erolsson
Structural
- Aug 24, 2011
- 8
Dear all,
I'm studying the adhesion between elastic-plastic spherical particles by use of "surface based cohesive behavior" in Abaqus 6.9 and I'm now doing a parametric study on how the shape of the traction-separation law influences the results given that the surface energy remains the same.
I started with a bilinear traction-separation law (TSL) which worked well, and now I want to make simulations with the exponential law discussed in the Abaqus manual with the parameter \alpha = -3. The cohesive law is defined in tabular form.
The simulations don't give the results that I'm expecting and by studying the nodal displacements and the contact pressure I can get the traction-separation behavior out of the results. Here, interesting things can be seen. At the node at the center of the contact area, the TSL from FEM and the theoretical (programmed) one agrees well as seen in the picture "Node1". However, if we move towards the edge of the contact area the programmed TSL and the TSL out of FEM starts to differ notably as seen in picture Node20 (taken in the middle between the contact center and the contact edge).
With the bilinear TSL, the TSL out of FEM and the programmed one agreed over the whole contact area
My question is: what am I doing wrong? The programmed TSL and the TSL out of FEM should be the same at every node, I hope
I attach a zip-file with the python script that is used to run the simulations, the interesting part with the cohesive surface behavior starts at line 420, and the two pictures mentioned above
Thanks in advance
Erik Olsson
I'm studying the adhesion between elastic-plastic spherical particles by use of "surface based cohesive behavior" in Abaqus 6.9 and I'm now doing a parametric study on how the shape of the traction-separation law influences the results given that the surface energy remains the same.
I started with a bilinear traction-separation law (TSL) which worked well, and now I want to make simulations with the exponential law discussed in the Abaqus manual with the parameter \alpha = -3. The cohesive law is defined in tabular form.
The simulations don't give the results that I'm expecting and by studying the nodal displacements and the contact pressure I can get the traction-separation behavior out of the results. Here, interesting things can be seen. At the node at the center of the contact area, the TSL from FEM and the theoretical (programmed) one agrees well as seen in the picture "Node1". However, if we move towards the edge of the contact area the programmed TSL and the TSL out of FEM starts to differ notably as seen in picture Node20 (taken in the middle between the contact center and the contact edge).
With the bilinear TSL, the TSL out of FEM and the programmed one agreed over the whole contact area
My question is: what am I doing wrong? The programmed TSL and the TSL out of FEM should be the same at every node, I hope
I attach a zip-file with the python script that is used to run the simulations, the interesting part with the cohesive surface behavior starts at line 420, and the two pictures mentioned above
Thanks in advance
Erik Olsson