Hi there
Please could you pass the following suggestion to the relevant personnel?
The option to define stiffness in orthotropic soil materials in Plaxis 3D may be useful. For example when approximating using the Mohr-Coulomb soil model (which exhibits an overestimation of stiffness and therefore swelling-like effects in unloading/excavation stages) for a soil volume, defining different values of E' (considering Drained material in this case) may assist with this.
The inputs of E' may be as simple as defining a value in each of the orthogonal directions x, y, and z.
This may help forego using more "thorough" soil models that require additional testing, preliminary data etc.
Just a suggestion, if you'd also like to share your thoughts
Thanks!
Hi James,
Well, the problem is more fundamental: Mohr-Coulomb assumes soil behaviour is elastic whereas the HS model takes into account elasto-plastic behaviour. In case of an excavation it would be elasto-plastic deviatoric behaviour that makes the difference.
But if you're only interested in how much difference orthotropic stiffness makes for deformation and you're sure failure is not an issue you could try with the Jointed Rock model. This model has orthotropic stiffness .... just fill in high strength for the stratification direction so that it won't fail and in principle you would have a linear elastic model with orthotropic stiffness.
My guess is that you can get pretty descent results, but only as back calculation. Because making a prediction of the increased vertical stiffness may not be so easy. So you would have to guess and play around, is that better than making an educated guess for HS parameters based on for instance correlations?And furthermore, you may be able to reduce the swelling and even turn it into a settlement....but how far would that settlement extend beyond the wall? Generally with elastic models one would still find settlements at very large distance from the excavation while in reality it's more localized.
With kind regards,
Dennis Waterman
Hi Stefanos,
Thanks for the response
I recognise that the Hardening Soil and HSSmall models would be more suitable in excavation / unloading situations. My suggestion was purely speculative, perhaps for cases in which testing would not be available and a preliminary analysis with conservative inputs are required. Using a Mohr-Coulomb model would surely overpredict swelling up after the excavation stage, however if we were able to alter the stiffness in the global vertical direction, how much different would this be to when using the Hardening Soil soil model.
Just a thought in the shower I had.
Let me know your thoughts
Dear James,
Thank you for your feedback.
Can you please elaborate on why would you use the Mohr-Coulomb model in excavation (unloading/reloading conditions) cases and not a more suitable constitutive mode such as the Hardening Soil?
The input parameters for Hardening Soil require laboratory tests such as a triaxial test and pedometer test which most of the times are performed. Also, even a rough estimate based on previous experience or published data can be useful as input rather than using a single stiffness for all conditions the soil experiences.
I agree that advanced constitutive models may have many and sometimes difficult-to-define material parameters, but the Hardening Soil, even though an advanced model, it is not that complex to understand and define.
I understand your request on having a single stiffness but different in each direction, but I would like some more information on how that would help get a realistic result on the deformations.
By the way, I have switched this post to a discussion to fit better to what we are doing (not a question post).