Choice of Nkt for estimating undrained strength of clay from CPT

[dgillette]

Most commonly, undrained strength, Su, is estimated as (qt - sigma-v) / Nkt. What is current preferred practice for selecting Nkt? Obviously, it depends on what strength test Nkt is referenced to, and to some extent on geology of the deposit.

I usually look at Aas, Lacasse, Lunne, Hoeg in the 1986 ASCE In Situ conference (Blacksburg VA). They have a plot of Nkt referenced to "lab" strength, which is the average of triax compression, triax extension, and DSS, as a function of PI. By that, I should be using 14-18, given PI in low 20s to low 30s. Is there something else I should be looking at? I'm dealing with a fairly sensitive clay, that MIGHT have been leached by fresh water after saltwater deposition. (I'm checking on that.) NC to lightly OC.

Cheers!
DRG

 

[Sudburyboy]

I've seen vanes, typically a Nilcon, turned in borehole adjacent to one of the CPT probes. This allows you to adjust Nkt to match the profile. Not sure if it is applicable for marine clays though.

 

[dgillette]

Thanks. We do have VST next to several of our CPT holes, but VST has its own issues, commonly requiring an empirical correction based on field performance, as a function of PI (Bjerrum), or indirectly as a function of OCR (Aas and friends). The correction wraps a whole lot of things into one fuzzy number (anisotropy, strain rate, and the fact that the failure surface around the vane bears only the slightest resemblance to the cylinder we generally assume for peak strength). With Nkt = 14, the CPT values match up reasonably well with UNCORRECTED VST, but I would probably need to increase Nkt to ~17 to match the CORRECTED VST. Given all of the issues the VST has (and the very high strain rates an earthquake would cause), I don't want to consider VST, corrected or not, to be the touchstone by which CPT gets "calibrated."

Sudbury ON, home of big nickel operations?

 

[coneboy]

In the past we have conducted numerous Nilcon vane borings adjacent to CPT tests in order to determine the site specfic Nkt factor. Some of the tests in very soft, sensitive soils showed an Nkt of 11. Other tests in soft to firm clays showed Nkt values in the 11.5 to 13 range. Recent tests in firm to stiff clays revealed an Nkt of 13 to 13.5 Your value of 14 seems reasonable.

I have seen the Su correction chart based on PI values and have discussed that issue with some of fellow Geotech Engineers. I get the impression that most people in this area of the west coast do not use the PI correction method, nor do the use the OCR correction.

Your uncorrected vane value correlation of 14 would seem to be a reasonable value.
Perhaps start with the Nkt of 14 and see if your Su values make sense. Did you plot the estimated OCR values using the Nkt of 14 and an appropriate value for Su/EOS?
Did the OCR values seem to make sense?

Just some other thoughts:
Did the vane operator seem to know what they were doing? Did the circular vane plots correctly show the rod friction? Did you use a low capacity electric cone like a 5 ton? A big huge 20 ton cone borders on useless in soft to very soft clays.
I will see if I can dig up a past example to post.
Coneboy

 

[dgillette]

Thanks, coneboy. I should have stated that the stress path I'm most interested in is more nearly DSS than triax compression (TC), and so a priori we expect lower strength and higher Nkt than if Nkt is referenced to TC (whether CIUC or CAUC). Therefore, I would expect Aas and friends' Nkt correlation referenced to "lab" strength (average of TC, TE, and DSS) to be more realistic for this case than their other one referenced to TC alone. (Typically DSS strength is roughly midway between CAUC and CAUE.) Just wondering if there was something newer that would supersede that.

When I put all the pieces together (lab consol data, lab shear data, CPT, etc.), 16 or 17 seems to fit a little better than 14, especially if I apply the Bjerrum correction for VST. The lab undrained strength ratios are coming out lower than typical, and unfortunately, a lower Nkt isn't going to help me a whole lot.

I have pretty high confidence that our people did the VST and CPT as well as they can be done. Measuring rod friction is SOP, and they knew the cone was going to be pushed through some pretty soft stuff, so would have brought the right equipment. (I don't know whether they have different cone bodies, or if they swap out load cells like the Wissa cone I used in the '80s.)

Best regards,
DRG

 

[Sudburyboy]

Bang on with the operations. My hometown.