1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Ansys Workbench- Max Principal Stress Error

  1. Jul 12, 2012 #1
    Hi all,

    I did simple problem, in which assembly is subjected to single load & it is fixed at other end. I used Higher Order Tet element (& tried with Hexdominent Method also) for the casting body.

    The vonmises stress shows the true results but when i was looking for the Max Principal stress, it was showing max value one element below the surface element. I wondered how it is possible that max principal stress value is below the surface.

    Is it related to some mesh discontinuity or something else.????

    Can anybody help me out....!!!!

  2. jcsd
  3. Jul 12, 2012 #2


    User Avatar
    Science Advisor
    Gold Member

    You're going to have to post some pictures for us to understand what your geometry looks like. Without that, all I can guess is mesh discontinuity...
  4. Jul 12, 2012 #3
    Pls check the attached file for the model. The image shows the cross sectional result for the Max Principal Stress. (Mesh- HexDominent)

    Attached Files:

    • 1.PNG
      File size:
      46.7 KB
  5. Jul 13, 2012 #4


    User Avatar
    Science Advisor
    Gold Member

    The error you're referring to is definitely due to the mesh density (or lack thereof). If you want a better look in a complex geometry location, you'll need to refine the mesh in that area.

    Have you done a mesh convergence study?
  6. Jul 14, 2012 #5
    Yeah i did that.... n also refined that surface because i knw the max value will come on to that area. The transistion between the elements were good. I also used aggressive meshing for shape checking. But the results are same...

    Is it possible that due to results extrapolated n averages at the node, sometimes the stresses within the body may dominates the surface stresses.........???
  7. Jul 14, 2012 #6


    User Avatar
    Science Advisor
    Homework Helper

    You have to be a bit careful interpreting "max principal stress". For example if the structure is in compression at the surface, you could get the situation where the "max" principal stress is zero, but you were probably more interested in the minimum (negative, compressive) principal stress.

    I have seen software that plots the "worst principal stress" (i.e. the one with the biggest absolute value), but that can have discontinuities where it jumps from positive to negative, which can also be confusing.

    For most types of finite element, the calculated stresses are discontinuous across the element boundaries, and the graphics output usually includes some sort of smooth interpolation. Some post processing software tries to do this in a mathematically consistent way, other programs go more for the "never mind the quality, just look at the pretty pictures" approach.

    I don't use Ansys so I can't comment on your specfic output. I suggest you look at the physical stress components (in the global X Y and Z directions), or a function like von Mises stress that is a mathematically "smooth" function of the stress field, to see if the issue is really with the model or just with the post processing.
  8. Jul 14, 2012 #7


    User Avatar
    Science Advisor
    Gold Member

    How big is the difference actually, in absolute & relative terms?

  9. Jul 16, 2012 #8
    Actually that surface is subjecting tensile stresses n body is casting. so i have to luk for the max principal stress.

    @PerennialII: At the surface the value is 57000 & the max is 64000..... so the variation is of 7000 within the single element. Also The ultimate tensile strength is 65000.....
  10. Jul 16, 2012 #9


    User Avatar
    Science Advisor
    Gold Member

    And what were the results of your mesh convergence study? It looks to me like the mesh is not fine enough to get detailed information out of the corner you're looking at...
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook