Shear strength for slope stability analysis

[Okiryu]

Hi, I am doing a preliminary slope stability analysis and would like to hear your opinions.

The slope is 1:2 (V:H), about 10 meters in height and covered by vegetation. Visual inspection indicates that the slope is stable. Water table at 10 below the top of the slope. Soils are sandy silts (ML) however they have about 23% of clays. Sands are about 41%, the remaining are silts. Moisture contents are about 27% with Atterberg limits from NP to LL=43 and PL=30. Degree of saturation is 73%~86%. I did some unconfined compression tests (qu=130~160 kPa) and some UU triaxial tests (c=38~55 kPa and phi=12~28 degrees).

My question is about how do you consider the shear strength for this analysis:
#1. c-phi soil
#2. Total stress/short term analysis (Su= "X" , phi = zero)
#3. Effective stress/long term analysis (Su= zero , phi = "Y"). For this case, perhaps the fully softened strength may be appropriate but based on the actual conditions of the slope, I doubt that wetting-drying process can affect the slope as it is properly covered with vegetation.

Anyways, I am inclined to use #1 above because of the proportions between sand, silts and clays. I was also looking at shear strengths using #2 and # 3, but the factors of safety resulted in less than 1. That is one more reason to use #1.

Talking with one of my seniors, he also suggested to use #1.

What do you think?

 

[GeoPaveTraffic]

I would use #3. If the calculations show a FOS less than 1, then you know your strengths or stratigraphy are incorrect. Note that I would not always use a Su-zero, would be small; but depending on my past experience with the materials and the laboratory results I might use a nominal value of 25 to 50 psf. Sorry, I just don't think in terms of kPa.

Mike Lambert

 

[Okiryu]

Mike, thanks for the reply. I was trying #3, but unfortunately I did not do drained tests, so I was using the fully softened friction angle from correlations with LL and clay content from Professor Stark. I also used cohesion=zero. The FOS that I got based on these parameters were less than 1 so as you mentioned, these strengths are incorrect. I think that because my soils are not purely clay soils the fully softened strength cannot be used.

Is any correlations to get drained strength parameters for sandy silt soils? Most of the correlations that I know are for clays and involve the PI values.

 

[Okiryu]

Mike, to follow up my last post, I think that the fully softened strength (FSS) is not appropriate in this case. This became more obvious after have seen the FOS less than 1 from the analysis using the FSS parameters. I also recall that the FSS is more applicable to stiff fissured clays and highly plastic clays that may be subjected to cycles of drying and wetting.

 

[GeoPaveTraffic]

Instead of using FSS parameters, you could use peak strengths from other published correlations. There is one based on plasticity index that has been used for many years and is available in several different sources. I use the chart included in Lambe & Whitman, but I believe the same data is available in USACE's EM on slope stability. Let me know if you can't find it and I'll track down the specific reference.

Mike Lambert

 

[Okiryu]

Mike, could you indicate the page number where the chart is in L & W book? Thanks again.

 

[GeoPaveTraffic]

Sure. The full reference is Soil Mechanics, 1969. Chart is on page 307.

Mike Lambert

 

[geotechguy1]

I'd go with No.3.

I use correlations between atterberg limits and peak/residual friction angles as appropriate. There are many correlations available in the literature. I normally use a friction angle that is one standard deviation below the 'average' value determined by the correlation (assuming that I have that info for the particular correlation) for design purpose.

 

[Okiryu]

geotechguy1, thanks for your reply.

Mike, I found that chart. See attached. Thanks. I guess that using this chart will be conservative since it says that it is for normally consolidated soils and my soils are overconsolidated.

Bottom line, also from geotechguy1 response, looks like there is an agreement on using the drained strengths. I will look at the peak strengths and run some analysis. Thanks again for your responses.

 

[ATSE]

If you don't already have the Duncan Wright Brandon text (Soil Strength and Slope Stability), you should get that now.
It is not either/or. Do both #2 (short term) and #3 (long term), with FSS values.

 

[PEinc]

I agree with ATSE, but for my projects, #3 usually controls because a little cohesion goes a long way.

http://www.PeirceEngineering.com

 

[Okiryu]

Thanks for your replies.

I am doing both analysis: #2 and #3. For short term (#2), I am using soil parameters from UU triaxial tests that I have done. Since my soils are unsaturated, the UU triaxial tests are giving me c and phi values. I will use both. The USACE manual for slope stability indicates that results from UU tests can be used also for partially saturated soils. Using c and phi values, FOS are larger than 1.5.

For long term (#3), I found that the FSS given by the Stark's correlations are very close to the peak values from the correlations shown in Lambe & Whitman's book as suggested by GeoPaveTraffic. Looks like the L & W correlation is a bit conservative. c will be zero. As PEinc indicated, this case controls.

BTW, I already ordered Duncan's book. I wanted to buy this book form before but forgot about it.

Thanks again for your inputs.

 

[fattdad]

We'd require drained and undrained analyses and would not allow the use of cohesion in the drained analysis.

For stiff-fissure clays (e.g., the Potomac in Fairfax, we'd limit the drained strength to fully-softened or residual. Don't want to type too much on that.

We'd also require stability to consider circular and non-circular analyses and in design-build, we'd require reliability analyses ala Duncan.

f-d

ípapß gordo ainÆt no madre flaca!

 

[HENRYZAU]

It doesn't look right to have phi > zero from UU triaxial, it should be zero. Clay tends to lose c' in long term so I would use option 3.

 

[Okiryu]

HenryZau, I think that you are correct about phi=zero if the soils are saturated. My soils are unsaturated and therefore phi>zero. There is a graph in Holtz and Kovacs book about UU test for unsaturated soils showing this. Agree that c' is very low and can be considered zero for long term (drained) conditions.

 

[fattdad]

For UU testing, we'd require (typically) that the lab prepare three specimens and confine the UU test at three different sigma 3 values (even though unconsolidated). In this manner, you can actually measure whether there is a phi>0 condition at low stress levels - i.e., until confinement realizes full saturation.

f-d

ípapß gordo ainÆt no madre flaca!