Cyclic triaxial testing 1

[talves]

I am a Geothecnical engineer that is now working on submarine slope stability. I have a cyclic triaxial apparatus that is prepared more for rock testing than soil testing. The thing is that the triaxial tests I perform do not simulate exactly the cyclic effect of an earthquake (i.e. considering an earthquake lasting 1 minute with 50 cycles I can only perform those cycles in 1 day of work). My question is simple: does anyone knows is there is something published and/or studied on how much 'strength recovery' happens with time in cyclic testing of soils? I would like to underline, though, that on the marine slopes I am interested in, most of the layers triggering instabilities are clay-rich beds, most of the time expansive and very plastic, that have nearly negletable porosity. Thank you all for your tips.

 

[Focht3]

You may need to modify your test setup. With a pneumatic 'bellofram" (sp?) system to apply the loads and a load cell to measure the force applied, you can get load cycles in the 1 Hz range - perhaps higher.



Please see FAQ731-376 for great suggestions on how to make the best use of Eng-Tips Fora. See faq158-922 for recommendations regarding the question, "How Do You Evaluate Fill Settlement Beneath Structures?"

 

[talves]

Thank you Foch3. My question is perhaps a bit more complex. I am actually interested in knowing the effects of testing my soil samples using a smaller frequency (i.e. for long periods) and if that will substancially alter the test results. I keep looking at bibliography on cyclic strain testing and none of the presented formulas are actually time dependent. My mistake or the differences in the test results have already been quantified??! Thanks

 

[Focht3]

Hmmm,

The key issue to look at is the consolidation t₁₀₀ time as it relates to the loading rate. If the loading rate is so slow that drainage has plenty of time to occur, then slowing the rate further will have little effect on the reuslts.

If the loading rate is much faster than t₁₀₀ then the loading rate will be a significant factor. If this is closer to your problem, then I suggest that you do a literature search on resonant column research. As I recall, Drnevich (sp?) at the University of Kentucky (?) did a lot of research on this topic. At least one of his former graduate students (Ron E.!) is an [green]Eng-Tips[/green] participant; hopefully he will see this thread and reply...



Please see FAQ731-376 for great suggestions on how to make the best use of Eng-Tips Fora. See faq158-922 for recommendations regarding the question, "How Do You Evaluate Fill Settlement Beneath Structures?"

 

[talves]

Nice to hear from you again foch3. Thank you for your tips. I would like to mention one thing: as the marine samples I am intending to test are at normal conditions saturated with sea water (and oftentimes in relatively shallow depths below the sea bottom) there is a tendency to apply Unconsolidated-Undrained tests to them. No drainage of porewater pressure will occur then...I suppose this will have an effect on the tips you wrote in your previous email, won't it?
Thank you in advance

 

[Focht3]

Yes and no. If the samples are pretty strong (which is typical of UU samples for offshore projects), they will tend to dilate at low strains. In theory, if you run cyclic tests on two identical samples at 'slow' and 'really slow' rates, the strengths will be the same.

If run the cycles at a low strain (0.1%) for 50 to 100 cycles before shearing, you will get some strain hardening in the soil's initial 'elastic' modulus. Evaluating the low strain modulus is very useful for evaluating (large) shallow footing settlements in overconsolidated clays -

Have you done any research on Drnevich et al? What did you find?



Please see FAQ731-376 for great suggestions on how to make the best use of Eng-Tips Fora. See faq158-922 for recommendations regarding the question, "How Do You Evaluate Fill Settlement Beneath Structures?"

 

[talves]

Thank you once again for your reply. I am actually happy for knowing that the strength of the samples won't be altered by testing them at 'slow' and 'very slow' loading rates. I don't know about the increase of loading I will aply, though and wonder if there is a difference between 'fast rates' (<1 Hz) and 'very low' rates. Whatever loading gradient I use from cycle to cycle, it was to be as close as possible to the recorded earthquake dynamics of the Greek areas I will be working on. Also, since the layers of interest are relatively young in age and shallow, it is expected them to be normally or under-consolidated.

I have done some research on Drnevich et al. and even sent him an email. Now, I just have to wait for a reply...

Thank you once gain for your help. Regards

Tiago

 

[harney]

talves--

A main issue regarding loading rates applied in cyclic trixial tests appears to be stress and porewater redistribution during shear. "Back in the day" Seed and his boys used to test at frequencies of 1 Hz (and faster) to simulate EQ loading conditions. A problem with this is that the test cannot be idealized as 'elemental' since porewater pressures are not allowed to equilibrate throughout the specimen, and the pwp being measured is likely not that induced on some potential plane of failure.

Many researchers now seem to be using slower loading rates, especially for liquefaction studies. These rates appear to be on the order of having periods of 10 to 30 seconds (Clarence Chan recommends 30 sec, I think). As long as the rate is slow enough for pwp redistribution and equilibration, there likely is not an effect of running the test "too slow" (within reason).

I also agree with Focht3 that you should account for the limitations of your apparatus. Pneumatic systems in particular seem to be problematic in achieving very high frequency loadings, due to their inability to push/relieve enough quantity of gas in short periods (for stress-controlled systems).

If you are looking for professional (consulting) advice, I'd have to recommend Steve Kramer at the University of Washington. He's at the forefront of EQ Engr and liquefaction research, and I know he has also done work (research and consulting I think) on submarine slope stability.

 

[Focht3]

[blue]harney[/blue] is right - don't try to measure pore pressure at "high" load cycling rates! My comments were directed at total stress approaches when dealing with 'dynamic' loadings, although I failed to note that in my comments.

And pneumatic systems are quite limited in the loading rates they can apply. You will need to match the loading system to the soil stiffness, test frame and desired load cycling rate. Pneumatic systems do offer some advantages, though - the load can be relieved more quickly than load cells or proving rings. (No 'coiled spring' effect.) And the load is not dependent on a specific amount of sample movement.

No loading system is perfect -

On the issue of thixotropy:

My question is simple: does anyone knows is there is something published and/or studied on how much 'strength recovery' happens with time in cyclic testing of soils?

I haven't seen one (that I recall.) You could also contact Steve Kramer per [blue]harney[/blue]'s recommendation; you might also talk to Ken Stokoe at Texas or Don Coduto at Cal Poly Pomona. One of those guys ought to be able to put you on the right track.



Please see FAQ731-376 for great suggestions on how to make the best use of Eng-Tips Fora. See faq158-922 for recommendations regarding the question, "How Do You Evaluate Fill Settlement Beneath Structures?"