I had a question revolving around the friction hardening law based on UBCSAND model utilized in CY model embedded in FLAC3D-V5. Apropos of CY model, the users are supposed to define a table of the mobilized friction angle in terms of the plastic shear strain. Inasmuch as the ensuing equations, i.e. Eq.1 and Eq.2, were posited for the frictional soils, I have tended to think of modifying the aforesaid equations so as to harness in the cohesive frictional soils medium in light of the fact that the friction hardening table in CY model ought to be adjusted to reflect more realistic behavior for a cohesive material.
Eq.1
Eq.2
Having perused HS model in PLAXIS plus CY model in FLAC3D, the formula concomitant with the elastic shear modulus has been rectified for the clayey soils as follow:
Eq.3
As respects the plastic shear modulus formula, can I have your say whether or not Eq.1 is expected to be modified with the intention of adding the mobilized cohesion?
The book "Disturbed soil properties and geotechnical design" by Andrew Schofield is dedicated to show that all soil is cohesionless. Clay "appears" to be cohesive only because of very slow draining rate. He concludes so because Clay is "cohesive" only for a short time - for the duration required for pore pressure to be lost.
As cohesive soil properties are not well understood, one must be careful when making assumptions regarding them.
Do you have real experiment results to refer to? Does the "improved" model "mimic real life" soil patterns and properties? Within what limits, with what margin of error? These questions are unavoidable in an objective dialogue.
From the few encounters I had with fine silt - it appears to remain equally cohesive, albeit softer, while liquefying in my hands. (I always play with excess material, to get a "feel" for it).
In a triaxial test silt behaves like a "very slow sand". Pore pressure dissipates very slowly, silt behaves like undrained sand event at relatively slow loading cycles. Extrapolating with respect to particle size one might assume clay could behave like a "very slow silt" then... which makes me doubt validity of Eq.3, where cohesion based normalizing is present.
I have good examples of stiffness changing proportional to confining stress (p) in sand. Sand is different from clay in many ways - but the observation do support G is proportional to p/pref (Eq.2.) - at least partially, in at least some cases.
Eq.1. appears to be the most "Physically" correct of all 3 - based on numerous triaxial tests of drained and undrained soils, ranging from silt to well graded gravel
In general - work in laboratory developed deep skepticism towards existing soil models. Models have strict limits of application which must be followed scrupulously. By combining two models one risks accumulating the errors, rather than improving solutions.
P.S. I' miss formulae which normalize G with strain (ε) history. ε history has a tremendous effect on sand stiffness. One can literally control G by manipulating nothing but ε during a triaxial test.
- Badges
- Science topic
- Similar topics
- Engineering
- Civil Engineering