Triphasic Boundary Conditions

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts
  • birdave
    Junior Member
    • Dec 2014
    • 6

    Triphasic Boundary Conditions

    Well, this seems simple enough. The only parameters to set are the boundary conditions, and yet I somehow managed to screw it up. I want to swell spheres in NaCl. I set initial conditions and effective pressure on the outer nodes of this 1/8 sphere, but the energy and residuals are nan, or alternatively negative Jacobian throws error. How do I set the initial concentration inside the sphere, isn't that the fixed charge parameter in triphasic material? Anyhow, this is my first FEBio file, so it should be easy to fix. Please see Gel.prv file attached. Any suggestions welcome.

    David
    Attached Files
  • ateshian
    Developer
    • Dec 2007
    • 1824

    #2
    Hi David,

    Your initial conditions need to be consistent with the prescribed boundary conditions at time t=0. In your case the prescribed effective concentrations and fluid pressure start at zero, but the initial conditions are non-zero.

    You should also make sure that your initial effective fluid pressure pe is consistent with initial effective concentrations ce. In particular, pe = -R*T*Sum(ce). You use R=8.314e-06 and T=298, and you have two solutes, each with ce=2 initially. This means that pe=-0.00991 (not -0.7). So you need to adjust these values (either ce or pe).

    Finally, you should set up your problem so that there isn't much swelling occurring at the first time step of your analysis, i.e., the swelling should progress steadily over time so that the analysis can converge properly. (This is equivalent to saying that the loads in a finite elasticity analysis should not be too large initially or else the deformation may be too great and the analysis would not converge.) This can be done for example by ramping up the fixed-charge density from zero to the desired value using a steady-state analysis.

    Best,

    Gerard

    Comment

    • ateshian
      Developer
      • Dec 2007
      • 1824

      #3
      Hi David,

      Your initial conditions need to be consistent with the prescribed boundary conditions at time t=0. In your case the prescribed effective concentrations and fluid pressure start at zero, but the initial conditions are non-zero.

      You should also make sure that your initial effective fluid pressure pe is consistent with initial effective concentrations ce. In particular, pe = -R*T*Sum(ce). You use R=8.314e-06 and T=298, and you have two solutes, each with ce=2 initially. This means that pe=-0.00991 (not -0.7). So you need to adjust these values (either ce or pe).

      Finally, you should set up your problem so that there isn't much swelling occurring at the first time step of your analysis, i.e., the swelling should progress steadily over time so that the analysis can converge properly. (This is equivalent to saying that the loads in a finite elasticity analysis should not be too large initially or else the deformation may be too great and the analysis would not converge.) This can be done for example by ramping up the fixed-charge density from zero to the desired value using a steady-state analysis.

      Best,

      Gerard

      Comment

      • birdave
        Junior Member
        • Dec 2014
        • 6

        #4
        Gerard,

        It is indeed a real honor to be addressed by a professor of your distinguished standing! I finally found the curve editor where I could set the initial concentration/pressures. I have entered some values previously prescribed to no avail (still Nan). Perhaps you have an idea about what is going wrong here. Please see file attached.
        Attached Files

        Comment

        • birdave
          Junior Member
          • Dec 2014
          • 6

          #5
          I noticed the "fluid density" for the material was at "0". Could this be the cause of "Nan"?

          Comment

          • ateshian
            Developer
            • Dec 2007
            • 1824

            #6
            Hi David,

            Here are some things you can do to get your analysis running:

            - Your initial conditions should be applied to the entire domain, not just the surfaces on which the boundary conditions are prescribed. Use the "Select Part" icon in PreView to select the domain. This is a major issue.

            - You specified R=8.314e-6 and T=298, with concentrations of 2000 for both solutes, this means that the effective pressure for your initial and boundary conditions should be -9.910288 (this is slightly different than the values you used). This is a minor issue.

            - You selected 4 time increments of 0.25. This means that your first time step is at 0.25, which may be too great of a change in boundary conditions in one step. If you fix the above issues and still can't converge, I recommend using 10 time increments with a step of 0.1.

            The fluid density should not affect the analysis.

            Best,

            Gerard

            Comment

            • ateshian
              Developer
              • Dec 2007
              • 1824

              #7
              Hi David,

              Here are some things you can do to get your analysis running:

              - Your initial conditions should be applied to the entire domain, not just the surfaces on which the boundary conditions are prescribed. Use the "Select Part" icon in PreView to select the domain. This is a major issue.

              - You specified R=8.314e-6 and T=298, with concentrations of 2000 for both solutes, this means that the effective pressure for your initial and boundary conditions should be -9.910288 (this is slightly different than the values you used). This is a minor issue.

              - You selected 4 time increments of 0.25. This means that your first time step is at 0.25, which may be too great of a change in boundary conditions in one step. If you fix the above issues and still can't converge, I recommend using 10 time increments with a step of 0.1.

              The fluid density should not affect the analysis.

              Best,

              Gerard

              Comment

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