Page 4 of 4
Posted: Mon Jan 13, 2020 7:47 am
that's what i said twice ... let him start!
Posted: Sun Feb 09, 2020 10:31 pm
This is Chris Rackauckas, developer of DifferentialEquations.jl and author of StochasticLifestyle.com. Wordpress stats led me over to this community and I found there's some fun discussion. Let me know what your community would need out of differential equation solvers and modeling tools to best support your activities. We'll soon be releasing a bunch of blog posts on PDE-constrained optimization and automated learning of partial differential equations with universal differential equations (showing more of the sparsity techniques to go beyond https://arxiv.org/abs/2001.04385
). I'm not much of a CAD guy but if you let me know what you guys need we can work it into our development roadmap and research! Indeed, a lot of our DAE work started because Hilding Elmqvist (inventor of Modelica) transitioned fully to Julia (Modia.jl) and he had a lot of questions to support tooling for his next gen Modelica. If you let us know what FreeCAD needs, we can work it into the Julia Lab's research and our next round of GSoCs
Posted: Mon Feb 10, 2020 1:44 am
ChrisRackauckas wrote: ↑
Sun Feb 09, 2020 10:31 pm
...If you let us know what FreeCAD needs, we can work it into the Julia Lab's research and our next round of GSoCs
You can see previous posts about Julia topics.
preCICE, DifferentialEquations.jl, DiffEqFlux.jl, SUNDIALS, CasADi, Modia
As you can see, those topics have received very few replies. It's not that we aren't interested in Modelica and efficiently solving mechanical and optimization problems, the issue is that currently there is no concrete implementation done in this aspect. Saso has posted this topic of Modelica and Julia more or less to spark interest, but so far nobody in our community has written the code to interface our CAD software with those tools. We need somebody to do this first, and then we can see what else can be done.
How can we connect FreeCAD with Modelica and Julia libraries? If you tell us how, then we can advance this further.
What does FreeCAD need? We need to know how to start because currently we have done nothing.
Posted: Mon Feb 10, 2020 9:32 am
ChrisRackauckas wrote: ↑
Sun Feb 09, 2020 10:31 pm
This is Chris Rackauckas, developer of DifferentialEquations.jl and author of StochasticLifestyle.com. Wordpress stats led me over to this community...
Chris, Wow! So good to see you here
those wordpress stats were probably mostly me reading and re reading all of your posts in the past weeks
It is true that both Modelica and OpenModelica projects are adding support for Julia (or maybe even moving to it), as said before Modia.jl and here are some slides from a recent OpenModelica presentation https://www.openmodelica.org/images/M_i ... mpiler.pdf
But for FreeCAD the potential for collaboration is IMO much bigger, on many different levels and I would even not put Modelica (and related/similar development to it) on the first place, so it would probably also be better to continue this discussion under the Julia topic https://forum.freecadweb.org/viewtopic.php?t=42400
Posted: Sat Jul 18, 2020 8:29 pm
Really nice to see this topic discussed!
Being mainly experienced in CAE for mechanical, control, and some hydraulics (OpenModelica, Adams, Simulink, "raw python"), I've read this with great interest.
I don't want to limit the discussion, but my take is that if such a thing has a chance it would need to "pick the low hanging fruits" first.
Since FreeCAD is first and foremost the tool to take care of geometric definition of systems (i.e. mechanical and dimensional properties), a primary focus on that would probably yield the fastest results. In that respect I feel that @looo seems on to something with the topic "node editor".
When modelling a 3D mechanical system in OpenModelica (i.e. chiefly using the Mechanics libraries Multibody, Translational, Rotational) the main benefit of FreeCAD would be to define coordinates, body types, joint types. In fact, especially bodies with non-trivial geometries are almost impossible to model in OpenModelica directly, so FreeCAD would be a big win there, even if it were only in a simplified manner with lumped masses and inertias for a start. Then of course animation of the results is what we are doing it for, but this should be "the least difficult", since the difficult topic of kinematics would be taken care of entirely by OpenModelica (the simulation result) in this case, so you would "just" need to force result coordinates and attitudes on the single CAD parts for every time step during animation.
Another topic (which was mentioned here as well) could be for geometrical properties of hydraulic piping, but there I suppose it will be more of a pre-processor use case.
Then again, for multi-body problems the approach @josegegas has taken with MBDyn seems to have taken a great head start, and come to think of it the things MBDyn needs as an input are quite comparable to the things OpenModelica would need, so synergies would be promising there I suspect. I have been wanting to find the time to take his toolbox for a spin...
Posted: Wed Jul 22, 2020 9:44 pm
I'm sure developers like adamLange would be interested too, his n-axis inverse kinematics solver in modelica, he has been using it to generate FreeCAD Path G-code for his 5 axis machine...https://forum.freecadweb.org/viewtopic.php?f=15&t=43224
He today released a very detailed step by step tutorial from FreeCAD to Modelica and back, showing that at the moment the process is still very manual, but clearly showing a potential for some integration.
Posted: Sat Oct 24, 2020 6:58 pm