Hello,

In general, you should plot the stress-strain at nodes. For elements, you need to know the integration points in the element. When you know the position of the integration points, you should be able to plot.

"In general, you should plot the stress-strain at nodes."

I totally and emphatically disagree. Finite element programs calculate stresses and strains at integration points; extrapolated values at nodes may be off by a factor of 2, for example in plastic materials.

This is two different questions.

Your comment is regarding the accuracy of the extrapolation from integration points to nodes.

You will find the answer for the first question if you look for instance at the Command Manual in the User Documentation of OpenSees and then the NDMaterial Command Lists:

Let's propose that you have assigned "PressureDependMultiYield" as a material model for the 2D soil column. the arrangement of the stress and strain components are as following:

σxx, σyy, σzz, σxy, ηr

and

εxx, εyy, γxy

These components will be recorded for one Material point or Gauss point once you call the Element Recorder Command for each soil element.

Now if you have soil elements type "SSPquad Element" then you have for each element one single integration point which records the response of your material.

http://opensees.berkeley.edu/wiki/index.php/PressureDependMultiYield_Material

thank you Ahmad :)