Hi i have the following Problem.
I know, there is a multipipe feature but the generated Geometry is not that what is need. Fillet also does not work. The resulting geometry will differ as well.
Extruding 3 Cylinders along 3 Paths, joining them with boolean union and filling the hole with a “pyramid” is the way to go here. This worked perfectly fine in a paramatric modeller. But in Rhino the boolean union is not possible. I tried it with surfaces and with solids. Everytime i get message that union is not possible.
I added an example file. The clump of solids is from rhino and the surface body imported from solidworks.
Would be nice if anyone has an idea how i can weld those bodies in rhino.
1 Like
WOrking on it. If i click upload my pdf reader starts an update and blocks the upload… Whatever…
Our IT departement has the magical talent to turn every high performace workstation into a banana.
1 Like
Try drag and drop the file into a post
1 Like
If it is that easy 
.
Copied via USB to my Smartphone and uploaded…
BooleanUnionTest.3dm (476.3 KB)
1 Like
Try the boolean union again but this time just the curved segments of the pipes. Coincident segments are most likely causing the issue here.
1 Like
Yeah. Thats right. But then it will fail in the next step.
My assumption: Rhino is not the right tool for the job?
1 Like
I split the unioned arc segment pipes and the straight pipes with a surface in the XZ-plane.
Exploded everything and split the curved surface segments with isocurves, then joinedd everything and capped it.
Another approach to model this T-pipe is by extruding a rectangle with concave corners. The fillets need to be done one by one. Done in one go, Rhino seems to dislike some coincidences.
BooleanUnionTest.3dm (1.2 MB)
1 Like
Thank you. Thats a lot of work. But good to see that it worked 
But unfortually that will be very hard to put in a script for a parametric FEA model.
at least my job is safe modelling this stuff 
Edit: your second approach is great. very good idea! That is also pretty good to mesh!
So if you’re meshing the whole thing for FEA, why isn’t multipipe good enough?
The deviation between both geometries is at maximum at the point where the maximung stress. So the inaccuraccies will stack up at this point. This is a tunnel and i will build brick elements with a python script arround this shell.
Red is the multipipe approch. For large Elements it is okay. But for smaller elements it is not the well.
Interesting. Post some images of the process.
I using a plugin called griddle from itasca. Basicly i offeset the mesh into multiple shells explode the meshs into single faces. extrude them to tiny quaders and mesh every single quader with 1 by 1 by 1 element… Unfortunally you cant open those files in rhino nor you can see the result. praying is the way to go
I’m only doing the model work and try to reduce the workload on our staff that is calculating. I dislike the calculating part 
This might also work for you.
