Hi Mohammadjavad Abdollahzadeh

Here are some suggestions:

For your first question; I suggest you apply manually the boundary conditions (and do not use the default boundary conditions by Abaqus).

In symmetrical boundary conditions, the translation perpendicular to the symmetry plane and the in-plane rotations should be fixed.

Ufortuenetly I could not understand where is your plane of symmetry. But imagine that your plane of symmetry is plan z-x. Then, you should fix Uy, URx, and Urz. Do the same of the other plane of symmetry.

(Pay attention that you should also fix all translations of one point (which might be the corner) to avoid rigid body motions).

For your second, question, it would be better if you put a screenshot of the errors/warnings from your solution (job/monitor). Then, we can discuss better the numerical issues. But You are right. In the "Step" you can reduce the increment size (initial, minimum, maximum) to solve the problem of convergence. So try to reduce the “minimum/maximum” to get a better time discretization.

You can also check what kind of interactions you defined between the layers. If you want to see the detachment, you may need to check the failure criterion and other parameters. I suggest you first solve the problem without any failure and then make the problem more complex.

Finally, your structure’s dimensions are quite good for using shell elements. Why don’t you use shell elements and model your structure as a laminated plate (like ice/interface/aluminum)?

Hi Mohammadreza Moeini

Thank you very much for your comments.

Based on your comments I checked the symmetrical boundary conditions. in YZ Plane I considered Ux = URy = URz = 0

Also, in YX plane I considered Uz = URx = URy = 0

unfortunately, I did not understand the following sentence that you mentioned:

(Pay attention that you should also fix all translations of one point (which might be the corner) to avoid rigid body motions).

Do you mean that I should fix one point (In one corner) of my geometry?

Second, I put a screenshot of the warning. There is not any error.

Also, In the status file, I got many warnings about the distortion of some elements. I think this is important but I am not sure about it.

Based on some papers I found that the failure stress for Ice is between 0.7 to 3 MPa. Also, it seems that the average value that is mostly used is around 1.4 MPa. As a result, I put this value.

Finally, as you suggest using the shell for aluminum instead of solid, I changed the aluminum to shell. unfortunately, still I cannot get any result.

I put some screenshots from the material and other parameters that I defined. Maybe they can help you to figure out the problem.

(If you have time, I can send you the inp file of the Abaqus, so you can check my model)

thanks a lot

Hi Mohammadjavad Abdollahzadeh

Thank you for the screenshots and for sharing more clarifications.

I think what you said about your symmetrical boundary conditions is correct.

I mean we should always pay attention to rigid body motion once we want to apply boundary conditions consider to the loading (otherwise, the F=KX can’t be solved). So yes you can fix the corner. But maybe it’s not needed, consider your in-plane loading (considering the symmetrical boundary conditions that you have applied the corner has been already fixed for in-plane translation/rotation).

We can live with distorted elements (just there are three questions, how much is the percentage in “mesh/verify mesh”? second, your geometry is very simple, I was wondering why you have distorted elements? Can you improve your mesh?)

From the screenshot that you sent, did you define contact interaction? If yes, you should apply over the master element (otherwise, the nodes can be over-constrained = more equations than the unknowns).

In fact, I did not understand why you defined it? (which is related to my lack of knowledge in this field). But yet, there are two questions: can you try another contact algorithm? Second, is it possible to solve without any contact interaction between the layers? (because contact problem would be very difficult to converge)

My previous suggestion was to model with shell elements and laminated plate (something like the attached figure) with three isotropic materials.

Yet, I suggest you model without any contact/failure mechanics and maybe with 3D elements to check whether or not the stress between the layers exceeds the maximum value that you mentioned (0.7 to 3 MPa). Then, at least you can be sure about the failure.

Your model in fact is more complex than I thought. Sorry, I am not sure that I can help you since I am not very familiar with model fracture mechanics and failure mechanics (which in this case I think would be the delamination between ice and aluminum.)