0 Replies Latest reply on Sep 25, 2012 2:39 PM by Santos Halper

    Micron scale static analysis accuracy??? Simulation not close to test results.

    Santos Halper

      Hey everyone,

       

      I've been trying to develop an FEA model to simulate a long slender wire (~.004" diameter) under a compressive load. The wire ends are offset, so when the load is applied, the wire buckles and acts as a spring.  The load is actually being applied in my model by displacing the free end axially, and not by applying an actual load.  The free-end axial displacement is only a few mils.  My goal is to determine the reaction forces and stresses for given axial displacements of the free end of the wire. The problem I'm having is that the FEA results I'm getting are not even close to physical test results.  I'm getting reaction forces that are more than double what testing has shown and stresses that are close to the materials yield strength which also does not correspond to physical testing.  The deformed shape of the wire matches pretty closely though.  Can you guys give me some insight on how my results could be so different from reality?  Is Simulation less accurate when modeling small-scale parts/assemblies?  I'm near the point of accepting the FEA results and just applying a scaling factor or adjusting the modulus of elasticity to get me closer to the test results.  Would this be bad practice?

       

      My model:

      -I've modeled the wire on its own and as an assembly for more accurate restraints (contact and friction) but the reaction forces and stresses are still too high.  The boundary conditions are as close to reality as they can get when modeled as an assembly.

      -I've quadruple-checked the material properties and they are all correct.

      -I've been running linear studies with the large displacement flag checked as well as non-linear studies with similar results. A non-linear analysis seems to give slightly better results and model the contact a little better.

      -The mesh is refined enough to give 4-5 elements through the thickness of the wire.

       

      Thanks.