Slope stability evaluation existing landslide

[TaviB]

Hi all,

The company where i work recently got involved in design of a landslide stabilization by piles. The geotechnical report gave the following parameters for the soil profile:
1. Deluvial soil (mobilized): peak phi=20.5 / peak cohesion=5kPa, residual phi=16 and undrained shear strength = 50 kPa
2. Stiff clay: peak phi:23.5 / peak cohesion = 40 kPa, undrained shear strength = 150 kPa
For the project an expertise from a local authorized person was done, in which a stabilization by piles is demanded.

Using the residual value for deluvial soil and peak ones for the stiff clay the calculated failure plane with FOS<1 from SLOPE/W is similar with the one measured on-site by inclinometers.

The ag of the site is 0.15g. EC8 states that for slope stability a kh of 0.5*ag should be used, but when concrete retaining structures are used kh=ag. With a change in slope angle and drainage we managed to ensure a FOS of about 0.95 under seismic loading, using kh=0.5*ag. So, a shear force of about 300 kN used as reinforcement in SLOPE/W is enough for FOS to be greater than 1. The problem is if we use kh=ag, the FOS without reinforcements drops to about 0.7, and shear force needed for increasing the FOS more than 1 is about 550 kN, which is almost double.

Our questions are:
1. Under seismic loading the expert considered the soil as a phi-c soil, so we did also. Considering the soil is a clayey one, for seismic loading would not be more correct using undrained shear strength? And for long term the fi-c values (residual and peak)? Using undrained shear parameters, 50kPa for mobilized soil and 150 kPa for stiff clay, the 300 kN are enough for kh=ag.
2. I know that other codes like AASHTO, and a lot of papers on this topic rely on a 0.5*ag values for kh if a displacement of about 1 to 2 inches will occur. Does anybody know why in EC8, when using concrete structures, the "r" parameter is only 1?

 

[bdbd]

First of all, your parameters are a little bit problematic. If you didn't have the existing landslide and you are designing a slope, OK. Bur for an existing slope, you should split the failure zone and other soils. Why? For the failure zone or shear band, the undrained response does not make sense. Imagine a cyclic shear box, strains are too high, and when you apply additional axial loading, you will still have drained residual parameters for the shear band. So, all parameters at the shear band should be drained and without cohesion.

Do not consider your soil reinforcements, bolts or anchors as concrete structures. They are not concrete structures. Piles or diaphragm walls can be considered as concrete structures. The theory behind 0.5k is something about the ability of soil mass to deamplify and tolerance to settlements, if I remember correctly.

 

[TaviB]

Thank you for your response, but i think i maybe was miss understood by you.

1. The lithology is the following:
0-6 m --> mobilized deluvial soil: phi = 16, determined with torsional ring shear
6-9 m --> stable deluvial soil: phi 20.5 and cohesion 5 kPa
9-20 ---> stiff clay: phi 23.5 and cohesion 40 kPa
These are the parameters used for slope stability analysis, and as I told it gives an FOS < 1 on the same plane as inclinometer, so the on-site confirmation is more than enough to confirm the soil parameters used in slope stability.

2. As i told in the beginning of the message, the soil will be stabilized with piles, so soil reinforcements, bolts or anchors are not to discuss for the project. The question 2 was about kh value adopted between American standards and European ones. I will put here for a better understanding:
Eurocode 8 - picture

NCHRP Report 611:
The decision whether to use the 0.5 factor currently given in AASHTO will depend on the amount of permanent movement of the nongravity cantilever wall that is acceptable during the design seismic event. If the structural designer reviews the design and agrees that average permanent wall movements of 1 to 2 inches at the excavation level are acceptable, the seismic coefficient used for design (after reducing for scattering effects) can be further reduced by a factor of up to 0.5.

The same considerations are given for anchored pile wall.

In our case, we have no problem with some permanent movements, but, the EC8 does not specify conditions like the one in NCHRP. And there are a lot of papers that describes that using 1 x PGA for kh is a bit too conservative.

 

[fattdad]

I don't have much to add, but for my uncertainty. . .

