Development of an ANN-Based Predictive Model for an Airfoil with a Trailing-Edge Flap

This work makes use of a data-driven approach to predict the aerodynamic performance of the wind turbine airfoil equipped with an active trailing-edge flap. The NACA 4412 airfoil is considered to investigate the effect of the trailing-edge flap angle on the efficiency, measured in terms of the lift-to-drag ratio (Cl/Cd). A vast set of aerodynamic data is generated to consider the angle of attack (α) from −20° to 20° and various flap angles (TE). Based on the generated database, an Artificial Neural Network (ANN) model is formulated to predict the value of Cl/Cd in terms of the angle of attack and the flap angle. The ANN model was trained using 81 iterations with 75% of the dataset used for training, 15% for validation, and 10% for testing. The predictions of the ANN model are then compared to the reference solutions from the Computational Fluid Dynamics (CFD) simulation in the form of systematic plots. The high level of agreement between the predictions of the two approaches emphasizes the validity and accuracy of the proposed data-driven ANN model. The model presented in this work provides an approach to efficiently and accurately overcome the repetitive simulation of aerodynamics and represents an attractive tool in the analysis of smart airfoils with active trailing-edge flaps. The ANN model provides an R2 value of more than 0.9898 on the test data.