Multi-parameter statistical monitoring and diagnosis method in cylinder score fault diagnosis of a reciprocating compressor

DANG Lu1， JIANG Zhi-nong1， FENG Kun1， ZHANG Zao-ping2， ZHANG Jin-jie1
（1. Diagnosis & Self-Recovery Engineering Research Center，Beijing University of Chemical Technology，
Beijing 100029， China；
2. Oil Refinery， China National Petroleum Corporation Jilin Petrochemical Company， Jilin 132022， China）

Abstract： In order to raduce the accident of reciprocating compressor cylinder score fault and other malignant accidents，the principal component analysis（PCA） method was applied to fault diagnosis of reciprocating compressors. After analysis of cylinder absolute vibration acceleration of one petrochemical enterprise，multiple parameters were extracted. Due to that different characteristic parameter's mechanical fault sensitivity was different，every two parameters were put in a group，and then the combination pictures were analyzed and compared with the picture that comes from the PCA method. The results show that the principal component analysis method will gain a stable work effect to reciprocating compressor，and always reflect equipment's working condition effectively. Thus parameter selection problem is solved and the cylinder score fault early warning is realized.
Key words： reciprocating compressor； cylinder score； muliparameter statistical monitoring； fault diagnosis； principal component analysis（PCA）