Announcement

Collapse
No announcement yet.

ABAQUS VUMAT for the Neo-Hookean model

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • ABAQUS VUMAT for the Neo-Hookean model

    A single-precision Abaqus VUMAT subroutine for the Neo-Hookean (NH) model. The subroutine is an example of how to write a VUMAT. The subroutine only works for plane-strain, axisymmetric, and 3D-elements.
    Attached Files
    Jorgen Bergstrom, Ph.D. PolymerFEM Administrator

  • #2
    Vumat

    Hello Doctor;
    I view the VUMAT, but when I have run Abaqus, it did not market, is that you can send me the file (.INP), for your example.
    Thank you
    doctorant polymer

    Comment


    • #3
      You can get exemplar input files here.
      Jorgen Bergstrom, Ph.D. PolymerFEM Administrator

      Comment


      • #4
        VUMAT condition

        Hello Dr;
        You below the condition that I applied to initialize the FE that correcpond in my simulation gradien elastic.
        Abaqus compile when I got started in the calculation and stops immediately, and when I calculate this condition normally, but on this condition is necessary to get good results.
        So is what you think is the right way to proceed?

        IF (totTime.EQ.0.0) THEN
        C setup FE (upper diagonal part)
        FE(1,1) = 1.02
        FE(2,2) = 1.039
        FE(3,3) = 1.059
        FE(1,2) = 1.079
        if (nshr .eq. 1) then
        FE(2,3) = 0.0
        FE(3,1) = 0.0
        FE(2,1) = 0.0
        FE(3,2) = 0.0
        FE(1,3) = 0.0
        else
        FE(2,3) = 1.0099
        FE(3,1) = 1.029
        FE(2,1) = 1.0049
        FE(3,2) = 1.079
        FE(1,3) = 1.809
        end if
        END IF
        doctorant polymer

        Comment


        • #5
          At time=0, should not FE = unit matrix ?
          Why do you have non-zero values for FE?
          Jorgen Bergstrom, Ph.D. PolymerFEM Administrator

          Comment


          • #6
            Hello Dr;
            Does it can introduce the Norton parameter (K and n) without Abaqus has gone through a routine for even the effect rate?
            Strain-rate = (sig/K)**n
            Thank you
            doctorant polymer

            Comment


            • #7
              The flow equation that you mention actually work rather well for many different types of polymers. Abaqus has two a model called "Two layer viscoplasticity" that I believe supports that flow rule. I have not had much luck with that model, however.

              Another approach is to use the *Hysteresis model in Abaqus. That model is also based on that flow rule.

              A third option is to write your own user-material.

              - Jorgen
              Jorgen Bergstrom, Ph.D. PolymerFEM Administrator

              Comment


              • #8
                Abaqus

                Dear Doctor
                Can you tell me what the « Lemaître » model exists in Abaqus?
                I introduced the viscosity parameters of in Abaqus, but I did not understand what "m and f?
                Thank you
                doctorant polymer

                Comment


                • #9
                  I don't recall a Lemaitre model in Abaqus. If there is such a model, can you provide a reference?

                  -Jorgen
                  Jorgen Bergstrom, Ph.D. PolymerFEM Administrator

                  Comment


                  • #10
                    damage parameter "D".

                    Dear Dr ;
                    What I realize in order to Abaqus
                    To calculated of the damage parameter "D".
                    Is that there is a possibility to calculate this parameter "D", for me it’s the lemaitre model
                    In attached you find a page on the model.
                    Thank you
                    Attached Files
                    doctorant polymer

                    Comment


                    • #11
                      Thanks for the info. For some reason, the thumbnails are very small and and not easily readable.

                      I am not sure I understand what the "D" parameter is, so I don't know how to calculate it.

                      -Jorgen
                      Jorgen Bergstrom, Ph.D. PolymerFEM Administrator

                      Comment


                      • #12
                        Ok once again i posted too quickly by searching in the archives i found some interesting post i have to understand first.
                        My understanding is that anything rigid body is controlled by abaqus explicit if one wants by using U, and it is equivalent to F and V in this framework.

                        original message:
                        Dear Jorgen. I am trying to understand this vumat before wirting my own (TNM based). But i am struggling at one point.
                        By going backward in the code.
                        I am all ok for:
                        c Stress = mu/J * Dev(Bstar) + kappa*log(J)/J * Eye
                        and for
                        c Bstar = J^(-2/3) F Ft (upper diagonal part)

                        But my feeling by reading the code is that instead of calculating F Ft you take the square of the stretch tensor. I would have no problem with that if ABAQUS was proposing the left stretch tensor ( F = V*R => F * Ft = V* R * Rt * Vt = V² V being symmetric) but i read ABAQUS documentation and it says that VUMAT communicate the RIGHT STRETCH TENSOR F = R * U !

                        So i don't understand something. Can you explain ?

                        By the way the ABQ doc says that the user should use U instead of F to write hyperelastic law but i don't really see how.
                        Last edited by Ajaj; 2012-11-02, 07:42.

                        Comment


                        • #13
                          Originally posted by Ajaj View Post
                          Ok once again i posted too quickly by searching in the archives i found some interesting post i have to understand first.
                          My understanding is that anything rigid body is controlled by abaqus explicit if one wants by using U, and it is equivalent to F and V in this framework.

                          original message:
                          Dear Jorgen. I am trying to understand this vumat before wirting my own (TNM based). But i am struggling at one point.
                          By going backward in the code.
                          I am all ok for:
                          c Stress = mu/J * Dev(Bstar) + kappa*log(J)/J * Eye
                          and for
                          c Bstar = J^(-2/3) F Ft (upper diagonal part)

                          But my feeling by reading the code is that instead of calculating F Ft you take the square of the stretch tensor. I would have no problem with that if ABAQUS was proposing the left stretch tensor ( F = V*R => F * Ft = V* R * Rt * Vt = V² V being symmetric) but i read ABAQUS documentation and it says that VUMAT communicate the RIGHT STRETCH TENSOR F = R * U !

                          So i don't understand something. Can you explain ?

                          By the way the ABQ doc says that the user should use U instead of F to write hyperelastic law but i don't really see how.

                          Hi Ajaj,

                          I had the same question a while ago - I think this document would answer your question regarding using U^2 instead of V^2.

                          http://imechanica.org/files/appendix3-vumat.pdf

                          page A3.29.

                          Due to the corotational frame, you are actually calculating the corotational portion of the Cauchy stress:
                          sigma = R * sigma_corot * Rt

                          using the corotational portion of the left Cauchy-Green deformation tensor:
                          B = F * Ft = R * U * Ut * Rt = R * U^2 * Rt

                          Comment


                          • #14
                            Quick question for everybody !
                            Why a high Bulk modulus over shear modulus ratio leads to high kinetic energy in an Explicit analysis and unstable results ?
                            I'm using Dr. Bergstrom's subroutine (NeoHookean) to model indentation. I works and I'm really thankful as this subroutine helped in understanding the basic process of using a typical VUMAT.
                            However the elements under the indenter which suffer high compression start distorting and the analysis stops. The built-in NeoHookean doesn't show this problem.
                            The higher the Bulk modulus over shear modulus ratio, the smaller the strains at which the distorsion occurs. However the kinetic energy becomes significant even for little mass scaling.
                            I would be glad if someone can explain the mechanism that triggers that and also discuss about what the example subroutine provided may be missing.
                            Thanks
                            Christos

                            Comment

                            Working...
                            X