Adaptive Peak Decomposition Approach for the Fault Diagnosis of Reciprocating Compressor based on General Frequency

Vibration signals of reciprocating compressor mainly include multiple impact and noise signals, traditional fault diagnosis approaches could not reveal the fault feature accurately. In this study, a novel notion of general frequency is proposed and some simulation signals are used to compare its performances with the traditional frequency and the instantaneous frequency. Based on the notion, an adaptive peak decomposition approach is investigated and applied to the fault diagnosis of vibration signal for reciprocating compressor gas valve in four states including normal valve state, gap valve state, fractured valve state and bad spring valve state. The results show the effectiveness and the superiority of the proposed approach in extracting the fault feature of reciprocating compressor.