3 Replies Latest reply on Jul 15, 2014 3:46 PM by Jared Conway

    shell mesh and bolt connectors revisited

    Mark Scholten

      Hi everyone,

       

      i am setting up a static analysis of a sheetmetal duct, consisting of bolted plates. i am using shell mesh, no penetration contacts and bolt connectors. from a previous very similar project i have learnt that the stresses in bolt connectors are dependent on mesh density, especially when the length of the bolts is less than 5 times the radius. in this case that ratio is a lot smaller than 5 for all of the bolts. KB S-058593 and S-059588 suggest to use solids at least locally, but that is not an option for me, because there are very many components in the assembly. As is, the study does not complete.

       

      given that i am refining the mesh around the bolt holes, would it make sense to:

      1. decrease the shank radius so that the bolt length over shank radius ratio is smaller than 5, so that the study is able to complete (the shank cross-sectional area is now way too small, so that reported stresses and strains become way too high)
      2. set the head and nut diameters to the real values to maintain the behaviour of the material around the bolt hole
      3. define a custom bolt material using a value of Enew=((old radius / new radius)^2)*Eold (this should make the strain of the bolt with reduced shank radius equal to the strain of the original bolt
      4. check bolt forces afterwards to make sure none of them break (necessary because the ultimate stress of the bolt material was unchanged).

       

       

       

      does this sound reasonable or should i quit tinkering benaeth the hood and go with the work-around i used previously (switching to bonded contacts, extracting interface forces and then calculating bolt forces manually)?

       

      thanks for your thoughts,

      mark

        • Re: shell mesh and bolt connectors revisited
          Nicholas Luyster

          Hi Mark,

           

          In this case, I would recommend that you create at least two simulations.  I often do this is with large simulations that have many parts.  By creating multiple simulations, you can devote your computational resources to the locations of interest and get much more accurate results in a shorter amount of time. 

           

          In one simulation switch over to solid elements and refine the mesh at the bolt holes.  It's okay to leave the rest of the mesh relatively coarse.  Then, in subsequent simulations, you can coarsen the mesh at the bolt holes and devote your computational resources elsewhere. 

           

          All the best,

           

          Nick Luyster

          Gosimulation.com

          Online SolidWorks Simulation Training

          • Re: shell mesh and bolt connectors revisited
            Shaun Densberger

            define a custom bolt material using a value of Enew=((old radius / new radius)^2)*Eold (this should make the strain of the bolt with reduced shank radius equal to the strain of the original bolt

             

            Can you provide additional explanation for how you came up with this relation?

             

            In theory, what you need to do for this to work is modify the material properties of the smaller shank bolt such that it has a stiffness matrix that matches the large shank bolt. This is where it gets tricky.

             

            SW uses rigid links to connect the two ends of a beam element to the respective shell/solid elements. While I haven't found a specific reference to it, I'd be surprised if SW didn't use Timoshenko beam theory for their beam elements. A beam element has 6 DOF per node for a total of 12 DOF (although I think in the beam element for a bolt connection the on-axis rotational DOF is released, but we'll ignore that for the sake of this), so the stiffness matrix has 144 terms.

             

            Some of the terms in the stiffness matrix are a function of the length of the beam (L), and some are a function of L, E (Young's Modulus), I (Second Monument of Inertia), G (Shear Modulus), A (cross-sectional area), and k_s (shear area factor). The coefficient for the entire stiffness matrix is a function of all of these variables. Now you see the problem.

             

            This is an interesting work-around, but I think the more important question here is why the KB suggests to use solid elements when the ratio is a lot smaller than 5. Understanding this is required in determining whether your work-around (or any other for that matter) will in fact work, or whether it'll just lead to inaccurate results. If determining all of this isn't feasible, then I suggest going with previous methods that you know work for you.

             

            I'd also like to point out that, at the end of the day, SW bolt connection tool is just an idealization; even under the best conditions theresults are not going to be highly accurate (at least the stresses around the bolted material...shear and tensile forces can be fairly reasonable).

            • Re: shell mesh and bolt connectors revisited
              Jared Conway

              what is the goal of the analysis? can you post a picture of what you're working with?

               

              if you're just trying to hold things together and you don't care about the bolt forces, then there are better ways to do it

               

              if you want the bolt forces, working around bolts isn't going to be a good idea. setup the analysis differently and extract the bolt forces a different way

               

              if you want to go with your workaround, build a test case and test it. that is the only way to know if it works.