Welcome to the forums. Please post in English or French.

You are not logged in.

#1 2017-09-08 06:41:05

winksaville
Member
Registered: 2017-08-04
Posts: 19

Test von Mises stress (signed)

A month ago I asked "von Mises stress is not the absolute value of von Mises stres (signed)" on "Code_Aster usage" because I found in a simple test of a cantilvered beam where some time the absolute value of "von Mises stress (signed)" was not always be equal to "von Mises stress" as I point out in that post.

Looking at the tests I do find some tests that where VMIS and VMIS_SG are tested such as ssvn112n.comm, ssvn112a.comm and wtnv144a.comm and in those tests I do see that the absolute(VMIS.VALE_CALC) is equal to VMIS.VALE_CALC.

So my question, is there a test that compares every VMIS.VALE_CALC against absolute(VMIS_SG.VALE_CALC)? or maybe my understanding is incorrect and they are not always equal.

-- Wink

Offline

#2 2017-09-08 17:01:00

Thomas DE SOZA
Guru
From: EDF
Registered: 2007-11-23
Posts: 2,853

Re: Test von Mises stress (signed)

Hi Wink,

VMIS and VMIS_SG components from SIEQ_* fields should always be equal in absolute value.
Indeed VMIS_SG is just VMIS multiplied by the sign of the stress tensor trace.

Hence, the existing tests are correct. Not all testcases will verify that this equality but because of the way these quantities are built, verifying them in a few testcases is enough.

If you happen to have a calculation where this is not the case, it may be a bug and I encourage you to attach files to reproduce the error.

TdS

Offline

#3 2017-09-08 20:20:58

winksaville
Member
Registered: 2017-08-04
Posts: 19

Re: Test von Mises stress (signed)

Thanks for the reply!!!

For a possible error please see "von Mises stress is not the absolute value of von Mises stres (signed)" on "Code_Aster usage", but since that post I have a simpler test.

I created a square beam 30mm x 30mm x 90mm in OnShape {1} and imported it into SimScale {2} where I meshed it and ran a simulation with the material being aluminium, one end of the beam fixed and the other having a -1000N load on the Z axis.

I then downloaded the results of the simulation to my computer and used Paraview to view the results. The graphic results look quite reasonable, but if I look at the spreadsheet view of the data, filtering out all but "von Mises stress" and "von Mises stress (signed)" to make it easier to view, you see that the point[15].VMIS = 2.21281e+06 and point[15].VMIS_SG = 315597:

vtkOriginalIndices,Points:0,Points:1,Points:2,__vtkIsSelected__,von Mises stress,von Mises stress (signed)
0,0,-0.015,-0.015,0,1.92423e+07,-1.92423e+07
1,0,0.015,-0.015,0,1.96056e+07,-1.96056e+07
2,0.09,0.015,-0.015,0,1.30009e+06,1.30009e+06
3,0.09,-0.015,-0.015,0,1.22461e+06,1.22461e+06
4,0.09,0.015,0.015,0,1.34788e+06,1.34788e+06
5,0.09,-0.015,0.015,0,1.28705e+06,1.28705e+06
6,0,0.015,0.015,0,2.06716e+07,2.06716e+07
7,0,-0.015,0.015,0,2.17607e+07,2.17607e+07
8,0,-0.015,0.013125,0,1.72486e+07,1.72486e+07
9,0,-0.015,0.01125,0,1.46712e+07,1.46712e+07
10,0,-0.015,0.009375,0,1.17931e+07,1.17931e+07
11,0,-0.015,0.0075,0,8.36993e+06,8.36993e+06
12,0,-0.015,0.005625,0,6.77667e+06,6.77667e+06
13,0,-0.015,0.00375,0,5.32268e+06,5.32268e+06
14,0,-0.015,0.001875,0,3.1453e+06,3.1453e+06
15,0,-0.015,-9.71445e-17,1,2.21281e+06,315597

I'm a total noobie so I've very likely done something stupid, in anycase let me know if there is anything else I can do.

Links, remove the space after "https " and copy/paste into browser:
{1}: https ://cad.onshape.com/documents/6630a295f0826388c6d4790f/w/d439940fdf48e2bbadbd2f51/e/e5f086039571a74a7d520668
{2}: https ://www.simscale.com/workbench?publiclink=3c265235-c284-4747-93ec-71392324eb15

Offline

Board footer