CU shear test confining pressures 5

[pelelo]

Engineers,

I have Shelby tube recovered at 15 ft below surface, which contains clay (CH).Water table was detected at 10 below ground.

It is my understanding that for CU triaxial tests, we need to provide to the lab the confining pressures, usually are 3 pressures.

My question is, how do you compute such confining pressures?, Do you just use K at rest x S'v.

Naturally, the confining pressures "envelop" that you provide to the lab is supposed to cover the computed confining pressure.

Please let me know your thoughts. Or if you know of a reference, please let me know,

Thanks.

 

[geomane]

For isotropically consolidated CU tests, I typically see confining pressures starting around effect overburden, then to total overburden, then to some increment higher depending upon your loads. That’s for effective stress envelope.

If you’re trying to determine undrained shear strengths, it should be based on your effective overburden.

 

[fattdad]

Here's the excerpt from our manual of instructions (VDOT).

CU-bar (consolidated undrained triaxial compression strength with pore pressure readings – ASTM D4767) testing provides the total and effective friction angle and cohesion intercept of soils. Consolidation cell pressures shall consider the effective overburden stresses, the proposed effective stress and a third effective stress condition such that the stress range in the laboratory replicates the anticipated stress conditions in the field. Soils for strength testing shall be classified in the laboratory. When preparing remolded samples for CU-bar testing, the sample shall be remolded at the minimum acceptable dry density and a moisture content that is wet of optimum.

f-d

ípapß gordo ainÆt no madre flaca!

 

[geomane]

fattdad,

Is the total stress envelope only used for rapid drawdown cases of dams? Well the Rbar total envelope parameters are converted to a separate set of parameters. Soil Strength and Slope Stability (Duncan) and the USACE Slope Stability both echo this, at least that’s how I interpret it. They state the total stress envelope doesn’t model any real world scenario.

 

[fattdad]

From a CU-bar test, you get effective and total stress behavior and strength. To that end, you can use the data for those analyses that rely on such values. What's the end game?

f-d

ípapß gordo ainÆt no madre flaca!

 

[pelelo]

fattdad/Geomane,

Thanks a lot for your responses.

Fattdad:

1 - So the c and phi values will be based on overburden stresses?. I thought the lateral stresses (Ko x S'vo) at some point would had role.

2 - When the "VDOT" manual states "the proposed effective stress". Does it mean my current effective overburden stress + the applied load?. If so, would you use the current effective ob stress + the change of stress (due to boussinesq)?.

3 - I was reviewing a lab test result. It was proposed 3 effective confining pressures: 25 psi, 30 psi and 35 psi. Sample depth was 35 ft. WT at 5 ft below surface. Sv'o ~ 2616psf = 15 psi. At that time the intent was to build a fill for a road. Do you think the test was useless, since SV'o doesn't fall in the 25 - 35 psi confining pressure range?. Soils were sandy material.

Please let me know,

 

[geomane]

fattdad,

This is the paragraph I was referring to. See attached photo.

 

[geomane]

Well that didn’t work. Try this one.

 

[GeoPaveTraffic]

peleo,

I think you are making this more complicated than need be. If you are using a c / phi model for the soil strengths; then the confining stress really doesn't matter since the model is a straight line that is applicable at all confining stresses. Here is how I would pick the confining stresses.

First confining stress at just above the estimated maximum past effective vertical stress at sample location. You want the stress a little higher than the past maximum so help reduce disturbance due to sampling.

Second confining stress just above the maximum planned effective vertical stress.

If you run a third test, and this is often not done in practice, then run it at whatever stress makes you happy.

Note that the above assumes you have three different samples to test. If you are doing a multistage test, i.e. running more than one test on the same sample, then the second and third stage pressures would be different. The second stage would be double the first stage and the third stage would be double the second stage. This is done to reduce sample disturbance due to the earlier stage(s).

