Learn multiscale modelling of composites

  1. Danish Handa

    Respected Sir,

    I am doing research on machining of composites. However, I have seen papers regarding multiscale modelling. I want to learn the same as I am working on woven composites. What is the best way to start the multiscale modelling? Please guide me with this thank you.

  2. Wenbin Yu

    You can start by reading the following two papers and ask questions. And also try TexGen4SC on cdmHUB.

    1. Liu, X.; Rouf, K.; Peng, B.; and Yu, W.: “Two-Step Homogenization of Textile Composites Using Mechanics of Structure Genome,” Composite Structures, vol. 171, 2017, pp. 252-262.
    2. Rouf, K.; Liu, X.; and Yu, W.: “Multiscale Structural Analysis of Textile Composites Using Mechanics of Structure Genome,” International Journal of Solids and Structures, vol. 136-137, 2018, pp. 89-102.
  3. Danish Handa

    Thank you so much. Sir, can you suggest any book or what else I need to read in order to work on failure analysis of woven composites. I am basically working on machining of composites and need to model machining processes using woven fibers. I will go through the papers and get back to you. If possible, can you suggest anything else if you feel I need to go through at the initial stage?

  4. Wenbin Yu

    I don’t have particular suggestions for books. ---- Emailed forum response from
  5. Danish Handa

    Thank you sir. I will read this papers and get back to you for further discussion. Thank you for the help.

  6. Matias Sessa

    PLEASE UNSUBSCRIBE ME. The unsubscribe link does not work. Thank you, Matias -----Original Message----- ---- Emailed forum response from
  7. Wenbin Yu

    Dear Matias, I am sorry that the unsubscribe link does not work. I have reported this to our IT team and it should have been taken care of. -----Original Message----- ---- Emailed forum response from
  8. negin rostami


    im trying to learn working on texgen4sc. so to satart learning , i modeled the woven plate according to this paper:

    ''Two-step homogenization of textile composites using mechanics of structure genome''

    so i used these properties for yarns:

    E1(GPa): 167.702 ;   E2(GPa): 22.762   ;    E3(GPa): 22.762   

    ;  G12(GPa): 5.101    ;  G13(GPa): 5.101     ;   G23(GPa): 5.136   

    nu12 : 0.2300  ;     nu13:0.2300    ;    nu23:0.490

     and for matrix

    E(Gpa) : 4.51    nu:0.38

    and also i used the 3D model and periodic boundary in all 3 directions .but the results are different from what is reported in the paper.

    please help me to  figure out my mistake which i've probably made. .





  9. Xin Liu

    Hi Negin,

    Thank you for your interest in texgen4sc. I guess you are trying to compare effective properties of plain woven composites using MSG solid model with PBCs. The results are listed in Table 3 in the paper. In this case, your SG should be a single layer plain woven SG. If you want to get 3D properties without PBCs in the thickness direction, your SG should be a 3-layer plain woven SG. Please make sure you have the same mesh density as reported in the paper.




  10. negin rostami

    thanks for replying. i appreciate it.

    yes im trying to model the SG model  (single layer plain woven SG with PBCs in the thickness direction ) exactly same as the model in the paper. but my main question is that why my results are different from the results in  tht paper. i cant find out my mistake which  has caused different results.  i will be thankful if you help me  find my mistake. because i aim to learn working with  TexGen4sc and i wanna make sure i do every step correctly.



  11. Xin Liu

    Hi Negin,

    You need to provide more information about your modeling. You may get very different results with a different parameter defined by you and in the paper. For example, you should check if the mesh in your model is identical to the paper. Moreover, you need to disable the "add 10% height" to make sure the geometry of your model is identical to the model in the paper. I attached the sc file of the model in the paper (please delete the .txt if you want to run the sc file using SwiftComp). If the sc file you generated using texgen4sc is different from the attached file, you may want to double check the parameters you used for modeling.

  12. negin rostami

    dear pro.f Xin Liu 

    thanks very much for your help. it means a lot to me.

    can you please tell me more about generating mesh using texgen4sc. because i just found the number of voxel option in texgen4sc during characterization properties. is there any thing else to set the mesh during this process ?


  13. Xin Liu

    Dear Negin,

    You are welcome. I am not a professor. You can call me Xin.

    Voxel mesh is the only option in texgen4sc for generating the sc file. The only mesh parameter you can tweak is the voxel count as you highlighted in the screenshot.

    Although there are other mesh options in texgen, I would suggest you to use voxel mesh to avoid possible distorted elements. You can read the mesh relevant threads on the texgen forum.

  14. negin rostami

    thank you very much dear Xin. and sorry for my mistake i just saw your resume and i found you are a Phd student. so congratulation  anyway.

    just one more question. what is the because i cant find any file with this format using texgen and because i lunch texgen4sc online using cmdHUB i cant access the export files . how can i get that file,then?

    and also how can i see the generated mesh using voxel tool? like the number and shape of elements?


  15. Xin Liu

    You are welcome.

    The ".sc" file is the input file for SwiftComp. Texgen4sc uses texgen to created meshed unit cell of textile composites and internally generate ".sc" file for SwiftComp to run homogenization. You can open the ".sc" file with the text editor which pops up with the homogenization results (see attached screenshot).

    You can see the number of elements and nodes in sc file. Please refer to the SwiftComp manual "SCtheory" ( If you want to view the mesh in a post-processor, you can generate a inp file with the same voxel counts and import to Abaqus, or write a simple script to convert sc file to any post-processor you are familiar with. The voxel mesh in texgen4sc is C3D8 elements.


  16. negin rostami


    can any one help me with modeling a RVE with aPBC using texgen? how can apply this boundary condition in texgen?

  17. Xin Liu


    Please see the screenshot.