Towards Aerodynamic Shape Optimization Using an Immersed Boundary Overset Grid Method
Brandon M. Lowe, Chase Ashby, James R. L. Koch, David Craig Penner, Jeffrey A. Housman, Jared C. Duensing · 2024
Traditional Reynolds-averaged Navier-Stokes grid methods applied to aerodynamic shape optimization can struggle with the deformation of surface and volume grids at component intersections, such as at wing-fuselage junctions. To overcome this, we propose an approach which utilizes curvilinear overset grids for the discretization of the domain, with the presence of the body modeled using an immersed boundary method. This approach handles complex geometries without the need for their explicit integration into the grid. The goal of this approach is to reduce grid generation time and allow for greater geometric freedom for component-based aerodynamic shape optimization. Two different methods are presented: a source-term-based and a ghost-node-based immersed boundary method. Flow analyses and adjoint solutions obtained using the proposed methods show promising comparisons with standard body-fitted grid methods. Preliminary aerodynamic shape optimization results obtained using one of the immersed boundary methods are also presented.