8 Replies Latest reply on May 6, 2014 12:14 AM by Jared Conway

    Large Displacement and Remote Loads

    Dermott McHugh

      Hi all,

       

      I am using a remote load in a simulation and am interested how the load changes in simulations using the Large Displacement solver.

       

      The help file seems to imply that the direct transfer is only suitable for small displacement. Does that mean that the rigid connection option takes in to account the displacement of the point that the remote load is applied to, and the direct transfer option does not?

       

      Thanks,

      Dermott

        • Re: Large Displacement and Remote Loads
          Jared Conway

          Probably Best to provide the full scope of your application.

            • Re: Large Displacement and Remote Loads
              Dermott McHugh

              I'm really just asking to try and better understand how the remote loading function behaves in simulations with large displacement.

               

              My model is basically a cantilevered beam (Stiff material: E ~ 100 GPa) on which I will be applying a remote load to the end face. The root of the beam is "potted" in much less stiff material (E ~ 1 GPa). There is a bonded connection between the beam and the potting material and the outside of the potting material is fixed.

               

              The load is applied to another member (not modeled) which is attached to the end of the beam. I am only interested in the stresses in the beam and the other member is sufficiently rigid, so I would like to use a remote load.

               

              Since the potting material deforms significantly, there are moderate displacements out where the load is applied, and I am wondering whether the remote load will account for that displacement, or whether it is ignored (as in small displacement). As end face of the beam rotates, will the load direction rotate or stay constant with respect to the global coordinate system?

               

              -Dermott

                • Re: Large Displacement and Remote Loads
                  Jared Conway

                  load direction never changes in static analysis regardless of whether you use large disp or not

                  and actually even with nonlinear it doesn't update unless you tell it to update

                   

                  so i think you actually have 3 questions

                  1. what remote load to choose (not sure which you chose)

                  2. does large displacement support remote loads (i believe it does)

                  3. does the large displacement solver include load direction updating (negative)

                    • Re: Large Displacement and Remote Loads
                      Dermott McHugh

                      Alright, thanks, that is helpful. I was forgetting the difference between changing the direction of the load and changing the location of load application.

                       

                      I was able to run a model with the large displacement option, and both types of remote load, so they both appear to be supported.

                       

                      I guess my other question is whether the displacement of the point of load application is taken in to account in large and/or small displacement studies. The load applied to my beam has a fairly large component along the long axis of the beam. So as the simulation runs in large displacement, the bending moment in the first step may be significantly smaller than in the last step (due to displacement of the point of load application). Does that make sense? If that makes sense, do you know if that would be accounted for in large/small displacement studies?

                       

                      Thanks,

                      Dermott

                        • Re: Large Displacement and Remote Loads
                          Jared Conway

                          what effect do you think it will have on the result?

                           

                          its really a moot point because you don't have any control over the solver in static

                           

                          but i'm not following what you think the effect will be

                            • Re: Large Displacement and Remote Loads
                              Dermott McHugh

                              If the point at which the load is applied does not displace with the faces it is connected to, I imagine it would not produce realistic results in studies with large displacement. I used a coordinate system to get the direction for the remote load. It seems like if the location that the remote load is applied through did not displace (but the rest of the model does), then the loads and moments transfered to the model would be inaccurate.

                               

                              In any case, I ran a few test simulations, and it appears that it is indeed a good idea to use the rigid connection option in large displacement, as the direct transfer option started to cause some artifacts on the face the load was applied. The test study I used was simply applying a load to the end face of a cantilevered beam (one end face fully fixes). The beam was 1x1x10", and I applied the remote load to a point 5" off the end of the beam at a 45 degree angle to the long axis of the beam, putting it into compression and bending. I tested the rigid connection and the direct transfer options, and compared them to a model which had a 1x1x5" rigid body bonded to the end of the beam, with the same force applied to its end face.

                               

                              In large displacement studies, the rigid connection remote load gave exactly the same results as actually using a rigid body to replace the missing geometry with a force applied to it. The direct transfer option gave a slightly different max stress, and also had some stress artifacts at the loaded face.

                               

                              Curiously, I tried to compare it to a non-linear model with the rigid body, and the rigid body did not rotate, which gave clearly erroneous results. Have you ever had a problem like that with rigid components in non-linear studies?