Single- and Multi-Objective Optimization of a Low-Speed Airfoil using Genetic Algorithm

Aerodynamic optimization is undoubtedly an important part of design due to its effect on an aircraft's performance. Objectives of such optimization problem usually involve black-box function of computational simulation, which will not fit the use of conventional gradient-based optimization method as it needs information of derivatives that only well-defined functions are able to provide. The following research presents an airfoil optimization using gradient-free technique called genetic algorithm (GA). The algorithm mimics the concept of genetic inheritance and Darwinian natural selection in living organisms. From a random initial population, GA will generate new individuals iteratively until a desired solution is found. The objective is to minimize the coefficient of drag from a low-speed airfoil of NACA 0012 using PARSEC parameterization technique and a low-fidelity CFD solver XFOIL, with an addition of minimizing the absolute value of coefficient of moment for multi-objective optimization problem. The airfoil is successfully optimized using the GA with the final result of a reduced drag coefficient by almost 50%, and a set of optimum solutions with varying trade-off for each objective is obtained from the multi-objective case.