I’m working on Conjugate Heat Transfer (CHT) problems and would like to discuss possible approaches. As far as I know, there are two main options:
1- Use the chtMultiRegion solvers available within OpenFOAM.
2- Couple OpenFOAM with external solvers (e.g., CalculiX) through preCICE.
For example, consider the case of simulating an industrial oven with numerous solid components and internal cavities where thermal radiation must also be taken into account.
How would you approach this kind of CHT problem and why? What has your experience been like using either the built-in chtMultiRegion capabilities or external solver coupling?
I’m open to all kinds of responses, but I’m particularly interested in hearing from those who have practical experience using both methods. If you have strong opinions or insights about one approach versus the other, please share your thoughts!
Hi,
I don’t have an extensive experience in doing CHT, but I tried chtMultiRegionFoam and preCICE (only in OpenFOAM-OpenFOAM coupling). Some results are here: BFS with preCICE. My feeling is that chtMultiRegionFoam can be faster but lacks versatility and I don’t think it is capable of capturing a transient well. On the other hand, configuring preCICE to have fast convergence can be tricky (at least it has been for me, as I have more experience with FSI). If you want, for example, to recover thermal stresses, coupling with CalculiX is a better way to perform those computations.
I hope it helps a bit.
Claudio
Thank you for sharing your experience with chtMultiRegionFoam and preCICE. Your insights, especially from your paper, are exactly what I’ve been looking for. (Haven’t read all yet but I intend to.)
Similar to you, I also have more experience with CalculiX-OF-preCICE in FSI and chtMultiRegionFoam in OpenFOAM, but I haven’t yet tried preCICE for CHT problems. I’m working on finding a streamlined method to apply any CHT problem with industrial geometries involving numerous solid and fluid regions.
With chtMultiRegionFoam, meshing complex geometries often fails when splitting into regions. A coupled approach where I mesh each region independently could simplify this, allowing structured and SHM meshes as needed with non-conformal interfaces. However, configuring the coupling correctly across multiple regions remains a challenge.
Hearing about your experience, I’m now considering trying preCICE-OF-CalculiX for CHT myself. Do you have any tutorial or material recommendations to help me get started? (I’m thinking of starting with this heat exchanger tutorial.)
Dear @Umut
I agree with what you say. I haven’t gone very deep in exploring CHT, the case I was considering is geometrically straightforward: a single planar surface, I didn’t have multiple regions or non-conforming meshes. I suggest starting with the tutorial you mentioned and playing around with coupling schemes. It will give you an idea of the complexity of the coupling.
Claudio