地铁车辆牵引电机超温故障预测技术研究

Research on overtemperature fault prediction technology of metro vehicle traction motor

地铁车辆牵引系统健康评估与故障预测是目前城市轨道交通领域的研究重点之一。采用数据挖掘技术,对地铁车辆牵引系统健康度评估进行相关研究。主要研究方法为利用无监督学习方法,建立机器学习模型。首先提取原始特征,采用方差过滤和主成分分析法进行特征降维,选用高斯混合模型得到类概率值作为车辆亚健康状态的判定依据。进而采用多层感知机模型训练牵引电机超温故障预测,取得了较好的效果。

Health assessment and fault prediction of metro vehicle traction system is one of the research focuses in the field of rail transit.Using data mining technology,the health evaluation of metro vehicle traction system is studied.The main research method is to establish a machine learning model by using unsupervised learning method.Firstly,the original features are extracted,the feature dimension is reduced by variance filtering and principal component analysis,and the class probability value is obtained by Gaussian mixture model as the judgment of vehicle sub-health state.And then the multi-layer perceptron model is used to train the overtemperature fault prediction of traction motor,and good results are obtained.