Learning Forum

[pau.riera-i-portillo]

Hi all, 

I am trying to model an iced, rectangular, metallic plate being hit from underneath with a cylindrical impactor. The plate is clamped on its short ends and modeled using *MAT_ELASTIC. To simulate cohesive fracture within the ice, I use peridynamics: the nodes of the block of ice are disconnected and the elements use *SECTION_SOLID_PERI and *MAT_ELASTIC_PERI. The block of ice is attached to the plate via a tiebreak contact. While the parameters in *MAT_ELASTIC_PERI are quite straightforward and the relationships to micro-modulus and critical stretch are well-defined, it is not the case with *SECTION_SOLID_PERI and the "Dr" parameter. The issue I am facing is that I am not sure which value to use for Dr, since changing it affects the results significantly. Moreover, I do not understand why, under the hood, LS-DYNA performs a check on the surrounding neighbors of every point to make sure there are at least 12 and at most 136, and then tunes the horizon size accordingly. Does that mean that the used horizon size is not the same as the one calculated using "Dr"?

For reference, I am attaching a picture of the model (green part is ice, the other part is the plate, the impactor hits the plate on its center from the back), and two snapshots of two identical simulations at the same time after the impact, the only difference being the "Dr" parameter used. The first one has Dr=1, and the second one, Dr=2. The snapshots represent a top view of only the ice, where the effective plastic strain (ie peridynamics damage) is fringed. As shown, there is some difference to the crack patterns that emerge from the impact just by changing this parameter.

I would appreciate any advice regarding how to pick this value or at least minimise its effect on the final results, as well as any improvements that could work in this case. Many thanks in advance!

[Reno Genest]

Hello Pau,

Why are you using DR=2 while the LS-DYNA User Manual Vol I recommends DR = 0.6 - 1.2 ?

 

The LS-DYNA solver will try to adjust DR so that the number of neighbors at a point is between 10 and 136. Search for Warning messages in the messag file similar to:

"Warning. The maximum neighbor number is 36

     the minimum neighbor number: 9

     which violates the threshold: 10~136 (Min~Max).

     The horizon size will be tuned, and new search starts"

 

 

I would pick a value of DR so that the number of neighbors at a point is between 10 and 136.

 

Let me know if this helps.

 

 

Reno.

[Reno Genest]

Hello Pau,

Here is one slide from our peridynamics training course:

 

This might help understand better the DR parameter.

 

Let me know how it goes.

 

Reno.