Choice of failure criteria

[lavecchiasignora]

Hi all,

Lets say there is a composite structure which is exposed of an impact of a sphere. Should I use the maximum strain criterion to see if the laminas will fail due to the impact of that sphere?

And if I have the same structure but now the structure is loaded by a pressure over a large area. Is it then better to use Tsai-Wu? Can anyone suggest why it is better to use Tsai-Wu?

Im very confused which failure criteria to use in those two load cases.

Thank you so much.

 

[SWComposites]

"Lets say there is a composite structure which is exposed of an impact of a sphere. Should I use the maximum strain criterion to see if the laminas will fail due to the impact of that sphere?" > you are probably going to have to use fracture based approaches to predict matrix damage and delaminations. Predicting fiber failure will likely required a progressive failure damage mechanics approach with fiber stress strength properties which are a function of strain rate.

"And if I have the same structure but now the structure is loaded by a pressure over a large area. Is it then better to use Tsai-Wu? Can anyone suggest why it is better to use Tsai-Wu?" > no, Tsai-Wu does not correlate well to test data; this has been discussed in detail in previous posts in this and the Aircraft engineering forum; do a search. Recommend using max strain with properties derived from laminate (not lamina) tests.

 

[ESPcomposites]

It can be quite confusing.

As SW stated, ply based criteria are not applicable to impact damage. If you actually know the state of the damage (which you may not) then you can employ a fracture technique as SW has stated. Short of this, the other approach is to typically obtain a pseudo-property called CAI (compression after impact). Because of this known issue in composites, it is well documented and often tested for.

For general criteria questions, this is a recent thread:

http://www.eng-tips.com/viewthread.cfm?qid=294778&page=1

Brian

http://www.espcomposites.com

 

[lavecchiasignora]

Thank you for your answer.

I cant do any kind of tests. And I need to simplify things as much as possible. Progressive failure damage is not considered in my problem and I cant use that right now.

So, two load cases; One impact of a sphere and one pressure over the whole area. I need to roughly estimate which failure criteria to use. I now I need test data and stuff but that is not available right now.

Thanks

 

[ESPcomposites]

SW and myself have given you the accepted methods. Short of this, you could make an engineering estimate. However, a good estimation would probably require prior experience with the problem and familiarity with typical impact strengths and failure criterion.

You are asking to solve a very difficult problem without using the proper methods. One thing that may help is that while compression strengths can be very low due to sublaminate buckling, tension strengths after an impact may not be significantly affected.

Also, why not start by reading the paper from the previous link? It may give you some insight into failure criterion. You may want to start thinking beyond Tsai-Wu.

Brian

http://www.espcomposites.com

 

[lavecchiasignora]

Is there any link that you could recommend to read that deals with fracture based approaches?

 

[SWComposites]

start here:

http://firehole.com/

probably the best composite analytical prediction methods and software out there

(and I have no affiliation with them, just have seen some results)

 

[yusf]

SW Could you please explain the following you mentioned?

"Recommend using max strain with properties derived from laminate (not lamina) tests."

Why not lamina don't we use lamina properties while we model finite element models?

 

[SWComposites]

Typically many of us use lamina stiffness properties for analysis to predict strains in the laminates, and then apply strain design values which are derived from laminate (multi-directional) tests using the max strain criteria applied to each lamina to predict strength. Laminate derived strain values are used because strength (strain) values from lamina ([0], [90], [+/-45]) tests typically do not accurately predict laminate strength (don't believe everything you read in a text book). Search this forum; this has been discussed in more detail before.