modeling test damage

[Cucurbitacea]

good evening to all, I am struggling with the modeling of a composite material audition (in particular of a plain weave fabric) in patran and I am struggling with some doubts. in particular in one of the 4 layers of the test there is a cut in the fibers (this area affected by the cutting was modeled as a laminate with 3 layers with the properties of the composite and 1 layer with the properties of the matrix alone as it seemed reasonable to argue that in the cutting area the load is transferred only thanks to the presence of the matrix). from the analysis I would like to reconstruct the position of the damage using the strain field in the various plys of the laminate but the latter are equal in all the plys (logging on I have attributed this result to the fact that since linear movements in the thickness of the laminate the strains will be constant). therefore to grasp any differences between the strains of individual plys would be necessary a model with 3d elements instead of using 2d shell elements?

 

[MekIng]

Hi.
First I would like to bring my perplexity, it is not clear to me because you thought about finding the location of the damaged area using the field of deformation and not the move.. do not lose "information" by evaluating the deformation being such a field obtained by differentiating that of the shift? I'm probably wrong, but it's just my curiosity.

returning to your question, even I think using 3d elements is a good idea only that at that point you should shape the real damage (nothing difficult)
at this point I would go to evaluate the deformation at the top of the elements by getting a 3d map that allows you to see where there are the damaged areas with the only flaw that you will get a singular trend in the area of your interest

Are you developing a damage monitoring system?

 

[Cucurbitacea]

First of all, thank you for the answer, I actually didn't think about moving instead of strains (as you say I should have at least a linear pattern of travel) I'll try to see what I can pull out.

As for modeling 3d I was evaluating a little how to do because there are solid elements "laminated" but both at the level of documentation and use does not seem too clear and intuitive to use, so I evaluated to model each single layer and assign them as properties of a 3d anisotropic material obtained "ruoting" the propitiousness of the composite. to model the real damage you intend to assign also in this case matrix property to the elements involved in cutting?

Yes, in fact I should model with different fidelity the presence (note) of the damage for a training of an artificial intelligence algorithm .

 

[MekIng]

training of artificial intelligence through simulation, I imagine it's for the graduation thesis!


I would also model every single layer with 3d elements to which to apply the properties of anisotropic material.
As for modeling the damaged area I did not change the properties but only put a discontinuity in the volume/area (modeling a clique).

As for the use of travel I don't understand why you refer to a linear trend, you mean using linear elements you expect that trend in general? or even close to damage?
If I do not remember badly, near the discontinuity, you should find a trend proportional to r^0,5 where r is the distance from the apex of it. the interesting thing is that such trend in the case of discontinuity type clique is always the same for any type of anisotropy, that is always equal to the case of the isotropic material even if then it differs in absolute value of a constant, you can find it and recover all the material already present in literature regarding the isotropic materials

 

[Cucurbitacea]

(b) the Commission's proposal for a Council Directive on the approximation of the laws of the Member States relating to

I will also try this approach to insert a discontinuity into the material not "shaping" the knots between them.

with a 3d model (with 2 elements along the thickness and with the properties corresponding to the "ruoted" material) for the individual layers I still get strains very close to the same value (but also here by reasoning the strain of the individual layers because it should be different if there is a perfect between the layers themselves?). I remember from the clt that the movements of the movements were linear along the thickness.
in proximity to the damage I still get the greater strains however in case it is convolved a layer "inside" like the 2nd or 3rd the maximum strains do not find them in these two layers but respectvametne in the 1st or 4th as if the trend of the strains in this area is linear.

You're saying about the tension at the top of the clique? I did not think about this aspect (and in the various courses for fatigue we treated especially metals so in composites I did not know the same reasonings)

p.s. good August