Muscle Material Activation

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts
  • ambaus
    Member
    • Jun 2008
    • 43

    Muscle Material Activation

    Hello,

    I am trying to use Syliva Blemker's muscle model. However, when I try to specify activation, can't get the problem to solve in FEBio. The log file says I have negative Jacobians. Is my geometry problematic? I've attached what I have done. Any suggestions would be greatly appreciated.

    Thanks,

    Anne
  • maas
    Lead Code Developer
    • Nov 2007
    • 3400

    #2
    Hi Anne,

    A couple of things to keep in mind. This model has been implemented very recently, so there might still be a couple of bugs in it. Although it passed our initial tests, some new issues have shown up in the last few days. We are currently working on ironing out all the kinks. This also applies to the tendon material.

    Another thing to keep in mind, is that we also have reasons to believe that the formulation itself is not always stable. We have found some cases where the formulation does not predict realistic results. (E.g. uni-axial incompressible compression predicts zero stress). This may also be a part of the cause why the models does not seem to run well.

    I will look at your file later this afternoon, but I've found that this model runs best using full-Newton. Try setting the max_ups parameter in the Control section to zero to use full Newton and see if that helps.

    Cheers,

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

    Comment

    • ambaus
      Member
      • Jun 2008
      • 43

      #3
      Hi Steve,

      This helped me to get Sylvia's model working but it doesn't help mine.

      Anne

      Comment

      • ambaus
        Member
        • Jun 2008
        • 43

        #4
        Muscle Activation not working?

        Hi Steve,

        I also tried using the same geometry as Sylvia's model, just a long rectangular prism, with no prescribed deformations, just activation. I get the same Jacobian errors (please see attached files). Is the activation of the muscle material not working?

        Thanks,

        Anne

        Comment

        • maas
          Lead Code Developer
          • Nov 2007
          • 3400

          #5
          Hi Anne,

          For the first model (MTU_contract), I believe your activation parameter is too big. Without activation, the stresses are in the range 1-2. With full activation (=1) you are adding a fiber stress of the order 1e+5, which is too big for this problem and is why you are getting inverted elements. Try setting the activation to a level where your fiber stresses are in the range 1-2. E.g. 1e-6.

          For the second problem (muscle_twitch) I believe you are experiencing the problem with the model that I had mentioned earlier. In Silvia's model neither the matrix nor the fibers resist uniaxial compression and I believe this to be the cause why your model is not running. Understand that this is not an FEBio problem, but a problem with the actual constitutive model. I would like to do some more testing on your problem to verify this. I will keep you posted.

          Cheers,

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

          Comment

          • ambaus
            Member
            • Jun 2008
            • 43

            #6
            Hi Steve,

            This helped. When I turned down the activation, I got the model to run. Thanks a ton. Let me know if you figure out any other nuances that may help me in future models with the muscle material.

            Anne

            Comment

            • Adam Baker
              Junior Member
              • Jan 2008
              • 29

              #7
              No contraction stronger than 150 kPa?

              As a follow-up to this... running the muscle_twitch.feb file with a variety of activation parameters, things seem to break down after applying 150 kPa of active stress. Is that pretty much not going to change? (i.e. no matter how gradually I apply the force)

              The thing is, 350 kPa is the active stress that everyone cites in the literature. Zajac 1989. Is that just not possible with the Blemker model?

              If not, back to isotropic hyperelasticity!

              Adam

              Comment

              • maas
                Lead Code Developer
                • Nov 2007
                • 3400

                #8
                Hi Adam,

                The implementation of Blemker's model has some issues. It might just a bug. Once I find some time, I'll take a look at this. I'll keep you posted.

                Cheers,

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

                Comment

                • HeikoStark
                  Junior Member
                  • Mar 2009
                  • 26

                  #9
                  Muscle and tendon material

                  Thanks for the very good software 'FEBio'.

                  Is there already an update for the muscle model? I have found some mistakes in your user manual related to the muscle and tendon material. The element 'lambda' was not found and I think it should be 'lam_max'. Why is your bulk modulus lower than the values in the article 'Blemker 2005 A 3D model of muscle ...'?

