test with algor express 2010

[mir]

Hi.
I did a test with algor ... a simple section hollow stainless steel aisi 304 alloy on one side with 1000 n distributed (I do not understand if it is a distributed load or many loads from 1000 newton!).

I agilely get the result in the annexed pdf ... and I find the critical conditions where +o- expected (some millimeter inside the beam perhaps by question of approximation of the binding surface).

everything very simple and quick... but after I try to overload to see what happens to break.... and I find that you do not "break" and deform the beam infinitely.... I do not have experience with the fem though if you ask me the type of material possible that does not make considerations on the break sigma or at least on the yield sigma?

Thank you.
♪

 

[cacciatorino]

Hi.
I did a test with algor ... a simple section hollow stainless steel aisi 304 alloy on one side with 1000 n distributed (I do not understand if it is a distributed load or many loads from 1000 newton!).

I agilely get the result in the annexed pdf ... and I find the critical conditions where +o- expected (some millimeter inside the beam perhaps by question of approximation of the binding surface).

everything very simple and quick... but after I try to overload to see what happens to break.... and I find that you do not "break" and deform the beam infinitely.... I do not have experience with the fem though if you ask me the type of material possible that does not make considerations on the break sigma or at least on the yield sigma?

Thank you.
♪

You have to go through non-linear analysis to do what you're asking. remaining in the linear field, the software takes only two values of the material: elastic and coeff. module of polka dot, considering the material as linear elastic infinity.
If, on the other hand, you want to sow yield and plasticization, or whatever is not modelable as ;
== sync, corrected ==
Then you have to move to the non-linear, with its increases of difficulty and costs.

 

[mir]

but can I use the sigma obtained as pre-dimensioning (confronting with the yielding/breaking sigma)?

 

[cacciatorino]

but can I use the sigma obtained as pre-dimensioning (confronting with the yielding/breaking sigma)?

Yes and no...
you will have points of your model where the calculated sigma is greater than that of yielding, and all right. But you have to understand whether behavior is real or due to the concentration of tension in a corner, which in reality would be solved in a small localized plasticization and therefore the effort would be redistributed on adjacent fibers, while instead in linear numerical analysis produces peaks of tension.

If you have wide areas where the sigma is beyond the safety values you have to shelter the project, if instead these areas are very small areas in correspondance of edges, it should be evaluated by case. considers that also the way of making mesh strongly influences calculations.

 

[stef_design]

If you have wide areas where the sigma is beyond safety values you have to shelter the project....

ciao caccia, for "safety values"Do you mean the yielding and breaking values of the material? ! ?
Hi.

 

[cacciatorino]

ciao caccia, per "safety values"Do you mean the yielding and breaking values of the material? ! ?
Hi.

depends on what you're doing.... if an elevator, a clothes rack, a train cart, a nightmare... .

 

[stef_design]

depends on what you're doing.... if an elevator, a clothes rack, a train cart, a nightmare... .

Whatever the application you will always have to relate the value of von mises found with the disregard of the material.. ! ?

 

[cacciatorino]

Whatever the application you will always have to relate the value of von mises found with the disregard of the material.. ! ?

Yeah, it's one of the criteria. but the relationship between the acceptable von mises and the yielding sigma depends on the degree of safety required for the manufacture.

But as I said if you do the linear analysis, the value of von mises can also be misleading: you will always find singularities where the value will be very high because of a mesh not dense enough in a edge for example, but in real life this wouldn't happen because you would always have some localized plasticization phenomena that would bring down the punctual tension, things that you can't really take. Now I'm not a business fem, but from that little job I've done and seen it seems to me that things work like this.