一种聚类优化融合故障诊断方法及其应用

发布时间：2018-05-27 11:04

本文选题：聚类分析 + 权值矩阵　；参考：《中国机械工程》2016年15期

【摘要】：针对单一聚类诊断方法难以准确、全面识别不同故障状态的问题,提出了一种聚类优化融合故障诊断方法。分别利用社团聚类、K-均值聚类及粒子群聚类三种方法对故障进行识别,得出三种聚类方法对应的故障识别准确率,在此基础上构建初始权值矩阵,并通过遗传算法对初始判断矩阵与三种聚类方法进行优化,得到最优权值矩阵与优化的聚类模型,用于融合诊断。轴承故障诊断实例结果表明,该聚类融合诊断方法能够有效提高故障识别准确率。
[Abstract]:In order to solve the problem that single clustering diagnosis method is difficult to accurately identify different fault states, a clustering optimization fusion fault diagnosis method is proposed. In this paper, three methods of community clustering, K-means clustering and particle cluster clustering are used to identify the faults, and the accuracy of fault identification is obtained. Based on this, the initial weight matrix is constructed. The initial judgment matrix and three clustering methods are optimized by genetic algorithm, and the optimal weight matrix and the optimized clustering model are obtained for fusion diagnosis. The results of bearing fault diagnosis show that the clustering fusion diagnosis method can effectively improve the accuracy of fault identification.
【作者单位】：湖南科技大学机械设备健康维护湖南省重点实验室;苏州东菱振动试验仪器有限公司;
【基金】：国家自然科学基金资助项目(51575177) 湖南省教育厅优秀青年项目(14B057);湖南省教育厅资助重点项目(13A023)
【分类号】：TH17

【相似文献】