Dear Attiq,

generally speaking, changing the size of the mesh modifies the solution results too. You should investigate and conduct a (mesh) convergence analysis, namely use a lot of configurations of mesh density (sparse to dense) to evaluate if/and how your solution tends to converge. For the specific problem you can also consider use a locally refined mesh, e.g. bias seeding, in the vicinity of the impact zone. Moreover, the convergence and reliability of a nonlinear analysis like that depends on more parameters than mesh convergence, like proper contact definition, time step/mass scaling etc..

Regards,

Dimitris

Dear Dimitrios,

I am actually doing a validation study and the contact conditions has already been specified by previous researchers. I am using surface to surface contact with hard contact and Penalty friction co-efficient 0.47, Kinematic Constraint with small sliding formulation is used as recommended. That's why I was wondering if there is an issue with mesh sizing.

As seen in the graph, the impact load at 0.005 sec, 11kN is the maximum load observed in the tests. However, the further continuation of impact load beyond 0.005s is the issue. I have refined the mesh on slave surface to avoid any penetration but it still doesn't resolve the issue.

Yes I can see the graph. Do you need validation with the second graph that you enclose, right? I don't know if this is a mesh issue; Looks like your structure can not sustain any further increase of the load. Do you know a priori if plasticity effects are caused due to the impact excitation?

Dear Dimitrios,

I need validation with the first graph where the impact load drops to 0 at 0.005s. I am working on global plastic behavior of the structure in which elastic perfectly plastic material is considered with 500MPa yield stress and corresponding plastic strain of 0 and 0.3.

The impact load is 25.45kg with predefined velocity of 7m/s. This a two step analysis. In the first step, different levels of axial laod (70%,50% and 0% of steel tube capacity) are applied and failure is studied under these levels of load.

I was wondering about the meshing of impactor which changes the shape of element at the impact location, picture attached. In actually test, the sharp edge is slightly rounded. It might be possible that the shape of impactor has an effect on the excitaion.

Generally a fine mesh gives more accurate result. But, to be ecomical better to begin with coarse mesh and refine the mesh by cheching your results

For getting accurate contact forces, first we have to ensure that the mesh densities of the contacting surfaces should be matched.

Second we can use the soft options if they are shell elements as we can use in LS-DYNA.

Hi Kuldeep,

Are you suggesting that mesh size should be the same for both contacting bodies. Is this a requirement for Abaqus Explicit too like Abaqus Standard?

Actually when I tried matching the meshes, I was getting the errors regarding too much rotation for elements on slave shell surface.

What do you mean by soft options? I have been reading literature about impact surfaces and I never came across this term. Can you kindly elaborate?