QQ客服 官方微博 反馈留言

	一种基于性能退化数据的电子设备缓变故障预报方法
	何英 ; 周东华 ; 俞容 ; He Ying ; Zhou Donghua ; Yu Rong
	2010-05-12 ; 2010-05-12
关键词	性能退化 故障树 故障预报 时间序列分析 模糊神经网络 performance degradation,fault tree,fault prediction,time series analysis,fuzzy neural network TH165.3
其他题名	Gradual failure prediction of electronic equipment based on performance degradation data
中文摘要	针对电子设备的测试和故障识别提出了一种基于性能退化数据的缓变故障预报方法。重点针对如何体现电子设备功能模块状态与其性能退化数据和故障树之间的内在关系开展建模研究,将难于定量描述的映射关系通过模糊神经网络转化为定性描述的映射关系,结合时间序列分析对性能退化趋势进行评估和判断,为电子设备缓变故障的预报和故障源定位提供了一种便于工程实现的建模方法。以功能模块为基本的故障定位单位,以性能检测数据、可靠性数据、故障树和故障机理等为输入,建立了基于故障树的时间序列神经网络拟合模型,采用时间序列分析对性能退化数据的波动趋势建立分析模型,利用基于故障决策树的模糊神经网络对系统可能故障来源建立判断识别和预报模型,基于非参数密度估计和假设检验方法对预报可信性建立评估模型,从而提高了故障预报的准确性。; In this paper,we propose a series of models for electronic equipment gradual failure prediction through analyzing performance degradation data.As quantitative models are complicated and hard to be established,the authors put their focus on describing the relationships between performance degradation and fault symptom/future possible fault source through setting up half-qualitative models based on time series analysis and fault tree clustering with the aid of fuzzy neural network technique and fault inference method.Functional module and circuit board are regarded as the basic failure units in order to meet the needs of on-site maintenance.The data and information needed in the model include test results of the performance parameters,reliability data of the equipment in the functional module and circuit board level,specific fault tree information,necessary failure mechanism of the equipment and etc.The future degradation trends of the performance are analyzed through time series analysis using fuzzy neural network simulation;the possible fault functional module is recognized and predicted by modified evolutional fuzzy neural network and the prediction probability is evaluated by non-parametric density estimation and hypothesis test,all of which provide a convenient implementation approach especially suitable for engineering application and result in more accurate fault prediction.
语种	中文 ; 中文
内容类型	期刊论文
源URL	[http://hdl.handle.net/123456789/28611]
专题	清华大学
推荐引用方式 GB/T 7714	何英,周东华,俞容,等. 一种基于性能退化数据的电子设备缓变故障预报方法[J],2010, 2010.
APA	何英,周东华,俞容,He Ying,Zhou Donghua,&Yu Rong.(2010).一种基于性能退化数据的电子设备缓变故障预报方法..
MLA	何英,et al."一种基于性能退化数据的电子设备缓变故障预报方法".(2010).

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。