NEURAL NETWORK BASED FUEL SAVINGS OPTIMIZATION FOR TRANSPORT AIRCRAFT

The paper develops an adaptive performance optimizer using the technology of artificial neural networks (ANN). The adaptive performance optimizer is aimed at adjusting the aircraft control surface configuration to minimize the drag force which in turn will boost the fuel efficiency. The control surfaces involved in this optimization are the outboard symmetric ailerons which re-camber the wing. The initial flight scenario is the level flight at cruise. The ANN optimizer identifies the optimal symmetric aileron position based on flight conditions and aircraft states and sends this position command to the flight control system. The automatic flight control system uses this signal to command the symmetric outboard aileron accordingly. This results in a minimum drag aircraft control surface configuration for fuel savings. The ANN optimizer is trained off-line extensively with multiple flight conditions and multiple parameter variations to render the performance robustness. The ANN optimizer is also equipped with on-line learning capability for adaptation to modeling error and aircraft aging.