Set-Membership Fault Detection under Noisy Environment in Aircraft Control Surface Servo-Loops

Abstract Based on a consistent interface between a data-driven and a model-driven approach within an interval framework, the paper deals with the detection of two important electrical flight control system failure cases of aircraft control surfaces, namely runaway and jamming. Robust and early detection of such abnormal positions is an important issue for early system reconfiguration and for achieving sustainability goals. The motivation behind this work is the development of an original set-membership methodology for fault detection where a data-driven characterization of random noise variability (which is not usual in a bounded error context) is combined with a model-driven approach based on interval prediction in order to improve the accuracy of the overall detection scheme. The efficiency of the proposed methodology is illustrated through simulation results using data sets recorded on a highly representative aircraft benchmark.