KUBC results not smaller than Voigt ROM

1. 

   Imad Hanhan @ 11:35 am on 14 Nov 2017

   Hello all,

   As I was comparing all of my results, I was addressing the part of the problem that states:

   "Verify that the results from RVE analysis with homogeneous displacement boundary conditions are smaller than those from the Voigt rule of mixtures."

   However, for me, the results from the homogeneous displacement boundary conditions (E_1_KUBC=169.2422) are bigger than those from the Voigt rule of mixtures (E_1_Voigt=167.5805). Has anyone encountered this? Does anyone have any suggestions for what may be wrong?

   I have checked my boundary conditions in the RVE, and I have checked my calculations for the Voigt rules of mixtures. I didn't see any apparent errors. Does anyone have any suggestions?

   Thanks,
   Imad

   Report abuse

2. 

   Wenbin Yu @ 6:41 am on 14 Nov 2017 — Edited @ @ 6:41 am on 14 Nov 2017

   Are your KUBC results converged with refining mesh? ---- Emailed forum response from wenbinyu@purdue.edu

   Report abuse

3. 

   Kunal Samel @ 4:41 pm on 14 Nov 2017

   does reassigning the fiber and matrix sections to the respective cells work? also check the fiber volume fraction (radius)

   Report abuse

4. 

   Imad Hanhan @ 11:24 am on 15 Nov 2017

   Dr. Yu and Kunal,

   Thank you for your suggestions. Dr. Yu, I do believe my mesh had converged. To be sure, I re-ran a refined mesh (image attached) and it resulted in about the same result (E_1_KUBC=169.6201) which is a difference of about 0.2, and is still higher than the Voigt ROM.

   Kunal, thanks also for your suggestions.  I did go in and check the material assignments, I edited the sections (images attached) and it did not change the outcome.

   I also attached an image plotting S11 on the deformed shape.

   Do you have any other suggestions for what may be wrong in my mode?

   Thanks again for taking the time to help.
   -Imad

   cdmhub-ref.pdf

   1 MBClick to download

   Report abuse

5. 

   Kunal Samel @ 3:12 pm on 15 Nov 2017

   it seems the boundary conditions are active only at 2 surfaces or at the edges. The BCs in KUBC are applied to all the surfaces

    

   my deformed shape is like this

    

   

   Report abuse

6. 

   Kunal Samel @ 3:12 pm on 15 Nov 2017

   it seems the boundary conditions are active only at 2 surfaces or at the edges. The BCs in KUBC are applied to all the surfaces

    

   my deformed shape is like this

    

   

   Report abuse

7. 

   Akanksha Parmar @ 3:40 pm on 15 Nov 2017

   Hi Imad,

   I have same E1 value as yours, but the eigenvalues for KUBC are indeed smaller than those of Voigt.

   Did you calculate eigenvalues?

   Report abuse

8. 

   Imad Hanhan @ 4:27 pm on 15 Nov 2017

   Kunal and Akanksha,

   Thank you for your suggestions. Kunal you are correct, my model only imposed displacement on the x-faces for the case of epsilon_11, thank you for catching that error. I've gone in and fixed my boundary condition and now my E_1_KUBC=167.4519, which is now slightly lower than my Voigt ROM result (E_1_V=167.5805). Thanks!

   Akanksha, that's interesting, you are correct, I checked my eigen values and they were in fact smaller than those of Voigt.

   Thank you both.
   -Imad

   Report abuse

9. 

   Wenbin Yu @ 4:46 pm on 15 Nov 2017

   Although eigen value is a check, the diagonal terms of the stiffness matrix, Young’s moduli and shear moduli should also be ranked according to the specified orders. If one of these conditions is not true, something wrong with your results.

   Report abuse

10. 

    Kunal Samel @ 5:00 pm on 15 Nov 2017

    Hi Imad,

    Glad I could help!

     

    Report abuse