Calculate SIF using FESU

Questions & Discussions about ROHR2 and Finite Elements
Beiträge: 134
Registriert: Mi 14. Sep 2011, 10:18

Calculate SIF using FESU

Beitrag von r2user » Mi 19. Feb 2014, 17:07

Hi Rohr2 Support

I’m calculating the flexibility analysis of some piping, but some of them got the D/t>100, so i’m trying to calculate the “true” SIF(I0 Out plane and Ii In Plane) based on a theory that shows SIF(o or i)=SigTresca(o or i)/Sig Nominal(o or i), where ”SigTresca(o or i)” is the maximum out-plane or in-plane membrane stress of Tresca and “Sig Nominal(0 or i)” is the nominal stress(Sig = M/z) due to out-plane or in-plane bending moment applied M(0 or i).
How can I extract these bending moments M0 and Mi from the shell finite element analysis and how can I extract the maximum out-plane and in-plane membrane stress of Tresca or Von Mises from the shell finite element analysis?
Best regards

Rohr2 User
Sie haben keine ausreichende Berechtigung, um die Dateianhänge dieses Beitrags anzusehen.

Beiträge: 384
Registriert: Mi 14. Sep 2011, 08:23

Re: Calculate SIF using FESU

Beitrag von rohr2support » Mi 19. Feb 2014, 17:27

Hi Rohr2 User

The whole design of ROHR2FESU was done in order to avoid the trouble to calculate a SIF.
Due to the fact that the Finite Elements are substructures and these substructures can be reused as often as you like without extra
"calculation" cost as the Finite Element calculation is performed only once there should be no need to go back to a simple SIF
(or inplane and out-of-plane SIF).
Actually ROHR2FESU calculates "generalized SIFs" for each unit load case and each element ROHR2FESU will store the SIF.
Therefore you get exact results for all combinations of load cases and all elements in your FE Model at the same price that it costs
to calculate a SIF.

The problem with the SIFs is that you need to calculate the maximum intensification factor for the worst element but this may vary from load case to load case.
Therefore you may calculate a big SIF from a load component but you apply something different in your actual calculation so the results are still too pessimistic even though you have gone through all the efforts to calculate a SIF. The next problem is that you need to calculate the flexibility as well in order to have a realistic model of the piece replaced by the FE model. All this is done automatically if you reuse the FESU-Model.

The next problem is that your definition for the SIF is not well defined:
SIF(o or i)=SigTresca(o or i)/Sig_Nominal(o or i)
Is it SIF(o or i)=SigTresca_membrane(o or i)/Sig_Nominal(o or i) or
SIF(o or i)=SigTresca_bending(o or i)/Sig_Nominal(o or i)
The answer is that for the S1/SL equation it should be the membrane stress and for S3/SE equation it should be bending stress.
But even this is not strictly conservative. The stress analysis using piping code (e.g. EN13480-3 chapter 12) with S1, S2, S3, S4 equations assumes that the longitudinal stress due to internal pressure is half of the hoop stress. This is true for a cylindrial pipe, but not for a general shape. Therefore the SIF must include a safety-margin for this effect as well, if the pipe is subjet to internal pressure as well as bending moments. By using a FESU-component all this is taken care of, as there also is a factor for the pressure.
So even though the stress codes tell you to "calculate" the SIF they dont tell you how.

So the short answer is "don't do it"!

Now if you still want to calculate a SIF you can create "unit load cases" for inplane and outplane moments, by simply applying a single moment to the structure.
Then define a stress equation for membrane and bending stresses for each unit load case.
Then you calculate and look at the results.
Under Display options you can select "Loads at connection nodes" which will toggles the display of the forces and moments applied at the connection nodes.
Then you can select each stress equation and choose S(Pm) or S(Pb) in the result to display the maximum von Mises Stresses.

Best regards

Rohr2 Support
Sie haben keine ausreichende Berechtigung, um die Dateianhänge dieses Beitrags anzusehen.


Wer ist online?

Mitglieder in diesem Forum: 0 Mitglieder und 1 Gast