[RameshK]

Hi all

I have two questions in the context of Co-located Finite Volume method for NS equations

1) Regarding artificial dissipation scheme introduced by Jameson and Turkel while using Central difference in FVM.

Can this formulation be used in a grid with Cylindrical coordinates or curvilinear grids? if so how do we justify the amount of dissipation added by this method?

As their paper suggests they have used it on Cartesian mesh and many of the other people used it on Cartesian grids,

2) while using Implicit method using LUSGS, the flux jacobians are split and are used, these split jacobians are on the cell face (interfaces) and act as numerical dissipation, am I correct or wrong?

can any one give some Idea regarding the implementation, I have already in the mode of implementation but need to get some confirmation

any suggestions regarding this