von Mises stress/ Worst Case results -Simulation 2009

[Silum]

Using SW 2009 with Simulation Pro I get some very strange results.

I am modelling a very simple bicycle frame under loads. I have designed it using weldment structural members and meshed using beam meshing. My results have been fairly strange.

The default stress plot is STRMAX- Worst Case stress. This shows an accurate/expected reading of around 40MPa under the load I am subjecting it to. But when I list the results and get the von Mises stress table, my maximum vM stress is a measly 900N/m^2 or so.

My question is... whats wrong? Why such the massive difference between max worst case versus max von Mises?

A second question is, is there any way to refine beam meshing? I only have the option to 'create mesh' with an all beam component. Cheers for any help,

Robert

 

[kellnerp]

Can't you apply mesh refinement to a beam mesh?

How many elements are there on each section of the weldment?

TOP
CSWP
BSSE

http://www.engtran.com

http://www.niswug.org

"Node news is good news."

 

[Silum]

A whopping 250 elements....total. About 40 per beam. Now I know that beam meshing allows for much less elements compared to say solid meshing, but how do I actually have control over the number? I see no place to refine or even adjust the size of elements when doing beam meshing...what am I missing?

How do you have confidence in the results if you can't compare similar studies of different mesh sizes? Perhaps i'm missing something fundamental about beam meshing and its accuracy?

 

[kellnerp]

I wouldn't worry as much about beams as say tets or bricks unless there is a huge stress gradient in the beam or shear is a huge percentage of the stress which is not likely. It is a very simple element formulation and hard to screw up. It is the formulation that I and any other degreed analyst had to program in school.

Post a screen shot of the mesh menu. I don't have it in front of me right now.

TOP
CSWP
BSSE

http://www.engtran.com

http://www.niswug.org

"Node news is good news."

 

[Silum]

Essentially their is no menu.

See the screenshot, as soon as I click create mesh, it creates the mesh without further input or adjustment. It only does this on models with all beam elements.

That isn't the real problem though (although I would like more elements), the main problem is the massive difference between the plotted Worst Case stress of around 40MPa, and the tiny von Mises stress of a measly 900Pa at the same node.

Their doesn't seem to be any decent way of producing the results either for an all beam design. If I click 'Create Report', i get a warning that 'Reports cannot be generated for pure beam mesh studies'. I can only assume these problems are due to solidworks being unable to work with beam components properly. What's the reasoning behind not generating a report, just because a design uses beams doesn't make the data/results any less relevant. I'm left with 2 options, a graphical plot, and a inaccurate von Mises stress table.

 

[kellnerp]

I guess that's why I like running beam models in Cosmos/M which you should have. It gives a nice plot of the moments, shear, top fibre stress, axial stress, etc.

FEMAP has an even nicer presentation in which it shows a virtual outline of the beam and a cock's comb presentation of output values.

TOP
CSWP
BSSE

http://www.engtran.com

http://www.niswug.org

"Node news is good news."

 

[Silum]

I've had trouble getting CosmosM running stability but I will definately give it another shot if you think that is the key. Cheers,

Robert

 

[kellnerp]

Wow, for me Cosmos/M has been rock solid for years.

A little trick is to edit your .ses file so you can make a parametric model. Save it as a .geo file.

TOP
CSWP
BSSE

http://www.engtran.com

http://www.niswug.org

"Node news is good news."