You may have to refine the meshes and correct some tie constraints between the welds (especially the bottom ones) and plates (keeping in mind the surface assignment rules I’ve mentioned here). But more importantly, the loads should be corrected - moments can’t be applied directly to solid models - you have to apply them via the reference point of a rigid body constraint: https://www.youtube.com/watch?v=2upPxL45OZg
Also, your Surface Traction load of 9810 N is applied to the whole volume of the top plate. If you want to account for gravity, you should use the dedicated load for that.
I sent you a pm to discuss the details of your model and help you fix it step by step.
PrePoMax capable to spit surface even limited by two points. It can be useful when surface partition by extra edge or line is not supported in CAD. Alternatively, each shell parts left separated and connecting by tie constraint later.
below an example in principles,
compound (warning symbol in trees appears)
split surface by two points, need three segment to define (no warning symbol)
continuous mesh results
without compound mesh is separated and needed such as tie constraint later
If such two point are available then sure. But in most cases surface partitioning has to be done in CAD software. This is a standard part of geometry preparation for FEM. FreeCAD can easily partition surfaces with sketches thanks to its Boolean fragments tool. But the partition has to extend between the edges, splitting the face into subregions (it can’t be contained within the face because it won’t be recognized and won’t work for partitioning - as shown in Matej’s reply above). I also showed such an example here: Orthogonal shell mesh model - #6 by FEAnalyst
it seems partition is supported in modeling but the software or CAD kernel internally doing cleanup by removing extra edge when exporting to standard files, similar condition can be found at another CAD also.
I think the OP was talking about importing sketches themselves. We already clarified partitioning strategies with the OP. As I said above, partition lines may not be preserved if they don’t cut the entire face. Also, merging surfaces in FreeCAD is a bit tricky. Otherwise, all features created in FreeCAD should be shown in PrePoMax. It’s even using the same geometric kernel - OCC.
this is doing cleanup, probably.
Split Line should work, I remember that Matej mentioned using it. Unfortunately, I don’t have access to SolidWorks anymore. I can only check how it works in CATIA (especially v6 - 3DX).
Excuse me! Is this a job or some kind of research?
Because the weld seams are impossible! With a throat height of just 2 mm, you can’t achieve penetration. The joint won’t hold! It’s a waste of time checking for unworkable welds.
Before you think certain things, talk to a welder!
Perhaps the intention was actually to model the welds as convex. Concave welds seem rather uncommon. And it’s hard to mesh them, too.
More than in modeling, in reality. Perhaps with laser welding, but I don’t think that’s the case. It’s used for stainless steels. Simple rule: 0.7 times the minimum plate thickness. Welding with currents of a few amps to achieve such a thin weld throath results in minimal penetration into the base material. It’s not welding, but bonding!
I understand the need to model welds and bolts, but what’s the point if the reality is different?
Yeah, it might be better to exclude weld geometries from this global model since it won’t provide reasonable results for them, and fatigue is not considered. They only distort the meshes at the connections with this concave geometry.
We are discussing on two different levels!
But modeling must follow reality
Both aspects are relevant and related here - those welds are not realistic and cause issues with the connections in this model so they should be corrected or removed.
