首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于遗传小波神经网络的非线性动态自治网络故障诊断仿真算法
引用本文:谢宏,谭阳红,何怡刚.基于遗传小波神经网络的非线性动态自治网络故障诊断仿真算法[J].湖南农业大学学报(自然科学版),2013,40(1):65-69.
作者姓名:谢宏  谭阳红  何怡刚
作者单位:(湖南大学 电气与信息工程学院,湖南 长沙410082)
摘    要:为避免用状态方程计算和分析非线性动态网络的约束计算困难,特别是计算响应跨越边界时间的问题,针对非线性动态自治网络,提出一种基于规范式分段线性化总体表达式以及非线性网络的混合参数方程.求解该方程组可得到非线性动态自治网络的故障响应仿真算法,再由小波提取故障响应的特征.采用遗传算法对BPNN进行结构和参数优化,将得到的电路故障状态特征输入至遗传优化的BP神经网络进行故障诊断.仿真结果表明了该故障诊断算法的有效性.

关 键 词:故障诊断  非线性动态自治网络  遗传小波神经网络

A Simulation Algorithm of Fault Diagnosis Based on Generic Algorithm Wavelet Neural Networks for Nonlinear Dynamic Autonomous Networks
XIE Hong,TAN Yang-hong,HE Yi-gang.A Simulation Algorithm of Fault Diagnosis Based on Generic Algorithm Wavelet Neural Networks for Nonlinear Dynamic Autonomous Networks[J].Journal of Hunan Agricultural University,2013,40(1):65-69.
Authors:XIE Hong  TAN Yang-hong  HE Yi-gang
Abstract:When using traditional state equations to calculate and analyze nonlinear dynamic autonomous networks, there are some problems, such as limitations and calculation difficulties, especially the problem of calculating the time of the response crossing a boundary. To avoid these problems for the nonlinear dynamic autonomous networks, this paper presented a method based on canonical piece-wise linearization to obtain a set of canonical piece-wise-linear equations for dynamic units in nonlinear dynamic autonomous networks and the hybrid parameter equations of nonlinear dynamic autonomous networks. By resolving these equations, the simulation algorithm of fault response of nonlinear dynamic autonomous networks can be obtained. This paper used wavelet transform as a preprocessor to extract the fault features from the fault responses of nonlinear dynamic autonomous networks, and adopted Generic Algorithm to optimize the structure and parameters of BPNN. Then, the fault state features were fed to GABPNN to classify and determine the faults. Simulation results show that this fault diagnostic algorithm is an efficient analysis method.
Keywords:
点击此处可从《湖南农业大学学报(自然科学版)》浏览原始摘要信息
点击此处可从《湖南农业大学学报(自然科学版)》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号