I haven't seen any proper and direct discussions regarding this, so here's my attempt:
I have spent two years developing a complete 2D laser keyhole welding model in COMSOL. Although I have made significant progress, there are still a few forces that need to be added. I would like to express my gratitude to COMSOL for their interface, which displays equations while solving and provides the freedom to incorporate additional equations as needed. Building a 3D model in COMSOL is now a relatively quick task, but completing the simulation itself can still take around a month. While there are ways to reduce the runtime, it is not something that can be accomplished overnight or even within a few weeks.
Now, let's shift the focus to Flow3D. I have been using Flow3D for some time now, and it seems somewhat like a black box to me. In Flow3D, all the necessary physics are pre-implemented in the Flow3D-Weld module, enabling the creation of a 3D laser keyhole welding model with relative ease. Of course, mastering any software requires skill and expertise, particularly when it comes to numerical aspects. However, in Flow3D, adding physics to your model is no longer as a challenging task as it was in COMSOL.
For instance, applying Recoil pressure in Flow3D is as simple as selecting it from a drop-down menu, whereas, in COMSOL, you need to manually input the entire equation. Additionally, it is crucial to understand the underlying physics to appropriately apply this force, regardless of the software being used. Flow3D significantly simplifies this process.
I appreciate the convenience that Flow3D offers in terms of pre-implemented physics and user-friendliness compared to COMSOL. It is important to note that each software has its strengths and complexities, and becoming proficient in utilizing them efficiently requires experience and expertise.
I have been following your questions on laser welding using comsol multi physics software. I am also working in similar lines. I completed conduction heat transfer successfully. Then I started working with marangoni effect, but I couldn't able to figure out the solution for it. I got struct in the application of recoil pressure in key hole laser welding. Can you please help me?
I think that Flow 3D (AM Module) is a very promising solution to model the SLM.
Lots of simulations on the official Flow 3D YouTube channel concern keyhole formation (see video about powder bed fusion, PBF).
It provides lots of functionalities and it appears to be much easier than Comsol, at least for this kind of problem.
In addition, let's think that Flow 3D has been specifically used by different research groups to simulate AM techniques and other free-surface processes (where the Flow 3D formulation, based on cartesian grid works very well).
This is also the case for keyhole formation in PBF.
Another advantage: the cartesian mesh allows to impose the movement to the head (laser or extrusion head) which is very important when dealing with more complex simulations.
Finally, there are some support groups in the word that allow you to add specific add-on(s) to the base software, which is very imporant when considering additional physics.
About my experience with Comsol: In general, I tried to simulate other AM processes (especially FDM) in Comsol and it turned to be very difficult... I used Fluent to solve very simple deposition problems, but if I think to PBF, where lots of physics are involved the implementations through Comsol is challenging.
However, in Comsol you can implement whetever you want through the PDE module, by inserting the right coefficients in the exact PDE model.
Lot's of much simpler problems can be modeled through Comsol, and that the software is simply great when it comes to Multiphysics, but I think that a lot of work has to be made to allow users to use it for AM modeling, especially in PBF.
If you want a quick answer and result to your problem better to use Flow 3D but if you want to know the physics underlying the problem being solved and build your own model then COMSOL is the best option.