Mass source in chtMultiRegionFoam creates thermal wake
 

I would appreciate advice on the following difficulty in using chtMultiRegionFoam. I have set up a simplified 2-dimensional (wedge) representation of the SNO+ neutrino detector with three regions: the liquid scintillator volume (a sphere of 6 m radius with a cylindrical, upright neck of length 7 m and radius 0.75 m), the plexiglass wall that contains the scintillator, and the water surrounding the plexiglass and that extends laterally to a rock wall. All upper surfaces are held at a steady nominal laboratory temperature (say 293 K), and the rock wall and floor at an elevated temperature (say 303 K). Thermal inhomogeneity should set up a steady-state circulation in the cavity water and the scintillator, which are thermally coupled through the plexiglass wall. An additional factor affecting the flow is that the cavity water is extracted near the upper surface, chilled and returned at various points. I have represented this process by way of a single mass sink and four mass sources (whose locations are visible in the figure below). The total mass circulation rate is taken here as 6 kg/sec, and by virtue of the assumed azimuthal symmetry, the sources are actually closed circular tubes (whose planar cross-section is represented by a mesh cell set of radius 0.5 m).

Though a steady-state circulation is sought, for the time being I am using the solver chtMultiRegionFoam, unmodified. The figure shows the solution at time t=270 minutes from initialization (with the scintillator, plexiglass and water uniformly at 285 K). Of course given the substantial mass of fluid in the system, this is a very short time, and it is not clear whether the circulation pattern in the water is primarily a response to the mass sink and sources, though probably so (if the influence of the rock sidewall's higher temperature than the adjacent fluid were dominant, one would expect ascent along that face). But irrespective of that point, the solution shown suffices to illustrate the conundrum and motivation for this post. In the wake of the mass sink and of each mass source (each of which is shown on the figure as a coloured sphere) can be seen a thermal wake: a temperature deficit wake in the case of the mass sink, and temperature excess wakes in the case of the mass sources. This is physically undesirable and, particularly in the case of the source at the base of the cavity, an unacceptably gross effect.

I've checked through all tutorials for chtMultiRegionFoam and chtMultiRegionSimpleFoam and looked at every fvOptions file: I don't believe there is any case featuring a mass source that might set me straight. From an earlier forum discussion (https://www.cfd-online.com/Forums/op...ss-source.html ) I was alerted to the fact that, when introducing a mass source (sink) one should compensate for the fact that the appearing (disappearing) mass introduces (or takes away) thermal energy. Using the relationship between enthalpy and temperature I have tried by trial and error to "null out" this effect, but have not succeeded, and the effective one would need is ambiguous.


In the figure velow, the mass sink and mass sources are represented by coloured spheres. The mass sink (purple, near the top of the cavity water) has a strength of 6 kg/s. The mass source at the level of the equator of the sphere has a strength of 3 kg/s. The other three sources each have a strength of 1 kg/s. The vertical velocity (UZ, m/s) is auto-ranged distinctly for each region (scintillator, water).