基于MODWPT的包络阶次谱在滚动轴承故障诊断中的应用

Application of the envelope order spectrum based on the maximal overlap discrete wavelet packet transform to the roller bearing fault diagnosis

为了有效提取滚动轴承的故障特征,提出了基于MODWPT的包络阶次谱故障诊断方法.采用MODWPT将多分量的滚动轴承振动信号分解为若干个分量,对各个分量信号进行包络分析并对包络信号进行角域重采样;最后对重采样后的信号进行频谱分析,得到包络阶次谱,从而判断滚动轴承的工作状态和故障类型.采用该方法分别对仿真信号和实验信号进行了分析,结果表明了该方法的有效性.

The envelope order spectrum based on the maximal overlap discrete wavelet packet transform（MODWPT） is proposed and applied to the roller bearing fault diagnosis in this paper.By using MODWPT,a multi-component signal can be decomposed into a number of single-component signals whose instantaneous frequencies own physical sense.Therefore,MODWPT is particularly suitable for the processing of multi-component amplitude-modulated and frequency-modulated（AM-FM） signals.When a fault occurs in a roller bearing,the vibration signal of the start-stop process always display the multi-componennt amplitud and frequency modulated characteristics.In the method of envelope order spectrum based on the maximal overlap discrete wavelet packet transform,firstly the multi-component signal is decomposed into some single-component signals by using the maximal overlap discrete wavelet packet transform;secondly the envelope analysis is applied to each single-component signal;thirdly each envelope signal is resampled in angle-domain;finally the spectrum analysis is used to each resampled signal and the envelope order spectrum can be obtained.Furthermore,the roller bearing working condition can be identified according to the envelope order spectrum.The analysis results from the simulation and experimental signals show the effectiveness of the proposed method.