Hello community,
I'm trying to simulate cracks with cohesive elements (in Abaqus), but I would like to reproduce reality as much as possible. Therefore, I would first like to have elastic behavior after which plasticity comes which brings plastic strain. Then after some accumulated plastic strain, I would like an opening of cohesive elements like a formation of a crack.
I've started examining this case with perfect plasticity or with a slight hardening, I've also tried with proper hardening but can not get a cohesive zone. All dissipation goes in plastic flow.
Do you have any advice on what to do to get first some plastic flow (like in reality) and the formation of the cohesive zone that finally breaks like a genuine crack?
Material is steel:
E: 200000 Poisson: 0.3
Plasticity starts at 250 (perfect plasticity)
Cohesive zone_ Traction separation law
Elastic properties 6000000, 6000000, 6000000
Damage int: ????
Damage evolution0.0001
With Damage Initiation, I was trying different things 250 like where plasticity starts, or with hardening, and then put initiation on 251 or 255 and different combinations. Nothing works.
Thanks a lot in advance,
Domagoj
Dear Domagoj Uremović ,
first things first, I guess you have a stress-strain graph that defines your material and from which you have taken the material properties in order to define them in the software. Also, with "perfectly plastic", I assume you have simplified, at a certain extent, the constitutive relationship of you steel. Be sure that you have converted your stress-strain graph in its "true" form, from the engineering - i.e., experimental - values before entering the plasticity and material data into Abaqus.
That being said, if you want to simulate fracture with traction separation laws (TSL), you should be careful in defining also the damage initiation criterion (dmgini). I suggest you to use your true UTS, evaluated from the true stress-strain curve; hence, you should use a MAXPS dmgini criterion.
I remind you that you should work with the stress-strain graph in its true form, not with the "raw" engineering data. Also, be careful that the true form of the stress-displacement curve works well up until the UTS - so it is fine to use it for the computation of the damage initiation criterion; however, be careful if you have to be very precise with data after the UTS, since you should refine your true curve with a proper model. Finally, be sure to have properly evaluated your damage evolution criterion, even if the order of magnitude seems reasonable.
Hope this has been helpful,
Francesco
Thank you very much Francesco Cervellera, as I'm still building the model and not focusing on a specific material, I'm using parameters of steel.
Therefore, I think it's rather a possible problem in the setup of simulation or some constraint of cohesive elements that I'm facing rather than the material definition that I described in the original post.
Domagoj
- Badges
- Science topic