Consolidation Behavior of Clay Under Different Saturation Ratios 1

[bdbd]

My question is related to consolidation test behavior under different saturation ratios.

By standard, common oedometer tests are performed under 100% (approximately) saturation. They wet the soil for a period, then conduct the test. This will give full consolidation behavior.

However, in site, this is not true.

So, what will be the result, if I perform a consolidation test under a sample that is directly taken from the UD, not saturated in lab? Maybe you came across some graph in a book or paper, that shows in same graph, consolidation behavior of a clay in different saturation ratios?

What I am expecting? I expect that in less saturated clays most of the settlement will occur instantly.

Did you see any graph like that, or have you ever performed consolidation test in a specimen that is not saturated under lab conditions?

Regards.

bdbd

 

[oldestguy]

In my experience we never have any situation where compression of fine grained soil is of concern unless it is at or near saturation. Shrinkage and expansion situations come up, at less than 100% saturation, but they generally do not involve consolidation testing. The term "consolidation" usually refers to forcing water out of the voids.

 

[bdbd]

OG,

Thanks. So, did you just assume only elastic settlement will occur, for example, in a clay that has saturation around 70%.

I want to see the oedometer test behavior of these clays. I wonder if we have it in literature.

 

[bdbd]

I have no idea why I put this thread in Earth Retention. Mistake.

 

[fattdad]

at the start of any consolidation test, even when inundated, the sample may or may not be, "100 percent saturated. If you have any experience in triaxial testing, you know that often it requires quite a lot of back pressure to get all the air voids in solution. That backpressure is likely the head required by the natural soil to also become saturated. Imagine a sample from 10 ft below the water table (U=4.3 psi). If that sample requires 20 psi to be fully saturated, it'd be reasonable to think that full saturation may only be found at 46 ft below the water table. Doesn't change the test method though.

In the consolidation test, you inundate and load. You measure strain and time and you get various measurements of Cv. Cv is related to load increment and influenced by stress history and also degree of saturation. As the load increments increase and the strain accumulates, you eventually drive out all the air, rather than dissolve all the air. When saturation is found, you then get Cv for the saturated condition.

Maybe, I'm not being clear, but it's incorrect think that a soil under water table conditions by definition.

f-d

ípapß gordo ainÆt no madre flaca!

 

[bdbd]

fd,

Thanks for explanations, I understand your point.

But, aside from this discussion, do you have an idea about my question? Can you imagine the oedometer results of 6 clay samples with 0, 20, 40, 60, 80, 100% saturation ratios?

Should there be same amount of settlement with different time-settlement behavior? Or, should there be less settlement for less saturated. Latter one justifies the OG's method of calculation..

If there is no study like this, I will perform it.

 

[fattdad]

That'd be difficult research! To take a sample at 80 percent saturation and reduce it to 20 percent saturation would alter the state of stress and impose a stress history completely different from the parent sample.

In the below-80-percent range, I expect you could base the analyses on modulus. Problem is after the first rain (or the rising water table), the negative water pressure would go away, the stress history would remain and the presence of water would trump the elastic behavior - for a clay that is. . .

f-d

ípapß gordo ainÆt no madre flaca!

 

[aeoliantexan]

I have run quite a few oedometer tests on unsaturated loess at the existing water content. The e-log p curve looks classic. There is an "apparent" preconsolidation pressure that depends on water content, density, clay content, and stress history, then a "virgin consolidation" straight line. Most of the compression occurs quickly, but there is some time-compression that may look like secondary compression. I have seen some evidence that secondary compression settlement occurs in the field. I share fattdad's concern that unsaturated clay may exhibit a misleading preconsolidation pressure due to capillary action and might be more compressible in the field if it becomes saturated.

 

[Okiryu]

Similar to aeoliantexan, I have done some consolidation tests in unsaturated clays (80% or so) and the e-log p curves had the same pattern of the classical 1-D consolidation curves. I think that the main difference is that "consolidation" occurs very fast as air is expelling from the sample and not water which depends on the permeability of the soil. However, for unsaturated soils, I think that it is more important to get reliable values for soil modulus to be used in settlement analysis.

aeoliantexan, if you can expand more about your observations about secondary compression, it will be appreciated. I did not get it very clearly.

