Error from the CalculiX adapter for CHT coupling with heat-flux

Hello everybody,

I just started to transfer a coupled heat-transfer and fsi simulation from precice2.5 to precice3.3 and ran into some trouble with the calculix adapter. It fails with the following error:

terminate called after throwing an instance of ‘fkyaml::v0_4_2::type_error’
what(): type_error: The target node is not of a container type. type=NULL_OBJECT
From older threads, I was able to narrow down that the error might be an issue from an empty read or write statement on one mesh, which was mentioned in 2020 by @Makis No support for missing write-data / read-data · Issue #35 · precice/calculix-adapter · GitHub and has been solved by @KyleDavisSA in this issue Update master: code quality, configuration file checks by MakisH · Pull Request #36 · precice/calculix-adapter · GitHub

However, it seems that there might be a new problem. I also tried the tutorial case “flow-over-heated-plate-two-meshes”, where also temperature and heat-flux are coupled and it fails with the same error. It seems to me, the problem is, that for temperature and heat-flux different meshes (nodes and faces) have to be coupled, each with either a read or a write statement. The tutorial “heat-exchanger”, which uses heat-transfer coefficients and several read and write statements starts without problems. Can it be an issue of the develop branch of the calculix adapter? Does anyone have an idea how to fix it? Did anyone encounter a similar problem?

For completeness: I am using precice3.3, calculix2.20 and the develop branch of the CalculiX adapter in an Ubuntu 24.04 surrounding in a docker container. Maybe it is also an issue of the newer Ubuntu version? On precice2.5 and Ubuntu20.04 my case was running fine.

Please let me know if any information is missing.

I am grateful for any hint or idea :slightly_smiling_face:

Hi @SarahDi,

I can reproduce the issue in the flow-over-heated-plate-two-meshes tutorial. I will have a look tomorrow and add it to our system tests as well to prevent such regressions in the future.

Hi Makis, thanks a lot for your fast response! Good to hear, that you could reproduce the error and are working on a solution, thanks! :slight_smile:

@SarahDi the problem is clear and the fix easy:

You can now pull develop again.

(sorry for the delay, I only now managed to find some time to look at it)

Thanks a lot for the fast fix, everything is working as expected now :slight_smile: