Strength of Soil for Rapid Drawdown 2

[tq3610]

I have a small dam to be analyzed for rapid drawdown. No specific laboratory soil strength data available, but I have boring logs and gradations. The average N60 value through the embankment soils is 23 bpf. Gradation tests on the embankment soils show the material to be sand with ~10% gravel and ~25% fines. I've correlated a friction angle of 34 degrees via 5 or 6 published N-value/friction angle correlations.

The FS is ~0.1 below acceptable for the scenario of shear strength of 34 degrees.

Given the 25% fines and relative density, I am inclined to suggest there is some amount of cohesion in the soil. Applying just a miniscule 25 psf cohesion makes the stability satisfactory. Also, the dam has been in place for 60 years and shows no signs of instability.

So my question is: is it technically sound to apply a friction angle and cohesion value to this material (silty sand) for the undrained strength?

The fact that it's "granular" but yet still has a fair amount of fines puts me on the fence.

 

[EireChch]

If the fines are behaving like CLAY, based on their atterberg data, then 15%+ of fines could make the material perform more like a CLAY. So it could have some undrained behavior. However, I dont think there is a hard and fast rule. Its all about behavior and judgement.

For a drained analyiss, Adding 1kPa (25psf) of cohesion will be a big help but I think you would have a hard time getting it past a peer review especially for a dam structure

 

[cvg]

slope failures dont occur where the strength values are "average", they occur where they are on the low end of the spectrum. a sensitivity analysis might show that the FS is not acceptable

also, has the dam actually experienced rapid drawdown in the last 60 years?

 

[geomane]

If you’re using a friction angle (drained strength) your water level will affect your results. Where did you assume your water level to drawdown to for your analysis?

 

[GeoEnvGuy]

Silty sand does have cohesion it's just difficult to get a decent sample for testing in a triaxial or shear box. The undrained strength of the material is actually significantly higher in dilative materials. The justification to use a drained friction angle is that the material is sufficiently free draining that any strength gained in undrained shear due to pore pressure will drain.

Another general comment silty sand SM material typically has a lower friction angle than 34 assigned.

 

[cvg]

agree with geoenvguy, recent experience shows phi is more like 30 - 32 at best for acceptable material, excluding the high PI materials

 

[r13]

I think you should ignore the contribution of the fines, and just use the drained strength, as the retention of the fines under rapid drawdown is uncertain.

 

[GeoPaveTraffic]

I expect that you will have a very difficult time getting any reviewer to accept 34 degrees with any cohesion. A phi angle of 34 is on the high side for any silty or clayey sand. Furthermore when you then say that the material has cohesion, you are indicating that it is behaving, at least in part, as a fine grained material.

Which bring me to my real question. Is this material free draining? When doing a rapid drawdown analysis how you treat the piezometeric level in the embankment is at least as important as how you treat the strength. If the embankment is free draining, then it is very rare for rapid drawdown to be a problem.

Lastly, how are you performing the analysis? Are you using a two-stage analysis as recommended by the USACE or some other method?

Mike Lambert

 

[tq3610]

I've considered the material to be just a bit below the USACE recommended permeability to be free draining. So I'm assuming the material will need to consider the undrained conditions. Per FERC, the lesser FS using drained and undrained strengths were considered. NAVFAC DM7.02 assigns an SM material 420 psf cohesion for saturated strength and 34 degrees for the effective stress envelope (7.2-39 Table 1). If I run the analysis with that undrained strength (c = 420 psf and phi = 0 deg), the slope is very stable. It is the drained strength of 34 deg and c=0 psf that indicates a problem.

The dam has never undergone rapid drawdown (60 years in service), and nor will it likely in any reasonable manner. The only rapid drawdown that could reasonably occur would be a dam breach, in which case sliding of the upstream slope would be the least of the issues. The dam itself is a recreational dam. The analysis assumes full drawdown (not staged).

Water level was modeled from the normal pool or surcharge pool (depending on scenario), running horizontal to an 18" thick concrete core wall, and then dropping about 10 feet to the downstream side of the core wall, and then lowering in elevation from there to the toe of the downstream embankment. Borings on the downstream side of the core wall had water at that depth below grade, so I modified the modeled permeability of the core wall until the seepage analysis resulted in a similar down-stream water elevation as the borings.

I've run the analysis using SLIDE2 using the b-bar method.

As far as the phi value of 34 degrees, the material has on average 10% gravel, 65% sand and 25% silt. I agree the phi of 34 degrees is not "conservative", but I'm not sold on the value not being representative of the material. As I referenced above there are many correlations from N-value to phi that indicate 34 deg for the material; that said, I also understand those correlations are not specific to SM, and are often SP soils. That said, I appreciate your responses if you think this opinion is off base.

 

[r13]

Do not use cohesion, unless it is comprised of mainly clayey material, or lined. If the FS is off by one tenth only, it is acceptable without stretching the reasoning.

 

[GeoEnvGuy]

analysis using SLIDE2 using the b-bar method

I am more familiar with geostudio slope/w where the b-bar method is used to reduce the effective stress by adding a surcharge load without assuming steady state conditions. I typically use it as the end of construction case on clay foundations.

For the rapid drawdown condition of homogeneous fill I do a transient seepage analysis to get the phreatic surface elevated in the core and lower on both upstream and downstream sides. For cutoff wall or clay core dams with free draining shells I manual in a phreatic surface that uses the steady state phreatic surface in the core and drops down on both the upstream and downstream sides in the shell materials.

I find it hard to believe that the analysis is doing what you are expecting if a friction angle of 34 is a problem. You would require a really steep and high slope to have an issue. I would be looking at the slice info and charting the shear strength and effective stress for each slice to make sure the program is doing what you want it to be doing.