基于MCKD与OFSC的变转速下滚动轴承故障特征提取

MCKD and OFSC-Based Fault Feature Extraction of Rolling Bearing at Variable Speed

针对阶频谱相关(OFSC)法解决变转速下旋转机械故障诊断时易受背景噪声干扰的问题,通过最大相关峭度解卷积(MCKD)降低随机噪声的干扰,增强原始信号的故障脉冲;分析降噪后的信号,在三维空间中计算其阶频谱相关;依据降噪后信号的时频分布划定高能量频率区间,对三维OFSC进行集成运算,得到二维表达的OFSC,呈现故障信号的各类特征阶次。内圈点蚀故障模拟试验表明,二维OFSC能够清晰识别变转速下滚动轴承故障特征阶次,验证所提方法的有效性。

To eliminate background noise disturbance encountered by order-frequency spectral correlation(OFSC)method when used to diagnose the faults of a rotating mechanism,the paper combines maximum correlated kurtosis deconvolution(MCKD)with OFSC method to extract the fault features of a rolling bearing.It uses MCKD method to remove the random signal from original signal,thus obtaining denoised one with fault impulse intensified.After integration in the high energy frequency range delimited by time-frequency distribution of denoised signal,the three-dimensional OFSC can be expressed as two-dimensional form presenting various feature orders of the fault signal collected at variable speed.The simulated pitting failure test of inner ring validates the effectiveness of this new method.