                  In the description of the 'Transversely Isotropic Mooney-Rivlin'-Material you let contract a ligament material from 'Weiss 1996 Finite element implementation...'. This is a little bit confusing! Additional, I found a mistake in your formula. In 'Weiss 1996...' the formula of the strain energy is C5*lambda + C6. In which article can I found the description of the contraction model?

                  Why can I transform different material models? For example the Hyper58 model from Ansys to your 'Mooney-Rivilin'-model (Yücesoy 2002 Three-dimensional finite element ...).

                  Cheers, Heiko.

                  Comment

                  • maas
                    Lead Code Developer
                    • Nov 2007
                    • 3400

                    #10
                    Heiko,

                    The muscle (and tendon) materials have not yet been updated due to a lack of time. At this time I can't make it a very high priority so I'm not sure when I'll be able to take a look at this material again.

                    The correct name for the fiber straightening stretch is indeed lam_max. I will update the manual accordingly.

                    The active contraction model follows the approach detailed in

                    Guccione, "Mechanics of active contraction in cardiac muslce, part I" J.Biomechanicla Engineering 1993

                    I'm not sure I understand your last question. Do you want to convert material parameters? I'm not familiar with Ansys or the Hyper58 material, so I'm not sure how helpful I can be here.

                    Let me know if you need further assistance with anything.

                    Cheers,

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

                    Comment

                    • HeikoStark
                      Junior Member
                      • Mar 2009
                      • 26

                      #11
                      Compressibility

                      Hi Steve!

                      Thank you for your fast answer. The last days I have read a few explanations about hyperelastic materials. So I hope I can answer the last question myself.

                      But I have new questions about your material definitions. It is right, that your materials 'Mooney-Rivlin' and 'Ogden' are defined compressible (like it shows in your formula), but the incompressibility is enforced by a penalty method? Is this also true for the trans iso Mooney-Rivlin material? Why did you not use the formulas for incompressible materials (without W_vol(J))?

                      Cheers Heiko.

                      Comment

                      • maas
                        Lead Code Developer
                        • Nov 2007
                        • 3400

                        #12
                        Hi Heiko,

                        It is numerically difficult to enforce incompressiblity exactly. This is because for a truly incompressible material, the pressure doesn't follow from the constitutive relationship and has to be determined otherwise. The penalty method is a cheap and easy way to simulate (nearly) incompressible materials. For this method the pressure is simply defined as p = dU/dJ. This is the method that is implemented in FEBio and therefore, materials that are to be used in an incompressible formulation need the U term. FEBio also has a more advanced method to enforce incompressibility, namely the augmented-Lagrangian method, although I must admit that that method currently doesn't seem to work very well. I need to look into fixing it. However, that method also requires the volumetric term. Hope that answers your question.

                        Cheers,

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

                        Comment

                        • HeikoStark
                          Junior Member
                          • Mar 2009
                          • 26

                          #13
                          Hi!

                          In the muscle material I have found a undocumented parameter (G3). This is used to fix the zero-stress problem, but I can't find a description or values in 'Blemker 2005 A 3D model of muscle ...' or in the user manual.

                          Can you help me?

                          Cheers Heiko.

                          Comment

                          • maas
                            Lead Code Developer
                            • Nov 2007
                            • 3400

                            #14
                            Hi Heiko,

                            The G3 parameter was indeed added to fix a problem with the Blemker material model. Setting it to zero gives you the unmodified Blemker model. Setting it to nonzero adds a fiber stress proportional to the fiber stretch. I'll see if I can add some more documentation to the user's manual, but keep in mind that this fix has never really been validated so use it with caution. I hope to find some time in the near future to fix this material model.

                            Cheers,

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

                            Comment

                            • HeikoStark
                              Junior Member
                              • Mar 2009
                              • 26

                              #15
                              Thank you.

                              Cheers,
                              Heiko

                              Comment

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