[jas123]

Dear All, I am trying to make a 2D simulation of a desiccant dehumidifier. A mesoporous silica layer is placed inside an airflow channel. Humid air from inlet flows over the silica layer. During air flow, water vapor is adsorbed by mesoporous silica and dry air is produced at the outlet. Due to adsorption heat release, temperature of silica particle is increased and some of adsorption heat is carried out by air. So, air temperature at outlet is increased. As time passes, silica becomes saturated and adsorption capacity of the bed reduces. A details picture of my model is also attached in the first PDF file if you want to take a look. However, getting moisture transfer through wall from air to silica domain can't be done in Fluent. Alternately, I have found a reference paper (please refer to this link: http://repository.up.ac.za/bitstream...pdf?sequence=1) where they have simulated almost similar thing and they said, we can define the silica surface as interior instead of wall, moisture transfer could be achieved. problem is, I can't find interior in the BC panel for my geometry in such case. Can any body please give me suggestion about how can I create my geometry so that silica surface could be defined as interior instead of wall. I have been stuck at this point for more than a month and your co-operation will be highly appreciated. Thanks for your valuable time and effort.

[CeesH]

I assume from your question that the silica part and the rest of the domain are different parts in your mesher. This means you will have to specify both sides of the interface between them as an "interface" in your named selections (rather than interior), and match them in the mesh interfaces section (interior, as far as I know, only works within parts).

However, making it simply an interior will allow all flow to pas. is that what you want? Or are you going to make it a porous zone?

[jas123]

Dear CeesH, Thanks for your kind co-operation and reply. Yes you are right. Currently in my domain, silica part and rest of the domain are different and silica part is frozen. and I have specified both sides of the interface as "interface" in named selection and I have created mesh interface between them in Fluent. But that is the problem. I have two walls, one for air side and one for silica side. And for species transport, I can specify either zero flux or constant mass fraction as BC at the wall. And Fluent is not capable to diffuse moisture through wall using either of those two BC. That is why in this reference paper (http://repository.up.ac.za/bitstream...pdf?sequence=1) they said they have defined the brick surface as a interior to allow the moisture diffusion through the boundary. At the same time, they have defined porous zone within the brick. And you are also right, interior works within parts only. That is why I am wondering, how they have done it in the reference paper. Actually I want to do almost same thing like that paper. I want water vapor to be passed through the boundary to the silica zone and I want to define silica zone as a porous layer so that water vapor will be adsorbed by porous silica. If you kindly take a look to the reference paper and give me any suggestions about how to model my case, I shall remain grateful to you. Hope you will reply back. Thanks,

[CeesH]

you can join it to a single part I think. then, you can still use named selection on the bodies to specify the different cell zones.