How to use the pipeline postprocessing

About the development of the FEM module/workbench.

Moderator: bernd

thschrader
Posts: 1383
Joined: Sat May 20, 2017 12:06 pm
Location: Germany

How to use the pipeline postprocessing

Postby thschrader » Tue May 08, 2018 5:01 pm

Hi,
how must I set the filters in the FC-FEM postprocessing tools or in paraview
to see the concrete zones under tension alone? For adjusting the reinforcement.
In paraview you can enable the S1/S2/S3 stress plots. What does that mean?
The main tension/compression/shear stresses?
I played with the tools/paraview but I am not sure how to interpret the results.
File without results/cleared mesh. Advices welcome.
regards Thomas
rigid_frame_culvert_FEM.FCStd
(24.53 KiB) Downloaded 22 times
model.JPG
model.JPG (82.21 KiB) Viewed 677 times
FC_FEM_postprocessing.JPG
FC_FEM_postprocessing.JPG (148.13 KiB) Viewed 677 times
paraview.JPG
paraview.JPG (98.06 KiB) Viewed 677 times
User avatar
bernd
Posts: 7303
Joined: Sun Sep 08, 2013 8:07 pm
Location: Zürich, Switzerland

Re: How to use the pipeline postprocessing

Postby bernd » Tue May 08, 2018 6:57 pm

you gone try some very hard nut. Get the reinforcement out of a solid analysis. I really would like to do this, but I have never seen some practical workflow for day by day usage on simple geometry like the one you posted. With shell analysis it is easy, not yet eith calculux because we can not plot yet the needed moments, but with all the proprietary structursl analysis software.

Sorry thomas for posting something you probably know allready, and which does not help. But just wanted to say I'm very interested in this too.

bernd
UR_
Posts: 737
Joined: Tue Jan 03, 2017 8:42 pm

Re: How to use the pipeline postprocessing

Postby UR_ » Tue May 08, 2018 7:37 pm

thschrader wrote:
Tue May 08, 2018 5:01 pm
In paraview you can enable the S1/S2/S3 stress plots. What does that mean?
S1/S2/S3 are the so called "Principal Stresses".
Stress tensor (result from ccx: sxx,syy,szz,txy,...) describes stress components at a particular point (integration point of element).
All components are aligned to global coord system.
If you rotate this tensor (jaw, pitch, roll), one can find a orientation, where shear stresses disappear (= 0) and on the other hand normal stresses become their min/max. values.
Max. normal stress is called S1 and min. normal stress is called S3.
S2 is in between and is orientated normal to S1 and normal to S3
Positive values are tension, neg. values are compression.

So for not reinforced concrete you have to look for places, where S1 becomes positive values (tension).
thschrader
Posts: 1383
Joined: Sat May 20, 2017 12:06 pm
Location: Germany

Re: How to use the pipeline postprocessing

Postby thschrader » Wed May 09, 2018 12:44 pm

UR_ wrote:
Tue May 08, 2018 7:37 pm
If you rotate this tensor (jaw, pitch, roll), one can find a orientation, where shear stresses disappear (= 0) and...
Aaah.. yes, the shear stresses must disappear.
I should update my theoretical background :oops:

@ Bernd:
I am inspired by this book:
(the original text is in german: "Bemessung von Betontragwerken mit Spannungsfeldern")
https://www.ethz.ch/content/dam/ethz/sp ... er_inv.pdf
muttoni.JPG
muttoni.JPG (46.03 KiB) Viewed 616 times
tension_field.JPG
tension_field.JPG (47.13 KiB) Viewed 616 times
User avatar
HarryvL
Posts: 941
Joined: Sat Jan 06, 2018 7:38 pm

Re: How to use the pipeline postprocessing

Postby HarryvL » Wed May 09, 2018 2:38 pm

Thomas - from memory, ETH Zurich had a neat equilibrium based (lower bound) method for dimensioning concrete structures. These pictures brought back memories :)

Judging maximum tensile stresses as @UR_ suggests and the area they work on gives you a good way to determine the amount of reinforcement you would need to apply in that area. Using elastic stresses would make it the FEA equivalent of the ETH equilibrium approach. Don't forget shear failure and crushing though. This would require you to evaluate/plot a failure criterion - basically the "von Mises" equivalent for concrete. Simplest one I can think of is the "Mohr-Coulomb" criterion.
User avatar
bernd
Posts: 7303
Joined: Sun Sep 08, 2013 8:07 pm
Location: Zürich, Switzerland

Re: How to use the pipeline postprocessing

Postby bernd » Wed May 09, 2018 3:55 pm

the book from thomas is from the ETH guys you are talking about harry. Mean you guy talking about the same thing. BTW ETH is just a quater of an hour from my office. AFAIK none of these guys has combined this theory with a concrete volume analysis.

Ahh in the theory the tension material volume has no stiffness, in calculix it has.

Another one, why use this complicated theory. One could just integrate the tension stress and use the result for the reinforcement. It is what is done on any shell analysis.
User avatar
HarryvL
Posts: 941
Joined: Sat Jan 06, 2018 7:38 pm

Re: How to use the pipeline postprocessing

Postby HarryvL » Wed May 09, 2018 7:14 pm

Bernd, I was introduced to this in 1985 in a series of ETH lectures delivered in Delft. In those days it wasn't that common yet to do full FEA analyses in engineering practice, so it was an elegant method for hand calculations. Now with FEA as part of normal design work flow there is indeed less need for this; although it is still good to do some manual checks to make sure the FEA results are reasonable and understood.
User avatar
HarryvL
Posts: 941
Joined: Sat Jan 06, 2018 7:38 pm

Re: How to use the pipeline postprocessing

Postby HarryvL » Wed May 09, 2018 7:31 pm

Thomas, further to my earlier comment. The 2 failure criteria that need consideration / plotting:

Tensile failure: sig1>0 (as mentioned by @UR_)

Shear/Compression failure: (sig1-sig3)>(sig1+sig3)*Sin(Phi) + 2*c*Cos(Phi) (Mohr-Coulomb)

Where c=cohesion and Phi=friction angle. These can be derived from biaxial tests, but can also be estimated from:

Phi = 30-40 degrees (see literature)
2*c*Cos(Phi)/(1 - Sin(Phi)) = fc (uniaxial compressive strength

So a plot of Sig1 will tell you how to design the reinforcement (Sig1>0) and a plot of:

Sig_MC = (sig1-sig3)-(sig1+sig3)*Sin(Phi)-2*c*Cos(Phi)

Will tell your where the concrete fails in shear or compression (Sig_MC > 0). Sig_MC therefore plays a similar role as Sig_Mises for steel.

In principle this can be done with Paraview as a post-processing step.
User avatar
HarryvL
Posts: 941
Joined: Sat Jan 06, 2018 7:38 pm

Re: How to use the pipeline postprocessing

Postby HarryvL » Wed May 09, 2018 7:53 pm

EDIT: incorrect post deleted
Last edited by HarryvL on Fri May 11, 2018 5:00 pm, edited 2 times in total.
User avatar
HarryvL
Posts: 941
Joined: Sat Jan 06, 2018 7:38 pm

Re: How to use the pipeline postprocessing

Postby HarryvL » Wed May 09, 2018 9:06 pm

EDIT: incorrect post deleted.
Last edited by HarryvL on Fri May 11, 2018 4:57 pm, edited 1 time in total.