文章基本信息

标题：Binary Classifier for Fault Detection Based on Gaussian Model and PCA ⁎
本地全文：下载
作者：Tian Cong ; Jerzy Baranowski
期刊名称：IFAC PapersOnLine
印刷版ISSN：2405-8963
出版年度：2018
卷号：51
期号：24
页码：1317-1323
DOI：10.1016/j.ifacol.2018.09.564
语种：English
出版社：Elsevier
摘要：AbstractIn the field of Process Condition Monitoring (PCM), fault detection has been a fundamental and challenging problem. Particularly in the situation where new kind of operating mode is launched, detecting anomalies is quickly becoming difficult due to lack of sufficient prior knowledge of the newly imported operation. This article introduces a naive and low-complexity approach,Binary classifierForFault detection (BaFFle), which extract the operating mode in a naive representation and identify process conditions in unconformity with expectancy. The proposed method is devised on the basis of Gaussian model assumption and Principal Component Analysis (PCA). The validation of BaFFle is experienced on process data from Cranfield University. Experiments show that, BaFFle can reasonably handle the task of differentiation between normal and abnormal operations.
关键词：KeywordsProcess condition monitoringfault detectionPCAGaussian model