一种基于故障诊断的特征权重确定方法

摘要

基于主元分析法的故障诊断过程,提出了一种新的特征权重计算方法.首先利用现有数据离线建立主元模型,然后输入监测数据在线进行故障诊断,根据故障诊断结果,分析故障点处各特征的特征权重.经实验验证,该方法有效.%In this paper, a new feature weight algorithm is proposed, which is based on fault diagnosis using PCA. Firstly, PCA model is established based on existing data offline. Secondly, we can input monitoring data online for fault diagnosis. Lastly, according to the results of fault diagnosis, the weight of each feature on fault point can be get. Experimental results show that the algorithm is effective.