Hi, I am working on natural convection in closed domains with hot wall at bottom and cold side walls. The Ra is at order of 10e+10. I am using segregated solver and k-eps turbulent model. PRESO! for pressure-velocity coupling, QUICK for discretization methods. The fluid is dense, so I wrote all the properties as the function of temperature only. The case converged with residuals near 1e-4 and the heat balance is achieved. But the Mass is not conserved. The mass average temperature is much higher than the initiation temperature. The mass left in the domain is around 15% less than the initiation. I can not image how could this happen. And the simulated results fit the experiment measurement (temperature field measurements) well. Could someone help me about this? Thanks in Advance Asarum

Hi,

I still pose a few questions before answering, such as what is Cold and Hot wall mean? (relative to your initial condition...?) and Is your Fluid compressible? (not the flow but the Fluid) but I'll try anyway. I can see that you are working on a closed domain. This means your mass is a function of only density - assuming a non deformable mesh. As I recall you set the properties as a function of temp only, and I presume this includes density too. If that is the case then I wouldn't be surprised if you get a lesser ( a different mass than initial mass) - Residuals of order the value u mentioned may not automatically GUARANTEE convergence at all. My bottom line is that if your simulation is transient and the fluid is compressible then you may need to activate floating operating pressure - it is selectable when transient case is activated.

good luck

Melaku H.