基于MED-SK算法的行星变速箱故障特征提取

Fault Feature Extraction Method of Planetary Gearbox Based on MED-SK Algorithm

针对某型行星变速箱齿轮齿根裂纹类轻微故障的动态响应特征较弱,难以有效反映变速箱故障状态的问题,提出了基于最小熵解卷积的谱峭度算法,它能够抑制信号中的噪声并增强信号冲击成分,进而提取齿轮裂纹故障冲击,实现故障特征增强。采用最小熵解卷积对行星变速箱齿轮裂纹故障台架试验的实测信号进行预处理以增强信号中的冲击成分,通过计算信号峭度谱获取最优滤波参数设计带通滤波器对信号进行滤波,最后解调信号中的低频成分得到其包络谱,可清晰识别行星变速箱振动信号的齿轮裂纹故障特征。行星变速箱试验数据处理结果表明,该方法可以有效地提取行星变速箱微弱脉冲冲击类故障特征。

For a certain type of planetary transmission of crack minor failure root weak dynamic response characteristics,it is difficult to effectively reflects the problems of the gearbox fault status.The research proposed a means of Spectral Kurtosis method based on Minimum Entropy Deconvolution(MED-SK)algorithm that restrain impact noise and enhance the signal components,as well as extract the gear crack fault,realize the recognition of the planetary gearbox gear tooth crack fault vibration signals.The minimum entropy deconvolution was used to preprocess the measured signals from the bench test of planetary gearbox gear cracks to enhance the impact components in the signals.The optimal filter parameters were obtained by calculating the signal kurtosis spectrum to design a band-pass filter to filter the signal.The low frequency components in the demodulated signal were demodulated to obtain the envelope spectrum,which could clearly identify the gear crack fault characteristics of the vibration signal of planetary gearbox.The results show that this method can extract the weak impulse fault characteristics of planetary gearbox more effectively.