局域波时频域多重分形在故障诊断中的应用

FAULT DIAGNOSIS USING MULTIFRACALS IN LOCAL WAVE TIME-FREQUENCY DOMAIN

提出了一种振动信号时域与时频域的多重分形方法,并将该方法实际应用于故障诊断之中。信号局域波分解的时频域幅值矩阵经过网格划分后,通过最小二乘法计算出广义维数Dq。应用振动信号时频域多重分形方法在往复式压缩机填料泄漏故障进行了实例分析,并分别得出了对于该类故障较为敏感维数。实测往复式压缩机的振动信号分析表明,时域信号的多重分形只对早期填料泄漏故障较为敏感,时频域多重分形方法随着故障程度的增加其广义维数增加,能够较好的区分故障的严重程度。实践证明该方法在实际应用中切实可行。

Vibration signal multifactral method is studied,and this method is used in practical fault diagnosis.Generalized dimension D_q is caldulated from vibration signals in the local wave time-frenquency domain by the least square method.The multifracal method is applied to the packing leakage fault diagnosis of a reciprocating compressor.And sensitive demension of this fault is gained.The results shows that the vibration signal multifracals in time domain are sensitive only in the early stage of the packing leakage of the reciprocating compressor,and the generalized demension of the vibration signal multifracals in time-frenquency domain increases with the leakage level.It can indicate the fault level clearly.The example proves that this integrated method is feasible.