Diffusion Problem employing a Triphasic material model

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts
  • Squid
    Member
    • Oct 2012
    • 76

    Diffusion Problem employing a Triphasic material model

    Dear All,

    I am trying to run the attached T1.feb but it doesn't work. Can someone explain to me why? I really don't understand the reason.

    Thanks a lot.

    Best,

    Leo
    Attached Files
  • maas
    Lead Code Developer
    • Nov 2007
    • 3385

    #2
    Hi Leo,

    It's failing because your residual becomes a NAN. The question now becomes why does it do that? We'll look into it and let you know what we find.

    Cheers,

    Steve.
    Department of Bioengineering, University of Utah
    Scientific Computing and Imaging institute, University of Utah

    Comment

    • Squid
      Member
      • Oct 2012
      • 76

      #3
      Dear Steve,

      yes indeed...I am trying to figure out why but really I cannot understand. I really appreciate your help.

      Thanks a lot.

      Best,

      Leo

      Comment

      • ateshian
        Developer
        • Dec 2007
        • 1811

        #4
        Hi Leo,

        From a numerical perspective, the electric potential for triphasic materials is obtained from the solution of a quadratic equation whose coefficients depend on the solute concentrations and the fixed charge density. When some (or all) of these coefficients are zero (as occurs at the first time point in your model, since you set the initial concentrations and the fixed charge density in the triphasic domain to zero), the calculation of the electric potential evaluates the log of zero, thus producing NAN. Even if the values of the concentration are not exactly zero, starting with very small concentrations can produce negative concentration values (due to numerical roundoff errors) which also result in NAN when evaluating the logarithm. So it is best to use non-zero concentrations when performing triphasic analyses. This problem does not arise when solutes are neutral, since the electroneutrality condition is not used to evaluate the electric potential in that case.

        You can overcome this problem in your analysis by changing the initial and boundary conditions as follows: Instead of using 0 initial concentration inside and prescribing 1 on the boundary, try using 1 for the initial concentration inside and 2 on the boundary. This should resolve the issue.

        Best,

        Gerard

        Comment

        • Squid
          Member
          • Oct 2012
          • 76

          #5
          Dear Prof. Ateshian,

          thanks for the answer. I set:

          - FCD=1
          - initial concentrations (1 and 2)=1
          - initial fluid pressure=1 (I tried also with 0)
          - prescribed concentrations=2
          - prescribed pressure=-0.00247

          but i get always the same error. The modified version T1_V2.feb is attached. There could be another reason? Thanks a lot for the help.

          Best,

          Leo
          Attached Files

          Comment

          • Squid
            Member
            • Oct 2012
            • 76

            #6
            Dear All,

            sorry, I've found the reason of the issue. The material properties were inconsistent! I was using the drag coeff value instead of the permeability. Furthermore, I was using the diffusivity values in m2/s instead of mm2/s.

            It is working now!!!

            Thanks a lot for the help and sorry for the basic mistake.

            Best,

            Leo

            Comment

            • ateshian
              Developer
              • Dec 2007
              • 1811

              #7
              Hi Leo,

              Great, I am glad it worked out.

              Best,

              Gerard

              Comment

              Working...
              X
              😀
              😂
              🥰
              😘
              🤢
              😎
              😞
              😡
              👍
              👎