Long term stability of natural slope in soft clay

[JWB46]

Hi guys.
The 'textbook' philosophy for short term stabilty assessment of say an embankment is to use a total stress analysis, and to use an effective stress analysis for the long term assessment.
Suppose we have a natural slope in soft clay where there is little or no natural drainage. How should the long term assessment be made? Terzaghi & Peck seems to imply an undrained analysis.
Any thoughts or references would be very much appreciated. Thanks.

 

[Ginger]

If the natural slope has been there for a substantial length of time, then the exposed face will probably be the most likely drainage path. Therefore, you should follow Terzaghi & Peck's model. If the slope is to remain there long term then the same philosophy applies. the soil will eventually drain if the water table conditions allow it to.

Andy Machon

 

[JWB46]

Thanks Andy.

 

[Focht3]

Hmmm, maybe I'm misreading this message thread. I'm going to conditionally dissent (in case I've misunderstood.) I'd use drained parameters with an estimated water table.

I don't think that Terzaghi & Peck intended for undrained parameters to be used for long-term slope conditions.

 

[BigH]

This is always an interesting topic. If I remember my college prof correctly, for normally consolidated clays, the short term is the most critical. I would use undrained shear strengths myself. Though, keep in mind that there might be some conditions where the long term parameters (effective stress) should be checked. With embankments, the effective stress imposed on the sliding surface will be increased, thereby increasing the available strength to be mobilized. For very important slopes, you should probably do both analyses for completeness. (This isn't a cop-out, it is prudent although you will have to estiamate the pore-pressures to be applied if checking an intermediate stage).

For overconsolidated soils, the long term conditions are the most critical.

 

[Focht3]

BigH is correct, but we should also recall JWB46's original question. The fragment that I specifically responded to is in bold text in the following paragraph.

Suppose we have a natural slope in soft clay where there is little or no natural drainage. How should the long term assessment be made? Terzaghi & Peck seems to imply an undrained analysis.

For normally consolidated soils, the undrained (i.e. short term or "immediate&quot condition will typically control the slope design, particularly if a rapid drawdown condition can develop. For overconsolidated clays, the opposite is true: the long-term (i.e. effective stress) condition will generally control the slope design.

I understood JWB46's question as, in effect, "Is it okay to use undrained parameters for a problem where drained parameters likely have developed?" I conditionally disagreed because it seemed to me that JWB46 did not understand the difference between total and effective stress and how they should be used in design. I guess it was the "where there is little or no natural drainage" phrase that got my attention - by definition, a stable piezometric condition is always assumed for a "long term" condition. Some drainage will almost certainly have occurred in order to achieve a stable piezometric condition. This is fundamentally different than an undrained condition, where the piezometric condition is unknown - and not addressed at all.

One final thought: while most of our geotechnical literature addresses normally consolidated clays, most of the soils that we encounter (in the geographic sense, anyway) are overconsolidated. And a lack of understanding of overconsolidated clays and their potential loss of strength over time has led to some striking failures. (A good example is the Goodpasture bulkhead failure in the Port of Houston in the 1970's. Olson and Daniel (1982?) wrote a very good paper describing this failure.) We need to educate our clients, peers and staff on the dangers of relying on immediate strengths (UCs, pocket pens, whatever) for a soil that will lose its strength over time.

 

[davecooper]

The problem with undrained testing is the apparent cohesion that is typical of saturated quickly sheared or loaded soil samples. This cohesion under fully saturated, fully drained conditions will not be there. Or maybe I should say will not be there for long. In this condition all it takes is some external source of vibration or loading to cause the pore-pressure to go excess (increase beyond the background level) and the slope will fail. Once a fully saturated slope that cannot drain is loaded the clay particles will realign themselves from a random orientation to a more uniform orientation as the pore-pressure increases and cause failure. This is just my opinion. I am not an engineer but a geotechnical lab technician. I understand testing and I think soil mechanics from years of testing. Undrained testing is way overused in my opinion. Undrained analysis should probably be used for seismic loads or other types of fast loading conditions or situations where drainage cannot occur.

Dirtdoc1

 

[Focht3]

dirtdoc1:
The subject of apparent cohesion is a bit of a red herring in this message thread. Well educated and experienced geotechnical engineers do not use lab tests directly without applying their own experience and judgement.

