Eventos Anais de eventos
ENCIT 2020
18th Brazilian Congress of Thermal Sciences and Engineering
A STUDY OF DISCRETE ADJOINT-BASED GRADIENT METHOD IMPLEMENTATION FOR AERODYNAMIC OPTIMIZATION
Submission Author:
Gilberto Bueno Luque Filho , SP
Co-Authors:
Gilberto Bueno Luque Filho, Marco Aurélio Leonel Matunaga, João Luiz F. Azevedo
Presenter: Gilberto Bueno Luque Filho
doi://10.26678/ABCM.ENCIT2020.CIT20-0589
Abstract
The present work aims to perform an application of the implemented discrete adjoint optimization method to a wing optimization using low-fidelity aerodynamics computed with lifting line theory and an airfoil optimization in a transonic flow regime governed by the 2-D Euler equations. The first application was chosen for solving the optimization problem in a considerable short time, allowing exploration of coupled systems, validations and optimization method comparisons. The second one was selected to apply this knowledge to a problem that involves geometry parametrization, mesh deformation and non-linearities as the solution of the flow field has shock waves. The aerodynamic calculation, geometry parametrization and mesh deformation are computed by a Fortran code, which is, then, imported to the Python optimization process through a Python script and optimized using Scipy Python library. The results obtained are quite encouraging and demonstrate the applicability of the present approach.
Keywords
Adjoint method, Lifting Lifting Theory, 2-D Euler Equations, Algorithmic differentiation, Gradient-based optimization
DOWNLOAD PDF VIEW PRESENTATION
