Multiobjective Aerodynamic Topology Optimization using Quadtree Genetic Programming
Naruhiko Nimura, Akira Oyama · 2023
View Video Presentation: https://doi.org/10.2514/6.2023-3473.vid Multiobjective airfoil design optimization is developed using a new evolutionary topology optimization method combined with Reynolds-averaged Navier-Stokes simulation. The optimization method utilizes the quadtree image compression technique to reduce the number of design variables, making the optimizations highly efficient. The multiobjective design optimization of the trailing edge section of the NACA4412 airfoil was chosen as the validation problem. The optimization progressed monotonically and after 5000 solution evaluations the trade-off relationship of objective functions was observed and 36 non-dominated solutions were obtained. By exploring the solutions and flowfields around the obtained solutions, it is found that the optimization was successful in finding the geometrical characteristics that strongly affect the objective function value, such as the camber or taper of the trailing edge, indicating that the proposed method can be utilized for conceptual design or understanding of design objects.