[masands]

Hi all,

I am relatively new to CFD. I'm trying to validate a simple gas dispersion model in the presence of an obstacle. It is Thorney Island trial number 26.

It is a transient based simulation. The gas is freon 12 (31.6% w/w) and nitrogen (68.4% w/w) and is released instantaneously 50m upwind from a small building. The gas is denser than air with relative density of 2.0.

So I have enabled the species model and K-e standard model. I defined a log profile for the wind (Velocity Inlet) and ran a steady state with air only and converged. The wind speed is 1.9 m/s.

The side and up boundaries are Symmetry and the outlet is set as Outflow.

Then I patched the species to a cylindrical region, 50m upwind of the obstacle. I enabled gravity (-9.81 m/s^2) and enabled operating density (default). I ran the transient for 250 seconds with 1 second time step. I measured the concentration at the building (mol fraction of freon-12).

The shape of the graph looks exactly like the experimental which is good but unfortunately the %v/v of Freon-12 is exactly 10 times less than the experimental.

For example, the experimental peak is 2% at 50 s. Mine is 0.18% at 50 seconds.

I can't understand what's wrong.

Do I need to define any more boundary conditions?

Edit:

It appears that I have now discovered the source of my problem. The mass fraction of Freon I was patching was too low. I know that the total volume of gas released is 2000 m^3. Therefore based on the fact:

Density Freon = 1305.8 kg/m3
Density N2 = 1.135 kg/m3
Total volume = 2000 m3
Required relative density = 2

So

1.135 (1-x) + 1305.8x = 2*1.225

This gives vol fraction. From there I can calculate the total mass and hence mass fraction.

Although before I start the simulation, the highest density in the domain is 2 kg/m3 which is still less than the relative. How is fluent calculating the density?

[ramzan]

Greetings,
I am modelling the same dispersion experiment as yours and facing the same problem. Can you please guide me through this.