A Generalised Methodology for the Diagnosis of Aircraft Systems

An aircraft is made up of a number complicated systems which work in harmony to ensure safe and trouble-free flight. In order maintain such platform, many diagnostic prognostic techniques have been suggested, mostly aimed at components but some the system level. Together these form patchwork approach overall problem efficiently informing maintenance Original Equipment Manufacturers, operators /airlines, Maintenance, Repair, Overhaul organisations. It involves organisations having support several different approaches health management, therefore inefficient costly. current work, streamlined methodology put forward. This based on OSA-CBM (Open System Architecture for Condition Based Maintenance) can be applied any system. Integral with this use mRMR (minimum redundancy maximum relevance) feature selection, resulting symptom vector being used fault diagnosis. demonstrated three test cases: engine, environmental control system, fuel each case, digital twin setup, simulation conditions healthy faulty scenarios, diagnostics are detailed. Diagnostics carried out turn, using four machine learning supervised algorithms. The best performing algorithm will then subsequently vehicle level reasoner called FAVER (A Framework Aerospace Vehicle Reasoning), requires diagnoses as starting point reasoning ambiguity resolution.