中文摘要：

为满足牵引电动机状态监测中多维海量数据处理的需求,给出了一种基于改进主元分析的状态监测方法。该方法以均值化代替标准化对传统主元分析进行改进,在保留原有数据信息特征的基础上降低指标维数,消除变量关联,建立主元模型,利用SPE统计量和T2统计量判断电机运行是否发生异常。实验结果表明：基于改进主元分析的状态监测方法能够建立准确的状态监测统计模型并快速检测出电机异常情况,该方法在电机状态监测中是有效可行的。

英文摘要：

In order to meet the demands of multidimensional and mass data in traction motor state monitoring, a new kind of Principal Component Analysis （PCA） approach is proposed. This method, whose data preprocessing is im- proved, is an effective way which can not only reduce the dimension of motor index and eliminate correlation between process variables, but also reserve enough information of original data characteristics needed for modeling. Based on PCA model, a state monitoring experiment is carried out on a traction motor with SPE and T2 statistics. The experiment results validate that the approach can build an accurate monitoring model and detect abnormal state of motor effectively.