4 Replies Latest reply on Apr 25, 2013 4:59 PM by Jared Conway

    Large Deformation / Reality Check

    Tony Gatta

      Greets all.

      This 'experiment' was purely for kicks and I just wanted to share with the group to see

      1. if my assumptions are correct and/or

      2. I just got lucky

       

      I modeled a stainless steel rule (ie shop ruler, scale, etc).  250mm long x ~30mm wide x 1mm thick & ran a nonlinear

      simulation (shell elements) with one end fixed and the other with a prescribed rotation of 180* (3.14 radians).

       

      I'm sure most of you have seen this before.  I didn't go to 360 as am somewhat attached to my (still straight) ruler.

       

      I duplicated the test as best I could in my garage.. put the rule in a vise and used a force gage to bend it around

      180*.  I sort of eye balled things a bit, not very rigorous.. but again, just a curiosity.

       

      See image attached.

       

      SW Simulation predicted ~60 in-lbs reaction moment.

      My measurement = ~ 58 in-lbs

       

      So back to my question(s):

      1. the simulation shows a purely circular deformed shape.  I assume this is because I'm using shell elements,

      have a constant moment through the beam.. and no stress calculated throughout its cross-section.  ie.. only

      the reactions would be valid -- true / false?

       

      2. the 'real' test shows more of the 'parabolic' bending I'd expect.  I don't think the angle or overshoot of my

      force gage position should matter much in that I (think) I would always end up with the same reaction moment.

       

      The image may not look like 180 rotation though I'm quite sure I managed as much.  May be lens distortion from

      camera.. or I lightened up a bit while trying to shoot with the other hand, but you're seeing max force reading at

      ~180 degrees.

       

      Thoughts?

        • Re: Large Deformation / Reality Check
          Alin Vargatu

          I do not believe the fixture is the same. You might want to split the flat faces and add fixtures on a small area of them.

          • Re: Large Deformation / Reality Check
            Jared Conway

            Nice validation case! Pretty close. I think if you can match the way the load and restraints are applied you can get a bit closer but over all things are pretty close. Can you post your simulation results and file to see if there is anything else that can be adjusted. I think the restraint might not perfectly match like Alin suggested, but matching the way the load is applied will be hard. Maybe add one more part to get the motion exact or play with the material properties in the assumption they aren't perfect.

              • Re: Large Deformation / Reality Check
                Tony Gatta

                I don't think splitting the shell face and adding a fixed restraint on a flat face vs the straight edge

                should make any difference -- the first 'column' of shell elements can't move (no trans or rot) --

                but I can give it a try.

                 

                Though I think everyone would agree that you should never mess with a model thats giving

                you the answer you want.

                 

                I did try a force based analysis.. non linear / following force -- but I couldn't get the model

                to deform correctly.  And solution time took at least 10X that of displacement based.

                 

                I've attached my model and CWR file as a single ZIP if anyone is curious.

                 

                Jared -- not sure I follow about adding an addition part.  I certainly could tweak mat'l props to get

                the same exact answer -- but 1) that'd be cheating, no? and 2) i don't know the exact material of

                the ruler -- it just says "stainless" on it.  Nonetheless, any stainless will be about the same

                result +/-10% probably.

                 

                I'd just like to verify that my assumptions (about stress being meaningless but reaction force

                being correct) are true.  IE me and Simulation are making the same assumptions.  And I'm not

                just fooling myself with a lucky hit.

                 

                Again, any insight, greatly appreciated.

                  • Re: Large Deformation / Reality Check
                    Jared Conway

                    I don't think I understand why you don't trust the stress. The displacement, shape and reaction forces are what you expect, why wouldn't the stress be valid? Those are pretty good validations, the only other way to check would be a hand calc, but you've done better through physical testing.

                     

                    The reason I suggested a different pulling method is because based on the description of your setup, the displacement around an axis, I would expect a perfectly circular bend. In your real test you're pulling on it and restraining it slightly differently. The only way to get an exact match is with an exact setup.

                     

                    Overall, you could say you got lucky with the setup, but everything points to a valid analysis as far as I can see. You could throw some strain gauges on maybe for another validation point I guess, but I think you're good.