today i have been having problems getting the correct deflection on a simple Universal column (203UC46.3).
The stresses are spot on at 8.3 MPa in the middle of the beam. I have applied a UDL of 840.3 N/m. The beam is 5.95 m long.
I have midplane meshed this. The deflection however is wrong at 1 481 327 mm! It is about 1 million times too much. Hand calcs give 1.5 mm.
I have made sure my model units = my display units as I had a problem earlier this week with shrink fits.
Has anyone got some insight on this?
I opened up your fem file, set the plate thickness to be the midplane mesh thickness and material as Steel(ASTM-A36). I ran the analysis and got the maximum deflection as 1.5mm (which is what you expected). In short, I did not find any problems with your settings(unless you specified too thin a thickness or too weak a material). What version of the software are you using?
Perhaps you might want to update to service pack 2
I spotted the error thanks to your tip. I used the default alloy steel as a quick check. It has a youngs modulus of 0.205 N/mm^2. It should be a million times bigger, hence my million times error! This should be changed by you guys especially because its a default!
I'm using sp2 by the way.
Deflections are now correct. Thanks! By the way, If I select "use mid plane mesh thickness" does this pick up the different thickness in the web and the flange?
Yes, the Use mid-plane mesh thickness picks up the different thicknesses of plates provided that they are different surfaces. One surface can only have one thickness.
Either my memory or the Help documentation is wrong. But I thought the midplane mesh actually used the calculated thickness for each element, not just the same thickness for all elements on the same surface. So if you had a rolled beam where the flange thickness changed, you would see that "taper" by the varying thickness in each element. ("Results Contours > Other Results > Element Properties > Thickness of plate" in the Results enviornment.) Of course, the change in thickness would be a step change from element to element, not a smooth change.
This capability, if it exists, may be different for linear stress, MES, and thermal.
Simulation Mechanical user since Dec 1997
Thanks both of you for your info. I've been on a good learing curve with ASM these last few weeks. I've just installed sp2 on my home machine. I'm now trying to mesh a much larger group of assembled beams. I'm getting quite consistent crashes where the Autodesk error report comes up, sends a report and then it continues to mesh the model.
Has anyone seen this error?