The equations in the "large fluid simulation" is wrong
the navier-stokes is wrong, the correct form is
the divergence of the product of miu_{T} and the transpose of grad(u) is missing, it can’t be eliminated because miu_{T} is not constant

It is a well know simplication un case of Boussinesq approximation for the buoyancy, and the resultat are on golf agreement with experiment.

1 Like