Parts stressed beyong the material UTS should break

[bafuller]

In SolidWorks Simulation Premium 2009 sp 4.1 I am trying to analyse the effect of frame structures being subjected to impact by a large mass. I am using bi-linear elasto-plastic materials and getting good results but in some parts of the structure maximum stresses exceed the UTS, in the anylisis these members are still contributing to the strength of the structure whereas in reality they should break and take no further part. Does anyone know of a workable solution to achieve this?

 

[kellnerp]

Buy an FEA program with an explicit solver like ABAQUS or ADINA.

With a solver such as is found in SW (implicit solver) elements cannot detach from one another. All you will get is continued deformation of the element when you reach plasticity.

As an analyst it is up to you do determine when a portion of the model has failed based on plastic strain. The model will not detach.

TOP
CSWP, BSSE

http://www.engtran.com

http://www.niswug.org

"Node news is good news."

 

[gfbotha]

Never tried this, but I was wondering about the following possible work-around (by all means not 100% representative but just may be better than nothing...?): Would one get a solution if you replace the bi-linear SvS curve with say a tri-linear curve where the final linear section rapidly takes the stress down after say ultimate strain was reached? You will however not be able to use force control!

Please tell if you have given this a go.

Regards

 

[bafuller]

Kellnerp, many thanks for your answer. However, as a one man company I can not afford to invest in two FEA systems so I will have to consider sub contracting this part of the analysis.

gfbotha, many thanks for your answer. My understanding of defining a multi-linear stress-strain curve is that it has to be monotonic so the stress value must increase with strain and can-not turn downwards. The best it could do is become horizontal, but this would not solve my problem as the elements would still be contributing strength.

 

[rstupplebeen]

Depending on your geometry you could rerun with cracks where the failure is likely to occur. If you do this iteratively you might get good results. I hope this helps.

Rob Stupplebeen

 

[kellnerp]

If you were running in Cosmos/M you could kill the elements that were past failure.

However, I don't understand why you need a visual of the failure. Are you totally collapsing something far beyond what would normally be considered failure? Generally unconstrained plasticity is the point at which a structure has totally failed and there is no point in going further.

TOP
CSWP, BSSE

http://www.engtran.com

http://www.niswug.org

"Node news is good news."

 

[gfbotha]

In real life one actually do get (true) SvS curves which show softening, e.g. some hyperelastic materials and thermal effect in metals. But I am just not sure whether Cosmos would allow you to define a multi-linear curve this way...

If one is to try this out, I would not take the stress down all the way (to zero) and would consider displacement or arc-length control. Granted, convergence might prove more difficult.

Similarly one might also explore the NL elastic material model for a rough estimate - if true modelling of plasticity and behaviour during unloading is not important. With all the above I am assuming a first cut type of analysis where trends and order of magnitude are more important, also given your software limitations.

 

[bafuller]

Many thanks for all your responses.

I will consider these and make some trials and respond later.

Regards to you all