Failure to converge analysis with DruckerPrager

[chill]

Hi,
I have tried to include a drucker prager model into my analysis of a layered soil with different elastic stiffnesses into each layer. Each layer has 0 cohesion hence the sigmaY value in all drucker prager layers is 0 and each layer has a different friction angle (peak friction angle) hence the rho values for each layer are different (method for calculating rho is consistent with opensees wiki). The analysis is not converging and I have no idea why, is it possible if you can help please!

The model is attached.

[STKO Team]

The Drucker Prager model seems to have convergence issue. Could you try with another model?

[chill]

Hi,
I have attempted to use the Manzari-Dafalias model (I am aiming for a perfectly elastic-plastic model) however to check whether it is has the desired response before I implement it into the model I have set up a 1 element test. The point of this element test is to consolidate pressure in the x and y directions initially and then apply a z displacement to measure the deviatoric stress, deviatoric strain and mean effective pressure repsonse.

But, when I go to the post processor, I cannot do a gauss point plot so have no way to evaluate the stresses in the element. I have attached the one element model, is it set up wrong/how do I evaluate the stresses and strains in the element.

This element set up worked with the DruckerPrager model and gave the desired deviatoric stress and strain response.

Thanks.

[STKO Team]

The manzari dafailas in combination with a single point integration (SSP brick) does not follow some rules for printing the results, so that STKO does not understand the location of the gauss points.

If you use the stdBrick it will work.
However, I noticed that it takes way to much time in the second analysis, and this time is used in the integration of the constitutive equations, like if the material is trying to do tons of iterations to compute the plastic state. Make sure the parameters are correct

[chill]

Thanks for your help, I am now able to evaluate stresses. As the deadline to my project is looming and getting the Drucker Prager model working would be ideal, is there anything I could try to get the Drucker-Prager model to converge in the model on my initial post. Do you have any idea what the problem may be and is it a problem with my model (e.g. analysis parameters/model setup) or a problem with the DruckerPrager itself?

[STKO Team]

It's within DP itself. Try to use the material tester. Do a cyclic test, and test it with different numbers of increments.
You will see that the smaller is the SigmaY, the larger should be the number of increments, otherwise you will see strange hysteretic behaviors