Burst Pressure of a thick-walled Sphere

[HarryvL]

... and now it also becomes important to look at the plastic strain to judge potential rupture of the steel.

[HarryvL]

So in this case fcFEM over-predicts theory by 0.045% ... not bad

Is it possible to share your model?

Both uploaded here: https://github.com/HarryvL/fcFEM/tree/m ... eCAD_Files

[HoWil]

So in this case fcFEM over-predicts theory by 0.045% ... not bad

Is it possible to share your model?

Both uploaded here: https://github.com/HarryvL/fcFEM/tree/m ... eCAD_Files

Thanks.

[HoWil]

FYI I remember now, I saw this kind of routine in Freelem software. After installing the soft, for the first run, a stress test routine was testing 126 typical study cases. Then the soft told to the user a result "126/126" tests succed. Something like that.
http://www.freelem.com/qualif/index.htm

Hi alex, Hi HarryvL,
Hope such a testing can be part of this years Gsoc .... please see also https://forum.freecadweb.org/viewtopic. ... 28#p299023
ElmerFEM also includes tones of test-cases which could be used for comparison.... https://github.com/ElmerCSC/elmerfem/tr ... /fem/tests or http://www.nic.funet.fi/index/elmer/doc ... nonGUI.pdf
BR,
Howil

[bernd]

Hope such a testing can be part of this years Gsoc ....

let's say improved. Because we gone have some tests already. Not as much as there should be but there are a few ...

- start FreeCAD
- run code ...

Code: Select all

import Test, TestFem
Test.runTestsFromModule(TestFem)

See https://github.com/FreeCAD/FreeCAD/blob ... TestFem.py for some help

We even have real test cases for whole models, ok just one but we have ... https://github.com/FreeCAD/FreeCAD/blob ... manager.py This one is not yet a verification test. But as harry stated in oofem I made the examples verification tests too. See https://github.com/berndhahnebach/FreeC ... e07a9a11b0

Locally I have a few more, but did not have the time to push the to master.

BTW: harry, cool stuff Just for my understanding. By the use of calculix an non linear material we should get the same results, shouldn't we?

[HarryvL]

... just for my understanding. By the use of calculix an non linear material we should get the same results, shouldn't we?

In principle yes, but Calculix cannot handle plastic collapse with load control (yet). I asked Guido a year ago and he said it was on his TODO list, where (I think) it still is. One of the reasons I decided to start something that is fully in our control.

[HarryvL]

Fans of unit testing may enjoy this one. The work I did on optimising band width has introduced an error in boundary conditions.

punch_gone_wrong.jpg (79.79 KiB) Viewed 977 times

I now know that I have broken the two golden rules of profiling and optimizing:

1) Don't optimize too early - I did, making changes harder to manage
2) Don't assume you know where the bottlenecks are - I did, having spent a lot of time on band-width reduction, only to find that the bottleneck was in stress updating

I have now included cProfile in my code to tell me where the bottlenecks are (een ezel, in 't gemeen, stoot zich niet twee keer aan dezelfde steen)

[HarryvL]

Fixed:

punch_fixed.jpg (44.47 KiB) Viewed 970 times

[HarryvL]

... and now it also becomes important to look at the plastic strain to judge potential rupture of the steel.

With the critical plastic strain utility (https://forum.freecadweb.org/viewtopic. ... 93#p303392) it is now possible to judge the potential of ductile fracture for both designs.

The steel used for the pressure vessel is high strength steel and the ductility in the SMCS model should be chosen accordingly. In the absence of actual data I select alpha=1.0, which is at the lower end of what is reported in literature. However, without specific material data this choice may be unconservative !

Like in the case of the plate with hole, I determine the critical plastic strain ratio at plastic collapse:

fcFEM_Hollow_Sphere_L_u.png (29.23 KiB) Viewed 891 times

fcFEM_Hollow_Sphere_Epsc_Ratio.png (241.29 KiB) Viewed 891 times

The ciritical plastic strain ratio for the sphere is well below the value of 1.0 at which ductile fracture may occur.


However, the results for the hollow cube are quite different:

fcFEM_Block_with_Hole_L_u_2.png (28.31 KiB) Viewed 891 times

fcFEM_Block_with_Hole_Epsc_Ratio.png (132.51 KiB) Viewed 889 times

Now the critical plastic strain ratio far exceeds what is acceptable and ductile fracture is likely well before the ultimate load is reached.

Close inspection of the results shows that the stress triaxiality in the cube (T=3.7) is much higher than that in the sphere (T=0.7), leading to brittle behaviour.

[bernd]

BTW: Harry, How do you do the diagramms?