Parallel Advancing Front Method


The primary focus of this project is to design and implement a parallel framework for an unstructured mesh generator based on the advancing front method (AFM). In particular, we target large-scale Computational Fluid Dynamics (CFD) simulations of complex problems. The desire to perform better quality and high-accuracy numerical simulations more efficiently is enabled by: (1) the widespread availability of parallel machine architectures and (2) the ability of numerical solvers to take advantage of parallel machine architectures. The quality of the simulation produced by the solver depends on the particular mesh that is provided by the mesh generator. In order to accurately capture complex physical phenomena, a large-scale high-resolution mesh is required. Furthermore, for time-dependant flow problems involving moving bodies, aero-elastic analysis of aircraft undergoing surface deflections, interactive CFD/design analysis resulting in geometry perturbation, etc., re-meshing is performed during the simulation. In this scenario, the mesher and the solver co-exist at the same time and both compete for resources. Moreover, generating such a mesh on a single machine is a challenging task due to both CPU and memory limitations. A parallel mesh generator, with distributed memory is more suitable for this task.


1. Parallel Mesh Generation For CFD Simulations Of Complex Real-World Aerodynamic Problems
George Zagaris, Shahyar Pirzadeh, Andrey Chernikov and Nikos Chrisochoides. In Proceedings of 6th Symposium on Trends in Unstructured Mesh Generation (MeshTrends VI), 2007.


NASA (LaRC) GSRP Fellowship 2006, 2007

Related Links