Research On Airplane Fault Prognosis And Diagnosis System Based On Flight Data

The thesis has studied and developed the airplane fault prognosis and diagnosis system based on flight data. The users will get acquainted with the status of airplane system in time and well informed by the fault signs after decoding and analysis of flight data. It provides a Case-based reasoning expert system with the purpose of collecting, summarizing and propagating expertise, which consequently improves the quality of fault diagnosis, and takes the place of maintenance experts in a manner.The main achievements in this thesis are:1. The flight data decoding and application subsystem has been designed. The new decoding algorithm based on dynamic list can decode all flight parameters according to the user's requirements. The flight data are used in the ways of pilots' panels and display simulating, parameter curving, take-off and landing simulating, and paranormal reporting. In this way, the ground maintenance crew could know well about the airplane faults.2. A Case-based reasoning expert system for airplane fault diagnosis has been constructed. The thesis applies the Object Oriented technology for case representation and distributed structure for it in detail. The system improves case comparability algorithm, and uses the case retrieval technique which made up of the reducing in different levels according to some tactic and the nearest neighbor method. The case learning method based on case comparability is also introduced. This system not only well meets the requirements for fault prognosis and diagnosis, but also simulates the capacities of experts' fault diagnosis to a great extent. The deficiencies of current airplane fault diagnosis methods have been overcome in some way.3. The system prototype has been developed by means of Borland C++ Builder6.0 and Oracle 9i. Some airplane fault cases are added into the case database. The function of the system prototype has been validated by the use of the flight data from 66 scheduled flights.