[yeharav]

Hi all,

I am running the same case on 3 different servers, each with different OF version (2.2.2, 2.4.0 and 3.0.1). The cases resemble a stratified channel flow using a solver that was adapted from the BuoyantBoussinesqPisoFoam solver.

The simulations results differ (I examined vertical profiles of p_rgh in the center of the domain) after 10 time steps (0.5 sec, CFL~0.2) when 2nd order convection schemes are employed. I have examined SuperBee, vanleer, vanalbada and quick

When I Use first order schemes (upwind) or remove the non linear terms I get identical results for all OF versions.