Data-driven Methods for Fault Localization in Process Technology

• Auteur principal: Kühnert, Christian (auth)
• Format: Électronique Chapitre de livre
• Langue: anglais
• Publié: KIT Scientific Publishing 2013
• Collection: Karlsruher Schriften zur Anthropomatik / Lehrstuhl für Interaktive Echtzeitsysteme, Karlsruher Institut für Technologie ; Fraunhofer-Inst. für Optronik, Systemtechnik und Bildauswertung IOSB Karlsruhe
• Sujets: Time series; Signal processing; Data Mining; System identification; Causality
• Accès en ligne: DOAB: download the publication; DOAB: description of the publication

Résumé

Control systems at production plants consist of a large number of process variables. When detecting abnormal behavior, these variables generate an alarm. Due to the interconnection of the plant's devices the fault can lead to an alarm flood. This again hides the original location of the causing device. In this work several data-driven approaches for root cause localization are proposed, compared and combined. All methods analyze disturbed process data for backtracking the propagation path.

Description matérielle

1 electronic resource (XVIII, 194 p. p.)

ISBN

KSP/1000036427
9783731500988

Accès

Open Access