遗传算法优化的模糊神经网络在故障诊断中的应用  

Simulation Studies of Genetic Algorithm for Optimizing Fuzzy Neural Network in Fault Diagnosis

作　　者：刁帅[1] 刘磊[2] 

机构地区：[1]武汉船舶职业技术学院教务处,湖北武汉430050 [2]哈尔滨工业大学,黑龙江哈尔滨150001

出　　处：《武汉船舶职业技术学院学报》2010年第3期30-33,共4页Journal of Wuhan Institute of Shipbuilding Technology

摘　　要：本文针对船舶柴油机故障诊断系统提出了一种基于遗传算法优化训练的模糊神经网络诊断方法,介绍了这种模糊神经网络故障诊断系统的结构及其参数形式,通过遗传优化算法对它的权值和阈值进行了学习优化训练。这种方法可以有效地避免通常所选BP算法训练易陷于局部极值的问题,最后将该遗传算法优化训练的模糊神经网络系统应用到船舶柴油机的故障诊断中,通过仿真研究,说明了该方法的有效性。This article proposes a approach based on the Genetic algorithm optimization training fuzzy neural network,and introduces this kind of fuzzy neural network fault diagnosis system structure and the parameter form,through the Genetic optimization algorithm to its weight and the threshold value has carried on the study optimization training.This method may effectively avoid the question that usually chooses the BP algorithm to train easily to sink into the partial extreme value.Finally this Genetic algorithm optimization training fuzzy neural network system is applied in the fault diagnosis of the marine diesel engine.Through the simulation research explained this method validity.

关 键 词：模糊神经网络 遗传算法 故障诊断 船舶柴油机 

分 类 号：U364.121[交通运输工程]