利用模式识别对柴油机进行故障诊断的一种方法

Pattern recognition method used to identify faults

通过对柴油机的几种常见油管压力故障信号的分析 ,提出了一种提取模板向量和信息向量的方法 ,并且建立了五种模板库 ;利用欧式距离作为信息向量和模板向量的相似性测度 ,设计了一种基于最近邻准则的试探聚类算法来进行故障诊断 ;最后提出了一种可以通过人工干预、完善故障模板类以避免错误诊断的方法。实验结果证明这种方法可以有效的利用油管压力信号对柴油机进行故障诊断 。

This paper presents a method on extracting th e template vector and inf ormation vector and establishing five template libraries, which are based on the analysis of the familiar fault signal of the gas engine's vitta pressure. A ten tative clustering algorithm is devised to diagnose the fault based upon the neig hboring rule. At last, A method is brought forward, which can avoid the wrong d iagnosis by artificially intervening to perfect the fault template libraries. Ex periments proved that the method can effectively diagnose the faults of the gas engine using the vitta pressure and may gain the better result as soon as the te mplate libraries are perfected to a certain extent.