This paper describes a new approach to the development of a diagnostics and prognostic capability for an advanced cycle marine gas turbine by treating the problem of engine diagnostics as an optimisation problem. In order to detect a fault, the engine performance model would simulate all possible combinations of potential problems with varying degrees of faults and a comparison would be made with the actual parameters obtained from an engine. The difference between the actual and simulated parameters would be converted to an objective function and optimisation technique like genetic algorithm would be used to minimise the objective function. The application and advantages of this technique is discussed.
展开▼