Hi,

Is there anyone who succeeded in modeling contamiants transport(ex. CO2 etc.) in air with FLUENT in 3D Case? I tried to use the multi-species transport model in Fluent 5 and 6 to simulate CO2 transport in air in a ventilated 3D room ( a mixture of CO2 and He, so that it can be heavier or lighter than air, isotherm), the result is very bad compared with experiment data as reported in IEA Annex 20 reports(Flow Pattern in Buildings, Test Case F). I tried every possible option (turbulence models, density specification methods, segregated and coupled solvers, etc.), but even with the case that the density of the mixture equals that of the air, the result is far from experiment. Can anyone give me some hints what's wrong with this or if FLUENT is capable of doing such kind of simulations?

Thank you very much for your help in advance.

Regards,

Marc

can you briefly describe how you are modeling the experiment with particular attention to the modeling of the inlet diffuser performance and the contaminant source?

Dear Mr. Schwarz,

Thank you very much for your reply.

For the inlet diffuser, I represented it as a mass-flow-inlet with component directions(X, Y, Z) and a turbulent intensity of 10% together with its hydraulic diameter.

For the contaminant source(CO2+He), I tried first a momentum-less source (volumetric mass source) because the source is very small (30mm pingpong ball, with 6 evenly distributed holes of 1mm diameter to let out the contaminants in a room of 4.2x3.6x2.4 m, the flowrate is 0.025L/s; the ventilation flow rate is 0.0151m3/s)compared with the main flow field, but it doesn't work, and then I added also the momentum source(I separated 7 cells as the source cell and added 6 momentum and mass sources to each of the 6 cells outside and added a negative mass source to the cell at the center to account for the entrainment of air by the 6 small jets). I tried also the unsteady approach (URANS) and the coupled solver, for the latter I used a very small courant number (0.0005!), it never converged(after more than one week's running!). BTW, when I try to use the RSM model with quadratic pressure-strain option, it always diverges! I can send you the case file if you need it.

Hope to get your help soon.

Best regards,

Marc

to accurately model the jet penetration from the inlet diffuser into the room, you really need to consider the performance aspects of this diffuser, in this particular experiment,i believe the inlet diffuser consists of 84 nozzles rather than the single inlet you have described ... take a look at the following paper which talks about simulating case E (without contaminants)

http://www.tfd.chalmers.se/~lada/pos..._roomvent96.ps

and you should also consider looking at chen's report on simplified inlet boundary conditions for room inlet diffusers ...

http://www.confex2.com/store/items/ashrae/RP1009.htm

in regard to the way you modeled the contaminant source a simple volumetric mass/species source should work just fine.

in terms of convergence, you may have to try an unsteady approach since air jets are sometimes wander making steady-state solutions difficult in low ventilation flow situations like this, but i would certainly avoid using extremely low courant numbers or extremely low underrelaxation factors since those make it difficult to move the solution towards convergence.