5 Replies Latest reply on Oct 29, 2016 7:02 AM by Jason Smith

    Beam Bending Stress (Ms/Ss, Mt/St) vs. Render Beam Profile

    Jason Smith

      In order to verify beam simulation results on a simply supported beam vs. hand calculations, I'm modeling a piece of 3" x 3" x 3/16" angle iron, 120" long with 3 loads of 100lb each located at the centroid 30" apart, but acting on an inner face. When I run the simulation and "render beam profile" with bending stresses in DIR1 and DIR2 directions, the results are dead on to my hand calcs. But when I plot the "Bending Stress in DIR1" and "Bending Stress in DIR2" directions without "render beam profile", I get results which don't resemble anything I've calculated. Sometimes they are 2-3x higher, sometimes lower, depending on the model.

       

      In my model, the 100lb loads only cause a bending moment applied about DIR 2, the shear and bending diagrams in both DIR1 and DIR2 match hand calculations. The beam is restricted from rotation at its end faces, but can translate at one end.  According to the help file, my assumption would be that solidworks is plotting the max tensile fiber stress at an element cross-section when I don't "render beam profile". In my case, that should be 12699 lbf/in2 at max moment about DIR2, and solidworks plots this as 12650. Close. But then solidworks plots the bending stress about DIR1 as 21197 lbf/in2, despite the fact that there is no shear force to cause this, then gives a worst case value by adding those two. What?

       

      Can anyone explain this? Thanks.

        • Re: Beam Bending Stress (Ms/Ss, Mt/St) vs. Render Beam Profile
          Jason Smith

          So in light of the above, I ran a total of 5 beam load bending tests with different loadings and cross-sections. 3 were with the same loading as described above, one with 3"x2"x1/8" rectangular tube, one with 3"x3"x3/16" angle, and one with a 2"x1"x1/8" square tube with a 1.5"x2"x1/8" angle plug welded to the top. Using the angle and rectangular tube models, I made 2 more versions, adding a side load of 4x 100lb equally spaced along the length. In all cases, plotting bending stress in DIR1 and DIR2 with "render beam profile", the bending stress matched dead on to my hand calcs. In 4/5 cases, without "render beam profile", the stress values were total garbage, with no resemblance to anything calculated. In the case of the rectangular tube with top and side loadings, the plot did finally display the max fiber stress at the cross-section. However, that was the only correct -Ms/Ss or -Mt/St plot of the 5 models.

            • Re: Beam Bending Stress (Ms/Ss, Mt/St) vs. Render Beam Profile
              Keith Frankie

              My guess is that you're getting expected results when applying loads to a symmetric beam, where loads are applied along the beam's principal axis of inertia.

               

              I think the issue lies in loading profiles, like the angle iron, at an angle to the axis of symmetry.  This results in bending that happens out of plane with the loading.  This side bend can be seen when looking 'down the beam' in a results plot with deformation turned on.

               

              I found this video helpful: Introduction to Unsymmetric Bending (1/2) - Mechanics of Materials - YouTube

               

              Troubleshooting is also easier if you upload the SolidWorks files you're testing.  I think the attached file is a copy of your loading conditions.

                • Re: Beam Bending Stress (Ms/Ss, Mt/St) vs. Render Beam Profile
                  Jason Smith

                  Hi Keith,

                   

                  Thanks for your response. Yes, that model is pretty much the same as mine, only difference being your loads are at the centroid and mine are at the centroid in z, but on the face in y. But the stress results are the same as my model. Hand calcs put the extreme fibers at -17328 and 12699 lbf/in2. This is exactly what I get in your model as well, when rendering beam profile. Without render beam profile, I get 32000 lbf/in2 under "upper bound axial and bending" in the software.  This level of stress is nowhere to be found on the beam ... there's no axial force, and there is only the 3x 100lb shear forces causing bending about DIR2.

                   

                  Yes you're right about the lack of symmetry and bending about the principal axes. However, the neutral axis was located and proper equations were used to hand calculate the stresses, and again the software confirms it in "render beam profile". My issue is that the "bending about DIR2 stresses" listed in the software are totally unrelated to anything, and I'm wondering where they are coming from. The software gives 21000 lbf/in2, which even when comparing individual components of the stress values to hand calcs about the 2 principal axes is nowhere near correct.

                   

                  I find it somewhat disconcerting that these values are the first presented to the user, anyone not verifying all of the software's calculations by hand calculations needs to beware.

                    • Re: Beam Bending Stress (Ms/Ss, Mt/St) vs. Render Beam Profile
                      Keith Frankie

                      Ah ha!  I've figured this out.

                       

                      The values 21,138 and 12,620 are the stresses at the 'phantom' locations (2.180,0) and (0,2.180), where (0,0) is located at the centroid.  These are the locations directly above/below and left/right of the centroid, at the same distance as the extreme extent of the beam profile (in the x and y directions).  These points do not lie on the profile itself.

                       

                      As stated on page 273 of the book Intro to Static Analysis Using SolidWorks Simulation  (link) "These values are accurate for beam with cross sections that are symmetric in two directions.  These values are also conservative for all other cases""

                       

                      The 33,758 psi load is the sum of 21,138 and 12,620, as described in S-060255 from the knowledge base.

                       

                      I'm working on a more detailed write up.  Understanding what values SW displays is important...

                        • Re: Beam Bending Stress (Ms/Ss, Mt/St) vs. Render Beam Profile
                          Jason Smith

                          Good find, interesting, never would have thought of that. I'm still a little confused as to why they call it DIR1 and DIR2 though, I guess they just mean y and z?  With render beam profile, the DIR1 and DIR2 refer to the direction of the bending moment vector (component) causing the stress, as opposed to the position/location of the actual stress values.  It would have been nice if Solidworks would have included all of this in the help pages.  In this case, the actual max part normal stress is -17328 lbf/in2, whereas without verification one could interpret it as +33758 lbf/in2.