聚类算法在设备故障诊断中的研究与应用

Rearch & Application of Cluster Algorithm in Fault Diagnosis

本文针对复杂设备故障诊断难度大的问题,以聚类分析技术为基础,从故障诊断的本质出发,将故障模式识别问题转化为实现数据聚类问题,挖掘设备运行数据间的深层关系,完成故障诊断。以某型号柴油机燃油系统状态为实例,利用k-means算法进行模式识别,分类结果与试验样本所表征的个系统运行状态完全一致,验证了数据挖掘技术在故障诊断应用的可行性。

For concluding the difficulty of fault diagnosis of complex equipment,Basis in order to cluster algorithm,starting from the nature of fault diagnosis,fault pattern recognition problem can be transformed into data clustering problem,excavation the deep relationship between the operating data.Taking a model diesel engine fuel system status as an example,using k-means algorithm to pattern recognition,classification results and experimental characterization of samples of a system is running exactly the same state,verification of data mining technology feasibility of the application in fault diagnosis.