基于EMD的滚珠丝杠振动信号滤波技术研究

Study on Vibration Signal Filter Technic of Ball Screw Based on Empirical Mode Decomposition

针对磨削加工中滚珠丝杠振动信号低信噪比、非平稳特性,提出基于经验模式分解(EMD)的振动信号滤波方法。通过EMD将滚珠丝杠振动信号分解为多个平稳的固有模态函数(IMF),并选择包含振动信号特征的若干IMF分量重构,提高了振动信号的信噪比,实现了对振动信号的平稳性分析,得到了振动信号的本质特征,为磨削滚珠丝杠实时监测和故障诊断提供了有效的理论基础。并通过仿真及实验数据分析,验证了EMD滤波技术是表征振动信号的非平稳特征及提高信号信噪比的有效工具。

According to the vibration signal properties of low SNB and non-stationary in grinding processing of ball screw,the method of vibration signal filter was proposed based on empirical mode decomposition（EMD）.The vibration signals of ball screw were decomposed into lots of intrinsic mode functions（IMF）by EMD,and some of IMF were selected to reconstruct the vibration signals,then,the SNB was improved,and the stationary analysis of vibration signals was implemented,so the essential features of vibration signals were obtained,which offered more efficient theoretical foundation to monitor online and fault diagnosis of ball screw.By simulation and analysis on experimental data,it is validated that the EMD filter technic is a useful tool in reducing noise and depicting the non-stationary property of vibration signals.