General Mechanical

General Mechanical

Verification of results from Explicit dynamics

    • ganesh2516822

      I'm new to explicit Dynamics. I'm currently doing a simulation of a mechanical brake system for a robotic joint for my project. I'm having a hard time obtaining a consistant result. After every mesh refinement, the impact stress keeps on increasing.

      The spoked body hits the pin (fixed on one side) with high angular velocity. I've taken the spoked entity as a rigid body. Initially I thought I could verify the results analytically, equating rotational kinetic energy (0.5*I*W^2) with impact energy (0.5*F*deflection) and thereby obtaining obtaining value of impact force (assumed the deflection to be that of a cantilever beam FL^3/3EI). Later finding the value of bending stress and comparing it with the numerical value. Is it the right way to do it ?

      Mesh sizes and Bending stress at pin
      0.8mm >> 337.16 MPa
      0.6mm >> 351.88 MPa
      0.4mm >> 385.87 MPa
      0.2mm >> 512.67 MPa
      These were the bending stresses that I've got from the post processor. The bending stress I got analytically was 413 MPa using the aforementioned method. I assumed that the stress'd converge at  that point after mesh refinement. Can anyone tell me what I'm doing wrong here ? I've also attached the energy summary
    • SaiD
      Ansys Employee


      Based on the energy summary plot, it looks like the Hourglass energy is quite high. Hourglass energy is a numerical energy and hence should be less as compared to other physical energies (e.g. the Internal energy and Kinetic energy), say less than 5% of the physical energies. So the results you are obtaining are not entirely accurate.

      I think the cause of hourglassing here is that when the spoke body hits the pin, an edge of the spoke comes in contact with the pin (I am basing this on the image, correct me if I am wrong). Hence a line force is applied on the pin. Applying a line force or a point force may lead to hourglass behavior in explicit analyses. A few suggestions to fix this:

      1. If it is possible to modify the geometry of the spokes to have rounded edges rather than sharp edges, the contact force applied will be distributed over a small area rather than being applied as a line force. However, if this is the actual geometry in real life, this may not be a feasible solution.
      2. Try using Hourglass stabilization.
      3. If 1 and 2 don't work, switch to fully-integrated elements (by default, Explicit analyses use elements with reduced integration formulation for computational efficiency).

      Hope this helps!

Viewing 1 reply thread
  • You must be logged in to reply to this topic.