Question regarding the units of parameters in the input file

[xibaiw]

Hello Mr. Willberg and Mr. Hesse,
I’m currently studying the example examples/unified_bond_based, and I found the material parameters for Aluminium in the input file. However, they seem to differ significantly from the Aluminium material parameters we typically use (density and Young’s modulus). What is the reason for setting it up this way? By the way, I adjusted the Young’s modulus(70GPa) and density(2700kg/m3) to match the units in the documentation, but the results don’t look very good.
Models:
Material Models:
Aluminium:
Material Model: "PD Solid Elastic"
Young's Modulus: 3.0e11
Poisson's Ratio: 0.2
Symmetry: "isotropic plane stress"
Zero Energy Control: "Global"
Blocks:
block_1:
Block ID: 1
Material Model: "Aluminium"
Damage Model: Damage
Density: 27000
Horizon: 0.0032

[CWillberg]

Hello @xibaiw thank you for the comment. The naming is wrong. Its an arbitrary material. I think it might be an copy & paste error. Sorry for your irritation.

Be aware that the unified bond based was created based on a paper. It gives reasonable results. However, I did not have time to verify it. If you use it and you get good results, maybe a publication, please let us know. We will link to your results.

[xibaiw]

Thank you Mr. Willberg, I’m trying to achieve a good result. Additionally, I’d like to confirm that if PeriLab can set piecewise functions for boundary conditions.

[CWillberg]

@xibaiw Piecewise boundary conditions are possible

It looks like this

Boundary Conditions:
BC_1:
Variable: "Displacements"
Node Set: "Node Set 1"
Coordinate: "x"
Value: "0.1"
Type: Dirichlet
BC_2:
Variable: "Displacements"
Node Set: "Node Set 2"
Coordinate: "y"
Value: "0.1*t"
Type: Dirichlet
BC_3:
Variable: "Displacements"
Node Set: "Node Set 3"
Coordinate: "z"
Value: "if t>0.00001 0 else 0.1 end"
Type: Dirichlet

You can use x,y,z as well. It uses the basic coordinates.

[xibaiw]

And another question I have is, what unit is used for critical energy? Is it N/m?

[JTHesse]

Yes exactly. However, we are often using N/mm, because the model geometry is often in mm.

[xibaiw]

I understand, thank you very much for your reply!

[xibaiw]

Does this mean that in Perilab, the units for my parameters are not limited to those specified in the documentation, and I only need to ensure that all parameters are consistently converted to the same unit system? For example, I standardize length in meters (m) or millimeters (mm). Additionally, can the value of critical energy be interpreted as the fracture energy Gf from fracture mechanics? 2025-03-06 17:01:50 "Jan-Timo Hesse" ***@***.***> 写道： Yes exactly. However, we are often using N/mm, because the model geometry is often in mm. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

[CWillberg]

Correct. We defined a unit free system. Therefore, as in many FEM software tools, you have to define a conistent Parameter definition.

[xibaiw]

Hello, I recently obtained a good result and compared it with experiments from the literature. Now, I have a few new questions regarding PeriLab that I would like to consult with you about.
1.Still regarding the critical energy, I am not sure whether PeriLab uses the energy release rate or the fracture energy for the critical energy.
2.In PeriLab, what formula is used to calculate the critical stretch? I have used it several times for simulation, but the results have been very poor.
The figure below shows a comparison between my current calculation result and that from the literature.

[JTHesse]

Hi @xibaiw regarding the critical stretch formulation I recently found a bug. I will release a new PeriLab Version now.

[JTHesse]

The new version will be available in a few minutes: JuliaRegistries/General#126620

[CWillberg]

Hello,

the critcal stretch is given here [https://perihub.github.io/PeriLab.jl/stable/theory/theory_damage/#Damage-Models-Theory](Critical stretch).

The critical energy follows the definition of Foster. It is the critical energy release rate, obtainable, e.g. by an double cantilever test.

can you provide us with your poor results (also the input you used)

[xibaiw]

sure, i send my package to this email? [email protected]

[CWillberg]

I did not receive it. Try [email protected]

[xibaiw]

Im already sent my package to [email protected]

[CWillberg]

@xibaiw
I checked your result files and your critical stretch result looks not correct. However, if I run your input deck with the current version of PeriLab, I got the expected results.

I suggest you update to the current version.

[xibaiw]

Yeah, The result of the critical energy and the critical stretch are completely different. After converting the critical energy, the results are significantly different. Maybe the version I used in docker is too old ,I'll try it again in my julia!

[franciscoyapor]

The result of the critical energy and the critical stretch are completely different. After converting the critical energy, the results are significantly different.

Be aware that, in general, this is true given the way the two models are defined. For example, the critical energy release model has a direct dependency on the size of the horizon, whereas the critical stretch does not. This difference arises because the critical energy model integrates contributions from all bonds within the horizon, making it sensitive to horizon size and mesh refinement. On the other hand, the critical stretch criterion is a local condition applied directly to individual bonds, making it less influenced by the horizon size.

[xibaiw]

This is my understanding. According to the derivation formula of the critical stretch, the critical stretch is a function related to the horizon size, while the critical energy model is based on the energy release rate, which is related to the material parameters.

[JTHesse]

Try again with version 1.3.4

[xibaiw]

Yes, I tried it again with version 1.3.4 , I got a very good result! Obviously, the version I used in docker is out of time. Thanks a lot for you guys help!

[JTHesse]

The latest docker image is updated as well: docker pull perihub/perilab:latest