基于层次分段多尺度散布熵的矿井提升机主轴承故障诊断  

Fault diagnosis of mine hoist’s main bearing based on hierarchical piecewise multi-scale dispersion entropy

作　　者：董荣伟[1] 杨宁[2] DONG Rongwei;YANG Ning(School of Intelligent Manufacturing,Yancheng Polytechnic College,Yancheng 224005;School of Electrical and Information Engineering,Jiangsu University,Zhenjiang 212013)

机构地区：[1]盐城工业职业技术学院智能制造学院,江苏盐城224005 [2]江苏大学电气信息工程学院,江苏镇江212013

出　　处：《机械设计》2025年第1期94-100,共7页Journal of Machine Design

基　　金：国家自然科学基金(面上)项目(32171895);国家自然科学基金(青年)项目(31701324);江苏省产学研合作项目(BY2022434)。

摘　　要：针对层次多尺度散布熵(HMDE)粗粒化过程中存在的信息泄露及熵值计算不稳定的问题,文中提出了层次分段多尺度散布熵(HPMDE)的概念。结合极限学习机(ELM),提出了矿井提升机故障诊断的HPMDE-ELM方法。HPMDE采用分段粗粒化方式,解决了HMDE粗粒化过程中存在的不足。根据仿真信号对HPMDE的参数选择进行了分析,并与HMDE的结果进行了对比分析,结果表明:HPMDE的计算结果更稳定。通过矿井提升机驱动系统主轴承的故障诊断实例对HPMDE进行了验证和对比分析,结果表明:HPMDE对不同故障的可区分性更强,诊断精度更高。Since the hierarchical multi-scale dispersion entropy(HMDE)suffers information leakage and instability of entropy calculation in the coarse-grained process,in this article the hierarchical piecewise multi-scale dispersion entropy(HPMDE)is proposed.Combined with the extreme learning machine(ELM),a method is developed for fault diagnosis of the mine hoist based on HPMDE-ELM.HPMDE uses the piecewise coarse-grained method to overcome the shortcomings of HMDE in the coarsegrained process.The parameter selection of HPMDE is analyzed by the simulation signal;the HPMDE and HMDE results are compared.The results show that the HPMDE results more stable.The HPMDE results are verified and compared with those obtained from the examples of main-bearing fault diagnosis for the mine hoist’s drive system.It is shown that HPMDE is more effective to distinguish various faults and has a higher standard of diagnosis accuracy.

关 键 词：矿井提升机 故障诊断 层次多尺度散布熵 分段 

分 类 号：TH132[机械工程—机械制造及自动化]