Abstract
ABSTRACTThree issues are tackled in this study to improve the robustness of local remeshing techniques. Firstly, the local remeshing region (hereafter referred to as ‘hole’) is initialized by removing low-quality elements and then continuously expanded until a certain element quality is reached after remeshing. The effect of the number of the expansion cycle on the hole size and element quality after remeshing is experimentally analyzed. Secondly, the grid sources for element size control are attached to moving bodies and will move along with their host bodies to ensure reasonable grid resolution inside the hole. Thirdly, the boundary recovery procedure of a Delaunay grid generation approach is enhanced by a new grid topology transformation technique (namely shell transformation) so that the new grid created inside the hole is therefore free of elements of extremely deformed/skewed shape, whilst also respecting the hole boundary. The proposed local remeshing algorithm has been integrated with an in-house unstructured grid-based simulation system for solving moving boundary problems. The robustness and accuracy of the developed local remeshing technique are successfully demonstrated via industry-scale applications for complex flow simulations.
Highlights
In the context of computational aerodynamics, simulating flows around geometries that may change their shape and/or position with time has commonly occurred in many important aerospace industry applications, such as store separation, stage separation and flying vehicle maneuvering
The boundary recovery (BR) procedure (Chen, Zhao, Huang, Zheng, & Gao, 2011; Du & Wang, 2004; Si & Gärtner, 2011) is still the main practical challenge of developing a robust Delaunay mesher, in the context of local remeshing, where the grid faces can be greatly stretched during the grid deformation process, and some stretched faces may appear on the boundaries of the holes
To efficiently compute a flow problem that contains millions of grid points, the computational fluid dynamics (CFD) solver has to be parallelized by subdividing the computational grid into several parts (Karypis & Kumar, 1998), solving the discretized equations on each part of the grid concurrently and communicating flow data shared by more than one process at domain interfaces using the message passing interface (MPI) library
Summary
In the context of computational aerodynamics, simulating flows around geometries that may change their shape and/or position with time has commonly occurred in many important aerospace industry applications, such as store separation, stage separation and flying vehicle maneuvering. With respect to another type of mainstream unstructured meshing algorithm, i.e., the Delaunay triangulation algorithm, its termination problem for arbitrary complex geometries has been theoretically resolved Despite this key achievement, the boundary recovery (BR) procedure (Chen, Zhao, Huang, Zheng, & Gao, 2011; Du & Wang, 2004; Si & Gärtner, 2011) is still the main practical challenge of developing a robust Delaunay mesher, in the context of local remeshing, where the grid faces can be greatly stretched during the grid deformation process, and some stretched faces may appear on the boundaries of the holes. A store separation simulation of a fully-loaded F16 aircraft model is conducted to demonstrate the applicability of the developed system for handling complex geometries that are often experienced in industry
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have