Data-Driven Reduced Order Modelling for Aerodynamic Shape Optimisation

Multi-fidelity Data-driven Aerodynamic Shape Optimization of Wings with Composite Neural Networks

“…As a good alternative of CFD-based design optimization method, surrogate-based approach has been successfully applied to airfoil and wing shape optimization problems [3,17,18]. Recent progresses have shown that data-driven ASO approach can substitute CFD-based optimization while reducing the optimization time with the help of machine learning techniques [19][20][21][22]. Reliable data-driven optimization process relies on accurate surrogate models and compact parameterization methods.…”