I have a model to calculate the flowfield. After 20000 interations, the residual of X-velocity is about 0.05 and will stay the same for the next 10000 interations. Can I consider it to be a convergence result?Is it too big for convergence(In fact, the default residual is 1e-3)?
Re: Convergence problem
(1). You can plot some flow variables of interest to you vs number of iterations. This is the only meaningful way to determine the solution convergence. (2). If these flow parameters oscillate, then you have some convergence problems. (3). The more accurate way is to use contour plots between iterations. Then you can see areas of oscillation. (I think you can plot the residuals contour plot) (4). Convergence problem can be caused by many factors, among which, mesh quality is one of the most obvious one.
|All times are GMT -4. The time now is 12:22.|