9 Replies Latest reply on Apr 14, 2015 11:21 AM by Alexander Salimian

    Moment on part with fixed constraint.

    Adam French

      Hello All,

       

      I'm simulating a so leg made of steel tubing that is being glued down into a groove. (For a product we're deploying next week!)

       

      I need to determine the moment exerted on the glue joint (modeled by the fixed constraint in the attached file) by the loads at the top of the leg, which are modeled on a small face in this test part, but which will be exerted by a pin joint  in the actual assembly.

       

      Attached is a simplified version. I'm looking for the "free body force" as a moment around axis 1.

       

      I talked with tec support with my full assembly (which meshes and runs), and they said the process is select 1) axis1, 2) point 1, and 3) the constrained face. At the time on the phone I was using my actual parts, and when I said "that doesn't match my hand calcs" they said "we can't get into that level of detail, but that's how you use the tool."

       

      Now I've got a simplified part loaded up (attached), and it should be clear that a 10N force on a .5 meter moment arm should result in a moment about axis 1 of 5Nm.

       

      When I follow the tech-support procedure above I get Mz=-2.4Nm.That's not right.

       

      I've scratched my head on it for several hours, read the forums, all that, could really use a hand - this is for work, we're releasing this week, and we would really like confirmation that these parts were designed right.

       

      Thanks,

       

      Adam

        • Re: Moment on part with fixed constraint.
          Shaun Densberger

          I tried your model and noticed:

          1. I get very different results depending on whether I'm using shell or solid elements. I get the right answer if I use solid elements, although it seems SW outputs the moment values with respect to a new CS (instead of the WCS).
          2. It seems that the free body forces are half of the reaction force in the shell element model. This effect is not present when using solid elements.
          3. It shouldn't matter, but out of curosity I tried to place a point on Axis1 at the intersection point with the Right Plane to see if it changed the results at all. However, for whatever reasons, SW wouldn't let me do that (placing points in SW seems to be such a cumbersum task).

           

          Solid

          Moment Solid.png

           

          Moment Solid 2.png

           

          Shell

          Moment Shell.png

           

          Moment Shell 2.png

            • Re: Moment on part with fixed constraint.
              Adam French

              Shaun,

               

              Can I ask what your mesh settings were for the sample part I posted? I've had issues in the past with meshing tubing parts as solids, so had gone right to shells for this assembly.

               

              Appreciate you looking into this.

               

              Thanks,

               

              Adam

              • Re: Moment on part with fixed constraint.
                Adam French

                Shaun,

                 

                Turns out meshing the model as a solid was not a problem, so thanks for showing that's possible.

                 

                As for the results you get, all the other numbers except the moment around z appear random, so I don't think I can trust the z calc for non-obvious cases (like my actual parts).

                 

                Your numbers:

                Fx: 3.5N (small unexplained reaction force in x)

                Fy: -497N (really? the reaction force to a 10N load is -497N? Unlikely).

                Fz: .0357N (ok, close enough to zero to almost believe.

                Mx: -12.4Nm (what?)

                My: -0.0823Nm (ok, we'll call that noise to, but FEA should be much closer than this)

                Mz: -4.97Nm (yes, this is close to hand-calc 5.0, but .03Nm is actually more error than I would expect).

                 

                Any other ideas?

                 

                Next I'm trying:

                - installing SW 2014 on my machine and trying again.

                - other simple cases with free body and reaction forces.

                 

                Thanks,

                 

                Adam

                  • Re: Moment on part with fixed constraint.
                    Shaun Densberger

                    As for the results you get, all the other numbers except the moment around z appear random, so I don't think I can trust the z calc for non-obvious cases (like my actual parts).

                     

                    Your numbers:

                    Fx: 3.5N (small unexplained reaction force in x)

                    Fy: -497N (really? the reaction force to a 10N load is -497N? Unlikely).

                    Fz: .0357N (ok, close enough to zero to almost believe.

                    Mx: -12.4Nm (what?)

                    My: -0.0823Nm (ok, we'll call that noise to, but FEA should be much closer than this)

                    Mz: -4.97Nm (yes, this is close to hand-calc 5.0, but .03Nm is actually more error than I would expect).

                     

                    Any other ideas?

                     

                    Yea, there seems to be some unexpected numbers, although the -497 N is expected ([10 N * 500 mm / 10 mm] = 500 N).

                     

                    That being said, I have no idea why:

                    1. There are some weird numeric reaction loads. I'll try to play around with it in hopes of getting a better understanding, but so far I haven't got a clue short of a bug in the code.
                    2. Shell element give half the value that solid elements do. Again, all I can think of is a bug in the code.

                     

                    I tired this on both 2014 and 2013.

                     

                    I tested this in another package and got the expected results, so it's not an issue with the finite element method.

                     

                    Moment.png

                • Re: Moment on part with fixed constraint.
                  Jared Conway

                  version/service pack?

                   

                  have you tried just a simple tube or cylinder? just to make it as simple as possible?

                    • Re: Moment on part with fixed constraint.
                      Adam French

                      Jared,

                       

                      I'm on 2012, SP4.0.

                       

                      I've tried on simple shelled parts, with similarly strange results that ultimately led me to call tech support. I'll take another crack at it with solids. I had been using shells because I've had diffiulty meshing tubing parts in the past. I'll try that again, but any tips on mesh settings for tubing parts would be welcome.

                       

                      Thanks,

                       

                      Adam

                    • Re: Moment on part with fixed constraint.
                      Adam French

                      So for those following along at home, here are my best numbers yet on a sample part:

                       

                       

                      You'll notice that in this configuration I've selected Axis 1, the Point1, and Face<1>. Face 1 is actually the face where I'm applying the force - which doesn't completely make sense as to why that would work, but look at the numbers:

                       

                      Fx: -0.201 N (ok, small unexplained force along x at 2% of load, strange but not impossible).

                      Fy: 10N (Yes, that's the correct value).

                      Fz: -6.22e-5 (Yes, that's a nice small value for a Fz componenet introduced as a result of bending. Direction and magnitude seem reasonable).

                      Mx: 0.25Nm (hmmmm, that's not the product of Fx and the moment arm (.5m) so that's strange. 5% calculated Mz, so not insignificant).

                      My: 0.00505Nm (Sure, small unexplained value for My less than 0.1% of calculated Mz, I can live with that).

                      Mz: 4.99Nm (Yes, this is the correct answer, 0.02% variance from hand calc Mz.

                       

                      So this may well be the right way to use the Reaction Force/Freebody Force tool. Very strange paradigm. If it is, the challenge is now how to use this in an assembly, because in my actual assembly there is a pin joint through the tube behind Face<1> that ties this leg (and three others like it) to a flat surface.

                       

                      Next Step:

                      Adding an install of 2014 to my machine, running some more simple cases, and calling a buddy who's done a bunch of simulation. Welcome your thoughts, or frankly any solid examples of using the resultant force tool.

                      • Re: Moment on part with fixed constraint.
                        Alexander Salimian

                        When you view the free body force selecting an axis will present a second reference frame in the lower right of the view, it appears as a green cylinder with Z and R labeled. Picking an axis only gives you the moment about that axis in the direction of the local Z (along the axis). The sum in the X and Y directions appear to be residual moments which seem to decrease as mesh is improved.

                         

                        By not selecting a constraining axis, and only selecting a point and face (where your restraints are), you get the close to the expected results in X, Y, and Z. My problem was a 1 m long rod with 1 N load in the X direction and 1 Nm torque at the end in the -Z direction.

                        moments.png