10 Replies Latest reply on Apr 29, 2010 9:15 AM by 1-M4OM40

    Beams floating away

      Why are my beams floating away when i try to do an analysis. Some stay, some float away and are ignored for the use of the analysis. I never checked a box that said "please make my life a living hell and do everything that could be of no use to me" I also never checked a box that said "please make all my beams float away into space, i put them there just for the fun of it, not to be analyzed..."

       

       

      GRRRRR

      Someone help

        • Re: Beams floating away
          Dan Eldred

          Click Simulation->Help->Simulation Online Tutorial

          • Re: Beams floating away
            Derek Bishop

            Sounds like you may have some problems wth the contacts. Do you have bodies in the model other than beams?

            • Re: Beams floating away

              My guess is you have a funadamental problem with your boundary conditions. While SW Simulation can give you results, it only knows what you tell it. You must define proper boundary conditions with each and every part. As Derek mentioned, part of this includes contacts. SW SImulation doesn't "understand" two objects colliding with one another unless you tell it to look for this by defining contact sufaces.

               

              Without a screenshot or description of what you're modeling, it's really hard to hazard a guess at anything specific.

               

              If you have an assembly with multiple parts and do not define how they interact (with contacts for example), SW Simulation will apply loads and unless you have a contact or BC to stop a part, it will "fly off".

               

              Please note: MATES ARE NOT CONTACTS! Because you define two parts of an assembly as coincident when you built it HAS NOTHING to do with how they behave in Simulation. The default contact is bonded which means and touching pieces will have their nodes glued together. If you have a TINY gap between them and don't define a no-penetration contact, one part can and will "fly away".

               

              What you should consider doing is drawing a free body diagram of the item you're simulating. Take each part with applied loads and determine what forces will counter those loads. Those counter forces must be boundary conditions or contacts.

               

              Since you're getting parts that are flying off, you're missing an opposing force to keep it in place. This may be friction, a no-penetration contact gap, or a non defined fixture.

                • Re: Beams floating away

                  Hi,

                  yeah i was having a bad day and got a little vocal in my question I appologize for that.

                   

                  First off i double checked that i had no solid bodies, everything is either sheet or beams.

                  That all makes sense, I had done "global contact bonded", and I guess that doesn't always get everything for some reason? I manually put in contact sets and that seems to have tied everything down.

                    • Re: Beams floating away

                      I, too, have had issues with global contact sets. When I'm running a problem using the contact gap solution, I will pretty much always manually define them and suppress the global condition.

                       

                      If you're using beams in your model, that might explain it. I'm pretty sure that SW Simulation defines a beam as a single axial line and then performs simple beam equations on that line. This simplifies the meshing and calculation greatly. The problem is that this would remove the surface nodes as it's changed the body to a line. This may have broken the bonded condition you were trying to globally set: without touching nodes, they cannot be bonded and the beam would fly away.

                       

                      I bet that if you remesh and don't define your beam as a beam, but a body, that the bonded global condition would work fine.

                        • Re: Beams floating away
                          Derek Bishop

                          I've found that if I try to bond a gusset as a shell two beams it will not bond and I get an error message when trying to run the study. The only was I can do this is by making the gusset into a solid.

                           

                          Each beam elements have two nodes. These form a line of elements that make up the beam.

                          • Re: Beams floating away

                            Hi,

                            That's really logical, and my next question stems from it. My model solves okay now.

                             

                            When determining how much stress there is at the end of a beam, do you take the stress found at that point (the end of the line simulating the beam), and divide it by the surface area of the the end of that beam in order to get an accurate reading for how much stress there really is?

                             

                            This is a sheet metal tank with structure pieces crossing the gap at the middle. My highest stress seems unreasonably high, is that because I still have to manually divide that point load by the surface area of the end of the beam? Or has solidworks already taken the end surface area into account to arrive at this solution?

                             

                            Ive attached a picture of what I'm dealing with.

                             

                            Thanks again!!

                              • Re: Beams floating away

                                A beam element in SW Sim is two nodes and one beam. It's basically a line. I don't know how it treads the end connection, but my guess is that it treats it as a single node resulting in artificially high stresses where it connects to a surface. If you're connecting a beam end to a surface, I would model the beam as a body.

                                 

                                There are probably other ways to do this, but I don't know them (I rarely use beams as anything I can treat as a beam I typically do hand calculations on).

                                 

                                If you find an answer, please post back as I would like to know as well.

                                 

                                Another idea: try contacting your reseller for support. I did this the first time (I've had SW for 5 years now) a couple weeks back. It was VERY helpful and made me sad I hadn't used it sooner. I was talking with a real person within about 10 minutes of contacting them by e-mail. He was very knowledgeable and was able to describe how SW Sim worked in a couple of areas that I was weak in (built-in bolt connectors). He was able to explain how SW Sim treats these connections and answer a question very similar to yours.

                                • Re: Beams floating away
                                  Derek Bishop

                                  Interesting point (excuse the pun). Maybe try a simple model of a beam ending on a surface and test your results while you wait for your VAR to respond.

                                    • Re: Beams floating away

                                      Hi,

                                      It turns out that it was calculating the stress based on a infinitesimmaly small contact point at the wall creating unusuall high loads, because a beam is just two nodes and a line. You cant even accurately divide that number by the known surface area of the end of the beam because it still wont be right.

                                       

                                      VAR suggested creating a midplane surface in the beam representing its basic shape, then in the simulation you can define a thickness for it. Now, where it contacts the wall, you set the edges of that midplane surface bonded with the sheetmetal wall, and it takes the thickness of the material into account. Now my results are very accurate!

                                       

                                      Im still waiting for other more detailed answers from the VAR but for now this is plenty to practice with and at least get more realistic results. See the attached image, you can see how the wall behaves just like as if the pipe really were there, instead of just a single point.

                                       

                                      Cheers