Hello,

I am relatively new to CFD and am in need of some advice.

My problem is that I need to validate the mesh that I am using. I know that usually you basically reduce the size of the grid until there is no change, or an acceptable change in results. However my problem is that I am using t grid with the aim of adapting it. Is there a technique for validating adapted t grids.

My idea so far is to create a t-grid and gradually reduce cell size until I am happy that the solution is grid independant, and only then adapt the mesh to the flow. Is this right, or is there an easier quicker way?

I hope this makes sense

Thanks for your time

Chris

(1). Start with a reasonable mesh, the one you like. (2). Obtain the converged solution, and plot the flow field. (it can be velocity, pressure, total pressure, tke) Remember to set the level of contour plots (set the min. and max.) because you will need this for comparison. (3). adapt and refine the mesh (wall, or region), run the case again, and plot the results with the same contour level min/max. (4). Make comparison between the two solutions. (5). If the wall properties are important, you should also check the shear stress or skin friction there at the wall. (6). Repeat the same process, until the solution is no longer changing. (7). You can test out various options to adapt the mesh. Sometimes, it is easier to use local refinement. It depends on the type of the problem you are solving. (8). So, start with a coarse mesh, then refine the mesh locally until no change in solution. try some options along the way.

Thanks alot John, that is very useful.

Regards

Chris Rand