Nodal Stresses of the Last time step

[dolesch]

Hi all,

for my user element I need the nodal stresses from the last time step. How can I access them within an user element routine?

Regards
Dolesch

[Prof. S. Govindjee]

Nodal stresses are a globally computed quantity and are thus not easily accessible inside an element, where the Gauss point stress are normally used.

It seems unlikely that you actually need the nodally projected stresses inside an element.  You should be able to make do with the Gauss point stresses.

Nonetheless, if you really need the nodal stresses inside an element, you will first need to compute them globally, say, using the STREss,NODE
command at the Macro prompt.   Then you can find them in the arrary hr( * ) at location npnp+numnm (in blocks of length istv).  See the files
pparaview.f and p_paraview.f in the program folder to understand what you will find in hr( * ) and how to access the values.

[dolesch]

Thank you for your quick response. I am trying to compute the nodal error values by using an adapted version of the subroutine ster2d. In the manual isn't much information about the routine.
So the next question. Were do user elements write their s-file while in the process of isw == 11 ( hr(np(57)) maybe?) and can I adopt the same procedure of the stress projection for the error projection?

Thanks in advance

Dolesch

[Prof. R.L. Taylor]

The best way to follow the process is to start in 'pmacr1.f' and follow the projection of stress.  You will see that the steps use formfe.f for isw.eq.8 and then call pltstr.f which should do all the assembly of the p(*) and s(*,*) arrays.

Note that the pointer used for the s array is reset to np(60) which is the real area used for storing the element projection quantities.

The same path is followed for error projections.  However, as you note there is not much coded at this time.

[dolesch]

Thank you Prof. Taylor and Prof. Govindjee.