基于参数优化MPE与FCM的滚动轴承故障诊断

发布时间：2018-03-01 20:42

本文关键词： 滚动轴承故障诊断参数优化多尺度排列熵遗传算法微粒群算法模糊C均值聚类　出处：《轴承》2017年05期 　论文类型：期刊论文

【摘要】：为精确提取滚动轴承振动信号的故障特征,提出了一种基于参数优化多尺度排列熵与模糊C均值聚类的故障诊断方法。首先,针对多尺度排列熵算法的参数确定问题,综合考虑参数之间的交互影响,基于遗传算法与微粒群算法对参数进行优化;然后,利用参数优化多尺度排列熵对滚动轴承振动信号进行特征提取,并通过模糊C均值聚类确定标准聚类中心;最后,采用Euclid贴近度对故障样本进行分类。通过分类系数与平均模糊熵检验聚类效果,证明了多尺度排列熵参数优化的有效性;与单一尺度排列熵、样本熵结合模糊C均值聚类方法的对比分析表明,基于参数优化多尺度排列熵与模糊C均值聚类的故障诊断方法具有更高的故障识别率和更广阔的适用范围。
[Abstract]:In order to extract the fault feature of rolling bearing vibration signal accurately, a fault diagnosis method based on multi-scale permutation entropy and fuzzy C-means clustering is proposed. Firstly, the parameter determination problem of multi-scale permutation entropy algorithm is proposed. Considering the interaction between parameters, the parameters are optimized based on genetic algorithm and particle swarm optimization algorithm. Then, the vibration signals of rolling bearings are extracted by multi-scale permutation entropy. The standard clustering center is determined by fuzzy C-means clustering. Finally, fault samples are classified by Euclid closeness. The clustering effect is tested by classification coefficient and average fuzzy entropy, and the effectiveness of optimization of multi-scale permutation entropy parameters is proved. Compared with single scale permutation entropy, sample entropy and fuzzy C-means clustering method, it is shown that, The fault diagnosis method based on multi-scale permutation entropy and fuzzy C-means clustering has higher fault identification rate and wider application range.
【作者单位】：燕山大学河北省重型机械流体动力传输与控制实验室;燕山大学先进锻压成形技术与科学教育部重点实验室;燕山大学河北省工业计算机控制工程重点实验室;
【基金】：国家自然科学基金项目(51675460,51405426) 河北省自然科学基金项目(E2016203306)
【分类号】：TH133.33

【相似文献】