Apparent cohesion can have a lot of causes, only some of which have anything to do with the test method and/or test procedures. The failure envelope is seldom, if ever, truly linear - which leads to an "apparent cohesion" when triaxial tests are run at confining stresses that approximate realistic in situ octahedral stresses, and then plotted as though a linear relationship exists. This doesn't make the test - or the method of analysis - wrong. The graphing and terminology are design expedients - nothing more.

Drained tests are often inappropriate - they model behavior that is not representative of expected field behavior when undrained shear occurs. Here, too, the apparent misconception that the failure envelope is linear can lead to a potentially dangerous conclusion. It is not.

Regarding the following quote, " In this condition all it takes is some external source of vibration or loading to cause the pore-pressure to go excess (increase beyond the background level) and the slope will fail. Once a fully saturated slope that cannot drain is loaded the clay particles will realign themselves from a random orientation to a more uniform orientation as the pore-pressure increases and cause failure." You appear to be talking about apparent cohesion caused by negative pore pressures. While this could theoretically occur, it seldom does in practice. And your "random orientation to a more uniform orientation" during failure is nonsense.

My suggestion: stick to what you know.

 

[davecooper]

Focht3:

You seem to know your stuff! I have been trying to convince my clients to draw failure envelopes non linear so they don't get all the apparent cohesion. Most engineers are not aware of this. What I meant by the clay particles changing from a random orientation was that when the pore pressure increases the structure changes and the particles realign themselves. They have to or else it wouldn't fail. You are right - apparent cohesion is caused by negative pore-pressure. It's when it goes positive that it dissapears. Cohesion does not exist under fully saturated and drained conditions. That's why slopes fail. You are right I am out of my area of expertise disscussing engineering, I am a lab technician, but tuff! I'm having fun. I'll put my knowlege of testing and soil mechanics against yours any time! Come on big fella!

Best Regards,

Dirtdoc1

 

[Focht3]

What a challenge! I like your spunk! We all learn better when our "comfort zones" are challenged.

You are correct to a point: when the pore pressure change, the soil particles do reorient - because the structure must change for failure to occur. The "nonsense" portion of your statement had to do with the clay particles will realign themselves ... to a more uniform orientation.... I see no need for this to occur in order for deformation to take place. In fact, the deformation should occur when the minimum potential energy is expended to make the action happen. A "uniform orientation" does not seem to be a logical part of the transformation.

On the subject of apparent cohesion: be aware that there is still some disagreement about whether all cohesion is due to negative pore pressure, or whether a portion may be due to other factors. Generally speaking, the debate is not particularly important to most practicing geotechnical engineers. What is important is developing a reasonable estimate of the critical strengths that affect the slope under study.

Regarding the issue of linear vs curved failure envelopes: this is a sore subject with me. As universities have slowly reduced the requirements to earn BS and MS degrees over the last 20 to 30 years, geotechnical engineering classes have been forced to deal with abbreviated courses of study. Shear strength has been an area that has received steadily decreasing attention. I, too, have encountered far too many "experienced" geotechnical engineers that think that the failure envelope must be linear. What a shame.

 

[davecooper]

Focht3:

I think I get the idea of a uniform clay particle orientation from running torsional ring shear to determine residual strengths. The clay particles have to be aligned.
A while back I ran some stress path testing. The method was to take a triaxial sized sample (min. height = area x 2),back pressure-saturate it, then consolidate it to a Kc of 0.4 (inverse 2.5). Once consolidation was complete the test was failed by increasing the pore-pressure (0.1 psi/hour, slow enough to achieve a drained condition) while the axial load (sigma-1) and lateral load (sigma 3) were kept constant. This of course took some very sophisticated computer controlled flow pumps and load frame. The strain was monitored throughout the test. Depending upon the consolidation pressures the samples could take 8 to 9 hours (for 1 psi sigma-3), to a week +- (at 8 psi sigma three) for failure to occur. Basically the samples would sit there and do next to nothing (as far as straining) for the majority of the test then all of the sudden they would slump. Under this type of condition the particles probably didn't need to be algned in any particular orientation. They actually flowed. When the stress path was plotted on a P vs Q diagram the test progressed from right to left, horizontally, until the sample started to flow, then bent and headed for zero cohesion in all 40 tests (on overconsolidated adobe too!).

