Hello everybody,

While calculating velocities in finite volume method we use some schemes like upwind, QUICK etc. But to calculate flux across control volume face, simply the average value of the node velocities of the cvs to which the face belongs, are taken. Why a scheme a not adopted to calculate flux?

With regards

jika

It depends what kind of grid do you use. If you have staggered grid then you have the exact velocity value on the face. If you have collocated grid you have to interpolate the value between two CVs to which this face belongs. Staggered grids are only good for some primitive geometry with orthogonal lines. Collocated grids are much more flexible.

I think you mean "colocated," as in all variables for a control volume are located at the same place - Alternatively, forget the control volumes and say colocated variables are located at the nodes of the mesh.

"Collocation" is a mathematical term, and "collocated variables" are operated on in "collocation."

Thank you.