基于声音信号的托辊故障诊断方法

Fault Diagnosis Method of Rollers based on Sound Signals

为实现对远程带式输送机托辊故障的检测,研发基于声音信号的托辊故障诊断方法。结合远程带式输送机托辊轴承的特性,针对性提出多种分析方法相结合的方案。诊断系统中包括时域检测、快速傅里叶变换(FFT)峰值检测、功率谱检测、小波包分解与重构和希尔伯特(Hilbert)包络分析结合、经验模态值分解(EMD)检测方法,可实现采集数据显示、波形分析和故障诊断。其中,时域检测、快速傅里叶变换峰值检测、功率谱检测可初步判断托辊是否发生故障,经验模态值分解可以预估出带式输送机托辊发生故障的区段,小波包分解与重构和希尔伯特包络分析联合法可进一步提取托辊轴承故障频率,确定轴承发生故障的部位。最后,通过实验验证了基于声音信号的托辊故障诊断方法的有效性。

In order to realize the detection of the remote belt conveyor rollers,a set of fault diagnosis method was developed based on sound signals.A fusion scheme based on a variety of analytic methods was proposed according to the roller bearing characteristics of the long-distance belt conveyor.The scheme incorporated several detection methods such as time domain detection,FFT peak detection,power spectral density detection,wavelet packet decomposition,reconstruction and Hilbert envelope analysis,and empirical mode decomposition (EMD).It can realized acquired data display,waveform analysis and fault diagnosis.Among these detection methods,the roller failure can be preliminary judged by the time domain detection,FFT peak detection and power spectrum detection.The failure section of the belt conveyor rollers can be forecasted by EMD.The method combining wavelet packet decomposition and reconstruction with Hilbert envelope analysis can further extract the failure frequency of the roller bearing.With this method the location of the failed bearing can be determined.Finally,the experiment verified the effectiveness of this fault diagnosis method for rollers based on sound signal.