基于 PCA 的往复压缩机气阀故障异常监测方法

摘要

In order to satisfy the pressing needs of valve fault automatically detection, and according to the characteristic of valves’ temperature data,it’ s found that the suction ( or exhaust) valves’ temperature is consistent when the valves work,other-wise it’ s not.From this fact,the authors use the PCA ( Principal Component Analysis) to extract features to reflect the perform-ance based on valves’ temperature data.Simultaneously,with the establishment of radial basis model,it can achieve valve fault a-nomaly detection and automatic location of abnormal valve,and lay the foundation of early and quick warning of valve fault.%针对气阀故障异常自动检测迫切需求，针对气阀故障在温度数据的表现特点，即同类气阀正常工作时温度波动一致，故障时温度波动存在差异，采用主成分分析（ PCA）从气阀阀盖温度数据中提取故障特征参数，建立基于径向基函数的故障异常监测模型，实现了故障异常自动检测，并可进一步对故障气阀进行自动定位，为故障早期快速报警奠定了基础。