Fall 2002, METBD450 - FEA Applications II

Prof. Dave Johnson, dhj1@psu.edu, Penn State - Erie, The Behrend College

Homework Assignment: Submodeling

Concepts:

• Submodeling
• Advanced postprocessing (path operations, force summations)

  Problem:

  A simplified open-end 5/8" wrench has been analyzed with a fairly coarse model.

  Details of this analysis include:

  • modulus = 29x10⁶ psi
  • Poisson's ratio = 0.3
  • part thickness = 0.25 in.
  • Smartsize mesh key = 2
  • 809 PLANE82 elements
  • 2642 nodes
  • ANSYS 6.1: Maximum SEQV of 58,162 psi occurred at the bottom, inside corner of the wrench.
  • SEPC was ~17.3%
  • 0.075" fillets at the sharp, inside corners were ignored
  • a 125 psi. load acts over the last 1" of the wrench handle.
  • vertical constraints on "stiff" spring elements attach where the nut contact is assumed on the wrench.  On the bottom surface, 0.2" from the wrench tip and to the right is treated this way.  On the top interface, from the upper right corner of the nut and 0.2" to the left, rest on this stiff support condition.  (This makes the "nut" stiff, but not rigid, i.e., it has some flexibility).
  • a single horizontal constraint is imposed at the midpoint of the arc in the open end of the wrench.

  To run the coarse model yourself, use the ANSYS 6.1 input: Coarse Model  (You can cut and paste these commands directly to the ANSYS Input Window or into a text file to "Read Input From...")

To do:

Construct a submodel of the region around the high stress location.  (Use some of the geometry from the Coarse Model file).  If you use the coarse model database: after you clear the mesh, delete the spring elements, nodes, element type and real constant set for the nut approximation.

Include a 0.075 in. fillet in the corner between the bottom of the wrench face and the arc in the back.

DO NOT cut into the contact region unless you want to re-create that stiff spring boundary for your submodel.

Make the submodel boundary cuts through regions of low stress and low stress gradient on the coarse model.

Turn in:

• A plot of your submodel showing ALL the boundary conditions.
• A plot of vonMises stress on the submodel (with SEPC annotated on the plot, along with the % difference in max. stress for the submodel vs. the coarse model).
• Plots to compare the coarse model and submodel stress contours.  (Use the same contour scaling, focal point, and magnification for both plots - to compare, OVERLAY the submodel results on the coarse model results).
• Plots using "path results" to compare the coarse and submodel quality at the cut boundaries.
• Show the path geometry on a node plot of the submodel and of the coarse model.
• A POST1 listing for the COARSE model results: Nodal Calcs > Total Force Sum, showing the total moment acting on the center of the nut (which is 0.3125 from the left edge). Identify the moment that is the wrench torque.  Verify (and compare) the total force sum in the X and Y directions, as well as for the moment about the center of the nut.
• Comment on the comparison of stress results along the edge of the submodel to the same location in the coarse model, based on both the contour plot overlay method and on the path plot graphs.  Was your submodel cut far enough away from the stress concentration region to believe its results ?
• The annotated log file for your submodel construction, cut boundary interpolation, and analysis

  ---