this problem,we use an algorithm combining SMOTE,Bagging to improve the performance of the classifier.The results of simulation show that the G-mean value of the classifier composed of SMOTE algorithm,Bagging algorithm and voting mechanism is higher than that of Bagging_SMOTE_CART,Bagging_SMOTE_BPNN and Bagging_SMOTE_KNN algorithm.