Triaxial Testing - Which test to run at which situation? 3

[tocoadog]

CD, CU, Quick Test. I was asked by a junior engineer that this morning, and I couldn't answer with confidence.

Any comments would be appreciative. I have very little experience with clays, but in the past we have always performed drained tests on sands. Also, shouldn't you always run undrained on clays to model the potential for pore water pressures inthe future? Also, when to run consolidated as oppossed to unconsolidated? Potential fill materials for unconsolidated? Again, any help would be great. As you can guess, this is killing me trying to find the answer??

Btw, I did look for the topic in past threads, and while I may have overlooked a post, I didn't see anything that specifically answered the question.

Thanks again.

 

[MRM]

Hi,
It may be the best to see a soil mechanics text or three about this. That way, you'll have the necessary strength and pore pressure diagrams to see too, which are really helpful in understanding the differences between CD, CU, and UU testing. Once you understand those testing formats, the idea behind the Torvane shear strength device, pocket penetrometer, and the vane shear test will also become clearer. Additionally, you should become very familiar with Critical State Soil Mechanics.

I can offer the following facts on this topic and then maybe some others will chime in. We can fill in the gaps as the discussion unfolds, and I can be corrected where it is appropriate. I’ve always been interested in this stuff and I use it frequently, I’ve never tried to explain it to anyone before. Let’s give it a try…

1. These different testing types generally model different loading conditions. For example, a slope, with soil consolidated to a given stress level, failing progressively over time, may be doing so in a drained fashion. A slope suddenly cut in a saturated soft clay, also consolidated to a given stress level, may fail in an undrained manner. We all know about the potential danger in quickly loading a soft saturated soil; it will likely fail in an undrained manner. An overconsolidated clay that is suddenly cut to form a slope may actually have stronger undrained conditions, and then become more unstable over time, finally failing in a drained manner (again, CSSM is what helps explain this).

2. As I’ve alluded to above, if you're looking at natural soils in the ground, the only testing results you'll be interested in are of the "consolidated" variety. That's because the soil you're trying to model is consolidated to a certain stress level in the ground. That doesn't necessarily mean that the UU test is worthless though... See lower down.

3. The CD and CU tests are both consolidated to some effective stress level prior to the shear phase. The CD is sheared slowly as to maintain drained conditions, while the CU test is sheared quickly to maintain undrained conditions.

4. The CD test is an "effective strength" test. That is, it provides an effective friction angle. One problem and complaint with this test is that it takes a long time to run. Because of this, I’ve never run one, ordered one, or have actually heard of one performed in practice.

5. The CU test is handy because it can be run relatively quickly (compared with the CD test). If it is outfitted with a pore pressure gauge, then the results of a CD test can also be obtained. In other words, the CU test can either provide "total stress" parameters (Su) or both total stress (Su) and effective stress parameters (phi') depending on whether pore pressure is obtained during the testing.

6. When using the results of an effective strength test, such as CD or CU (when pore pressure info allows it), the strength for any depth within the NC zone of a soil of similar mineralogy can be determined. This is not so with a total stress test. With a total stress test, only the soil strength at the depth the soil was recovered is available. So a CU test (without pore pressure data) would likely need to be run on samples at multiple depths in order to use a total stress method for design purposes.

7. The UU test works on the premise that a sample of saturated NC clay pulled from a given depth, trimmed, placed in a triaxial cell, loaded to a total pressure without allowing drainage, and then sheared undrained will yield similar results to the Su gained during CU testing. This Su parameter is unique to the depth the sample was taken from. Again, we're back to only having strength info about the depth the sample was taken from with the "total stress test." But having said that, you can see that the UU test actually does represent a “consolidated” test-the consolidation pressure is the same average effective stress that was on the sample in the ground. This can partially be explained by the fact that the failure was actually governed by the effective stress within the sample. This is true whether we are able to measure it or not.

8. The UU test is one of the most misunderstood tests for the reasons in #7 among others. However, if you understand the theory behind it, it is still very useful. An even cheaper version of the UU test is available; the unconfined compression test. This test works the same way, although with a zero confining stress. And there is yet an even cheaper way to get an idea of the undrained shear strength at a given depth; the pocket penetrometer or Torvane taken on samples at varying levels. These tests do suffer from some forms of error and other limitations, but if you understand them, these tests will help provide useful Su data. Some of the error between these simple tests and the more expensive UU and CU without pore pressure measurements is even “offsetting” in my opinion provided a good clay sample is under consideration.

 

[dgillette]

Good summary, MRM.

My only quibble is the use of the term "total stress" in describing undrained strength. With SHANSEP, CSSM, Phi-cu, c/p', etc., we use effective stresses to predict undrained strength. I've seen some fairly senior engineers get fouled up by the customary terminology and do things that don't make much sense when examined in effective-stress terms. I think it's clearer to just avoid the term and use "undrained strength" or "undrained analysis" instead. (Exception: partially saturated materials, where total stress CAN be used to predict the strength, at least until there is a change in the moisture content. It's useful for end-of-construction stability analysis of large embankments.)

Best regards,
DRG