Hope that helps. Thanks for getting in touch.

]]>I realized that I don’t distinguish between stress concentration and stress singularities. Thank you for this clear and useful article.

Because I am dealing with concrete structure, I have always singularities in the connection between a plate ( 2D element) and a column ( 1D element). even when I do the nonlinear analysis I cant remove the singularity.

you mentioned that if the material nonlinearity is activated we can remove the singularity but it never happens with me, I cant overcome the singularity. do you have some advices maybe? ]]>

Also, thank you very much for reaching out to me in such a timely manner.

When performing the last exercise (the one about mesh sensitivity),

the normal stresses at the top surface will never truly be zero, but some really close values to zero.

Acin has kindly explained that this is due to the fact that stress values are computed at Gaussian points, which are located inside the elements.

So, the values that are stored in the nodes on the top of the beam are extrapolated from these values. As you refine the mesh in the thickness direction, the Gaussian nodes will get closer to the top surface, but never quite on the top surface.

Therefore, while the result will approach 0, but never quite have that value.

Thank you for your help Acin! ]]>