期刊文献+

基于自适应周期变异粒子群优化BP神经网络的旋转机械故障诊断 被引量:1

Rotating Machinery Fault Diagnosis Based on BP Neural Network with Adaptive Periodic Mutation Particle Swarm Optimization
下载PDF
导出
摘要 在旋转机械故障诊断应用研究中,BP神经网络在网络结构选择、网络训练精度和网络泛化能力方面还存在很多问题;粒子群优化(PSO)算法作为一种优化方法被应用于BP神经网络的参数优化,但存在早熟收敛的问题,影响了神经网络的优化性能和诊断效果;提出了一种基于自适应周期变异粒子群优化BP神经网络(APMPSO-BP)的算法模型,该算法通过改进粒子群自适应搜索策略来提高网络泛化性能,根据群体适应度方差以及当前粒子最优解来确定全局最优粒子的变异概率,变异操作增强了粒子群优化算法跳出局部最优解的能力,避免了算法的早熟收敛;通过实例验证和对比分析表明,采用APMPSO算法优化的BP网络在网络收敛速度、故障诊断准确度上都优于其它几种网络,证明该方法具有较高的诊断精度和效率。 BP neural network has many problems in the application research of rotating machinery fault diagnosis, such as network struc ture selection, network training accuracy and generalization capability. As a kind of optimization method, particle swarm optimization (PSO) algorithm has been applied to BP neural network parameters optindzation, however, this method leaves the problem of premature conver gence which will affect neural network optimization performance and diagnosis effect. Therefore, a model of BP neural network optimized by particle swarm optimization algorithm based on adaptive periodic mutation (APMPSO) is proposed. The algorithm improved network gener alization ability by improving adaptive search strategy, and determined the optimal particle mutation probability according to group fitness va riance and current particle optimal solution. The mutation algorithm enhanced the ability of particle swarm optimization algorithm to jump out of local optimal solution, and avoided premature convergence. Verified by examples and comparative analysis, it shows that APMPSO BP neural network is better than several other network models in convergence speed and fault diagnosis accuracy, and this method has higher di agnosis accuracy and efficiency.
作者 姚杰 李红伟
出处 《计算机测量与控制》 北大核心 2013年第10期2624-2626,2654,共4页 Computer Measurement &Control
基金 国家自然科学基金资助项目(51274171)
关键词 粒子群优化 自适应周期变异 BP神经网络 故障诊断 particle swarm optimization adaptive periodie mutation BP neural network fault diagnosis
  • 相关文献

参考文献9

二级参考文献34

共引文献48

同被引文献23

  • 1熊伟丽,徐保国.基于PSO的SVR参数优化选择方法研究[J].系统仿真学报,2006,18(9):2442-2445. 被引量:66
  • 2VEDAM H, VENKATASUBRAMANIAN V. PCA-SDG based process monitoring and fault diagnosis[ J]. Control Engineering Practice, 1999, 7 (7) :903-917.
  • 3GHATE V N, DUDUL S V. Optimal MLP neural network classifier for fault detection of three phase in- duction motor[ J]. Expert Systems with Application, 2010, 37(4) : 3468-3481.
  • 4SALAMA M A, HASSANIEN A E, FAHMY A A. Deep belief network for clustering and classification of a contin- uous data [ J 1. IEEE International Symposium on Signal Processing and Information Technology, 2010:473- 477.
  • 5HINTON G E. A practical guide to training restricted boh-zmann machines [ R]. Report of Momentum, 2010, 9 (1):1-20.
  • 6KENNEDY J, EBERHARTR. Particle swarm optimiza- tion[ C ]. Proceedings of IEEE International Conference on Neural Networks. 1995 : 129-132.
  • 7TIELEMAN T, HINTON G. Using fast weights to im- prove persistent contrastive divergence[ C]. International Conference on Machine Learning, 2009 : 1003-1040.
  • 8HINTON G E, OSINDERO S. A fast learning algorithm for deep belief nets[ J]. Neural Computation, 2006,18 : 1528-1554.
  • 9DOWNS J J, VOGEL E F. A plant-wide industrial process control problem [ J ]. Computers and Chemical Engineering, 1993,17 (3) :245-255.
  • 10徐乐华,凌卫新,熊丽琼.基于面向对象自适应粒子群算法的神经网络训练[J].计算机应用研究,2009,26(1):111-113. 被引量:6

引证文献1

二级引证文献21

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部