I'm also looking at a slope stabilization where piles at the toe of the slope make the most sense. What I'm noodling over, is what's the state of stress on the piles? If the pile is kept from rotation as the slope is stabilized and if they are at some spacing of 3d or 4d, etc. you'll have arching effects owing to friction. So, do you base the forces on the pile at Kp*gammaZ? Do you use Kp*phi/10*gammaZ (i.e., phi/10 is close to the Brinch-Hanson correction factor for arching)?

I also need to evaluate the overall design strategy for such piles.

f-d

ípapß gordo ainÆt no madre flaca!

 

[bdbd]

Oh I missed the pile part, 1.0 makes sense now.

But about your parameters: there are almost thousand or more combinations to fit the observed site data, inclinometers. I still suggest you to introduce an additional layer through the shear band and use values without c'. I am sure that you can fit only phi soils to the observed landslide

 

[TaviB]

bdbd, instead using the shear band, i used residual values for the soil above 6m (landslide depth),and of course without cohesion. I preferred this way, rather than a shear band, because on-site the terrain looks pretty messed up, and there is a lot of water. And yes, there may be thousand of combinations that can fit the observed site data, but here comes the judgement of the engineer to chose the one who simulates the best how soil behaves.

fattdat, regarding soil arching effect i like to use the method developed by University of Akron, professor Robert Liang, and get from the method the net force acting on piles, based on pile diameter and center to center distance. Regarding the forces on the piles, i also know of Ito and Matsui approach, or Poulos, but not used it.

 

[Mad Mike]

I'm only an engineering geologist, and I don't try to act smarter than I am. On a project I ran some 4 years back I identified a landslide on a property embankment. We had some space to play with so I recommended an earthworks solution to dig the whole thing out, about 250 feet of embankment length. The development team preferred the idea of installing piles at the embankment toe, contiguous wall; so I determined the failure geometry, materials etc. and handed it over to a geotechnical engineer for design.

The piled wall was 1m above the ground and 10m below the ground, most of that into weathered rock. Tensioned anchors of 20m length restrained the capping beam. Deep buttress drains were cut into the back-slope. It looked like major over-design...until it wasn't...

Today I am assessing survey data which show the capping beam to have deflected by 400mm. The piles are bent and rotated, holding on by a thread. All of the anchors snapped within a few months of construction.

Interesting and perhaps relevant to your assessment, is that after the problems arose, I took samples of the failure plane material and tested them in a drained shear box...one sample returned a Phi = 5 degrees, from memory the others were about 10 to 15 degrees.

My point is; a documented landslide is not the best place to be looking for budget cuts. The laboratory data you've been given may be accurate, or not accurate at all. The hydraulic regime through that slip plane may change significantly over time. If I could go back, I'd assume the worst case of everything and sleep well at night!

All the best,
Mike

 

[bdbd]

Mike, that is quite an experience. Would you be willing to give more details? I am more interested in procedure of getting the material from failure plane. Is it a batch sample, completely remolded or intact core sample?

 

[Mad Mike]

bdbd,

In the case of sampling from the failure plane, we cut perpendicular trenches down to the failure surface near where it came to daylight at the toe, then sent the laboratory guys in to cut suitable undisturbed block samples there and then. It was my second time doing this, and both times I've been lucky enough that the failure plane was sharply defined- visibly sheared, softened and with some water seeping through it. I've had other times where for the life of me I could not distinguish a failure plane precisely. In the latter case, I would have the laboratory run a remoulded shear box with a reversal to try and get the residual strength. Half the time the results came back dubious.

The subtleties of extracting the samples and keeping them intact are beyond me, but I'm assured it can be done reliably by an experienced technician.

Cheers,
Mike

 

[theCorkster]

Late response.

Given the amount of displacement and alignment of clay structure in slide plane, a fully remolded modeling of the material strength is appropriate. The repeated-direct shear requires numerous displacement and reversal cycles which can result in material squeezing from the shear surface between the top and bottom of the shear box. The torsion (ring) shear runs a fully remolded sample. As I recall the torsional shear result will be one to two degrees lower than the repeated direct shear.

 

[fattdad]

Regarding the strength of clay, here's the approach I like. I like to take the sample, hydrate it to the liquid limit and normally consolidate the soil in the odometer. Do that three times and you'll have some observation of peak strength and some observation of residual strength. Use the peak for fully-softened. The residual is the residual.

Workshop at Virginia Tech (CGPR) has information in this regard.

f-d

ípapß gordo ainÆt no madre flaca!