Multi-material Component - LPBF Tutorial Case?

[drlukeparry]

For our work in multi-material LPBF using our Aerosint system available at Nottingham, we are evaluating our long-term model choice for evaluating the melt-pool dynamics across these interfaces.

I wanted to check if there was an existing 3D example for multi-material LPBF test-case, or if not, whether it is trivial to simply extend the existing single-material lpbf_large tutorial case?

Presently, we are currently at a cross-roads, evaluating whether we develop further an existing in-house code built on Ansys Fluent and UDFs, which has similar capability but lacks adaptations to use multi-material and multi-laser capability.

Hopefully, you are able to provide an opinion on this,

Kind Wishes,
Luke Parry

[tomflint22]

Hi Luke,

yes it would be very straight forward to modify the LPBF test-cases to be multi-material. You would just have to modify the setfieldsdict with the volume fractions of the different compositions you were interested in and then use the multi-component solver.

We're just in a bit of a state of flux at the moment as we prepare to release Version 3 of our software with additional physics and improved ray-tracing algorithm (faster, better scaling). I'll leave this open, and one of us will look at making a multi-component tutorial for the multi-component solver once we get chance.

Thanks,

Tom

[drlukeparry]

Hi Tom,

Thank you for the quick response! Completely understand, and look forward to the future release. In the meantime, we can have a go at modifying this based on your suggestion, to get a feel of whether it will be sufficient for our research.

Generally, everything with laserbeamFoam looks good so far, and now runs on our HPC and Windows PCs.

[tomflint22]

That's great to hear! I'll re-open this as an issue as I think it would be god for us to show how to do this with laserbeamFoam.

Thanks Luke!

[drlukeparry]

Hi @tomflint22,

Firstly, we've been using the laserbeamFoam model and is starting to produce some nice result so far, and a viable alternative to Flow3D. The improvements to the codebase switching to OpenFOAM-ESI have been really great, so a big thank you to yourself, Prof. Cardiff and other contributors to this.

We are in the process of revisiting towards the multi-material LPBF model. The incompressible multicomponentLaserbeamFoam is not available anymore with the change to compressibleLaserbeamFoam solver. For LPBF, thew new model relies coupling of a compressible metal.vapour phase to the model to induce a pressure-recoil effect and limit the maximum surface temperature.

On initial tests, I am having difficulties with convergence and stability (Co = 0.5), with temperatures becoming negative under some situations. Runtime performance appears to be affected, but I have not confirmed this properly.

For LPBF_small_vapour, I have I tried without using metal.vapour phases in transportProperties, although this changes the behaviour of the model, leading to no thermal transport or creation of pressure.

I was wondering whether there could be option to retain the more simpler incompressible multi-phase solver, similar to the previous multicomponentLaserbeamFoam? Generally, the pressure recoil term is sufficient to account for the thermophysics we need to capture in LPBF.

pVap = 0.54*p0*Foam::exp(LatentHeatVap*Mm*((T - Tvap)/(R*T*Tvap)));

[tomflint22]

Hi @drlukeparry ,

ah I see. Yes that compressiblelaserbeamFoam solver is very numerically stiff actually. It's proving quite challenging to set up the cases there with this additional physics we have captured.

We have another solver that captures the volumetric dilation and mixing, just omits the compressibility in the bulk liquid and vapour phases. I think what I will do is add this to the repository. It's nearly as good as the compressible solver but a lot more robust.

[drlukeparry]

Hi @tomflint22

Thank you for your quick response. That's really good to hear there is an alternative solver that uses the incompressible behaviour.

I expect for some additive process (e.g. arc DED ) that accounts for compressibility, including propagation of shockwaves will be highly valuable to model the detailed behaviour of the melt pool interaction. For L-PBF, I expect including compressibility will not provide a huge advantage, especially when there are so many different phenomena involved. Most other challenges relate to solidification and mechanics within the consolidated material - a problem for another day 🙂

We are happy to test and provide you feedback for your alternative solver when this becomes available. We were very surprised how robust the immisicbleTwoPhase laserbeamFoam model has been so far. If this works, we will perform some multi-material prints to validate the track morphology, so it would be great to include you in any future publication.

Many thanks once again!
Luke

[drlukeparry]

Hi @tomflint22, did you have any plans to share the model that you had previously developed or was it still under further development and testing for the time being?