论文部分内容阅读
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 bearing fault diagnosis, in order to reduce the interference of noise on the wavelet transform, an integrated processing method is proposed which uses empirical mode decomposition and then wavelet transform to analyze the signal. Based on the adaptive decomposition of the empirical mode decomposition method, the automorphic function is reconstructed by using the natural modal function with high kurtosis value and high contribution rate. The autocorrelation function is calculated, and then the wavelet transform is performed to obtain the decomposed signal The concatenated spectrum of detail signals performs Hilbert demodulation on the best performing components. The method solves the problem that the noise has no obvious effect on the interference of the weak fault signal and improves the fault recognition rate and the efficiency of the wavelet transform. The experimental results of pitting failure of bearing rolling elements show that this method can effectively extract the fault characteristics of rolling elements of rolling bearings and has better effect than traditional envelope demodulation.