A concept for fault diagnosis combining Case-Based Reasoning with topological system models. Issue 2 (2020)

• Record Type: Journal Article
• Authors: Zinn, Jonas; Vogel-Heuser, Birgit; Ocker, Felix
• Is Part Of: IFAC-PapersOnLine. Volume 53:Issue 2(2020)
• Page Start: 8217
• Page End: 8224
• Publication Date: 2020
• Subjects: Fault diagnosis -- Expert Systems -- Case-based Reasoning -- Model-driven Reasoning -- Model-driven Engineering -- AutomationML; Automatic control -- Periodicals; 629.805
• DOI: 10.1016/j.ifacol.2020.12.1989
• Languages: English
• ISSNs: 2405-8963

Abstract

Abstract: Automated failure recovery plays an important role in improving Overall Equipment Effectiveness and is a building block of industry 4.0. However, in an increasingly dynamic market, failure recovery mechanisms need to be able to adapt to system changes. Starting with fault diagnosis in automated Production Systems for assembly and logistics, this paper proposes a novel approach to combining Model-based Reasoning on topological system models with Case-based Reasoning. The topological models are leveraged for case adaption, which significantly reduces the engineering effort of adding new fault types to the system, compared to signal-based methods. Furthermore, the approach does not rely on complete fault models existing in advance; thus, the case database can be continuously built up during operation.