摘要
在轴承故障诊断中,为降低噪声对小波变换的干扰,提出了先用经验模态分解、再用小波变换对信号进行分析的综合处理法。在用经验模态分解方法的自适应性对信号进行分解的基础上,选用峭度值优选贡献率高的固有模态函数重构信号,计算其自相关函数,然后进行小波变换,得到分解后细节信号的级联谱,对效果最好的分量进行Hilbert解调。该方法解决了噪声对弱故障信号干扰导致诊断效果不明显的问题,提高了小波变换的故障识别率和效率。轴承滚动体点蚀故障试验结果表明:该方法能有效提取轴承滚动体故障特征,与传统包络解调相比具有更好的效果。
In order to reduce the interference of noise on the wavelet transform in bearing fault diagnosis,the comprehensive method that firstly using Empirical Mode Decomposition(EMD) and then analyzing the signal with the wavelet transform is put forward.Based on signal decomposition by EMD self-adaption,the intrinsic mode functions of kurtosis value with preferably high contribution is selected to reconstruct signal from autocorrelation noise reduction and then conduct the wavelet transform to obtain the cascade spectrum of decomposed detailed signals as well as the best-performing component Hilbert demodulation.It solves the problem that the interference of the noise on the weak fault signal results in the obscure diagnosis effect,and improves the fault identification rate and efficiency of wavelet transform.The bearing rollers pitting failure test shows: this method can effectively extract the rolling bearing fault features,and has better results compared with the traditional envelope demodulation.
出处
《装甲兵工程学院学报》
2013年第3期39-43,共5页
Journal of Academy of Armored Force Engineering
基金
军队科研计划项目
关键词
故障诊断
EMD
自相关
小波变换
fault diagnosis
EMD
autocorrelation
wavelet transform
