25th Congress of International Council of the Aeronautical Sciences, 3 - 8 September 2006, Hamburg, Germany
Paper ICAS 2006-6.2.4

STATISTICAL FAULT DETECTION AND IDENTIFICATION IN AIRCRAFT SYSTEMS VIA FUNCTIONALLY POOLED NONLINEAR MODELLING OF FLIGHT DATA DEPENDENCIES

D.G. Dimogiannopoulos, J. D. Hios, S. D. Fassois
Univ. of Patras, Greece

Keywords: Fault detection and identification, aircraft systems, statistical methods

A statistical Fault Detection and Isolation (FDI) scheme for aircraft systems based upon the nonlinear modelling of relationships among flight data quantities is introduced. Relationship changes indicate the occurrence of system faults, and are reliably detected by statistical hypothesis testing. The scheme's effectiveness is confirmed via numerous test cases.

view full paper