Post Plasticity Analysis Failure

[captainhog]

I am currently trying to match the results that were obtained in an Instron machine within COSMOS Premium 2008 and have had some success. I am able to match the force necessary to reach the plasticity point within 5.7% but the data after that is not correct (See attached image)

http://files.engineering.com/getfil...4ca5-84b8-4acccdacd26e&file=Analysis_Plot.jpg

The study is a nonlinear static study and to mimmick the instron compression machine I have applied a reference geometry restraint that compresses the model constantly. I then used the resultant force result to extract the force for every time step and plotted this in excel.

Is there a way to obtain the correct force after the model has reached the plasticity point? The type of material is 6061-T6 aluminum and the data for this was obtained direct from cosmos material library. I would appreciate any insight into this issue because I have exhausted all other resources.

Thanks in advance.

 

[GregLocock]

What shape is the test item? It looks to me like you have a softening spring - either the part is cracking, or it has deformed so much that the moments have increased.

I'm assuming your model is non linear for both material AND geometry?

Incidentally your correlation even in the linear range is totally unnacceptable, I'd sort that out first.

Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[captainhog]

Greg,

Thanks for the reply and good call on the softening spring. It is basically an energy absorption link that undergoes a force which causes it to go to failure. Your assumption about the deformation increasing to the point where the moments have increased.

I have not been able to uncover what the error is in the linear range but I will attempt to fix that first. Within what range should I expect COSMOS to be able to predict the experimental results?

If I am able to sort this out, is there a way I can account for the increasing moment?

Thanks Again.

 

[captainhog]

Greg,

I have been trying to fix the mismatch in the linearity portion of the graph and had a couple questions for you.

-I am a little confused when you asked is the geometry non linear. I set up the study to be non-linear static and I have checked the large displacement option under properties. Is this accomplishing non linear geometry?

-For the material, I am using plasticity-von mises directly from the cosmos material library. Do I need to change anything else or am I missing something completely?

 

[GregLocock]

"-I am a little confused when you asked is the geometry non linear. I set up the study to be non-linear static and I have checked the large displacement option under properties. Is this accomplishing non linear geometry?"

Yes. I've never done non linear FEA in a conventional FEA package, so I don't know exactly how commercial programs control these things.

"-For the material, I am using plasticity-von mises directly from the cosmos material library. Do I need to change anything else or am I missing something completely? "

Well, that I don't know.

So far as accuracy, your predicted stiffness (the gradient of the linear bit) in the linear range should be better than within 10%, and if the boundary conditions are straightforward I'd have thought better than that should be easy.

Might be worth getting your material tested.





Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[ktop]

The first thing I am going to ask is what is your plasticity model? Kinematic, Isotropic or from material curve.

It looks like the post yield behavior would suggest a material curve be input as kinematic and isotropic models make a lot of assumptions.

Are the locations of yield in your model matching those in your test article?

Are you using a follower force setting in your boundary conditions or is your FEA force always pointing the same direction?

I have seen that small jog in the elastic portion at the onset of loading and always attributed it to slack in the test apparatus. However, the slope of the elastic portion is a tad less than that in the FEA.

KTOP