[panou]

At finite volume discretization we can use for inviscid fluxes at element interfaces the local lax-friedrich scheme (LLF) which has a artificial viscosity.
My question is:

At the faces which are boundaries of domain we use again the LLF flux(using the values of interion element and ghost element) or we use another scheme or simply central flux?

Thanks.

[cfdnewbie]

Originally Posted by panou

At finite volume discretization we can use for inviscid fluxes at element interfaces the local lax-friedrich scheme (LLF) which has a artificial viscosity.
My question is:

At the faces which are boundaries of domain we use again the LLF flux(using the values of interion element and ghost element) or we use another scheme or simply central flux?

Thanks.

You use the LLF as well, and apply the bc through the ghost cell value!

[panou]

ok. Thanks.
No i want to do a question for LDG scheme for flux viscous discretization if you know.

For applying the ldg scheme we must set at each face of mesh, a Right and Left element. Both neighbour elements which have the same face must use the same flux so that to be conservative scheme. In 1D case this is simply to define the R & L element. At 2d how can i determine the R & L of each edge? Can i do this randomly or i must keep a specific orientation ? A manner which i do this is with set a vector for all mesh and thus define the R & L of an edge by sign of vector dot normal vector.