Like I have said before, I am not an engineer, don't know the first thing about it. I have been told that cohesion can be very critical depending upon the depth in consideration. I have a feeling that the failure of many engineered slopes are due to an overestimation of cohesion. I ran this thought by an engineer with extensive experience and he agreed with me. I just don't know enough about the engineering end of it but I think cohesion is way over used. I know how precious it is to you engineers. I get calls from frantic engineers asking where's my cohesion (joke). I can take an overconsolidated adobe or a soft normally consolidated marine clay and run a triax-cu-pp, tx-cd, torsional or stress path and get zero cohesion if I am allowed to choose the confining pressures. Drained direct shear doesn't work unless the sample is fully saturated and it can only get that way naturally or by back-pressure saturation in the lab. That is why direct shear almost always shows some kind of cohesion and these other tests don't. The direct shear sample can only be inundated. The unsaturated portion of the sample creates a suction (negative pore-pressure) therefore creating an apparent cohesion. Cohesion can disappear under fully saturated conditions were drainage can occur if loaded. I don't believe that cohesion is not critical for slope stability. Slope stability equations don't work without it! There is something wrong here. I am walking way out on a limb here. Any engineer could knock me off pretty easily but I have a feeling that something isn't right. I am sure Focht3 will straighten me out though.

The testing was based on a paper written by Scott Anderson & Nicholas Sitar entitled:"Rainfall Induced Debris Flows". These guys along with guys like Stark are way out there. Their understanding of soil mechanics and engineering are so complete it probably comes as natural as breathing. I hate to hear people knock PHD's. They laugh and say stupid things like "oops a PHD must have gotten involved on this one" and love to knock their so called lack of real hands on experience and knowlege. These guys are so far beyond most that it is truely not funny but concerning. Uh Oh I'm getting on my soap box again so I'll shut up.

It is apparent that Focht3 and I are not going to fully agree on these issues. I would love to hear from some other engineers, their thoughts on slope stability and cohesion. Not that I don't thoroughly enjoy reading Focht3's thoughts. There is so much to learn and so little time. Every time I get to a point to where I think I know it all I realize that I don't know squat! And of course there are always those people who love to let you know that you don't know squat.

Best Regards,

Dirtdoc1

 

[Focht3]

Dirtdoc1:

First and foremost: I agree with much of what you have posted. Too many engineers rely entirely too much on "cohesion" with no appreciation for what causes it to occur. A very bad, sad circumstance.

A good posting - your description of the K₀ test failure by increasing the back pressure is very interesting. (While I find the test of intellectual interest, I have yet to find a practical need for it. Why was it run this way?) What do you mean by an "adobe" soil? Is this a CL material with a substantial silt content? (Based on this soil description, it appears that you are somewhere in the California / Nevada / Arizona area.)

As a side note, you described the stress path plot as a p-q; did you mean p-q' (effective stress plot)?

I am very comfortable with the concept of "cohesion" created by negative pore pressures, and recognize that this is the likely cause of most of what we call "cohesion." But there are other causes of "cohesion" that have little or nothing to do with negative pore pressures. Two specific causes come to mind: mild cementation, and the effect of soluble salts in the sample and pore water. Both of these can occur in a natural setting, but be destroyed (unintentionally) in the lab.

Remember that you use distilled water in your triaxial tests (in fact, all "normal" geotechnical tests use distilled water in accordance with ASTM.) The distilled water is slightly acidic due to the presence of CO₂, and will dissolve some cementation bonds in alkaline soils. And the absence of dissolved salts in the distilled water will cause some salts to be leached out of test samples, affecting their behavior. In these cases, the tests can show no cohesion while it actually occurs in the field.

