This paper proposes a novel neural network-based approach with Bayesian regularisation for analysing air foils and wings. Here, the established NACA0012 airfoil, a standard in aviation and aerospace engineering, is used. Empirical methods, various software, and computational fluid dynamics (CFD) simulations have been commonly employed in airfoil analysis. To forecast and model the aerodynamic characteristics of the NACA0012 airfoil and its wing counterparts, this study presents a datadriven approach utilising artificial neural networks (ANNs). Datasets are first generated using software xflr5, followed by AI-based Bayesian regularisation (AI-BR). Approximations in analyses were demonstrated using three data sets: training (80%), testing (10%), and validation (10%), with 20 neurons. Large-scale Simulink results on mean-squared error, error histograms, and regression analyses further emphasise the proposed AI- BR's competence, dependability, and accuracy. A direct and immediate way to assess the accuracy of models or measurements is to use an absolute error (AE) plot, which shows the differences between estimated and actual values. The findings support the effectiveness of integrating ANNs with Bayesian regularisation in aerodynamic analysis. The approach not only enhances prediction accuracy but also opens new avenues for future investigations into more complex aerodynamic models.
Fatima Noor Al-Mansouri (Mon,) studied this question.