Abstract: Component deteriorations of an aero-engine result in the instrument measurement deviations from nominal values. These measurable deviations can be used as a tool to predict the deterioration level of an engine. A state variable model, which takes the deterioration factors into account, is developed at first, and then these deterioration factors are treated as augmented states of the linear model. A Kalman filter is then designed to predict the component deteriorations. The unmeasurable output parameters are trimmed by the predicted deteriorations. Digital simulation results of the adaptive model are presented and compared with those of the nonlinear engine model. The real-time adaptive model is shown to be able to match the actual engine well.

Key words: aerospace propulsion system, real-time adaptive model, Kalman filter, state variable model