Generic Fault-Diagnosis Strategy based on Diagnostic Directed Acyclic Graphs using Domain Ontology in Automotive Applications
Konferenz: AmE 2019 – Automotive meets Electronics - 10. GMM-Fachtagung
12.03.2019 - 13.03.2019 in Dortmund, Deutschland
Tagungsband: GMM-Fb. 93: AmE 2019
Seiten: 5Sprache: EnglischTyp: PDFPersönliche VDE-Mitglieder erhalten auf diesen Artikel 10% Rabatt
Behravan, Ali; Meckel, Simon; Obermaisser, Roman (University of Siegen, Siegen, Germany)
In safety-critical systems the functionality in the presence of faults can be preserved using fault recovery in combination with robust diagnostic techniques. Fault-diagnosis is essential in many fault-tolerant control applications. In addition, low fault-positive and fault-negative rates are required in order to maximize the customer satisfaction and to reduce maintenance cost. The international standard ISO 26262 that is named functional safety for road vehicles is an adaptation of the international standard for electrical, electronic, and programmable electronic safety-related systems (IEC 61508). The automotive safety integrity level (ASIL) evaluates the failures based on severity, exposure, and controllability factors for assigning risk levels and is a key component of the ISO 26262. Embedded fault-diagnosis systems should be resourceefficient and simple for implementation with low-cost controllers, however, keeping a high diagnostic coverage. In the past, malfunction indicator lights (MILs) on the dashboard reported anomalies in a system but not the specific component. Nowadays, system-specific fault-diagnosis techniques are available, however, with the ongoing progress in vehicle’s technology coming up with more and more complexity, describing a generic fault-diagnosis strategy is still challenging. In this paper two main goals are addressed. The first one is to get a full demonstration of the Fault-Error-Failure propagation in a domain ontology such as automotive ontology application which is also useful for the proposed diagnostic appraoch. Also this paper provides a solution for generic fault-diagnosis based on diagnostic directed acyclic graphs (DDAGs), distinguishing system, subsystem and component relationships thanks to the system ontology and Fault-Error-Failure propagation model. A key point is that the diagnosis system itself must be robust and able to overcome missing diagnostic inputs, e.g. missing sensor data, by proper signal substitutions. The approach is a part of condition-based maintenance (CBM) which uses run time data for fault-diagnosis. The diagnostic results are required to initiate repair or maintenance tasks prior to a failure.