Soil testing for slope stability analysis

[longisland]

Hi,
I've assigned borehole on site to collect soil samples & data for slope design.
Unfortunately, I'm getting disturbed samples since the formation is very stiff sand to sandstone of SPT 50. What other tests should I assign?
In fact, is there a rule of thumb in soil test assignment for slope?
How to determine the overall stability of the slope?
P/S: Any pointers in regards to the geologist's report? What are the key elements I should be looking into?

thanx

 

[aharcourt]

I can only offer my own experience to confirm that you are doing valuable work. Last week I parked my van on a road. It had been raining for three days. In the morning I went to go out in the van and it had been swept across the road and into a wall by a big mud slide. Now it is wrecked, I have to buy a new van, and the insurance doesnt cover landslides. So keep up the good work and when you know how to stop landslides get in touch because the local government here will employ you.

 

[Rockjoint]

Dr. Skip Watts at Radford University (conduct a search of this forum for details) may be able to assist you with additional considerations regarding slope stability, particularly rock slope stability.
Rockjoint
rockjoint@yahoo.com

 

[Focht3]

I like aharcourt's post! (Although I'm sorry for his loss.)

longisland -
I'd try running direct shear tests on uncemented sand samples - to get a lower-bound indication of the angle of internal friction. Be sure to run tests over the full range of expected overburden values. And try to test those layers, seams, etc. that are most likely to give the slope its future stability problems.

Intact samples are more problematic. If you must have them, then use a backhoe and hand carve some block samples. Keep them moist - and test as soon as possible. If you can trim triaxial samples, try CU with pore pressure measurements. Or try to core with geotech drill rig set up to core soft to medium hard rock to get rock samples.

But remember that slope stability is not based on the greatest strength, rather the issue is the weakest link. Look in the geologist's report for whether the sandstone is massive or fractured, the bed tilt (adverse or favorable), weak seams, etc. It's the small, usually overlooked details that can cause an otherwise stable slope to fail. Get very detailed geotech logs and use 3D software to give you a spatial perspective. Review it with the geologist - he may have insights that you don't.

And don't forget to look at any slope failures in the area of your site.

 

[doc11]

Can't help you much with the test assignments, but maybe you should look at coring the material instead of trying to do drive samples. There are several methods that work well in hard soils and soft rocks. I prefer to use either HQ3 or a GeoBarrel, but there are other systems including 94mm or Pitcher Sampler. Core recovery and quality will also depend alot on the experience of the driller. Try to find a contractor that has done this type of work before, they may be more expensive but it will be worth it.

 

[BigH]

Life as a Geotech, eh? A few points - is your "slope design" one in cut? or in fill? How high is the slope to be? Water levels are important - especially in sands where break-out of water onto slope at toe can be detrimental.

Unless things have changed - a sand is NEVER very stiff. Dense, compact (or medium dense) etc. but never given a consistency term.

Important: I would look particularly carefully at whether there is any layering in your sand deposits. If the slope is to be a cut with high initial water levels (do put in piezometers or standpipes), there can be disastrous effects effects when the cut starts to "lower" the groundwater level. Instead of dropping in a "homogenous" fashion, the water level will be shunted out at fine interlayers thus undercutting the sand.

Another consideration - unless the slope is very high - say more than 15m or so as a rough initial guess, elaborate testing of the sand is likely not warranted. Estimates of phi values can be made and your slope may not be overly critical to variations. With such a very dense sand, it may be somewhat overconsolidated and may stand at a greater angle initially than later. Suction effects may be appropriate to check out. But, the groundwater regime is of utmost concern, though.

As for the sandstone, earlier response is correct if the sandstone plays an imporant part role in the slope (as opposed to the fdn) - especially if there are dips into the pit (assuming cut) or clayey infills of fractures.

If this is for foundation design of an embankment, the above still holds to a degree, but N=50 is a pretty darn good material for a foundation base. Be careful if a water dam. You will likely need cutoffs in foundation.

Hope this helps in addition to the learned notes earlier.

Best regards to all.

 

[PEinc]

I am disturbed by longlsland's questions, such as "how to determine the stability of the slope?" and "what other tests should I run?" and "what are the key elements I should be looking into?"

Someone assigned to perform a slope stability analysis should be a lot more familiar with the topic or should be working under the direct supervision of an engineer experienced in this type of analysis.

I highly recommend that longisland get help with this one.

 

[davecooper]

I cannot speak to the engineering aspects of a sand slope but I can address the testing issues to a small degree. I think that direct shear is way over used. It is not a very dependable test for many reasons. Much of the cohesion and friction come from the equipment itself. The shear plane is predetermined and fails the material where it may not want to fail, which yeilds higher strengths. The sample area decreases as the test progresses. The area loss can cause the top cap to tilt which I would imagine causes all kinds of weird stresses at the shear plane. The sample is consolidated while emerged under water until 100% consolidation is achieved in the consolidated-drained and consolidated-undrained test. This is not enough to fully saturate the sample. Back-pressure is the only way to fully saturate the sample. Since the sample is not fully saturated I believe false cohesion is created. Although there are those engineers who use undrained quick shears at field moisture. That method may have it's place but is way over used because most engineers do not understand testing or soil mechanics. Triaxial CU w/pore-pressure or even drained triax are better tests. Anderson and Sitar developed an interesting stress-path test to model debris flows. Stark wrote a paper on the torsional ring shear. That is the only way to get residual values but probably doesn't apply here unless you find some clay seams. I just hate to see the direct shear test used so much but I guess at least their doing some testing. Make sure you test the material in a saturated state. That is when cohesion or cementation dissapears and you get failures. Slopes are usually fine until they become saturated. Cohesion does not exist in fully saturated drained conditions. All it takes is one little stringer of clay to fail an overconsolidated or cemented slope. But I am getting way out of my area of expertise now. I agree with the last fellow. Get someone with more experience involved.

