ALGOR MESHING OF LARGE THIN WALLED MEMBERS

[mpotts]

HAS ANYONE HAD EXPERIENCE USING ALGOR TO ANALYZE LARGE WELDED STRUCTURES MADE UP
OF THIN WALLED MEMBERS. I AM CURRENTLY TRYING TO MESH A STRUCTURE THAT IS MADE
UP OF 2X2X1/8W TUBING. THE OVERALL WELDMENT IS 30 FEET LONG X 5 FT WIDE.
USING SOLID A SOLID MESH THE ANALYSIS KEEPS STALLING OUT. BECAUSE OF THE THIN
WALLS I HAVE STARTED MY MESH SIZE OUT AT 1/8". IS THERE A BETTER
WAY TO MESH A STRUCTURE OF THIS TYPE?

 

[rb1957]

phew, solids seems over-kill for this ...

why not beam elements (sounds like you have a strcuture made up of tube elements, yes?)

and, please, click off the "caps lock" ...

Quando Omni Flunkus Moritati

 

[mpotts]

Sorry for the CAPS, always reads like someone is yelling. When dealing with Algor that is usually what I feel like doing.
It is a big truss structure, you are correct. I am using Catia V5 as my cad system and I stp the model into algor. I would
really not like to redraw the structure in algor.

 

[rb1957]

well i guess it's pick your pain ...

easy to import the CATIA model, easy to mesh with solid, hard for the FEA to crunch all those numbers,
easy to import the CATIA model, less easy to mesh the surfaces, still a lot of number for the FEA,
harder to create a truss (stick) model (could you extract CLs from the CATIA model ?), much easier to mesh and FEA

Quando Omni Flunkus Moritati

 

[structSU10]

How many elements are in your model? What kind of PC are you running? keep a close eye on your RAM usage during a run, if it hits 100% then stalls out, you may have a memory issue. You also want to make sure you have plenty of space on the hard drive the model is stored on. My IT department had to reconfigure my systems paged memory to help run large models.

 

[mpotts]

Thanks guys, I appreciate the advice. I have broken the structure down into more manageable pieces, 7 total, and imported these
into algor. So far the first 3 smaller segments have meshed. Not quickly, but they have been successful. I plan to do the same
for the larger segments. I am hoping once I know that they can be meshed individually I will try to bring them in as one assembly
and try to mesh the whole thing.
It seems my biggest issue is how unforgiving algor seems to be for the slightest non tangency. Either 2 surfaces barely (.0005")
off from one another or curved surfaces running tangent to planer surfaces. I have removed all radius corners from my tubing in
order to avoid some of those errors.
I have been watching my RAM use and I moved the stress folder to my local hard drive. This helped with the processing speed quite
a bit.