Grid Skewness
 

Hi all, I am writing my own code for using non-othogonal collocated grids for finite volume method. Grids are structured. I did not scale the equatyions or the domain. What I am trying to solve is very narrow long channel that has a moving wall. The code crash due to grids skewness. domain length over width ration is about 500 (width about 1 mm) My question is what is the definition of skewness? Could this problem be solved by scaling the domain before reading and the equations?

Thanks

Re: Grid Skewness
 

"Skewness" is about angle deviation from orthogonality. The length/width ratio (usually termed "aspect ratio") is another measure of grid quality. Scaling therefore changes the aspect ratio but not the skewness. Try to build a grid which is closer to orthogonal.

Re: Grid Skewness
 

Actually "skewness" has nothing to do with orthogonality. Orthogonal meshes can also be "skew" Skewness is the deviation from the line connection the cell centres of the cells bracketing a face to the face centre. If this line cuts the face at its centre the mesh is not skew!