Best Regards,

Dirtdoc1

 

[Focht3]

Dirtdoc1:
Points well made about the direct shear box. I seldom use them - and I don't own one. But the results can give an indication of lower-bound behavior where no field performance can be used as guidance.

You can't run "undrained" direct shear tests on sand - period. Saturation is an issue, but apparent cohesion should not be a problem unless there is appreciable clay binder in the sample.

I do understand lab testing, and I have run a lot of direct shear tests myself. I also run CBRs, consols, triaxial CUs and CDs and Proctors. And I still run index tests periodically - moistures, Atterbergs, UDWs, gradations, hydrometers, pH tests, etc. Sometimes the lab test results don't tell the whole story; you have to get your hands dirty.

But let's not get too wrapped up in the testing. Remember that the point of doing the tests is to provide meaningful data to the designer. Slope stability design is different than many other design assignments: the weak zones are the key, not the "average" condition. Hence my interest in the lower-bound sand strength results.

PEinc:
You have a good point: the original question did not indicate that longisland has a great deal of experience with slope design. I chose to give longisland the benefit of the doubt based on some of his other postings, and assumed that he was having trouble because he wasn't getting good samples. That may have been a bad assumption on my part.

BigH:
As usual, a very helpful posting. It is well worth reading.

 

[longisland]

Hi all,
thanks for the input.
I'm working under a geotech engineer with 20-year-experience. Testing is not my field to begin with. I was assigned to monitor the field tests on site. To my dismay, most borehole(mazier sampling) came out with very poor samples, low recovery & low RQD. It's frustrating to collect only disturbed samples when I need intact ones for triaxial tests.
I've consider collecting block samples but the layer of interest is around 6 to 10 m.
In short, I'm looking for ways to collect undisturbed samples. I've discussed this matter with the geotech engineer in charge & some of my geotech friends; in fact we have something in mind.
On the other hand, this forum has been one of the most helpful resource; I believe there's always some better ways to solve a problem.

again, thanx a bunch

 

[doc11]

Longisland
The triple tube should have worked, and probably still is your best bet for getting an undisturbed sample. I had to do some looking to find out what the mazier sampler is,it is very similar to the Christensen 94mm sampler we use here in the states. If I'm not mistaken Atlac Copco bought out Christensen. There are a couple of reasons why you had disturbed samples:

1 The shoe on the end of the inner barrel is too long (this is the most likely reason). Try a shorter one.
2) You have the wrong type of bit on the outer barrel for the formation you are drilling. In formations that have a tendancy to wash-out i.e.; dense sand and soft sandstone, you should use a face discharge bit.

Also there is an option on that barrel to use a thin wall "shelby" tube. This may work but you probably will have a hard time extruding the sandstone part if the sample.
Core recovery in soft and broken formations is as much technique as anything. Like I said in my earlier post try to find a driller that has experience coring in this particular formation, if you can.

 

[doc11]

One more thing - try shorter runs say .5 meter to .75 meter instead of 1.5 meters.

 

[Focht3]

longisland:

I empathize with your problem. doc11 has given some good suggestions - sounds like he has good experience with similar equipment and techniques.

Since you were in the field during the sampling, do you think the poor quality samples were due to drilling techniques, "loose" soil conditions, or trying to sample within the shear zone? Or was it due to the lack of enough fines to provide a binder for the samples?

Keep in mind that your inability to obtain "good" samples may be telling you something (indirectly) about the site -

 

[touma]

I had a very similar problem but in our case the intended cut was suppose to be between 21 and 26m high. Access was very poor as the terrain was pretty much virgin, it was tough enough to walk. Anyway we managed to get a rig to the location (don't ask how).

I have the following to share with you guys and hope it will be of help:

1. I do echo the problems in sampling in terms of disturbed samples etc... and SPTs were mostly refusal with some weaker layers every now and then.

2. I very much favor Focht3's opinion in terms of getting to know the strength of the weaker layers as the system is as strong as its weakest link.

3. Must keep a close eye on the driller as he could botch things very easily if he is impatient or inexperienced.

4. Initially I ran several direct shear tests on the relatively weaker layers. No matter what the complaint is about such a test it is still a practical and useful test and in my opinion quite indicatif. Unfortunately Geotechnical Engineering is still very much an art and as such we should try to favor the practical side of things rather than the scientific side (although I have nothing against it, but sometimes too much theory pins us down).

5. For the cemented layers, I sent a double tube split barrel 101mm dia. with a diamond bit, slowed the RPM and the only pressure applied was basically the weight of the rods, shortened the run length and no water pressure. It was not a great success but did manage to get a couple of samples for a U.C test (perhaps a triaxial test would have been better but I had no access to one at the time, but shortly we will own one. That established sort of an upper bound (although it is expected to be higher than that).

6. Once the lower and upper bounds were established, I ran a sensitivity analysis (basically an if-so type analysis) and each case was considered separately. Whenever the weaker zones were of significant thickness (in the order of 2m and up) I utilized the lower bound parameters, otherwise I used average + standard deviation.

 

[doresh]

Using a wireline drilling rig might help on sampling relatively undisturb samples. In addition to direct shear tests one m,ay consider unconfined tests also.