[brsutherland]

I have been using openfoam to simulate buoyant plumes in uniform density fluid for 2 years now. No problem. Hot fluid originates from a chamber at the base of a cylindrical domain with fluctuating near-constant velocity at the inlet and an outlet at the top of the domain using "matchedFlowRateOutletVelocity" with OpenFOAM-v2012 using buoyantBoussinesqPimpleFoam.

In new simulations I have prescribed the initial ambient temperature using setFieldsDict (with boxToCell commands) to be non-uniform so that the plume would rise and then hit a temperature interface at a height 1/3 from the top of the domain above the nozzle. In this case, there is flow leaving the nozzle, as revealed by the velocity and a passive tracer field "St". But the temperature of this fluid is the same as that as the ambient fluid, even though the fluid in the nozzle is hot. What really baffles me is that after running for 9.2s simulation time, the hot fluid in the nozzle starts to come out together with the pre-existing jet flow finally creating a buoyant plume, and then this addition of buoyancy changes the flow.

How can the initial condition (the ambient temperature profile) have such an unphysical effect upon the initial nozzle temperature flux, but not the nozzle momentum flux? Why does hot fluid from the nozzle suddenly (discontinuously) rise into the ambient fluid after 9.2s, though time steps are deltaT=0.001s?

I hope someone can help.

Bruce