[Dreoasteh]

I wanted to start a threat to discuss about the present state of conjugate heat transfer solvers in OpenFOAM.

Recently I have been trying to carry out a research project based on cht and in reviewing literature I have found this:http://www.tfd.chalmers.se/~hani/kur...eport_Turo.pdf
Here, the author clearly states that cht solvers in OpenFOAM are considerably worse than other commercial options in terms of convergence speed.

Furthermore, in the OpenFOAM.org funding article (https://openfoam.org/news/funding-2017/) they explicitly mention that "Heat transfer, including conjugate heat transfer (CHT)" require "significant code refactoring and/or rewriting" and will be addressed in upcoming versions.

With this said, do you believe that state of cht codes are good enough to invest time and effort into serious research in conjugate heat transfer using OpenFOAM?

Thank you for your opinions.

[juho]

Originally Posted by Dreoasteh

Furthermore, in the OpenFOAM.org funding article (https://openfoam.org/news/funding-2017/) they explicitly mention that "Heat transfer, including conjugate heat transfer (CHT)" require "significant code refactoring and/or rewriting" and will be addressed in upcoming versions.

With this said, do you believe that state of cht codes are good enough to invest time and effort into serious research in conjugate heat transfer using OpenFOAM?

If you have the opportunity invest serious time and effort to conjugate heat transfer research, you should talk to the OpenFOAM Foundation to see if your work could be part of the "significant code refactoring and/or rewriting" needed to address the issues that affect your work. Things won't advance unless someone puts in the work.

[Moopie]

Any views from users who have compared the latest different OpenFOAM branches?

My beginner's take on the state of things so far, which could be outdated:

• .com has implemented a coupled monolithic approach in 2022 (setting useImplicit = true; for the coupled temperature boundary fields, but currently with open bugs https://develop.openfoam.com/Develop.../-/issues/2536)
• .org has energy-only loops, also implemented in 2022 (nEcorrectors, https://github.com/OpenFOAM/OpenFOAM...cd3536ff31b254)
• foam-extend has conjugateHeat*Foam which is also a coupled monolithic approach (compared in 2017 in this report https://www.tfd.chalmers.se/~hani/ku...eport_Turo.pdf). I've read posts that this only supports one fluid and one solid region but not sure if this has now been extended to arbitrary numbers of regions?

Any recommendations for accelerating convergence in CHT simulations? My cases are very much dominated by inter-region coupling. I've only been testing with .com v2212 so far, able to get a reasonable converged solution but with far greater number of iterations than expected since I'm not able yet implement useImplicit = true;. Even by 120,000 iterations ~10 hours later it still has not fully converged with ~1 million cells, 3% of which is from 1 fluid region, the rest being from 3 solid regions. With the implicit setting on, I seem to get unphysical results in single core and parallel simulations - somewhat hesitant to spend too much time diagnosing given the open unresolved issue.