Predicting the Fluid Flow Around an Airfoil Using Machine Learning Methods

The aim of this paper is predicting pressure distribution around the NACA0012 airfoil by using machine learning models. The pressure distribution model is produced to be used in a simulation environment that is designed to illustrate the behaviour of the fluid around the airfoil in various atmospheric conditions. The model should work for all of these different conditions and finding the solution for all by using CFD (Computational Fluid Dynamics) algorithm is not time and cost efficient. Thus, instead of solving it for all conditions, a training set is formed with certain cases and the distribution for other conditions are predicted with minimum 91.8% accuracy by applying artificial neural networks, decision tree and K-nearest neighbours methods.