Beam connection as a bolt

Hi,

Could someone please explain to me some basics of modeling beam as a bolts, and how to correctly (and why) apply a beam connection as a bolt? After analyzing some tutorials, techtips, courses.ansys.com on bolt joints, still I have a lot of doubts.

I created a beam connection (contacts> insert beam). My basic questions are:

1) does it matter where reference and mobile are connected? I mean, is mobile should simulate the head of the bolt and reference thread, or there in no difference?

2) What value should have the parameter behavior (and why): deformable or rigid?

Thanks in advance.

Answers

  • @PiotrM

    #1 It doesn't matter which side you call Reference and which you call Mobile.

    #2 Make the behavior rigid since the flexibility is in the beam and the surrounding material of the joint.

    Did you watch my series of videos? https://forum.ansys.com/discussion/1217/bolt-pretension-clamps-two-plates-together

  • @peteroznewman

    Thanks for your anwser. Yes I saw your video but I also saw another where approach which is opposite of yours. There bahavior was set to deformable (https://www.youtube.com/watch?v=7Q61-622pWE) what is more there was a comparison of the stress results of the 3d element bolt and the beam connection, which look the same for the deformable setting for beam.

    To be more precise my connection is not exactly the same because there is a nut and bolt used to clamp and I have a thread in the bottom plate. On the top plate (black) there is bolt head - moved on 2/3 of the real bolt head height on pinned to the face . On the bottem plate (blue) there is thread pinned to the edge on the hole face. For both I am using deformable which should match with RBE3if I am right.


  • peteroznewmanpeteroznewman Member
    edited December 2020

    @PiotrM

    Thanks for the link to another video. The point he and I agree on is that if stress at a bolted joint is critical, it should be represented using solids for the nut and bolt, not idealized beams and constraint equations.

    I can give you an example where Rigid behavior will produce results that are closer to the results from a solid bolt head than a Deformable behavior on a Beam connection. The example is two plates made of plastic or any material that is hundreds or thousands of times softer than a steel bolt. In that case, a Rigid behavior of the bolt head provides a stress field in the plastic that is closer to the solid bolt head than the Deformable behavior, which will pull the edges of the hole down way too much.

    Yes, Deformable is like RBE3 from Nastran. I recommend you use the upper half of the hole in the blue plate as the scope for the Mobile connection to the Beam. That will represent the stiffness of the threads more accurately than scoping to an edge on solid elements, which will be too flexible.

    The length of the bolt that can stretch is somewhat excessive in your threaded bolt on the right because you moved the top up by 2/3 of the bolt head and the bottom goes all the way to the end of the threads. In reality, the length of bolt shank available for stretching is only the thickness of the black block. You would get a more accurate representation of the flexibility of the bolt by locating the top beam vertex at the bottom of the head and the bottom vertex at the top of the blue block. Most of the load is carried by the first two threads, so you could have the beam go a little lower than the top of the blue block.

    When there is a nut and bolt in two clearance holes, then the length available to stretch the bolt is the sum of the thickness of the two plates and that is correctly represented in the beam in the hole on the left.

  • Thanks @peteroznewman for the discussions.

    An offset 2/3 of the bolt head height is suggested by a more experienced engineer. Same as using an edge line to attach the thread. However, pinning the thread to the surface (as you suggested) has more advantages, including: stress not being generated at the edge but spreading over the surface which helps with convergence of the analysis. This solves some of my problems.

Sign In or Register to comment.