If you're using an MPP version of LS-DYNA you need to check that MPP1 is checked in your contact card and lower the bucketsort frequency. Activate optional card A and use soft 1 or 2 . Yo may want to increase your MAXPAR value to 1.05 

Thank you. I am not quite sure how to do that yet, but I will try and let you know. Thank you

Hello, in CONTACT_ERODING_SURFACE_TO_SURFACE card you have to define a slave and a master as you know. In that definition, what did you choose? For example, this kind of situtation can happen if you choose bullet as a part and nodes of hit surface as only nodes or segment. Then when only first layer of nodes were erroded, there wont be any contact definition between rest of nodes of target. Sincerely.

Thank you. 

In fact I didnt have the option MPP1 checked. Then in the new card, I have tried with SOFT 1 and I have reduced the BUCKET from 200 to 50 and LCBUCKET is at 0. This case give me failures for negative volume and at the end the result is the same: the bullet starts eroding but after some time it just seems to pass through the rest of the block (I have included now a cut to show it better).

I have tried with SOFT 2 but it does not run due to a memory problem.

Ali, I am setting the bullet as master and the block as slave. I understood that was the proper configuration, but from your message it is not so clear.

In fact, if I understood correctly, the CONTACT_ERODING_SURFACE_TO_SURFACE is symmetric and "the definition of the slave surface and master surface is arbitrary since the results will be the same" (copied from the manual).

Thank you

Actually your choice of bullet as master and plate as slave is correct. But I mentioned about SSTYP and MSTYP: did you set them part or segment or node list? If you only set the surface of your target as slave segment, then the code doesnot know about the rest of your part. You can may be upload some reduced input of your problem. Sincerely.

Thank you, 

I am setting both SSTYP and MSTYP as 3 (part). I think that it is the right choice, am I right?

https://www.dropbox.com/s/75d1ux4mu1cfngl/projectil.k?dl=0

Hi Pablo,

Having seen your model the first thing I noted is that your mesh is far too coarse for this type of simulation.  Have a look at this paper for comparative meshes. 

https://www.researchgate.net/profile/Michael_Saleh/publication/241096550_Ballistic_performance_of_multi-layered_metallic_plates_impacted_by_a_7.62-mm_APM2_projectile/links/0a85e5372bd8df01d5000000.pdf?origin=publication_detail

Article Ballistic performance of multi-layered metallic plates impac...

Hi Michael, 

I already knew that paper. I just gave you an example with a coarse mesh, but what I am doing is with something much finer. Ali asked me to upload a reduced input. I don't think that the problem is in there. In any case I will try. 

Assuming your mesh is fine, the contact you have chosen will only work for the surface segment faces as the initial search will only identify those. To allow for the interiors segment faces to take part you must set the IADJ value in your contact card to 1.   

I think that will be it. 

That might work! I will try. I missed that parameter

I agree with Michael. On the other hand, when I investigate your input, I see that you use solid element formulation 0. I am not familiar with this formulation but when I checked the keyword manual the following statement was written in the remark 3 of *SECTION_SOLID keyword;

"Element for Modified Honeycomb Material. Element formulations 0 and 9, applicable only to *MAT_MODIFIED_HONEYCOMB, behave essentially as nonlinear springs so as to permit severe distortions sometimes seen in honeycomb

materials. In formulation 0, the local coordinate system follows the element rotation

whereas in formulation 9, the local coordinate system is based on axes assing through the centroids of the element faces. Formulation 0 is preferred for severe shear deformation where the barrier is fixed in space. If the barrier is attached to a moving body, which can rotate, then formulation 9 is usually preferred."

And you defined your solid parts with element formulation 0. Then your material model does not work with your element formulation. As I said I am not so familiar with this formulation, sorry if I am wrong. Sincerely.

Thank you Ali, I will try that too. The IADJ didn't work completely. I will try and show you the results. Thank you again

I have been trying with different ELFROM. You are right that it was not correct. 

With ELFORM=1, the code works but the result is the same (only erosion in the first part). 

I also tried with ELFORM=3 and =-2 but it gives me some kind of problem related with memory that I dont fully understand. 

Just increase the memory on the command prompt or if you use gui, you can select the memory at where you select number of cpus.

Thank you again, but it is doing the same think. I must be missing something. I have configured:

- IADJ=1, to erode also interior elements.

- ELFORM: I have tried 1, 2 and 3. With 2 and 3 I get an error due to "negative or zero determinant". I am going to try a new mesh with tetrahedrons and solid formulation 10. 

UP to this point the best that I manage to get is this image in which the bullet reaches the block and starts eroding but at some point it just passes through the block. 

We have tried with different cases with finer meshes up to a point in which we are not able to solve it due to a memory error. The case that seems to work best is the one with IADJ=1  and ELFORM=1. But it behaves in the same way, it just erodes the first part of the block and then the bullet just passes through the rest. 

Any idea?

Hi Pablo, is there any progress? I checked your .k file again and could not see any constraint or boundary condition for the target plate. But it is still not an answer for your problem. And there is no hourglass definition in the keyword. For large deformations hourglassing maybe a problem and need to be accounted for. At least you can check it from matsum. But still not an exact answer for your problem. Material properties might be problematic. You may need to define a CONTACT_INTERIOR which is useful for large deformations of soft materials. I hope you could come up with an idea. If you could, please inform, i am very curious about what is wrong with this problem. Sincerely.

Thank you Ali, We still don't have any solution to the problem. I will try what you are proposing and I will let you know. 

Hi, Pablo, maybe you can try erod_node_surface contact type. Also, I noticed that you don't specify constrain boundary conditions on the edge, any reason for that? I saw similar situation on other K file.

I recently expericened this problem also.

One thing i found is when the bullet speed is @300m/s, it works, without any problem.

When I set it to 1600m/s, it doesnt work at all.

Pablo Fajardo have you solved this problem?

Pablo Fajardo

Try using CONTACT AUTOMATIC NODES TO SURFACE. I faced the same problem and it worked.

even I'm also having the same doubt.i used JH model for plate and JC model for bullet.for some cases(velocities) bullet piercing the plate,for some cases it is creating a small deformation at contact and still it is able to pass.i don't known how?

Better set IADJ=0.

Also set SOFT=2. Eroding_surface_surface will work. More expensive than automatic but it updates the contact surface when surface elements are eroded.

Also set single surface definition to the plate and see additionally.

I experienced the same problem, i.e..when the impactor speed is above 500 m/s, the simulation did not work as expected.

The card used in this following link was useful for the solution.

https://www.youtube.com/watch?v=7tgtrtHveb4

You may try AUTOMATIC SURFACE TO SURFACE, but use SEGMENT instead of PART of PART SET. In this SEGMENT, you define all the faces of all the elements in your target.