When and where
Thursday 21 February 2019, 12:00 Sala Consiglio, INM Headquarters Rome
Abstract
Adaptive grid refinement is the technique of locally refining a part of a computational mesh, during the simulation, according to the requirements of the solution being computed. My research concerns the creation of such methods which are powerful enough to function in a modern general-purpose fluid mechanics solver for complex flow problems. The first part of this talk will discuss the challenges faced and the opportunities created in the development of such a method.
A central issue for an adaptive grid refinement method is the refinement decision, the choice of which cells are to be refined. Here, the refinement criteria are based on metric tensors, which directly indicate the desired cell sizes in different directions. This approach provides flexibility for the refinement criterion. Finally, I show that the metric tensor context is a straightforward way of globally varying the cell sizes in the adapted mesh. Thus, adaptive refinement can be used to create the series of coarse to fine grids used in grid convergence studies for assessing numerical uncertainty, or in any other situation where simulations of different precision are required
Biography
Jeroen Wackers studied Aerospace Engineering at Delft University of Technology. He defended his Ph.D. in 2007, based on work at CWI, Centre for Mathematics and Computer Science, in Amsterdam. He then joined Ecole Centrale de Nantes where he is responsible for the development of adaptive grid refinement within the Navier-Stokes solver ISIS-CFD and the study of adaptation in the context of realistic flow simulation in hydrodynamics.