How to Accurately Extract C_ij Components from 2D Periodic Unit Cell in COMSOL?

Manoj Patlu Nagabhushana

Please login with a confirmed email address before reporting spam

I'm currently working on reproducing the 2D square microstructure stiffness data used in the GM-TOuNN paper to eventually extend it to 3D. However, even for the simple 2D square unit cell, the elasticity matrix ( components) I extract from COMSOL doesn't match the original values reported in the paper.

What I’ve done so far: I created a 2D square unit cell with a central void to represent varying volume fractions (v), controlled via void size. I applied 3 load cases (ε₁₁ = 1, ε₂₂ = 1, ε₁₂ = 0.5) and used average stress values to compute C₀₀, C₁₁, C₂₂, C₀₁, C₀₂, C₁₂. I performed a parametric sweep over volume fractions and saved average stress values for each load case in a CSV.

My request: Can someone please help me:

  1. Confirm the correct way to extract values in COMSOL for a 2D square unit cell with varying volume fraction?
  2. Clarify if using elasticity matrix is more accurate than average stress for this purpose?
  3. Point me to any reference or COMSOL-specific best practices for implementing this type of homogenization?


7 Replies Last Post Apr 13, 2025, 5:57 a.m. EDT
Amit Suresh Patil COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 3 days ago Apr 10, 2025, 6:00 a.m. EDT

Dear Manoj,

I looked into the simple example you attached, in that example the Cell Volume should not set to Automatic in the presence of voids, rather it should be User defined with volume set to 1[mm]1[mm]1[m]. What are the reference values for this example?

More references about material homogenization with COMSOL are 1. Homogenization of Material Properties 2. Micromechanical Model of an Octet-Truss Lattice Cell 3. Homogenized Material Properties of Periodic Microstructures

Regards Amit

Dear Manoj, I looked into the simple example you attached, in that example the *Cell Volume* should not set to *Automatic* in the presence of voids, rather it should be *User defined* with volume set to 1[mm]\times1[mm]\times1[m]. What are the reference values for this example? More references about material homogenization with COMSOL are 1. [Homogenization of Material Properties](https://www.comsol.com/support/learning-center/article/Homogenization-of-Material-Properties-80311?utm_content=bufferced1b&utm_medium=Social&utm_source=LinkedIn&utm_campaign=comsol_social_pages) 2. [Micromechanical Model of an Octet-Truss Lattice Cell](https://www.comsol.com/model/micromechanical-model-of-an-octet-truss-lattice-cell-132251) 3. [Homogenized Material Properties of Periodic Microstructures](https://www.comsol.com/model/homogenized-material-properties-of-periodic-microstructures-109241) Regards Amit

Manoj Patlu Nagabhushana

Please login with a confirmed email address before reporting spam

Posted: 3 days ago Apr 10, 2025, 5:42 p.m. EDT

Dear Amit,

Thank you for your reply . I'm trying to match the values as obtained in this graph , which is from paper I attached. Somehow the values are not matching for lower volume fractions.

Dear Amit, Thank you for your reply . I'm trying to match the values as obtained in this graph , which is from paper I attached. Somehow the values are not matching for lower volume fractions.


Manoj Patlu Nagabhushana

Please login with a confirmed email address before reporting spam

Posted: 3 days ago Apr 10, 2025, 7:09 p.m. EDT

Dear Amit,

Thank you for your valuable suggestion. I'm able to match the values now, except for one value C22 which is basically C33 in COMSOL (elasticity matrix) .I think it has got to do with loadcase settings .I'm attaching you the Boundary condition and Loadcases I'm using .This is for 2D study.Please let me know if any change is required in this .

Regards, Manoj

Dear Amit, Thank you for your valuable suggestion. I'm able to match the values now, except for one value C22 which is basically C33 in COMSOL (elasticity matrix) .I think it has got to do with loadcase settings .I'm attaching you the Boundary condition and Loadcases I'm using .This is for 2D study.Please let me know if any change is required in this . Regards, Manoj


Please login with a confirmed email address before reporting spam

Posted: 2 days ago Apr 11, 2025, 6:09 a.m. EDT

How did you determine the values of Cij?

How did you determine the values of Cij?

Manoj Patlu Nagabhushana

Please login with a confirmed email address before reporting spam

Posted: 2 days ago Apr 11, 2025, 8:15 a.m. EDT

Dear Said Bouta,

Thank you for your reply. I ran for 3 loadcases and measured Cij in Elasticity matrix voigt notation . I'm attaching screenshots for both.

Regards, Manoj

Dear Said Bouta, Thank you for your reply. I ran for 3 loadcases and measured Cij in Elasticity matrix voigt notation . I'm attaching screenshots for both. Regards, Manoj


Please login with a confirmed email address before reporting spam

Posted: 2 days ago Apr 11, 2025, 11:46 a.m. EDT

In curve 4 of GM-TOUNN... pdf, he plotted the curves for the case of X-shaped cells, it is not the same as your case.!

In curve 4 of GM-TOUNN... pdf, he plotted the curves for the case of X-shaped cells, it is not the same as your case.!

Amit Suresh Patil COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 8 hours ago Apr 13, 2025, 5:57 a.m. EDT

Dear Amit,

Thank you for your valuable suggestion. I'm able to match the values now, except for one value C22 which is basically C33 in COMSOL (elasticity matrix) .I think it has got to do with loadcase settings .I'm attaching you the Boundary condition and Loadcases I'm using .This is for 2D study.Please let me know if any change is required in this .

Regards, Manoj

Hi Manoj,

Yes it has to do with the loadcases and cell periodicity settings. Manually remove the auto created model nodes and regenerate again. My advice it first try to build a 3D model and then use that model as a reference to build a 2D model.

Regards Amit

>Dear Amit, > >Thank you for your valuable suggestion. I'm able to match the values now, except for one value >C22 which is basically C33 in COMSOL (elasticity matrix) .I think it has got to do with loadcase settings .I'm attaching you the Boundary condition and Loadcases I'm using .This is for 2D study.Please let me know if any change is required in this . > >Regards, >Manoj Hi Manoj, Yes it has to do with the loadcases and cell periodicity settings. Manually remove the auto created model nodes and regenerate again. My advice it first try to build a 3D model and then use that model as a reference to build a 2D model. Regards Amit

Reply

Please read the discussion forum rules before posting.

Please log in to post a reply.

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.