往复式压缩机填料密封的机械故障诊断方法

MACHINERY FAULT DIAGNOSIS OF PACKING SEAL IN RECIPROCATING COMPRESSOR

采用机械故障诊断学中的时域识别方法，通过对往复式压缩机填料密封气腔内气体压力波动的分析，对其建立自回归模型，根据所建模型的ＡＩＣ指标，从数值上识别填料密封的正常与异常状态；结合对压力波形的频谱分析，判断故障的类型，利用各组填料密封腔中压力的分配情况，判断出故障产生的位置．经实验证明，采用填料密封腔中气体压力信号诊断技术及３种手段（时域识别、频谱分析、压力分配识别）相结合的诊断方法。

An autoregressive model has been established by the time domain recognition method of machinery fault diagnosis to analyze the gas pressure fluctuation of packing seal chamber in reciprocating compressor. According to the AIC index of the model, the abnormal states of paking seal can be distinguished form the normal state. The fault position generated and its pattern can be judged form the pressure distribution in the seal chamber and the frequency spectrum analysis of the pressure wave form. Testing results indicate that the machinery fault of packing seal of reciprocating compressor can be exactly monitored by the diagnosis of pressure signals in the packing seal chamber and the three adopted techniques(time domain recognition, frequency spectrum analysis, and pressure distribution recognition).