Learning Forum

[Luca Tricarico]

Hi, I'm simulating a point source in a fluid domain via udf. The UDF implements the thermal source in the centroids of the cells of the domain from a txt file. For better accuracy in simulating thermal gradients I would like to use fluent's adaptive mesh refinemet option. Is there a way to implement AMR in fluent in this case? This is because by applying the adapttime mesh refinement I would need to recalculate the source in the centroids of the mesh which is modified by the AMR method.

[Rob]

Depending on how you set the source you may find the refinement just adds the source into a smaller volume (cell) which means the temperature gradients are still not resolved unless you further refine the mesh which could again add the source into a smaller cell etc.....  

How often are you updating the source positions? An alternative option could be to split a cell(s) out after meshing and then add the source to the new cell zone. You can adapt in that region without changing the source as volume as it's a zone rather than a cell. Have a look at mesh separation, and you can do that using a register (in the adaption tools).