On the subject of PhDs: some of what you hear is "education bashing" - the jealousy of another's advanced education. But the criticism of PhDs is not always for this reason. My biggest complaint with most PhDs is that they expect their "book learning" is a perfect substitute for field experience. It most certainly is not. You learn things in the field about the geology and macro structure of the soil / rock that does not appear in any textbook or equation. This knowledge requires one to get hot (cold, wet, bug bitten, etc.) and dirty because it requires working behind drill rigs and being present for substantial periods of time on construction sites. I know too many PhDs that think they are "above" this kind of experience. Karl Terzaghi called this part of a geotechnical engineer's work the "observational method." Dr. Terzaghi recommended this approach as the principal method of dealing with large construction projects.

As long as some PhDs maintain this attitude, their education has been long, expensive, difficult - and incomplete. And their judgement is potentially flawed. In the words of a Hindu friend of mine, "PhD no sacred cow."

 

[davecooper]

Focht3:

Wow! I got you going. I will try to answer your questions sequentially:

I realize that apperant cohesion can be caused by things other than negative pore pressure caused by dilation or partial saturation or cementation. I also know that I have allot to learn!

1)You should know what adobe is! It is a fat clay (CH).

2) The stress path test was run to determine the potential for debris flow. It is not a Ko test. (by the way how did you type the sub "o"?). It was a project meant to implament a model for slope stability by the USGS and we were hired to do the testing. Whether it ever gets implemented I don't know. Probably not due to the difficulty of the testing procedures.

3)I did mean effective stress. The total stress pq plot is not very usefull in my opinion.

4)Where am I located? I would rather remain anonymous. That way I can vent and make an ass of myself without losing clients.

5)We do not use distilled water in strength or permeability tests for the reasons that you mentioned.

6)I completely dissagree with you about the PhD being so far removed from real life engineering that their opinions are questionable. They are doing the research. They understand testing and soil mechanics. You may have been correct in Terzaghi's time but not now. I find just the opposite. I find most engineers lacking a basic understanding of soil mechanics and testing and rely upon practical experience too much! I think the pendulum has swung. But I have to admit and remain humble in the fact that I am not an engineer and know less than you. I probably should not be scolding anyone but myself. The only problem is that I have engineers asking me how to interpret our test results! That is scary. You my friend are much more knowlegeable than I, but I could still teach you a thing or two about testing and soil mechanics and I am sure you could teach me a thing or three about all of the above. My question to you is: do you think that you have more to learn?

I have got to lay off these eggnog and brandies!

Best Regards,

Dirtdoc1

 

[Focht3]

Dirtdoc1:
The last question first: I'll quit learning the day I die! And I have no doubt that you could teach me a few things -

I get "wound up" sometimes; not angry, just animated.

I understand your position on PhDs. I did not mean to malign them all; just those that think that education trumps experience. It doesn't - you need both. You also need to apprentice with a really good geotechnical engineer in order to make the best use of your education and experiences.

Please keep in mind that I have had opportunities to disagree with a few PhDs on a professional plane; you might not be too surprised to learn that I am thorough and win quite a few of these disagreements. (I even had one of my (well known) grad school professors effectively "surrender" when I challenged his analyses and arguments during a litigation assignment about 10 years ago.)

Unfortunately, you are often not in a position to challenge an engineer when he/she claims expertise based solely on "experience." (This is unfortunate.) I find that the claim of "experience" without an adequate technical argument to support their position is usually a smoke screen for ignorance...


Regarding adobe: I'm very cautious about local soil descriptions. In particular, the term "caliche" has caused me a lot of indigestion over the years. Just wanted to know what you meant by the term.

On the use of subscripts, underlines, bold text, etc. - these are built into TGML (Tecumseh Group Markup Language). TGML is proprietary to these forums and allows you to format your posts, enabling bolding, italicizing, and so on, as well as typing in color and changing your text alignment (center, right alignment). (Text in italics copied from a Javascript pop-up panel on the Eng-Tips website. You can open the panel by clicking on the link just above the Submit Post control button at the bottom of the "Your Reply" window. Emoticons/Smileys are described in the adjacent link.)

 

[BigH]

I love it!!!! PASSION PASSION PASSION The three words that say it all!

Re: Ph.D.'s Had a chance the other day to talk to a professor at one of India's top Engineering schools - we discussed wick drains vs sand wicks. Asked him about cost and installation; he said that they teach the theory but doesn't know anything about the cost or installation problems. This is the problem with a lot of professor types - not the practising Ph.D.s. They love their degree . . .