Eventos Anais de eventos
COBEM 2019
25th International Congress of Mechanical Engineering
An Efficient Multi-Objective Optimization Strategy Using the Self-Organizing Maps for Airfoil Shape Designs
Submission Author:
SungKi Jung , SP , Brazil
Co-Authors:
Octavio Silva, SungKi Jung, Luiz de Siqueira Martins Filho
Presenter: Octavio Silva
doi://10.26678/ABCM.COBEM2019.COB2019-1283
Abstract
In this study, an efficient multi-objective optimization strategy is presented to circumvent a time-consuming problem of the optimization, which is combined with the evolutionary algorithm and the computational fluid dynamics (CFD). The strategy is based on the two approaches; a readjustment of initially designed ranges of parameters and construction of the rapid response model replacing the CFD calculations. The self-organizing map method and the generalized regression neural networks are employed for the readjustment of the design ranges and the response model, respectively. For the multi-objective optimization, the adaptive range multi-objective genetic algorithm was used. Lastly, present approaches are tested to design the airfoil shapes, which have a better aerodynamic performance than a reference airfoil in terms of a maximum lift coefficient and a lift-to-drag ratio.
Keywords
self-organizing map, generalized regression neural networks, adaptive range multi-objective genetic algorithm, PARSEC method, Computational Fluid Dynamics
