基于振幅熵和功率谱重心的转子振动故障诊断

发布时间：2018-04-21 23:39

本文选题：转子 + 聚类　；参考：《中国工程机械学报》2017年02期

【摘要】：对信号进行特征提取是故障诊断的关键,为了提高转子振动故障诊断的准确性,根据转子振动的特点提出了基于振幅熵H(A)与功率谱重心C的转子振动故障诊断方法.通过计算功率谱的重心得到表征功率谱变化的功率谱重心特征,计算振幅的熵值得到反映幅值分布特征与振动集中程度的振幅熵特征,组成二维特征量(H(A),C).然后通过转子故障模拟实验采集数据,对其进行DBSCAN聚类、K均值聚类、层次聚类、网格聚类4种聚类分析.结果表明,基于振幅熵H(A)与功率谱重心C的二维特征量(H(A),C)能够作为评价转子振动状态的综合特征指标.通过对传统的二维特征量(偏度、均方根值)、(裕度、标准差)运用网格聚类法进行转子振动故障诊断识别,结果表明,(H(A),C)的选取较于传统特征量的选取能更好地对转子运行中出现的常见故障进行区分.
[Abstract]:Feature extraction is the key of fault diagnosis. In order to improve the accuracy of rotor vibration fault diagnosis, a rotor vibration fault diagnosis method based on amplitude entropy (HPA) and power spectrum center of gravity (C) is proposed. By calculating the center of gravity of the power spectrum, the barycenter characteristic of the power spectrum is obtained. The entropy of the calculated amplitude is worth the amplitude entropy characteristic reflecting the distribution of amplitude and the degree of vibration concentration. Then, through the rotor fault simulation experiment to collect data, DBSCAN clustering K-means clustering, hierarchical clustering, grid clustering four kinds of clustering analysis. The results show that the 2-D eigenvalue of the power spectrum center of gravity C based on the amplitude entropy (H) and the power spectral center C (C) can be used as a comprehensive characteristic index to evaluate the vibration state of the rotor. By using the grid clustering method to identify the rotor vibration fault diagnosis, the traditional two-dimensional characteristic variables (bias, root mean square value) (margin, standard deviation) are used to diagnose and identify the rotor vibration fault. The results show that compared with the traditional feature selection, the selection of the HGV / C) can better distinguish the common faults in the rotor operation.
【作者单位】：南昌航空大学航空制造工程学院;
【基金】：国家自然科学基金资助项目(51365040) 航空科学基金资助项目(2013ZD56009) 江西省自然科学基金资助项目(20151BAB206060) 江西省研究生创新专项资金资助项目(YC2015-S314)
【分类号】：TH17

【相似文献】