基于改进欧氏距离的水轮发电机组振动故障诊断方法研究

Study on vibration fault diagnosis method of water turbine generator set based on improved Euclidean distance

本文针对水轮发电机组振动故障原因复杂、故障特征参数多的特点,提出了改进欧氏距离分析法。首先利用主成分分析对故障样本特征参数进行降维分析,选取若干主成分(包含>85%原始样本信息),计算出主成分分析下故障样本新的特征参数值,然后进行欧氏距离层次聚类分析。计算结果表明,该方法可以用于解决具有小同物理量振动故障特征参数的故障诊断。

In this paper,the improved Euclidean distance analysis method is proposed in view of the complicated causes of the vibration fault of the hydroelectric generator set and the many characteristic parameters of the fault.First,we use principal component analysis to analyze the feature parameters of fault samples,select some principal components（including〉 85% original sample information）,calculate the new characteristic parameter values of fault samples under principal component analysis,and then carry out hierarchical clustering analysis of Euclidean distance.The calculation results show that this method can be used to solve the fault diagnosis of the same physical quantity with small vibration fault feature parameters.