基于ART-2神经网络的行星轮系故障模式识别  

PATTERN CLASSYFICATION OF FAULTS FOR A PLANETARY TRANSMISSION BASED ON ART-2 NETWORK

作　　者：唐力伟[1] 肖志松[1] 康海英[1] 

机构地区：[1]军械工程学院一系,石家庄050003

出　　处：《机械强度》2007年第5期862-864,共3页Journal of Mechanical Strength

基　　金：国家自然科学基金资助项目(50375157)~~

摘　　要：探索复杂行星轮系故障诊断方法。对振动信号作功率谱分析,提取故障特征,并基于ART-2神经网络提出行星轮系故障模式识别方法。该方法不仅能够对待检工况的模式作出正确的判断,而且对未知模式具有较强的自适应分类能力,通过实例分析验证方法的可行性。A fault diagnosis method for the complicated planetary transmission is explored. Through power spectrum analysis for the vibration signals, fault features are picked-up, and a method of fault paltem classification for a planetary transmission based on ART-2 network is presented. This method is not ouly able to judge the patterns for the examined states correctly, but also provided with better ability of adaptive classification for the unknown pattern. The validity of this method has been validated through an example.

关 键 词：神经网络 行星轮系 模式识别 

分 类 号：TH132.425[机械工程—机械制造及自动化] TP183[自动化与计算机技术—控制理论与控制工程]