自适应粒子群优化的HMM故障诊断方法及应用

郭森,王大为,张绍伟,张学成

振动与冲击 ›› 2021, Vol. 40 ›› Issue (20) : 264-270.

PDF(2057 KB)
PDF(2057 KB)
振动与冲击 ›› 2021, Vol. 40 ›› Issue (20) : 264-270.
论文

自适应粒子群优化的HMM故障诊断方法及应用

  • 郭森1,王大为1,张绍伟1,张学成2
作者信息 +

A fault diagnosis method with application of HMM Based onadaptive particle swarm optimization

  • GUO Sen1, WANG Dawei1, ZHANG Shaowei1, ZHANG Xuecheng2
Author information +
文章历史 +

摘要

针对传统隐马尔科夫模型(hidden Markov model, HMM)解决学习训练过程中参数容易局部收敛的问题,采用一种自适应粒子群算法对HMM模型进行优化改进。在基础粒子群算法中加入惯性权重因子,并根据算法迭代结果对算法各因子大小进行动态控制,增强该算法的全局搜索能力。改进后的粒子群算法融入HMM模型训练过程参数学习的优化过程,能够提高HMM的训练精度。将其应用到油机电站的故障诊断当中,通过对其实测振动数据进行分析,与标准算法相比,对油机处于正常、供气不足、进气门间隙异常状态的分类准确率都有所提升,整体诊断精度达到97.3%。结果表明,基于自适应粒子群优化的HMM故障诊断方法能够有效解决传统模型的局部收敛问题。

Abstract

Aiming at the problem that parameters learning algorithm of hidden Markov model (HMM) easily converges to local optimal solutions, an improved HMM was proposed based on adaptive particle swarm optimization (PSO).An inertial weight factor was joined in the basic PSO.It was controlled dynamically based on the iterative results of the algorithm.Global search capability of the new algorithm thus was improved.The adaptive method was integrated to parameters learning algorithm of HMM, which contributed to optimize the initial parameters of HMM.The new model was applied to fault diagnosis for diesel engine with normal state, lack of air supply and intake valve clearance fault.Actual engine vibration data were analyzed.The correct classification rate of the proposed method reaches 97.3%, better than the traditional method based on PSO.The results verify that the proposed fault diagnosis method has a better performance than the traditional algorithm.

关键词

粒子群优化 / 自适应方法 / 隐马尔科夫模型 / 故障诊断

Key words

particle swarm optimization / adaptive method / hidden Markov model / fault diagnosis

引用本文

导出引用
郭森,王大为,张绍伟,张学成. 自适应粒子群优化的HMM故障诊断方法及应用[J]. 振动与冲击, 2021, 40(20): 264-270
GUO Sen, WANG Dawei, ZHANG Shaowei, ZHANG Xuecheng. A fault diagnosis method with application of HMM Based onadaptive particle swarm optimization[J]. Journal of Vibration and Shock, 2021, 40(20): 264-270

参考文献

[1]邱志明, 罗荣, 王亮, 等.军事智能技术在海战领域应用的几点思考[J].空天防御, 2019,2(1):1-5.
QIU Zhiming,LUO Rong,WANG Liang,et al.Some thoughts on the application of military intelligence technology in naval warfare[J].Air & Space Defense,2019,2(1):1-5.
[2]BAUM L E, EAGON J A.An inequality with applications to statistical estimation for probabilistic functions of Markov processes and to a model for ecology[J].Bulletin of the American Mathematical Society,1967,73(3): 360-363.
[3]崔天宇.基于HMM的语音识别系统的研究与实现[D].长春:吉林大学, 2016.
[4]王恒, 季云, 朱龙彪, 等.基于HDP-CHMM的机械设备性能退化评估[J].振动、测试与诊断, 2018,38(4):733-737+871-872.
WANG Heng, JI Yun, ZHU Longbiao, et al.Performance degradation evaluation of mechanical equipment based on HDP-CHMM[J].Journal of Vibration Measurement & Diagnosis, 2018, 38(4):733-737+871-872.
[5]王刚, 陈捷, 洪荣晶, 等.基于HMM和优化的PF的数控转台精度衰退模型[J].振动与冲击,2018,37(6):7-13.
WANG Gang, CHEN Jie, HONG Rongjing, et al.Model for the positional accuracy degradation of NC rotary tables based on the hidden Markov model and optimized particle filtering[J].Journal of Vibration and Shock, 2018, 37(6): 7-13.
[6]李奕江, 张金萍, 李允公.基于VDM-HMM的滚动轴承磨损状态识别[J].振动与冲击, 2018, 37(21):61-67.
LI Yijiang, ZHANG Jinping, LI Yungong.Wear state recognition of rolling bearings based on VMD-HMM[J].Journal of Vibration and Shock, 2018, 37(21): 61-67.
[7]ADITIYA N A, DHARMAWAN M R, DAROJAH Z, et al.Fault diagnosis system of rotating machines using hidden Markov model(HMM)[C]∥International Electronics Symposium on Knowledge Creation and Intelligent Computing.IEEE, 2017.
[8]耿宁.基于粒子群优化的隐马尔科夫模型的复合攻击预测方法[J].通信电源技术, 2015, 32(3): 69-71.
GENG Ning.Approach to forecasting multi-step attack using hidden markov model based on particle swarm optimization[J].Telecom Power Technology, 2015, 32(3): 69-71.
[9]江金源.基于HMM的滚动轴承故障诊断方法研究及硬件实现[D].哈尔滨: 哈尔滨工业大学, 2018.
[10]丁超然, 刘三明, 王帅, 等.基于改进连续隐马尔科夫模型的风机齿轮箱故障诊断[J].电力学报, 2019, 34(1): 68-78.
DING Chaoran,LIU Sanming,WANG Shuai,et al.Fault diagnosis of fan gearbox based on improved continuous hidden markov model[J].Journal of Electric Power,2019,34(1):68-78.
[11]朱嘉瑜, 高鹰.基于改进粒子群算法的隐马尔科夫模型训练[J].计算机工程与设计, 2010, 31(1): 157-160.
ZHU Jiayu, GAO Ying.Adaptive particle swarm optimization for hidden Markov model training[J].Computer Engineering and Design, 2010, 31(1): 157-160.
[12]KENNEDY J.The particle swarm: social adaptation of knowledge[C].Proceedings of IEEE International Conference on Evolutionary Computation, Indianapolis, Indiana, 1997.
[13]张金玉.装备智能故障诊断与预测[M].北京: 国防工业出版社,2013.
[14]何栋磊, 黄民.基于遗传算法优化HMM的刀具磨损状态监测研究[J].机床与液压, 2017, 45(15): 106-108.
HE Donglei, HUANG Min.Research on tool wear state monitoring based on optimized hmm by genetic algorithm[J].Machine Tool & Hydraulics, 2017, 45(15): 106-108.
[15]张西宁, 雷威, 杨雨薇, 等.采用自适应基因粒子群算法优化隐马尔科夫模型的方法及应用[J].西安交通大学学报, 2018,2(8):1-8.
ZHANG Xining, LEI Wei, YANG Yuwei, et al.Adaptive genetic particle swarm algorithm for optimization hidden markov models with applications[J].Journal of Xi’an Jiaotong University,2018,2(8):1-8.
[16]王霞.非平稳信号特征提取方法研究及其在内燃机故障诊断中的应用[D].天津: 天津大学, 2015.

PDF(2057 KB)

Accesses

Citation

Detail

段落导航
相关文章

/