[jrrelx]

Hi all.
I'm trying to simulate the pre-cooling of an air stream by means of evaporative cooling in Fluent. I intend to validate the CFD model with the results from the experiments I have conducted in a wind tunnel facility.

In my simulation, unsaturated air flows into the tunnel and is mixed with a an air stream carrying away small water droplets. As a result of the difference between the partial pressure of vapor at the interface and the vapor concentration in the bulk gas, water evaporates and the air temperature is reduced.

I'm using the DPM model along the k-epsilon turbulent model for the continuos phase. The thing is that for my low water mass flow rate injected experiments, the cases converge and the results match the physics of the problem (water droplets temperature reaching the inlet air wet-bulb temperature), humidity levels in the expected range (no greater than 100% in relative humidity) and temperature levels in fair agreement with the experimental data. However, for the high water mass flow rate injected experiments, the results of the simulations make no sense at all. The humidity levels increase above 100% and sometimes even the particle trajectories deviate from the expected path.

I was wondering if I was defining ok the physiscs of the problem. I understant that when droplet evaporation is activated in Fluent, the vaporization law is enabled and it uses the appropiate correlation to calculate the saturated vapor pressure, which is crucial to evaluate the evaporation since it's the driving force for the evaporation process. But having obtained realiable results with the same set-up, I no longer think that is related to it.
Instead, I've reading that there is a limitation when using the DPM model.

The second phase should be sufficiently dilute that particle-particle interactions and the effects of the particle volume fraction on the gas phase are negligible. That occurs for low volume fraction, usually less than 10-12%. In my simulations, the main aire stream volumetric flow rate is about 0.2 m3/s. At the injection section, I have 0.004 m3/s of air and 1.26e-6 m3/s of water (DPM). I don't think my cases are over that limit, but I can't see any reasons why the simulation is failing to converge.

The mixture of both humid air streams doesn't result in supersaturated conditions. I've attached some screenshots of the simulations that converged succesfully.

Any help would be appreciated.
Thanks in advance for your time.

Cheers.