[FLASH-USERS] possible errors in viscous diffusive flux in hydro unsplit

[Fabien Thiesset]

Dear all,

I double checked the viscous diffusive fluxes for the momentum equation 
in the hydro unsplit solver. I use version 4.7.1 and the file under 
consideration is 
source/physics/Hydro/HydroMain/unsplit/hy_uhd_addViscousFluxes.F90.

In theory, the diffusive flux of velocity component i in direction j 
should write

2 mu S_{ij} - \mu/3 \theta \delta_{ij},

where S_{ij} is the strain rate tensor, \delta_{ij} the kronecker delta, 
\theta = div(u) and \mu the dynamic viscosity.

If I am not mistaken, I found 2 errors in the computations of the 
viscous fluxes of u along x direction (it is called F02XMOM_FLUX in the 
code)

- the first appears at line 106-108, to compute du/dy that constitutes 
the diffusive flux of u in x direction. I think a 1/2 factor is missing.

- the second appears at line 122-124, to compute dw/dz that constitute 
the diffusive flux of u in x direction. There is some mix between 
VELY_VAR and VELZ_var to compute the gradient. It should be VELZ_VAR 
everywhere.

Can anyone have a look and confirm ? If it is indeed the case, is there 
a git repo for flash to raise an issue and provide the fix ?

I will tomorrow have a look also the diffuse flux for the energy 
equation and the viscous heating. I need also to check the Diffusive 
fluxes for the momentum equation for the split solver. I did it quickly 
this morning and it looked ok. I will anyway write a follow-up.

Thanks everyone !