Adding even a little bit of cohesion makes a huge difference in slope stability.  Dropping the cohesion from 190 to 7 should have a huge detrimental effect, as you have seen with your safety factor.

www.PeirceEngineering.com

Interesting, jg11.

As high as your drained strength is and as low as your FS is, I have to conclude that your slopes are quite steep, and/or you have a homogeneous dam without any internal drainage (hence, seepage exiting the downstream slope).

190 kPa is numerically equal to 7 kPa + 243 kPa tan(37).  At 100 kPa your assumed drained strength is less than half of the undrained.  Is it reasonable to expect that drained strength is so much lower than undrained strength at sigma' less than 243?

The fact that the soil has been mixed with cement begs a few questions - how much cement? and what "kind" of soil - crushed aggregate base type material?  Did you make compressive cylinders for testing or . . .   

assuming this is "soil cement" than what does "undrained" mean? well constructed soil cement may resemble concrete more than it does soil, permeability may be essentially nil and therefore, can the soil cement really be saturated?

Hello

Thank you for the replies, the mix is 5% cement and has a tested k=3.9x10-7.

The dam was 3:1 upstream and 2:1 downstream. I have changed the downstream slope to 2.5:1 and the FS is still below 1.5.

The dam is homogeneous.

If I follow correctly my assumed c' is too low for a soil with an undrained shear strength of 190kPa?   

Thanks,

OK, now I understand.  The 2.5:1 slope has FS<1.5 because there is seepage emerging from the downstream slope of the homogeneous dam high above the toe.  That is the only way FS could be that low assuming 37 degrees and 7 kPa.  Of course, you still may have an unrealistic strength, considering the amount of cement.  You are getting drained triaxial, or undrained triaxial with pore pressure measurements?

Ordinarily, I would tell you that unfiltered homogeneous dams should never be built above where people live, but it is POSSIBLE (I don't know for certain) that 5% cement might make the fill resistant to internal erosion.  I'm thinking primarily of erosion in cracks or at poorly bonded lift lines, and that the actual mixing is never as thorough as in the lab, even with a pug mill.  Anybody have experience with internal erosion in cement mod'd soil, or know of research data?

engineering properties of soil cement are related to a) mix design b)construction quality control. Given 5% cementitious materials mixed with clean sand and gravel alluvium, you could easily get a material resembling concrete with 3,000 psi compressive strength within 1-year. A dam or levee constructed of this material may be analyzed as a mass concrete dam. It may take 8 - 10% cement mixed with silty or clayey sand or even more with clay to get the same strength. Likewise, the structure must be constructed with good compaction, proper curing and without cold joints at the lifts or you may get seepage or delamination along the lifts.

That's a point I was raising - RCC is basically a cement based base course material (typically 100 to 120 kg/m3 of cementitious material).  We pay especially attention to joints. Typically, too, there is a membrane at the back side (such as Carpi) for principal seepage control but of course, one also has redundant systems when dealing with water.  So, again, what was the matrix soil?

there should be a toe drain.  Permeability of homogenous dams has no affect on the flow net.

f-d

¡papá gordo ain't no madre flaca!

Thanks again for the replies

There are no buildings/infrastructure within the flood zone. The matrix material which is borrow from nearby is Sandy Clay SILT 20/20/60. Good point about the strength increase of the cured mix, so the dam will probably be strong enough after 28 days or less. I will add a toe drain and see what happens. However the phreatic line is about halfway up the downstream face.