Learning Forum

[milad.mahjoo]

Hello everyone.

Right now I'm working on a spray cooling case targeting a hot wall utilizing DPM method. Droplets reaching the wall are forming a liquid film with defining a Lagrangian wall film (LWF) boundary. The domain consists of a large continuous fluid phase (air, in which the water droplets are injecting in) and a solid box with a heat energy source term. Between these two there is a coupled wall interface in order to simulate conjugate heat transfer between the wall and the formed film (with non-conformal mesh).

The problem is as the simulation begins and the droplets reach the target, the wall film does not form correctly and it's thickness begin to rise to false values highly on the center point. When wall's DPM boundary is defined other options (like rebound) everything works fine but LWF model along with coupled wall interface doesn't result to a certain valid values and the simulation diverges.

Does anybody have faced same problem before? or on contrary, have anybody implemented these two conditions (LWF and Coupled wall interface) along each other successfully?

Regards. 

[Rob]

It's not listed as a limitaiton https://ansyshelp.ansys.com/account/Secured?returnurl=/Views/Secured/corp/v242/en/flu_ug/flu_ug_lagr_wallfilm_bc.html  but I don't think I've ever used LWF with a non-conformal wall. The energy terms are tightly coupled so some caution may be needed. 

When you say the film has "false values" what is happening? Note, if you have a very fine surface mesh or large parcels (note parcel not particle) you could be adding a relatively high liquid mass to a relatively small facet. 

[milad.mahjoo]

 

Dear Mr. Rob.

Thank you so much for your prompt attention and response to my question. I really appreciate the time and effort you put into helping out.

To provide more context about my issue regarding your answer:

 

• Yes you’re right. This situation is not mentioned as a limitation in ANSYS Fluent User guide for Lagrangian wall film. (although the other model for simulating wall film which is eulerian wall film is mentioned as a model that is not working correct on non-conformal interfaces)
• Since I was curious about the energy terms too, I tried to run the simulation without even solving the energy equation (deseleting it from equations menu) but the problem still remains.
• Here’s a visual representation of the problem, as seen in the attached image. In left side where there is only a fluid cell zone and the target face is a normal wall boundary, the wall film is forming correctly over time. On contrary, on the right side where there are two cell zones (connected with an interface between non conformal meshes) and the target face is the interface, wall film is taking odd shape and heigh values (as its depicted with near 1 mm value which is incorrect)

 

• I changed cell characteristic lenght value for both solid and fluid meshes in various cases to check your point about face and parcel sizes but problem still remains.

If there’s anything specific I might have missed or further details you’d need, please let me know. I’m eager to resolve this and truly appreciate your guidance.

Thanks again for your support.

Best regards,

 

 

[Rob]

 

Please can you replot the left hand image with node values off? That looks very much like a mesh & parcel size issue in that there is a large amount of mass hitting a cell. LWF is transient, so how many particle release points (or streams) are there, and how big is the particle time step? 

To add, the film models are technically facet based rather than cell based so the nonconformal maths may become a little complicated. 

 

[milad.mahjoo]

Dear Mr. Rob.

Once again I really appreciate your kind attention. This is the same contour with node values off.

 

 

Left one is the correct case and the right one is the case that doesn't form correct wall on the target face (with non-conformal mesh). Number of streams are 100 and particle time step is 2e-5 (both of these values are selected with caution and get result independent)

[Rob]

The one on the right looks very odd. On the left isn't great and suggests you need more particle streams and to review the mesh. I'm guessing it's a hex map mesh?

[milad.mahjoo]

yes since the geometry is simple both cases are hex orthogonal mesh. Left image is not accurate as you've mentioned and its depicted just to show that the LWF works fine (even not accurate in this case) when case consists of only one zone. But the right one as it's obvious lacks any wall film formation process. So this is my main quastion: " Is non-conformal interfaces are limitation for Lagrangian wall film model utilization?" If Yes, I suggest to add this on Fluent's next version user's guide and what is the alternative suggestion to form wall films on such cases? If No, then what is the problem in this case?

[Rob]

I would avoid the interface. Waiting on colleagues to confirm if it's a documentation or solver issue: that's going to take a few more days.