[Shibi]

Hello to all,


I would like to know if there is any built-in function in OpenFoam to calculate the shortest distance between a face and a certain patch. In addition, any ideas on how I can filter the e.g., 20% faces further from a patch?


The image bellow is a resume of what is intended.

[olesen]

First thoughts. I think you would need a modified mesh wave similar to this
https://www.openfoam.com/documentati...e.html#details

but with the face centres for the tracking quantity. After that, you will need some global reduction to know min/max limits. After that your fun really starts. How do you bin things to ensure that you pick up your 20‰ of the faces? How do you coordinate this in parallel? Do you consider all faces, or pre/post-filter to only consider faces mostly parallel to the patch? Do you handle curved patches?

[Shibi]

Originally Posted by olesen

First thoughts. I think you would need a modified mesh wave similar to this
https://www.openfoam.com/documentati...e.html#details

but with the face centres for the tracking quantity. After that, you will need some global reduction to know min/max limits. After that your fun really starts. How do you bin things to ensure that you pick up your 20‰ of the faces? How do you coordinate this in parallel? Do you consider all faces, or pre/post-filter to only consider faces mostly parallel to the patch? Do you handle curved patches?

Hello Mr. Olesen,


Thank you for the reply.



My idea was to have a field to store this information before the solver started its thing. The user would supply the patch name, afterwards I would need to calculate the shortest distance between each face of any other patch and the defined patch through its face coordinates. After getting this field, I would have to check which cells would fulfill the requirement of being the furthest from the input patch.


For bit more context, this comes from a previous post (InterFoam letting one phase leave the domain) where a boundary condition was defined for use in InterFoam to let one of the phases leave the domain. However, the convergence is not good and I tough of defining part of the domain with a no-slip boundary condition and the rest with a dynamic outlet/wall boundary at least during the beginning of the simulation.



Best Regards