采用模糊聚类分析和BP网络相结合的高压加热器故障诊断

Fault Diagnosis of High Pressure Heaters by Using Fuzzy Clustering Analysis in Combination with BP Neural Networks

针对高压加热器系统各点参数耦合性强的特点,提出采用两次故障模式识别方法进行高压加热器系统的故障诊断,从而避开了建立整个系统故障特征知识库的问题。对模糊聚类分析理论及其在高压加热器第一层故障模式识别中的应用进行了分析;在分析高压加热器常见故障的基础上,介绍了利用BP网络进行第二层故障模式识别的方法。

Noticing the intimate correlation among the parameters, recorded at different points of high pressure heater systems, it is suggested to use double-deck fault mode identification for diagnosing faults of high pressure heater systems, wherewith the establishment of a fault repository for the whole high pressure heater system can be avoided. First of all, the fuzzy clustering analysis theory and the possibility of applying it in the first deck fault mode identification of high pressure heaters is. being analyzed; after which and based on analysis of generic high pressure heater faults, a method of second deck fault mode identification with BP neural networks is presented.