针对往复压缩机信号具有典型的非高斯脉冲特性,提出了一种基于α稳定分布模型的往复压缩机故障诊断方法。通过该方法估计了往复压缩机在气阀正常、阀片有缺口、阀片断裂和弹簧损坏4种状态下的信号分布参数,提取了估计参数下各状态的分数低阶统计量特征。结果表明:α稳定分布可以有效表征往复压缩机气阀信号状态,为其故障诊断提供了一种有效方法。
Considering the non-gaussian pulse characteristics of reciprocating compressor vibration signals, a fault diagnosis approach based on α-stable distribution model was presented. Making use of this approach, the signal distribution parameters of reciprocating compressor with normal air valves, with gapped valve sheets, with cracked valve sheets, and with fractured valve sheets and damaged spring were estimated respectively, and their characteristics of fractional low-order statistics for each state were extracted. The results demonstrate that the α-stable distribution can reflect air valve' s signal state and can benefit the fault diagnosis.