CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   On Gradient Grid Adaption (

David May 1, 2005 17:22

On Gradient Grid Adaption
Hello CFD community,

I am modelling a laminar pulsatile flow in a conical tube 0.01 m long *(x-direction) with a varying radius between 0.0065 and 0.0125 m (y-direction). I am modelling it as 2d for now and using Fluent 6.1.

Today I tried for the first time the dynamic grid adaption based on gradient (dx-vel/dy). I used the scaled condition with refining treshhold at 0.7 and coarsening treshhold at 0.3 (default values). Adaption was done at every time step.

When I did so I was surprised to see the solution process much faster than without this option.I observed repeatedly the following message by fluent:

Grid size (initial/adapted/changed) cells (100/ 100/ 0) faces (300/ 300/ 0) nodes (200/ 200/ 0)

The values below are used for simplification purpose. I would like to draw your attention to the "0" values obtained for cells, faces and nodes count.I tried examining the mesh visually but could not find any significant change to the mesh.

A few basic questions: 1) Has the mesh by modified in any way? 2) Is the dynamic adaption responsible for this increase in processing speed although the above Fluent message would suggest no change?

Regards David

Luca May 2, 2005 03:14

Re: On Gradient Grid Adaption
Sorry but are you using hanging adaption? in this case you can easily see is the grid has changed. Otherwise you can save a new cas file, reload it in Fluent and see while loading the messages that appear: node number, cell number. Yes I think that you may speed up your solution convergence history and maybe shorten a little the time for each iteration, but until a certain number of cells... Luca

David May 2, 2005 15:50

Re: On Gradient Grid Adaption
Thanks for your response. I will try it.

All times are GMT -4. The time now is 12:14.