Prediction of Airfoil Efficiency by Artificial Neural Network

The estimation of the right airfoil for an aerial vehicle is a crucial step in the preliminary design and can affect the aerodynamic performance of the vehicle. The parameter to measure the airfoil performance is aerodynamic coefficients. The problems associated with the determination of CL and CD are addressed by using different algorithms and models of artificial neural networks (ANN). In this paper, ANN model is used to calculate and verify the aerodynamic efficiency of an airfoil, which includes the estimation of the CL and CD of different airfoil geometries at various angle of attack (− 10° to 10°), Mach number, and Reynolds number. The ANN has been widely used for its ability to solve exceedingly nonlinear problems in the real world. This study includes training the dataset to form a suitable model that can predict the aerodynamic coefficients in less time and have more accuracy. The dataset is obtained through CFD analysis which was used to train the model and validate its efficiency. The artificial neural network (ANN) model that is developed allows an Mof range 0–0.7 and works on any airfoil. This model gives highly accurate results as compared to other numerical methods with the added benefit of minimizing time and high computational costs.