irregular high peak of stresses alona cut off wall

Hi,

by displaying the principal stresses of a line cross section in the middle of a cut off wall an irregular peak of stress keeps showing. The Plaxis Reference suggests adding interfaces to solve the problem of stress oscilations at the corners of stiff structural elements. Doing so didn't eleminate the irregular stress.

Any ideas on what reasons could be behind this irregularity and how to fix it?

Thank you all in advance.

Parents
  • Dear Ahmad,

    Extending the interfaces is to overcome stress oscillations in interfaces, but you mention you have a peak stress in a cross section. So indeed, extending the interfaces won't do much.
    Please note that stresses in cross sections are based on first extrapolating stresses from the stress points to the nodes of the elements through which the cross section is drawn, and then interpolating the nodal stresses onto the cross section. This means that cross section stresses are not always so accurate. Peak values may occur if for instance when a cross section is only a very small part inside an element or passes very close to a node or stress point in which case the extrapolation/iinterpolation procedure suffers from numerical roundoff. 
    Hence, please first check if the peak stress really exists by checking the cartesian soil stresses in the soil elements where you draw the cross section. If the peak doesn't  exist in the soil probably slightly moving the cross section would help...

    With kind regards

    Dennis Waterman

Reply
  • Dear Ahmad,

    Extending the interfaces is to overcome stress oscillations in interfaces, but you mention you have a peak stress in a cross section. So indeed, extending the interfaces won't do much.
    Please note that stresses in cross sections are based on first extrapolating stresses from the stress points to the nodes of the elements through which the cross section is drawn, and then interpolating the nodal stresses onto the cross section. This means that cross section stresses are not always so accurate. Peak values may occur if for instance when a cross section is only a very small part inside an element or passes very close to a node or stress point in which case the extrapolation/iinterpolation procedure suffers from numerical roundoff. 
    Hence, please first check if the peak stress really exists by checking the cartesian soil stresses in the soil elements where you draw the cross section. If the peak doesn't  exist in the soil probably slightly moving the cross section would help...

    With kind regards

    Dennis Waterman

Children
No Data