Changing order in stress projection; installation corrupt?

[crepes]

Dear Feap team, Dear FEAP‌ users,

Please have a look at the attached screenshot of the programmer manual for feap8.5.

For some tests I changed

Code: [Select]

s(i,1) = s(i,1) + sig(1,l)*xg
s(i,2) = s(i,2) + sig(2,l)*xg
in a user element to

Code: [Select]

s(i,1) = s(i,1) + sig(2,l)*xg
s(i,2) = s(i,2) + sig(1,l)*xg
and used "plot stre 2" in the unmodified version. It did not yield the same plot as "plot stre 1" in the modified version. This should yield the same plots, shouldn't it? I guess that there is something wrong with my code. Or are there operations beyond this routine in the feap code which require that the order of the stresses is correct? Actually I wanted to use this routine to plot other stress-related quantities.

Kind regards

crepes

[Prof. R.L. Taylor]

It should give the same result.  Can you try this with one of the standard feap elements and their projection. 
Possibly the array storing the stresses is inconsistent between the two routines?

[crepes]

Dear Professor Taylor,

Thanks for your reply. Following your suggestion I changed in $FEAPHOME8_5/elements/solid2d/slcn2d.f some lines to

Code: [Select]

            s(i,1) = s(i,1) + sig(2,l)*xg
            s(i,2) = s(i,2) + sig(2,l)*xg
and rebuilt the new executable. Using a standard feap material model gave the same plots for "plot stre 1" and "plot stre 2" as it should be.

Do you mean the sig-array? I do not see how this array could be a problem...

I recently noticed a strange behavior: In the user element I used the same lines as above. It gave the plots in the attached figures. The disturbance is located at the borders of single blocks which are tied in the input file (cf. the draft in the third figure). The input file works fine for the standard-feap-materials.

Kind regards

crepes

[Prof. R.L. Taylor]

My guess is that there is a dimension inconsistency between the element where the stress is computed at each gauss point and the projection routine.  Make sure that you have set istv = max(istv, ..n..) (where "n" is the number you are projecting in your element under isw.eq.1; also make sure you set the 'iste' parameter in your projection routine to the number you actually project in THAT module.  There should be no conflict to what you call the actual stress number between different element types -- that is if '1" is sig_11 in one routine it cannot be sig_22 or a stress resultant in another one (unless you plot material type by material type).

[crepes]

Now it works. Thank you very much for your help. Some time ago I had changed something which actually was wrong afterwards. I did not notice this since in those old circumstances there was no difference in the results due to this change. Then I was too lazy to remove this since I was almost positive that this change has no effect. But I was mistaken. After removing this, it works very well. Unfortunately I did not provide the corresponding code lines so you had no chance to find this mistake, I apologize for that.

Kind regards,

crepes