Downwash and Upwash Prediction Model for Unconventional Lifting Configuration

The present work introduces a neural model to predict upwash and downwash characteristics of unconventional wing configurations, such as for boxed wing aircraft, due to the need to predict trim conditions for a drone in urban air applications. The neural network is trained according to a low fidelity database based on vortex lattice method aiming to cover all possible configurations for the considered application. Some training points are verified with computational fluid dynamic simulations. Results demonstrate the feasibility of the proposed approach to predict upwash and downwash as functions of geometrical parameters.