Group: sci.physics.computational.fluid-dynamics
From: Victor
Date: Tuesday, March 11, 2008 4:41 AM
Subject: A criteria for grid size.

Hi

I am solving a flow problem around a cylinder using finite differences
and I have a grid size question. The question is not about stability
(my grid is stable), but about accuracy. When I have a problem that
has one result (a certain number) I set my grid and my convergence
values so the first 3 or 4 significant digits of the wanted result.
However, in my case I solve a shear 2D flow around a cylinder and I
don't really have one resulting parameter. What are the common ways to
determine when the grid is fine enough?

Victor Chernov

Safety Articles | News in English | 20lbs in 30 days | Bluegrass | Usenet Newsfeeds