 

[bdbd]

Thanks for inputs.

You both mentioned the same thing. So, do we have oedometer proval of accepting NO CONSOLIDATION for saturation ratio below 85%? It is over-simplification, no doubt. But for practice, we should do that. I see people calculating consolidation settlements no matter if clay is saturated or not, and we have to make it clear.

I wonder what you guys, aeoliantexan and okiryu, do in practice? A year ago, I asked the same question but couldn't get a good answer. Imagine a soil profile, lets say 20m. GW is at 10m. Soil below is fully saturated, and soil above lets say 50% saturated. It is not possible to have same Eu and mv values but lets say it is.

So what do you do here?

a) Elastic at upper, elastic+consolidation at lower
b) elastic and consolidation at both.

This is one question I couldn't find answer for a long time. My current practice is, if GW is not changing or there is no risk of changing until the building is completed, I do not calculate consolidation for unsaturated or partially saturated layers.

 

[Okiryu]

Similar to you, I do not consider consolidation for unsaturated or partially saturated clays. So, I use a) in your question above. Sometimes, I have used b), but found it overconservative. The reason why I have used b) was because we did not have enough data, just SPT N-values (low N-values).

For a), the main issue is how you can calculate the appropriate modulus for elastic settlements. If you are off, say 30% in you modulus value, you will be getting +/- 30% for elastic settlement values. In my area, for our medium overconsolidated clays, people normally use unconfined compression tests or correlations with SPT for getting modulus (perhaps not the best way to get the modulus parameters, but it is the common practice here). There are also some correlations between Eu and PI from Mike Duncan.

Also, even for saturated clays, if you have reliable modulus values, you may use an elastic approach (Hooke's law) dividing the layer in several sublayers and then integrate all them to get the total settlement. Perhaps this is more applicable for overconsolidated clays. For normally consolidated clays, I will use values from the oedometer.

I am sure that more experienced engineers in this forum will provide better inputs and am also curious to hear from them...

 

[aeoliantexan]

okiryu, It was a long time ago, and I can't access the project files, but asI remember, there was a lot of immediate settlement, not much of an "S" curve, and then a linear log-time compression relationship like secondary compression. If secondary compression starts in a few minutes, you get to see several log cycles.

If the loess was moist enough (say, above optimum) and sigificant earth fill and/or structure loads were planned, we often preloaded with an earth surcharge. There was considerable immediate settlement and some settlement with time, and we would monitor with settlement plates to determine when settlement was essentially finished so we could remove the surcharge. This was nearly always less than 3 months.

We founded a grain elevator on collapsible loess after overexcavating and replacing about 10 feet of it. We monitored the settlement during the first loading, maybe the second season too. I forget how much it settled, probably between 8 and 12 inches. I went back out of curiosity several years later, and it had settled a few more inches. I took that as secondary compression. Later, we drilled more borings for a second structure, and found the water table had risen up into the loess because the farmers nearby had started irrigating the fields. So---who knows?

bdbd, I would rarely put footings on clay at a pressure that would cause virgin consolidation. For your example, yes, I would calculate consolidation settlement in the deep saturated soil if virgin consolidation appeared likely, and elastic settlement in the shallow soil.

I seldom considered recompression settlement using consolidation tests because it seemed to overestimate the settlement greatly.

 

[Okiryu]

Aeoliantexan, thanks for the info. BTW, what is the soil classification of your loess , and also what are their typical water contents and atterberg limits?

 

[aeoliantexan]

Most commonly, the loess was CL. Close to the source, such as the bluffs overlooking the Missouri river valley, it might be ML. Sandy loess might be found in eastern Colorado or near the sand hills. 100 miles or so downwind of the source, it could be CH, but probably not collapsible.

On well-drained hilltops, especially with trees, water contents might be roughly 8% to 15%. In areas with some infiltration, usually in the high teens to mid-twentys. Where underlain by glacial till or other clay, it might be saturated - around 30%.

We did not run a lot of Atterbergs, but PL either side of 20 and LL 30s and 40s.

Based on 20-year-old memories.

 

[Okiryu]

Thanks aeoliantexan. We do not have collapsible loess here so I just was curious to see their properties. Thanks again...