Extended Validation of an Immersed Boundary Method in the CODA/Cassiopee CFD Environement for Unstructured Cell-based AMR grids
Please login to view abstract download link
In this paper an extended validation of the immersed boundary method (IBM) in the CFD Software by ONERA, DLR and Airbus (CODA) and Cassiopee CFD environment based on an unstructured cell-based adaptive mesh refinement (AMR) grid will be presented. To highlight the most recent CPU- and memory-efficient mesh generation for an arbitrary complex geometry, a succinct description of the algorithm sub-steps is provided. In addition, a brief overview of the IBM pre-processing, which generates the data required to apply the immersed boundary conditions (IBC), will be presented. These numerical method discussions are required as they address the four major bottlenecks of the previous implementation: memory constraints, CPU time/efficiency, code redundancies, and load balancing. The core of the paper is centered around the validation of the most recent and significant improvements brought to the IBM methodology, while a companion paper addresses the numerical details of the implementation. The test cases considered include the 2D NACA 0012 in the subsonic and transonic regimes, 2D MEA, and the 3D ONERA M6 wing. Body-fitted and Cartesian IBM solutions are considered to determine the accuracy of the current approach. Figure 1 shows an excellent agreement between the Cartesian IBM solution and the current AMR IBM solution for the subsonic 2D NACA 0012. This negligible difference between the two IBM approaches is also present in the other test cases currently considered.
