CFD for falling film evaporator. The liquid Refrigerator R134a is flowing from an inlet onto a horizontal pipe. And then some of the refrigerator vaporizes,while others travelling around the pipe and forming thin film.

I'm wondering whether I have to use the UDFs(Actually I don't like to use this) or just the general multiphase models will be enough for this problem. If so, which models should I use?

Thanks a lot!

[hitzhwan]

Originally Posted by Binglu Ruan
;135758

CFD for falling film evaporator. The liquid Refrigerator R134a is flowing from an inlet onto a horizontal pipe. And then some of the refrigerator vaporizes,while others travelling around the pipe and forming thin film.

I'm wondering whether I have to use the UDFs(Actually I don't like to use this) or just the general multiphase models will be enough for this problem. If so, which models should I use?

Thanks a lot!

You can print the UDF here.

[vinerm]

Whether the in-built models will work or not depends upon the phenomenon. If the pressure is high, then you can use wall boiling models. If the pressure is lesser than, say, 5 bar, then use thermal phase change model. Do note that this is available only in newer versions; I suppose starting with 18.0. As far as the film is concerned, if it is thin, then you have to use EWF and evaporation also has to be modeled using EWF itself.

[hitzhwan]

Originally Posted by vinerm

Whether the in-built models will work or not depends upon the phenomenon. If the pressure is high, then you can use wall boiling models. If the pressure is lesser than, say, 5 bar, then use thermal phase change model. Do note that this is available only in newer versions; I suppose starting with 18.0. As far as the film is concerned, if it is thin, then you have to use EWF and evaporation also has to be modeled using EWF itself.

(1)why the pressure has effect on the model that I should choose?
(2)what is the wall boiling models and thermal phase change model, where can I set them in fluent 18.0?
(3) I see some people use the Tanasawa phase change model(see attachment), which is different from the in-built model( Lee model) and more suitable for falling film, do you use this model before? I have try to edit it, but I failed many times, can you help me with it ?

[vinerm]

1. Pressure physically controls the phenomenon of phase change. Wall boiling models available in Fluent are meant only for high pressure phase change phenomena.

2. Lee model is thermal phase change model

3. You will have to use UDF to implement this model. It is doable but implementing phase change models are usually difficult. Most of these are numerically unstable. There are hundreds of models out there but only a handful of those are robust enough for numerical modeling of generic nature.

[hitzhwan]

Originally Posted by vinerm

1. Pressure physically controls the phenomenon of phase change. Wall boiling models available in Fluent are meant only for high pressure phase change phenomena.

2. Lee model is thermal phase change model

3. You will have to use UDF to implement this model. It is doable but implementing phase change models are usually difficult. Most of these are numerically unstable. There are hundreds of models out there but only a handful of those are robust enough for numerical modeling of generic nature.

1. where can you see the applicable condition for the two models?
2.Can you edit the UDF, I can give you some fee for you to help me .thank you. It is very important for me.

[vinerm]

You can refer Fluent's Theory Guide or search online for the pressure limits for RPI and Critical Heat Flux models. CHF is extension of RPI wall boiling models.

If you have already written a UDF, I suppose you'd have tried it. If you have not, you can start from a simpler one and develop it. Support from the forum will help you as well as all of us will learn something more. This is not a professional service, hence, no payment is involved.

[hitzhwan]

Originally Posted by vinerm

You can refer Fluent's Theory Guide or search online for the pressure limits for RPI and Critical Heat Flux models. CHF is extension of RPI wall boiling models.

If you have already written a UDF, I suppose you'd have tried it. If you have not, you can start from a simpler one and develop it. Support from the forum will help you as well as all of us will learn something more. This is not a professional service, hence, no payment is involved.

1. What is the RPI?

2. I have tried the Lee model, but it is not applicable. Have you use some udf that can calculate the film heat transfer on the gas-liquid interface? Is it a good result?

[hitzhwan]

The following limitations exist for the Eulerain Wall Film model:

The Eulerian Wall Film model is available for 3D geometries only.

Many models (for example, VOF multiphase flow or radiation) will not interact correctly with the film model without first modifying the boundary conditions using UDFs.

Wallfilm behavior at junctions with baffles (zero-thickness walls) may lead to unphysical results due to the local mesh topology (wall, wall-shadow) and should be used with care, that is, isolated film walls must not share a common edge.

The wall film can only be single-component, but not multi-component.

The Eulerian Wall Film model is not compatible with non-conformal interfaces.

The Eulerian Wall Film model is not compatible with mesh operations in Fluent (such as mesh adaption, cell separation, face zone extrusion, cell zone type change, and so on).

Release 18.0 - © SAS IP, Inc. All rights reserve



1. what does the meaning "The wall film can only be single-component, but not multi-component.",but I see the example you sent to me has water and air, there are two component.

2.what is non-conformal interfaces?

[vinerm]

Components mean material of the film. This has to be a pure material, e.g., water or oil. It cannot be a mixture of, say, two oils. In reality, it could be, but in Fluent it has to be modeled as a pure material. The example has two because only the film has to be single component. But around the film, there is a gaseous environment. So, there are gas and liquid.

Non-conformal interface implies existence of two separate boundaries at the location where two cell zones meet. Usually, this is supposed to be conformal. But sometimes non-conformal is required. Some users define non-conformal as a mesh where mesh in the zone 1 and zone 2 do not have a node-to-node match. However, non-conformal mesh can be non-conformal even if there is node-to-node match. As long as each cell zone has its own boundary, the mesh is non-conformal.

[hitzhwan]

Originally Posted by vinerm

Components mean material of the film. This has to be a pure material, e.g., water or oil. It cannot be a mixture of, say, two oils. In reality, it could be, but in Fluent it has to be modeled as a pure material. The example has two because only the film has to be single component. But around the film, there is a gaseous environment. So, there are gas and liquid.

Non-conformal interface implies existence of two separate boundaries at the location where two cell zones meet. Usually, this is supposed to be conformal. But sometimes non-conformal is required. Some users define non-conformal as a mesh where mesh in the zone 1 and zone 2 do not have a node-to-node match. However, non-conformal mesh can be non-conformal even if there is node-to-node match. As long as each cell zone has its own boundary, the mesh is non-conformal.

Can you give me an example of Non-conformal interface?

[hitzhwan]

I have a problem：
My computer has 4 cpu processors, and I use 3 of them to calculate in fluent by parallel, but the cpu used shown in the task do not increase to 75%, and the calculation speed is slow, do you know why?