基于粒子群优化的改进EMD算法在轴承故障特征提取中的应用

发布时间：2018-01-12 02:33

本文关键词：基于粒子群优化的改进EMD算法在轴承故障特征提取中的应用　出处：《振动与冲击》2017年16期 　论文类型：期刊论文

【摘要】：经验模态分解(Empirical Mode Decomposition,EMD)作为一种数据驱动的自适应信号分解方法,在轴承故障特征提取中有着广泛应用。针对EMD自身存在的模态混叠、端点效应以及三次样条插值带来的过冲/欠冲问题,同时考虑到有理Hermite插值方法具有一个形状控制参数,为选择最优的插值曲线提供了可能,基于此,提出了一种基于粒子群优化(Particle Swarm Optimization,PSO)的改进EMD算法,选定频率带宽作为IMF优劣评判准则,并以此作为PSO的评价函数;在筛分过程中,从众多不同形状控制参数对应的分解结果中寻找最优IMF从而确定最优形状控制参数;在每阶分解结果中都能保证所得IMF是最优的,从而达到更好的自适应性及更高精度。为验证所提出方法的有效性,采用传统EMD、EEMD与该算法对仿真信号进行处理、对比,并通过计算相关技术指标进行了验证。最优将其应用于滚动轴承故障特征提取,并与传统EMD算法、EEMD进行对比,包络谱结果显示,改进后的EMD算法具有更好的分解效果,抑制干扰并能提取出更多故障信息。
[Abstract]:The empirical mode decomposition (Empirical Mode Decomposition, EMD) adaptive signal decomposition method as a data driven, there is extensive application in bearing fault feature extraction. For the existence of EMD modal aliasing, the end effect and three spline interpolation to bring the overshoot or undershoot problem, taking into account the rational Hermite interpolation the method has a shape parameter control, to select the optimal interpolation curve may be provided, based on this, puts forward a method based on particle swarm optimization (Particle Swarm Optimization, PSO) of the improved EMD algorithm, the selected frequency bandwidth of IMF as evaluation criteria, and as a PSO evaluation function; in the process of screening. To find the optimal IMF to determine the optimal shape control parameters from the decomposition results of many different shapes corresponding control parameters; in each step of decomposition results can ensure the IMF is optimal, so as To better adaptability and higher accuracy. In order to validate the proposed method, compared with traditional EMD, and the EEMD algorithm, the simulation signal, and verified by calculating the related technical index. The optimal applied in rolling bearing fault feature extraction, and compared with the traditional EMD algorithm, EEMD comparison of envelope spectrum results show that the improved EMD algorithm has better decomposition, inhibition of interference and can extract more fault information.

【作者单位】：北京航空航天大学宇航学院;中国民航大学电子信息与自动化学院;
【基金】：国家自然科学基金(10972019)
【分类号】：TH133.3
【正文快照】： 轴承作为旋转机械中运用最为广泛且关键的部件,众多故障皆来源于此[1],同时,它的运行状态也直接影响了整台设备的产能以及精度。在实际运行环境下,轴承通常有以下几种故障形式:外圈故障、内圈故障、球故障以及几种复合情形。而这些故障原因多由于滑油污染、过载[2]、脉冲宽度

【相似文献】