中国机械工程 ›› 2013, Vol. 24 ›› Issue (20): 2753-2757.

• 信息技术 • 上一篇    下一篇

基于ICPSO优化的极限学习机在故障诊断中的应用

高斐1,2;李洪儒1;许葆华1   

  1. 1.军械工程学院,石家庄,050003
    2.95918部队,广水,432700
  • 出版日期:2013-10-25 发布日期:2013-10-25

Applications of Extreme Learning Machine  Optimized  by ICPSO  in Fault Diagnosis

Gao Fei1,2;Li Hongru1;Xu  Baohua1   

  1. 1.Ordnance Engineering College,Shijiazhuang,050003
    2.95918 Army,Guangshui,Hubei,432700
  • Online:2013-10-25 Published:2013-10-25

摘要:

极限学习机(extreme learning machine,ELM)的分类性能受随机产生的输入权值和隐层阈值的影响,为此,提出一种改进的混沌粒子群算法(ICPSO),用以优化输入权值和阈值,得到基于ICPSO优化的ELM故障诊断模型。仿真和实验结果表明,ICPSO算法改善了ELM网络的学习效率和诊断精度,可有效应用于故障诊断。

关键词: 极限学习机, 改进混沌粒子群算法, 故障诊断, 液压阀

Abstract:

Extreme learning machine with random determination of  the input  weights and hidden biases may lead to non-optimal performance.Therefore,a approach of ICPSO was proposed to select input weights and hidden biases for ELM algorithm.Simulation and  experimental results show that the ELM improved by ICPSO can be applied  to diagnosis with better learning efficiency and precision.

Key words: extreme learning machine(ELM), improved chaos particle swarm optimization(ICPSO), fault diagnosis, hydraulic valve

中图分类号: