Safety modelling for the time limited dispatch of aircraft

• Main Author: Prescott, Darren Richard
• Published: Loughborough University 2007
• Subjects: 629.1333490289
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.487689

Time-limited dispatch (TLO) allows aircraft to dispatch with known faults present in the engine control systems. These systems govern the thrust control of engines and, in order for a TLO scheme to be certified, the failure rates to loss of thrust control (LOTC) must lie within or below prescribed bounds. The aim of the work presented in this thesis was to develop a tool that could be used to model the time-limited dispatch (TLO) of aircraft and compare this with currently recommended techniques for modelling TLO. Currently techniques for modelling TLO require the failure rates to LOTC from various dispatchable system configurations, i.e. system configurations where a number of faults are present. These rates determine how long the system may be dispatched with certain faults present and also a failure rate to LOTC can be computed for the whole system. A number of approximations are used within the models, intended to make them simple to use. The concern is that these approximations might not accurately model the system behaviour and that the results obtained might as a consequence be unreliable. Monte Carlo simulation (MCS) was identified as an analysis method that could deal with the intricacies and complexities involved in the application of TLD. Software, written in C++, allowed TLO to be modelled using MCS. Full use was made of the flexibility of MCS and many TLO maintenance scenarios were considered as the code was developed. The MCS code was written in such a way that all failure rates to LOTC, i.e. those representing the system LOTC and the dispatchable fault LOTC rates, would be produced as outputs. Results obtained using the developed software and the currently recommended techniques were compared. Finally, the MCS code was embedded within an optimisation procedure in order to demonstrate how such a tool could be used in the design process for a system. A simple genetic algorithm procedure was employed to carry out this optimisation process.