PressureIndependMultiYield Material and angle of friction

[ChrisS]

In the documentation of PressureIndependMultiYield Material (note 1) the shear strength is related to angle of friction (and cohesion). It is also stated that "the yield surfaces are of the Von Mises type". There seems to be a contradiction.

I did some research on this and my findings can be found in the following document:
https://www.dropbox.com/s/wsh850b4w1bqy ... d.pdf?dl=0

I'm looking forward for your corrections and comments.
Chris

[skamalzare]

What sort of contradictions are you concerned about? I can't open your dropbox file.

[ChrisS]

The von Mises criterion is independent of mean stress (a cylinder). The given equation gives bigger shear strength for bigger mean pressure (like a cone).

I tried the dropbox link in another computer and it works (had to wait several seconds for it to load) .

[skamalzare]

Chris,

I went over your dissertation, looked into the source code and checked the equation system with FE book of Profs. Potts and Zdravkovic. I have to admit you are absolutely right. Good catch!

The yield surface behind PIMY is Drucker-Prager form, and its size changes with confinement.

The equation system in the source code, as you've pointed out, is wrong for the case with phi>0. We should correct line 00863 to be:
residualPress = 2.* cohesion / Mnys;

Again, thank you for pointing this out, and I hope Frank can help us with the corrections.

P.S. In your dissertation, in equation 2.48a, you need to take J2 out of square root. You have used the equation correctly in subsequent equations.

bests,
Soheil

[ChrisS]

Dear Soheil

You have spotted correctly the problematic line of the source code. And you have proposed a rewrite that makes the source code in agreement with the documentation.

However in my original post I argued that both source code and documentation had mistakes. I have reworked the proof, this time with only geometry and definitions and I ended with the same formula. Please have a look at it:
https://www.dropbox.com/s/ixnt29pg3ujw6 ... q.pdf?dl=0

My proposal is the following line:
residualPress = cohesion * cos(frictionAngle * pi/180.)/ sinPhi;

Thank you for your attention.

Kind regards
Chris