A Data-Driven Method for Aircraft Engine Model

This article introduces a data-driven method for engine performance modeling. Firstly, the steady-state performance model of the engine is created by Lagrangian linear difference. Then, transient-state data from test flight and NARX network are used to correct the transient performance based on the steady-state performance model. Finally, the engine performance model is obtained. According to simulation result, the overall prediction result of the engine model basically meets expectations. However, there will be a jump in the speed response prediction when the throttle lever is first changed. As shown in the results, network is needed to be optimized for predicting throttle lever change rate in the future.