Poisson's ratio from CPT output

[GoodnightKiwi]

Hi all - I'm having trouble finding a correlation / equation that provies a relationship between Cone Penetration Test data and Poisson's Ratio. Is anyone able to steer me in a direction where I may find it, if the relationship does exisit? Thanks in advance.

 

[Ron]

Not sure there is a good one. Poisson's ratio for most soils will be between .35 and .45. Much like specific gravity of soil, can be guessed as accurately as tested.

 

[dgillette]

Off hand, I wouldn't expect a unique relationship between CPT and Poisson's ratio. Peter Robertson doesn't mention anything like that in GUIDE TO CONE PENETRATION TESTING (which you can download from greggdrilling.com).

We commonly use 0.3-0.35 for static loading or dynamic loading of unsaturated material, and 0.45-0.48 for dynamic loading of saturated material, whose compressibility, or rather the lack thereof, is governed by the high bulk modulus of water (low in proportion to the shear stiffness). The latter actually fits fairly well with theory, backfiguring P's ratio from shear-wave and p-wave velocities.

 

[Mccoy]

Hasn't the value of drained Poisson Ratio been re-evaluated recently?
Paul Mayne reccomends as a general rule (also based on previous studies) 0.1 to 0.2 for granular and unsaturated soils.
It also appears to show, like other elastic moduli, a dependence on strain.

 

[Mccoy]

Sorry it took a while to find this reference:

http://casehistories.geoengineer.org/volume/volume1/issue1/IJGCH_1_1_3.pdf

For realistic values of Poisson’s ratio, the values are quite similar. In fact, D' = E' at ν'=0. Using local internal strain
measurements, recent research has shown the operational value of Poisson's ratio is generally quite small at working loads
levels such that 0.1 ≤ν' ≤ 0.2 (e.g., Tatsuoka, et al. 1994; LoPresti, et al. 1995), and consequently D' ranges from 1.05 E' to
1.11 E'.

 

[GoodnightKiwi]

Thanks for your feedback everyone.

Yes, I read the same thing about the values of Poisson' ratio being between 0.1 and 0.2 for all soil types, with v'=0.5 for undained conditions (refer the attached paper by Mayne and Poulos).

The paper then goes on to suggest that because of original adoption of v' between 0.25 and 0.45, the method of Schmertmann has an unconservative error of some 18%. This is because the peak strain is modelled by Schmertmann at 0.5 x B based on the older v' values, but it should actually be at about 0.2 x B based on the revised values of v'.

What I was hoping to achieve was to use the method of Mayne & Poulos to determine more realistic influence factors to replace the rather idealised and simple triangle distribution put forward by Schmertmann, and use this to get a general feel for settlement using CPT trace data. I figure we all have computers, so why not take the next step away from idelaised triangles.

From looking in the attached paper, their isn't a dramatic change in the strain influence factors between v' of 0.1 or 0.2 so my intial question is probably no longer important.

On a side note, do you think the revised values of v' between 0.1 and 0.2 would have much effect on other aspects of geotechnical engineering such as the Boussinesq bulbs?

 

[Mccoy]

That article from Mayne and Poulos is a very good update of the classic Schmertmann method and even includes revised values of the old 'Fox equations' corrections for foundation depth.
Only drawback (which has been duly outlined) is the approximation in case of layers of different rigidity, with tables which allow to know wether the approximation is over- or under-conservative.

 

[Mccoy]

Going back to the spreadsheet issue, yes a 0.15 value for Poisson would be OK, although there may be some exceptions (but I would use larger values only where such exceptions have been shown). There is a graph in a Mayne, Schneider et al. document which I'll have to dig out.
Probably there is no problem to a small Poisson where the soil shear strain or stress mobilization involved is relatively small, whereas with a substantial mobilization, objections to a small Poisson value might be plausible.