[aladdincham]

Hi guys,
How to generate a grid for CFL3D with ICEM CFD,and the grid is 2D,how to set in the CFL3D`S inp to compute the 2D grid,i find a question,the 2D grid of ICEMCFD is not fit for the CFL3D,WHY?
THX

[ramkr]

Hello, really sorry that you are receiving this reply a long time after your question. Still, considering the fact that it might be useful to anyone new with the same question,
CFL3D is a finite volume RANS solver, hence requires grid in volume format, even for 2D grids. Hence there is a requirement of atleast 2 points in the spanwise direction of your grid, which would yield volumetric cells.
By setting i2d=1 in the .inp file, CFL3D effectively reads the grid as 2D.

[naffrancois]

Just to avoid adding confusion, a finite volume solver does not need a grid extended in spanwise direction to perform 2D computations. You can perfectly handle real 1D-2D-3D grids with a finite volume solver. The fact that CFL3D or other softwares need the user to expand their meshes is more related to a coding strategy or related to old programming language limitations.

[ramkr]

Originally Posted by naffrancois

Just to avoid adding confusion, a finite volume solver does not need a grid extended in spanwise direction to perform 2D computations. You can perfectly handle real 1D-2D-3D grids with a finite volume solver. The fact that CFL3D or other softwares need the user to expand their meshes is more related to a coding strategy or related to old programming language limitations.

I'm sorry I didn't realize this valid point while posting the answer. Yes, the basic reason why 3D grids are supplied to CFL3D is not to do with finite volume method, but the fact that the code was originally written for 3D grids (as its name says). With time, feature to do simulation for 2D grids was included in its standard input file, by means of slicing the grid in the plane of interest.