Prediction of Flow Variables over a Serrated Double Delta Wing Aircraft using Neural Network model

Multidisciplinary design frameworks developed for aeronautical applications require significant computational power, which increases exponentially with the use of higher fidelity tools. The goal of this paper is to construct and apply a neural network model to predict vortices from CFD simulation data for a serrated double delta wing aircraft. The input parameters for the neural network are the wing, airfoil geometry, and flight condition. The model is built using the sequential API which allows for a linear stack of layers to be added one after another. Neural network predicted vorticity and CFD simulated result of vorticity over absolute pressure, strain rate, turbulent kinetic energy, and turbulent dissipation rate findings corresponding within 7 to 8, 3 to 4, 1 to 2 and 0.5 to 1% of each other. DOI : https://doi.org/10.52783/pst.1744