Aerodynamic Analysis of an Optimized Serrated Double Delta Wing at Different AoAs

Abstract

The aerodynamic performance of an optimized serrated double delta wing has been analyzed extensively at various angles of attack (AOAs) and Mach numbers up to 1.3. This research focuses on evaluating the aerodynamic parameters, such as lift, drag, and moment coefficients, using computational fluid dynamics (CFD) simulations and validating the results through experimental studies. The optimized design incorporates serrations on the leading edge of the double delta wing, aiming to enhance performance characteristics by modulating vortex flow dynamics. newlineTo predict and optimize the aerodynamic performance, an artificial neural network (ANN) model has been developed and trained on CFD-generated data. The ANN predictions align closely with CFD results, demonstrating the efficacy of the machine learning approach in aerodynamic studies. Principal component analysis (PCA) and k-nearest neighbors (KNN) techniques were employed to analyze the influence of key design variables on performance metrics, ensuring robust data interpretation and dimensionality reduction. newlineThe research includes detailed vortex dynamics analysis, wherein vortex strengths were computed for the wing ahead, over the surface, and in the wake region. The Rankine vortex strength was calculated for both serrated and non-serrated wing configurations to assess the influence of serrations on vortex stability and aerodynamic efficiency. The results show that serrations significantly alter the vortex structure, leading to improved lift-to-drag ratios at specific AOAs and Mach numbers. newlineMonte Carlo simulations were conducted to evaluate the probabilistic behavior of aerodynamic parameters under varying operational conditions. The outcomes reinforce the reliability of the optimized design in achieving superior performance across a range of scenarios. newlineThis study not only validates the aerodynamic advantages of serrated double delta wings but also highlights the potential of integrating advanced computational methods, machine learning models, and experimental techniqu