Cohesion intercept of clays-fact or fiction

[MRM]

Please offer your opinion about the following. Correct, disagree, or make fun of me as required…

My current understanding and opinion are that "cohesion" should not necessarily be used for design purposes and it should not really even be recommended in a geotech report, if the actual strength parameters are to be used.

Some of my reasons for this are as follows; first, "cohesion" as it is plotted on a chart is sometimes an artifact of the curved Mohrs envelope, and second, cohesion for soils is not as reliable or dependable as it is for real rock. That is it may change with time. Cohesion is not easily defined, universally. Hmmm, perhaps we’ll hear more on that…

Therefore, effective stress parameters for a stiff clay (for example) should include a friction angle only. But very importantly, that friction angle should be unique to the stress range in question. The same holds true with a soft, NC clay-an effective stress friction angle should be reported if effective stress conditions are desired.

I’m not saying that you won’t arrive at the “right” answer in a bearing capacity calculation if a phi/c parameter duo is used. What I am saying is that I don’t believe the phi/c duo is as fundamentally correct as many would think.

What do you think?

 

[GeoPaveTraffic]

I strongly agree with your comments about the phi angle being applicalble over some range of normal stress. I have seen many examples of tests conducted at normal stress far above or far below the stress of interest.

I disagree however, that clay soils have a long term, i.e. effective cohesion of zero. If you take a block of medium stiff clay, say 10 by 10 by 10 feet, it will stand for a very long time. If the effective cohesion were zero, the block would fail. I have heard some of the theories as to why the block will stand, but from a practical stand point the block stood and cohesion is a relatively simple way of making the theory fit the observed facts.

In the time I worked as a consultant, about the only time that the long term strength of clays was important was in slope stability calculations. I have worked enough landslides and done backin analysis that I'm convinced that most clay soils exibit cohesion in the long term.

 

[Mccoy]

I'd like to add a couple of points related to this discussion:

It is a known fact that the linear approximation to the failure curve in the Mohr model may entail large errors in particular conditions.
The restriction of measurements within the desired stress range might improve the method reliability, though it would be interesting to determine which would be this optimum stress range; too much a wide one: error due to wrong linearity assumption; too much a narrow one: possible difficulties in obtaining a good correlation in least squares fit. But I wonder, it shouldn't be too hard to come up with a generalized curvilinear model suited to most soil samples. Are there practical problems in that? It has been a long time I didn't have a chat with the lab guys.

2nd point: It's a known fact, often reported in the specific technical literature, that cohesion values (both effective and undrained) exhibit always a significant scatter, larger then friction angle values, thus being intrinsically less reliable.

 

[MRM]

I think Geopave's example of a 10 foot cube of medium stiff clay could be interesting to expand on. Maybe we need to first describe the clay and where it came from in case the behavior under zero stress conditions is dependant on that. I'm interested to see if anyone thinks there is a difference in the approach to that example if the clay's history were one the following two:

Scenario 1: The medium stiff clay came from a relatively shallow depth. The clay, at this shallow depth, was overconsolidated by desiccation or by other means. The effect of this overconsolidation is what gave the clay a medium stiff consistency.

Scenario 2: The medium stiff clay came from a very deep level below grade. Insitu, the clay was very close to normally consolidated. The effect of this high overburden pressure is what gave the clay a medium stiff consistency at this depth.

Assume that both scenarios involve removing the clay (magically) without changing its stress conditions until it reaches the surface and is set down.

Would both clays act the same when they are brought to the surface?

 

[gandersen]

This is an interesting discussion. I have always considered that the selection of c/phi depends upon the stress level. I also agree that based on "fully saturated" soil strength tests, unless there is some cementation, cohesive soils do not have a true cohesion. Jim Mitchell did a research project for NASA several years ago looking at the strength of cohesive soils at low effective stresses. Unfortunately, I no longer have a copy of the report.

I agree that looking at long term slope stability, it is necessary to consider cohesive soils with a cohesion in order to rationalize the observed performance. The way that I have thought about it is that the apparent cohesion could be related to negative pore water pressures in unsaturated conditions. In fact, we just worked on a slope in the greater Cleveland, Ohio area where in order to rationalize a stabilization approach, I had to use a suction-induced cohesion. The slope has been in place for awhile and is doing fine.

Glen Andersen, Sc.D., P.E.
BBC&M Engineering, Inc.
Cleveland, Ohio