Applying Bayesian optimization with Gaussian process regression to computational fluid dynamics problems

Bayesian optimization (BO) based on Gaussian process regression (GPR) is applied to different CFD (computational fluid dynamics) problems which can be of practical relevance. The are i) shape in a lid-driven cavity minimize or maximize the energy dissipation, ii) wall channel flow order obtain desired pressure-gradient distribution along edge turbulent boundary layer formed other wall, and finally, iii) controlling parameters spoiler-ice model attain aerodynamic characteristics airfoil with an actual surface ice. diversity problems, independence approach from any adjoint information, ease employing solvers loop, more importantly, relatively small number required simulations reveal flexibility, efficiency, versatility BO-GPR applications. It shown that ensure finding global optimum design size up 8, less than 90 executions needed. Furthermore, it observed does not significantly increase parameters. associated computational cost these affordable for many cases