基于GMM训练与HMM变换的波纹管振动信号分析  

Vibration Signal Analysis of Bellows Based on GMM Training and HMM Transformation

作　　者：赵亚文 范剑红 陈金国[2] 涂志松 曹存岚 张玉龙 ZHAO Yawen;FAN Jianhong;CHEN Jinguo;TU Zhisong;CAO Cunlan;ZHANG Yulong(New Engineering Industry College,Putian University,Putian 351100,China;School of Mechanical and Electrical Engineering,Putian University,Putian 351100,China)

机构地区：[1]莆田学院新工科产业学院,福建莆田351100 [2]莆田学院机电与信息工程学院,福建莆田351100

出　　处：《四川轻化工大学学报（自然科学版）》2023年第5期33-40,共8页Journal of Sichuan University of Science & Engineering(Natural Science Edition)

基　　金：福建省自然科学基金面上项目(2020J01918);福建省中青年教师教育科研项目(JAT220297);莆田市科技计划项目(2021G2001ptxy06)。

摘　　要：在分析了波纹补偿器异常与正常振动信号后,提出了一种基于GMM训练与HMM变换的振动信号分析方法。首先进行波纹管振动采集实验,保存振动数据,并进行初步时域分析;其次将在自然语言处理领域广泛应用的隐马尔可夫模型应用于波纹补偿器的振动数据分析;最后对提到的故障特征进行基于GMM-HMM算法的波纹补偿器的故障诊断模型构建。通过对实验数据进行测试验证,故障识别率高达96.7%,而传统的算法分析波纹管振动故障,其识别率最高仅为86.7%。此结果表明该算法实现了对波纹补偿器运行状态的准确识别,保证了故障诊断的合理性与高效性。After analyzing the abnormal and normal vibration signals of the corrugated compensator,a vibration signal analysis method based on GMM training and HMM transformation is proposed.Firstly,a bellows vibration acquisition experiment is conducted to record the vibration data and perform initial time domain analysis.Secondly,the Hidden Markov Model,commonly used in natural language processing,is applied to analyze the vibration data of the corrugated compensator.Finally,a fault diagnostic model is constructed for the corrugated compensator based on the GMM-HMM algorithm,focusing on the mentioned fault features.Experimental data test verifies that the fault identification rate reaches 96.7%,compared to only 86.7%using traditional algorithms for analyzing bellows vibration faults.This result demonstrates the algorithm's ability to accurately identify the operating state of the corrugated compensator,ensuring the rationality and efficiency of fault diagnosis.

关 键 词：波纹补偿器 监测 GMM-HMM算法 故障诊断 

分 类 号：TH113.1[机械工程—机械设计及理论]