I am trying to use a size function to ensure unstructured cells close to a moving boundary remain well refined whilst allowing the average size to gradually increase with distance from said boundary. I have been trying for months but have thus far been unable to achieve a satisfactory degree of control. Can anyone offer any advice? Thanks, J.

What exactly are you trying to do? Without having any info on your problem, here's some general info on sizing functions.

There's the three sizing functions (fixed, curvature, and proximity). Fixed and curvature are the most common used. Fixed starts with a certain size mesh, while curvature chooses the mesh size so that the angle between any two face normals is less than the value you specify. You pick what faces (or volumes, or edges, or vertices) you want to use as a source, what volume (or face) to attach the sizing function to, and then the start size (or angle if using the curvature SF), the growth rate (less than 1.2... maybe more if you're far enough away from the body, but I would just stay below 1.2), and the maximum size.

Decompose your body before using sizing functions. They take a long time when there's a big difference between your start size and your maximum size. Also, there's some errors to watch out for... after it creates the background grid it's going to list some details on the background grid... what you want to look for is the maximum number of cells hitting background tree depth... make sure this is 0! If it isn't, hit undo, then in the defaults, increase your TOOLS.SFUNCTION.BGRID_MAX_TREE_DEPTH. Keep doing this until you don't have any cells hitting your background tree depth.

Hope this helps, and goodluck Jason

Thanks Jason, that helps my understanding of meshing in Gambit. However my problem is with the dynamic remeshing in Fluent. No matter how good my original mesh is Fluent always introduces large cells close to my moving boundaries. At present I am combating this by creating an initial mesh with only very small cells in a remeshed domain which closely hugs the moving boundary (a passenger car), but it still introduces large cells! Fluent support insist that the size function remeshing parameters will control this but I can't seem to get it working. Any ideas? Cheers, J.

Sorry, I've never worked with Dynamic Meshing. I've got more than enough work backed up with steady-state problems that don't move!

I wish you luck, and there's a lot of people out here who are good at this. Hopefully one of them can help.

Jason

jason , is that reference fram prob solved,

Do you mean that question I posted a few days ago about the rotating reference frames? If so, then no... I got some examples, and did a lot of research on here, and I got some help from FLUENT, but it's still not coming out right. I have tail fins with a cant angle on them. When I hold the roll at 0, there should be a large rolling moment, and if I allow the body to roll in the direction it wants, the rolling moment should decrease until I hit it's steady-state roll rate (I'm actually inputing a roll rate). But that's not happening... in fact, nothing is happening... the only difference between the rolling reference frame sol'n and the non-rolling sol'n is that when I plot vectors of reference velocity, then they are drawn in a circular pattern around the axis of rotation (in the rolling reference frame sol'n), but there is no change in any of the forces! That's just not right, but can't figure out why... The B.C.'s I'm using are the fluid is a rotational moving zone with a specified roll rate, and the body is rotational moving but with a roll rate of 0 relative to the adjacent cells. Any suggestions?

Thanks, Jason

Have you tried to use the Meshed Size Function? And what is the growth rate you are using?

Meshed Size Function is new to Gambit 2.3. Have you used it yet? I'm in the process of upgrading and haven't gotten a chance to use them yet... any advice on them is appreciated.

Thanks, Jason

Yes, I am using it and it seems to be a pretty good extra tool. You mesh first a surface and create afterwards the Size Function that is build up on the meshed surface.

I am trying to get a good mesh for impingement jet on semi-cylinderical plate with holes in it which simulates the effects of leading edge(318K) cooling by impingement jet(293K). Now I am unable to get a nice mesh near the impinging area, as it is very important as the temperature gradient is very high near the wall.

Is the size function as option for it?