4 Replies Latest reply on Nov 28, 2012 7:10 AM by jerry towne

    Beam Deflection

    jerry towne

      I am simulating the deflectiopn of a beam with fixed supports at each end with a concentrated load at the middle.  The beam dimensions are 200mm long, 10mm wide, 10mm high.  The load is 100N.  E=2.1e11 N/m^2.  The Simulation result with default mesh is .003429mm.  The theoretical deflection is .00296mm.  +15.8% error.  The deflection equation for a centrally loaded beam fixed at both ends is dy=(P*L^3)/(192*E*I).

        • Re: Beam Deflection
          Anthony Botting

          Hi Jerry: using your equation, I calculate theoretical = 0.0238 mm and ran a Simulation (default, solid mesh) = 0.0244 mm, for a 2.5% difference (compared to theoretical). Can you describe the model setup? I used half a beam, fixed one end. I put a "roller" on the opposite end, and 50 N on that same end perpendicular to the beam axis.

            • Re: Beam Deflection
              jerry towne

              Tony,

               

              I checked my arithmetic and you are correct.  I made an error in my theoretical calculation.  I also re-modeled the simulation as you did and duplicated your result.  My model is a 200mm x 10mm x 10mm solid.  I created a split line at the midpoint of the top surface to which I applied my 100N load.  Both ends of the beam use a Fixed Geometry Fixture.  Mesh type is Solid Mesh. I re-ran and got a deflection of .0245mm.  I deleted my original simulation so I can't explain how I initially got the erroneous result.

               

              Thank you,

               

              Jerry

                • Re: Beam Deflection
                  Anthony Botting

                  Thanks, Jerry. I have encountered similar discrepancies at least several times over the course of using the software (since the early 90's when it was called "Cosmos/M"). Sometimes the results come-out unbelievable, and there's no explanation. I have resorted to shutting down the software (even re-booting the computer, on occasion), then creating a whole new model and then, it works just fine. So, what I'm trying to say is, these things seem to happen with software and I'm glad you wrote-in to the forum. I have found a good policy is to always suspect the answer and to check with colleagues, books, conduct hand-calculations in an attempt to determine upper and lower bounds to the answer, even use other software to double-check.  You cannot be too careful. I'm glad it worked-out for you. -Tony