Mike Lambert

 

[fattdad]

to geopave: You are making a distinction between C and C', Phi and Phi', correct?

to the OP: If the end game is slope stability, we'd want to see both effective and total strength evaluations on strength.

f-d

ípapß gordo ainÆt no madre flaca!

 

[pelelo]

Engineers,

Thanks for your responses.

It seems engineers use different approaches to this:

1 - As you fattdad and geopavetraffic stated, the confining pressures are determined based on the vertical stresses.

2 - Some other engineers use the K at rest x S'v as the confining pressures. The only problem with this approach is that you need to know if the clay is overconsolidated or NC in order to select the right Ko value. If NC Ko is between 0.5 and 0.9, if overconsolidated Ko may be greater than 1.

I prefer the approach by VDOT (presented by fattdad) and geopavetrafic.

 

[Okiryu]

I use effective stresses rather than Ko x S'v. Also, I tend to agree with GeoPaveTraffic" "...the confining stress really doesn't matter since the model is a straight line that is applicable at all confining stresses...". So, GeoPaveTraffic's input may answer your question about useless of test data if the sigma3 is not in the range of the proposed stresses:

"I was reviewing a lab test result. It was proposed 3 effective confining pressures: 25 psi, 30 psi and 35 psi. Sample depth was 35 ft. WT at 5 ft below surface. Sv'o ~ 2616psf = 15 psi. At that time the intent was to build a fill for a road. Do you think the test was useless, since SV'o doesn't fall in the 25 - 35 psi confining pressure range?. Soils were sandy material."

Anyway, you can also compare your phi' values for clays (I am assuming you are doing CU-bar tests) with correlations of LL vs. phi'. See attached extract from Lambe and Whitman book.

 

[fattdad]

Just to add, that Phi (not Phi') is horizontal only when the initial sample condition is saturated. That will surely be the case in a CU-bar, as test requires saturation. (Edit to add: Not all samples below the water table are saturated.)

f-d

ípapß gordo ainÆt no madre flaca!

 

[Okiryu]

f-d, yes, good point and agree. Attached is an extract from Bishop and Henkel report "The measurement of soil properties in the triaxial test"

 

[Okiryu]

f-d, BTW, I have seen several posts from you indicating that "Not all samples below the water table are saturated"... I am curious about that. Can you expand?

 

[fattdad]

If you work in a lab and have context on b-values, it becomes obvious that the amount of backpressure to get saturation is quite different than what's in situ.

I think it was professor Duncan that explained that in embankment dams, the embankment below the phreatic surface is not guaranteed to be saturated. It just takes too much pressure to get the last air out of the system, even when inundated.

f-d

ípapß gordo ainÆt no madre flaca!

 

[LRJ]

Generally speaking (there might be some specific applications where this is not the case) the consolidation stresses in a CIU should be based on the mean effective stress, i.e. σ'_m0 = (σ'_v0 + 2*σ'_r0)/3 = σ'_v0(1 + 2 K₀)/3. Using σ'_v0 alone as the consolidation stress is only correct if K₀ = 1, which itself implies an overconsolidated state.

For CAU tests you can vary σ'_v0 and σ'_r0 separately given a K₀ (= σ'_r0/σ'_v0) profile, typically inferred from yield stress ratio (YSR; also incorrectly conflated with overconsolidation ratio, OCR).

Incorrect test conditions are one of the most common sources of error in lab results. The other major factor is sample disturbance for intact materials, which is very common when undertaking cable percussion drilling - I'm yet to get a reliable oedometer result from a sample that was obtained using cable percussion drilling. One means of correcting for sample disturbance in advanced triaxial tests, such as CU tests, is to 'overload' the sample to 80% of its former maximum stress state (as inferred from YSR), thereby recovering the effects of sample disturbance, though this is more relevant for CAU tests than CIU tests. These recent stress history effects have been studied by several authors and shown to have considerable impact on the subsequent shearing results. Of course, the trade off is extra test time - is quality worth the extra cost?