General Mechanical

General Mechanical

Transient Structural — Rotational Velocity Joint Load Will Not Converge

    • shermama96
      Subscriber

      Hi Everyone,


      So I'm in the process of analyzing a gantry CNC machine for a senior design project. My model is somewhat large but I've suppressed many elements so that I'm starting with a very simple model to get familiar with transient structural and even on the simplest thing I can think of, I am running into issues. I have a leadscrew that is support by a bearing block by a revolute joint that I am trying to drive with a rotational velocity joint load (like the axis drive of a machine). I just can not get this model to converge at all, hasn't solved a single time. At first I was getting wildly distorted elements in the leadscrew, it looked like the 1 inch diameter component was growing in diameter to 5 inches. At one point, it looked like the elements that should be rotating were "stuck" to the supporting elements. I changed some things to mitigate that distortion, but still no luck. The solution takes 50 minutes to produce a non-convergence error, and because the model is so simple I would have expected it to solve very quickly. What I have done to debug this (with no success):

      • Changed revolute to general joint and constrained specific degrees of freedom
      • Refined the mesh in the region of contact
      • I was originally trying to drive at like 50 rad/s but have since moved that to only 1 rad/s for debugging
      • Ramped velocity instead of stepping
      • Increased the number of sub-steps
      • Defined the joint using remote points instead of by surfaces
      • Double and triple checked all defining coordinate systems
      • Made sure large deflections is on
      • I tried plotting FE connectors and could not for the life of me get them to display
      • Plotted Newton-Raphson residuals
      • Tried driving the joint with displacements or accelerations to see if it was velocity specific, none of those would converge either

      I'm sure in the 10 hours I've spent "debugging" this I've tried a whole slew of other things with no avail. I've attached a photo and a workbench archive, any help is deeply appreciated because as of right now I am very stuck.


    • 1shan
      Ansys Employee
      ,nHave you described all the bodies as flexible or just 1 body? You may want to keep the component of interest (leadscrew) to be flexible and define rest of them as rigid(Geometry>Part>stiffness behaviour >rigid). Also, you will have to use fixed joints (instead of contacts) to connect rigid components together. The bed is too big, meshing that would be futile.nRegards,nIshan.n
    • shermama96
      Subscriber

      Hello @shermama96,Have you described all the bodies as flexible or just 1 body? You may want to keep the component of interest (leadscrew) to be flexible and define rest of them as rigid(Geometry>Part>stiffness behaviour >rigid). Also, you will have to use fixed joints (instead of contacts) to connect rigid components together. The bed is too big, meshing that would be futile.Regards,Ishan.https://forum.ansys.com/discussion/comment/112124#Comment_112124

      Thanks for getting back to me. I have tried switching between rigid and flexible to no avail. The plan is to model an entire CNC machine but I am starting with the simplest components just to get familiar with the solver. In theory I'd like everything to be flexible because I'll be using this model to establish geometric parameters for some of the cast components like the bed. For reference I'm not using any contacts for any parts, all bodies are coupled using joints.n
    • peteroznewman
      Subscriber
      nI recommend you idealize the leadscrew in this model to something much simpler, such as a Translation joint.nYou can have a fixed frame (bed) with a bearing block for the leadscrew and moving frame with a nut on the leadscrew. Then you just have a Joint Load of velocity to move the nut along the leadscrew.nIf you want the nut to only provide 1 DOF instead of 6 DOF to the moving frame, then you will want to put some additional links in the mechanism to free up some DOF between the nut and the moving frame so that the slide or bearings between the fixed frame and the moving frame can take some load. Or use a General Joint instead of a Translation Joint.n
Viewing 3 reply threads
  • You must be logged in to reply to this topic.