[Obad]

Hi guys, I started using Open Foam a couple of days ago and I have some fundamental questions about the set up of the fvSchemes file.

Don't get me wrong I learned about CFD in my lectures and I understand the theory behind FVM, Godunov's idea, the numerical flux calculation (Roe, HLL, ...), higher order schemes (MUSCL, WENO, ...) and so on. However, I don't quite understand what some of terms in Open Foam mean and what they actually do. I read the User's Guide, but it's no big help to me...

I would like to understand the following terms: Gauss integration, Gradient schemes, snGradient schemes, Laplacian schemes, Divergence schemes, Interpolation schemes.

It would be great if someone could help me with these terms. Here is what I think some of these schemes do and my detailed questions:

Gauss integration:
My understanding is, that the flux at the cell faces is constant across the whole face, thus the flux will simply be multiplied with the surface area of that face. Why would I need a numerical integration scheme?

Gradient schemes: e.g. for viscous terms
I understand that the viscous terms need a special treatment because of the gradient of the velocity term. This term can for example be modeled by a simple finite difference. Is that what the gradient scheme does? Does the keyword Gauss linear mean that the gradient is approximated by a finite difference (slope of a linear function)?

snGradient schemes: I have no clue what this is good for...

Laplacian schemes: e.g. if viscosity is constant
Is this just a special treatment of the viscous term (instead of using Gradient scheme) if viscosity is constant?

Divergence schemes: e.g. divergence of mass, momentum, energy flux
With the numerical flux scheme (e.g. HLL) I tell the solver how the mass, momentum and energy flux is to be calculated. The numerical flux scheme is specified at the top of the fvSchemes file with the keyword fluxScheme. However, at this point the order of the numerical flux approximation is not specified. Is that what the divergence scheme is good for, to specify the order (e.g. 2nd order MUSCL)? If this is the case, then what is the difference between linear and MUSCL? Is it even possible to use a first order piecewise constant numerical flux?

Interpolation Schemes:
For the gradient scheme, the divergence scheme and laplacian scheme it is always necessary to specify the method of interpolation. Why is it necessary to state a separate dictionary in the fvSchemes file for the Interpolation Scheme?

As you see I'm a little bit confused ^^

[mmohaqeqf]

Have you found the answers for your questions?
I am also trying to find out what these entries mean and what they do.