MrGerritson
New member
- Jul 8, 2013
- 1
Hello everybody,
I want to use a given material-model that has a DISPLACEMENT Based damage evolution formulation. The abaqus-documentation says the approach is used to get an mesh-independent damage-formulation.
Contrary to this promise, I get different stress-strain curves when I use a) a different elementsize or b) a different number of elements in line of the strain.
As far as I understood the formulation an element with 1mm edge length will deform the same amount as an element with 5 mm edge length after damage-initiation. I would understand this behavior if the damage-displacement was a measure for the "crack-energy" that has the same value no matter how long the sample is (though the stress-strain-curves look totally different after damage-initiation).
But if I stack two elements in a row with a total length equal to that of a single-element, the displacement before failure is doubled and so is the absorbed energy (ALLIE). So it seems to me the behavior is heavily mesh dependent and the formulation is only valid for a certain kind of elementsize.
But since abaqus claims the opposite I can not believe it.
Could someone please enlighten me about that and tell me where I am mistaken here (as detailed as possible)!?
Thank you very much
Gerrit
I want to use a given material-model that has a DISPLACEMENT Based damage evolution formulation. The abaqus-documentation says the approach is used to get an mesh-independent damage-formulation.
Contrary to this promise, I get different stress-strain curves when I use a) a different elementsize or b) a different number of elements in line of the strain.
As far as I understood the formulation an element with 1mm edge length will deform the same amount as an element with 5 mm edge length after damage-initiation. I would understand this behavior if the damage-displacement was a measure for the "crack-energy" that has the same value no matter how long the sample is (though the stress-strain-curves look totally different after damage-initiation).
But if I stack two elements in a row with a total length equal to that of a single-element, the displacement before failure is doubled and so is the absorbed energy (ALLIE). So it seems to me the behavior is heavily mesh dependent and the formulation is only valid for a certain kind of elementsize.
But since abaqus claims the opposite I can not believe it.
Could someone please enlighten me about that and tell me where I am mistaken here (as detailed as possible)!?
Thank you very much
Gerrit