Currently working on dynamic, quasi-static numerical simulations with composite materials, i.e. carbon fiber epoxy laminates. I'm initially trying to simulate delamination between plies through the insertion of a zero-thickness cohesive element layer between two plies with the same orientation [0], and subsequently between a non-symmetric laminate [45/-45]. I'd expected the latter to show off-plane vibrations, but the former was also quite noisy.
I've managed to reduce said noise with Rayleigh material damping combined with some mass scaling for the [0] laminate, but this was insufficient for [45/-45] laminate. I was wondering if there'd be any recommendations, or anything I'm missing. Also, I'll probably try to use cohesive surfaces next, but I'm still unclear on the reason why cohesive elements cause vibration.
I appreciate your time and attention.
Dear Fred Furst, these links may be helpful
https://fenix.tecnico.ulisboa.pt/downloadFile/844820067124323/ExtendedAbstract.pdf
http://www-mech.eng.cam.ac.uk/profiles/fleck/papers/276.pdf
https://eprints.qut.edu.au/80058/1/DYNAMIC%20SIMULATION%20of%20CFRP%20STRENGTHENED%20STEEL%20COLUMN%20UNDER%20IMPACT%20LOADING.pdf
https://www.google.iq/url?sa=t&rct=j&q=&esrc=s&source=web&cd=18&cad=rja&uact=8&ved=0ahUKEwip1pOllYbWAhUBkRQKHdjbAOQ4ChAWCFEwBw&url=http%3A%2F%2Fyadda.icm.edu.pl%2Fyadda%2Felement%2Fbwmeta1.element.baztech-article-BTB2-0063-0089%2Fc%2FGlodkowska.pdf&usg=AFQjCNH_XyP3CrzicORw11x-vK_Tx_Q60Q
Dear Fred, good day
what is the type of matrix the you used it?
may be the is no contact between the layers
Hello,
I appreciate Mohannad's suggested papers, although I'm not executing an impact simulation, nor strengthening concrete or steel with CFRP.
To answer Hakim's remarks, I'm using a TIE constraint to connect the two cohesive element layer surfaces with surfaces from adjacent plies. As for the matrix, I'm not sure you're referring to this, but it's a laminate with carbon fiber embedded in epoxy.
It may depend on the traction separation law you are using. If your cohesive element fails, then it will become unstable due to the large deformation. I suggest to read this.
https://www.researchgate.net/publication/316927229_Computational_fracture_mechanics
Technical Report Computational fracture mechanics
Hello Silvio,
Thank you for your attention.
As a matter of fact, the model is comprised of two plies modeled as different parts with the same fiber orientation. I've added a zero-thickness layer of cohesive elements to one of the parts, and tied the two plies together in the assembly. This is why the noise I'm getting is unexpected.
Regards,
Fred
which traction separation law are you using? If you don't have softening this may be the reason.
Hello Silvio,
Sound like that could be it, but wasn't softening a behavior defined only for post-failure iterations? Anyway, I'm using Elastic Traction without Temperature-dependent data, with 0 field variables and a long-term moduli time scale for viscoelasticity. Also included Quads damage, BK damage evolution and damage stabilization, as well as Density. As for the cohesive section, it's response is defined as a Traction Separation law, and it seems like the correct choice according to the Abaqus documentation since I'm modeling a negligibly thin layer of adhesive.
Of course, I'd greatly appreciate any pointers for research and modifications for the model.
Regards,
Fred
What solved this problem was unexpected: a reduction in penalty stiffness.
- Badges
- Science topic