I am not 100% clear on what you are trying to do. I work with weldment everyday so I am hoping I can help you.
Are your sketch lines for the HSS braces all on the same plane and do the center lines of the sketch end at the same point on the lines for the I beam? Also are your sketches fully defined?
Where the HSS braces are to be trimmed to the I beam and then to each other it is best to create the I beam first an then create the braces. Then it is simply having the trim done correctly to the way you want it. Creating the braces prior to the beam could cause the beam to trim itself to the braces.
When you create the trim function do not include the I beam as part of the trimmed list but do select it as the part to be trimmed to.
Thanks for your reply Gary,
This is a 2d sketch / structure. All end points, coincident lines etc have been constrained, but when I redrawed that short line and that solid I-beam section reappeared as a beam. Now model behaves as expected, no loose brace ends, no haunting solid stubs.
Next I tried your method to draw a single line for each I beam and then trim HSS braces to that surface (body). As side bonus it gives an oppoturinity to dimension convergence point value "e", which is not so easily available in my current "splitted skecth". We'd like to go with gapped structure (welding..).
The outcome was awesome, just what I expected:
For now my questions have been addressed / resolved by themselves.
Even question #4: apparently letter "i" warns about an non-trimmed stub of brace profile, which was hiding between I-beam flanges?
Is it so that with Beam1D I can have only 6DOF or 3DOF constraints, no springs, no elastic supports?
I'm using version 2014 (premium) for this customer.
i do not believe you can connect springs to joints if that is what you're asking
One more thing: what would you recommend if I need to model this whole building and add some stabilizing braces between vertical columns, can I make an assembly and use those 2d frames there (pattern, spacing 5000mm), or should I stay in one *.sldprt and make all profiles there? My first intuition says that I should stay in one part, since it feels more robust and I'll most likely avoid a bunch of connetion issues when compared to an assembly?
part is easier
but beams can work between assemblies
yep, apparently you're correct with this one: When I add at least one non-beam body to that model, Connetions menu activates and becomes selectable.
This leads to my next issue: My next concern is how to connect corrucated panels (mid ceiling) to horizontal I-beam, so that all frames are connected to each other. I tried to create a contact set between panel and I-beams, but none of Touching faces / Non-.. / Find Shell edge- .. options worked.
i'm not sure what the problem is
the interface doesn't let you do it
or you're not sure what to select
similar to your previous question, what fidelity are you looking for?
you should be able to choose joint > surface of the shell or beam > surface of the shell. but if you're trying to choose edges of the shell, that may not make sense to the solver because it doesn't know that level of detail about your beam. the beam is just 2d.
best to do some testing on some sample parts to see what the different connection behaviors look like and compare to solid > solid so you know which one is "best"
Thank you for all your help, Jared.
My VAR quoted SWX Database as follows: "
Is it possible to establish no penetration contact between a beam and a
shell or solid?
No. Beams get meshed as one dimensional elements which only have the
dimension of length, so it is not possible to accurately model contact
with the actual cross sectional beam geometry using beam elements.
For cases like this, most users would model the beam using shells
(using the midsurface feature in SolidWorks is best since this will give
the results closest to the actual geometry). Then, it is possible to set
up contact sets between the shell or solid and the face of the beam
which has been meshed using shells."
So this is a show stopper, but we more or less likely will also need to take a closer look at some bolted joints.
These reasons combined will force me to start from a scratch with surface models.
remember that you can mix and match, that is the most efficient strategy
is the joint the important part or the overall behavior of the structure?
if the former, i would go with solids/shells combo
if the latter, beams
the reason being that with beams you have really the joints and connections at the joints and the beams, not the fidelity of the connections between the flanges...etc which i think you are trying to achieve
I'll consult my senior colleague, and we'll see if we manage to avoid going too deep to joint details.
If we must, I'll follow your proposal and go with shell model.
More likely there will be some bolted joints and we must model at least them in more detail.
Thanks for this Jared,