On the discrepancy of modelling the heat transfer for pure natural convection

Abstract

In the present study, buoyancy driven flow fields in a bench scale apparatus are experimentally investigated by means of a non intrusive measurement technique, here the 2D-2C particle image velocimetry (PIV). An electrically heated copper block was used to induce a plume. As the temperature at the surface of the heating source is the main driver for the fluid dynamics, the surface temperature field was measured by thermography. The experimental results are compared to simulations with the Fire Dynamics Simulator (FDS). Simulations with a prescribed, homogeneous distributed surface heat flux and a 3D thermal heat conduction with a volumetric heat source were carried out. The impact of both approaches on the resulting flow fields showed no significant deviations, however, both do not agree with the experimental data. Whereas the surface temperatures of the heating source and its vicinity are in a good agreement for the 3D heat transfer approach. This leads to the conclusion that the modelled heat transfer mechanism does not provide valid predictions in this setup.