文章基本信息

标题：Diagnosing Intermittent Faults through Non-linear Analysis
本地全文：下载
作者：Samir Khan ; Takehisa Yairi
期刊名称：IFAC PapersOnLine
印刷版ISSN：2405-8963
出版年度：2020
卷号：53
期号：2
页码：10304-10309
DOI：10.1016/j.ifacol.2020.12.2765
语种：English
出版社：Elsevier
摘要：AbstractTo cope with certain exogenous stimuli, there have been inexorable advances of technology, with an increased focus and fascination with the accuracy of diagnostic equipment. This can become a difficult problem to solve as it warrants real-time monitoring whilst taking up unnecessary measures to improve overall system reliability and maintainability. Intermittent faults may be benign or malignant in nature and their overall impact on a system varies with mission objectives and operating conditions. Major failures can often be averted if these problems can be detected sufficiently in advance by observing them in dynamical behaviour. The phase space trajectory reconstructed from a time series is known to elucidate such behaviours however it is seldom applied for fault analysis. This article makes use of dynamic system theory and investigates its application for fault estimation by analysing non-stationarities which arise due to the changing dynamics under intermittent conditions. Intermittent fault detection presents a challenge for traditional fault diagnostic equipment as they do not manifest themselves all the time. The idea is to move away from the traditional approaches and investigate the use of non-linear analysis by building a reference trajectory using the phase space reconstruction. This is used as an objective measure for any deviations caused by intermittent phenomena. The method is validated using simulated data and shows promise. The implications of the study are to identify new fault isolation bounds necessary to improve diagnostic success rates and potentially lead to early diagnosis of intermittent faults in electrical equipment.
关键词：KeywordsCondition monitoringfault detectiondiagnosticsnon-linear analysisdecision support