Which phase change model for sub-boiling drying of a wet porous block?

[Bha Kuri]

I am modelling the drying of a wet porous block using the Eulerian multiphase model in Fluent. The block is initially wet, with liquid water patched into the porous zone. One face is held at 80 °C, the opposite face is exposed to ambient air. The surrounding air is not modelled; the ambient side is imposed as a boundary condition. The temperature stays below boiling throughout, so the water leaves by evaporation into unsaturated air rather than by boiling.

I have looked through the theory guide and a number of papers but have not been able to find a clear answer on which model applies here. Which option in the Phase Interaction panel should I use: Evaporation-Condensation (Lee model), Thermal Phase Change, or Species Mass Transfer?

[Rob]

I'd use Lee or Thermal Phase change, probably the latter. However, there is also a vapour partial pressure to consider that's not included in the base model. 

Species transfer is more intended for solvent extraction and the like. 

[Bha Kuri]

 

Thanks Rob.

Setup: simple rectangular porous block, hot face at 80 °C, ambient face as a pressure outlet. No Corey model, just the base porous formulation, porous zone initialised at 0.5 liquid volume fraction. Saturation temperature is obtained from the local vapour partial pressure and I have checked it is calculated correctly.

With both Lee and Thermal Phase Change, the relative humidity spikes in the first few time steps and goes above 100%, so the gas saturates and the liquid never dries out. I have been working on this for a long time now and still have not been able to sort out the RH spike.

My questions:

1. What causes the RH to spike like this in the within first few time steps?

2. Is sub-boiling drying of this kind actually achievable with the built-in phase interaction models?

3. Since the surrounding air is not modelled, what would you recommend for the ambient face so vapour can leave the domain — pressure outlet with backflow species, or a specified species mass fraction?

Thanks.

[Rob]

If there is no mechanism for condensation then RH>100% is "normal", so check how you're accounting for vapour pressure and rate of phase change.  You did code the UDF correlations in Kelvin? All UDFs are SI regardless of the unit definitions in the solver. It's a sensible set up, but can confuse now Expressions use units: it's also been confusing Americans for decades. :) 

Also, in porous media note porosity does not effect volume fraction. Double check you've got the amount of liquid you want in the domain. 

I'd move the outlet away from the porous media. Whilst pressure boundaries with a backflow setting are valid, how would you know if you should be drawing in fresh air or some small amount of vapour?