A Data-Driven Approach to Detect Faults in the Airbus Flight Control System. Issue 17 (2016)

• Record Type: Journal Article
• Authors: Goupil, Philippe; Urbano, Simone; Tourneret, Jean-Yves
• Is Part Of: IFAC-PapersOnLine. Volume 49:Issue 17(2016)
• Page Start: 52
• Page End: 57
• Publication Date: 2016
• Subjects: Aircraft -- Flight Control -- Fault Detection -- Diagnosis -- Oscillatory Fault Case; Automatic control -- Periodicals; 629.805
• DOI: 10.1016/j.ifacol.2016.09.010
• Languages: English
• ISSNs: 2405-8963

Abstract

Abstract: This paper presents a data-driven strategy for the detection of failures impacting the flight control system. Early and robust detection of Oscillatory Failure Case (OFC) allows the aircraft structural design to be optimized, which in turn helps improve the aircraft environmental footprint thanks to weight saving. Compared to existing model-based techniques already used on in-service Airbus aircraft, this paper studies a novel signal processing approach based on distance and correlation. It is shown that a mixed similarity index between Euclidean distance and logarithmic invariant divergence gives promising detection results. This paper details the proposed approach by insisting on practical constraints due to implementation in embedded real-time systems such as the flight control computer. Preliminary results obtained from a Verification & Validation (V&V) on-going campaign are presented.