Sliding Interfaces

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts
  • Joanne Kwon
    Junior Member
    • Jan 2008
    • 13

    Sliding Interfaces

    Hi, everyone.

    For a part of my PhD I am trying to model spherical nanoindentation experiment on hydrogel.
    Using FEBio, I created a Mooney-Rivlin based poro-elastic box to represent hydrogel and spherical rigid body to represent the indenter.

    Using sliding interface contact, I need to declare a slave and a master surfaces. In the manual it said "usually advisable to select the more finely meshed surface as a slave", but is it really important to do so?

    Moreover, would it be better to use the "two way algorithm" if I cannot figure which surface should be a slave or a master?

    Many thanks,
    Joanne
  • maas
    Lead Code Developer
    • Nov 2007
    • 3400

    #2
    Hi Joanne,

    It is usually advisable to pick the more finely meshed surface as the slave. However, if this is always the best solution depends on the problem. There is really no golden rule, but here are some tips from my personal experience.

    If one of the objects is rigid I usually pick the rigid object for the master and apply a single-pass. If both objects deform and have similar material parameters, I pick the two-pass algorithm. Which surface is the slave is off course not important then. If it is not obvious which approach to take, and you can afford to spent some extra time on it, try some different approaches and see which one works best. And by best I do not necessarily mean the approach that gives the fastest solution. Also inspect the results visually and look for penetration. If none of the approaches gives you satisfactory results, try increasing the penalty and/or decreasing the tolerance and see if that helps. The contact algorithm is very sensitive to these two parameters and it might take some trial and error to find a set that works. To start with I always pick a tolerance of 0.1 and a penalty that is close to the bulk modulus of the softer material.

    Also, FYI, we are planning to add some tools to FEBio to make all this decision making a lot easier in the future.

    I hope this helps. Let us know if you have any other questions.

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

    Comment

    • yantaaa
      Member
      • Jun 2011
      • 38

      #3
      Originally posted by maas View Post
      Hi Joanne,

      It is usually advisable to pick the more finely meshed surface as the slave. However, if this is always the best solution depends on the problem. There is really no golden rule, but here are some tips from my personal experience.

      If one of the objects is rigid I usually pick the rigid object for the master and apply a single-pass. If both objects deform and have similar material parameters, I pick the two-pass algorithm. Which surface is the slave is off course not important then. If it is not obvious which approach to take, and you can afford to spent some extra time on it, try some different approaches and see which one works best. And by best I do not necessarily mean the approach that gives the fastest solution. Also inspect the results visually and look for penetration. If none of the approaches gives you satisfactory results, try increasing the penalty and/or decreasing the tolerance and see if that helps. The contact algorithm is very sensitive to these two parameters and it might take some trial and error to find a set that works. To start with I always pick a tolerance of 0.1 and a penalty that is close to the bulk modulus of the softer material.

      Also, FYI, we are planning to add some tools to FEBio to make all this decision making a lot easier in the future.

      I hope this helps. Let us know if you have any other questions.

      Steve.
      Hi Steve,


      To start with I always pick a tolerance of 0.1 and a penalty that is close to the bulk modulus of the softer material.
      I have a question about the bulk modulus you metioned.

      If I use neo-Hookean material for solid phase in a biphasic model, the definition in preview for neo-Hookean material is E and v. I think E is for Young's modulus and v is for possion's ratio? As for the solid phase of the articular cartialge, I adopted 0.5 MPa for E and 0.02 for possion's ratio. As bulk modulus K=E/3(1-2v), K should be 0.1736. So as you metioned, I should adopt a penelty that is close to 0.1736?

      But even if I adopt 1 for penelty, there is overlap between the two contacting surfaces. Is 0.1736 too small? Or possion's ratio in the definition of K (K=E/3(1-2v)) is for the solid matrix rather than the two phases as a whole, because in this way, the penelty can be much higher.

      Thanks in advance.

      Regards,
      Jerry

      Comment

      • maas
        Lead Code Developer
        • Nov 2007
        • 3400

        #4
        Hi Jerry,

        For a biphasic problem the effective bulk-modulus of the mixture can be much higher than the elastic bulk modulus so a higher penalty factor is advisable. Also know that the suggestion of using a penalty factor that is in the order of magnitude of the bulk modulus is only a very rough guess. Other factors, such as element size, can have a large effect on the final penalty factor that works. The bottom line is that you'll have to play with this parameter a bit to find one that works for your problem. Hope this helps.

        Cheers,

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

        Comment

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