Prediction of embankment settlement on loose granular soil

[bvn]

I am searching for the preferred method of predicting embankment settlement over loose, saturated sand.

The embankment is a 100 foot wide roadway over about 10 feet of soft silt over about 100 feet of loose, saturated sand (estuarine/alluvial/former beach) to silty sand over stiff soil.

Available data:
CPT (10 to 100 Tons / sq. ft)
SPT (N = 4 to 25 uncorrected, alternates)
Lab consolidation of overlying silt layer

Problem:
Prediction of Settlement of sand

Discussion:
In the literature Schmermann's method is the most often recommmended method for "accurately" predicting settlement in sand; however, a strain influence depth of 400 feet (4B) probably did not fall within the original data set that this semi-empirical method was based upon.

May Schmertmann's method be applied to very wide "footings"?? What would be a good alternative analysis method? For the record I have very little faith in accurately determining the settlement of sandy soils beyond full scale testing, but...

 

[VAD]

How high is the embankment. Is it just the thickness of the pavement structure. Maybe I am missing something.

 

[eric1037]

You can use elastic methods to predict settlement of sand. Fang's foundation engineering handbook has a good description. You can use a variety of techniques for determining the stress increase at different points below the embankment.

 

[jdonville]

bvn,

If the project cost and initial settlement predictions warrant, you may want to perform a combination of CPT, PMT and flat-plate dilatometer testing.

I was involved in a job once where we had a 40-foot MSE structure going up on sandy soil near the Mississippi. Bedrock was anywhere from 80 to over 120 feet in the vicinity. Our initial settlement estimate based on published (Bowles) empirical relations with SPT N values was on the order of 4 inches or more. Based on the insitu tests, we were able to restate the settlement estimate significantly downward. I believe that the primary factor has to do with the fact that modern SPT hammers deliver more energy than those used to develop older correlations, thus an SPT N-value of, say, 8, indicates much denser soil than an SPT N of 8 obtained with a rope-and-cathead hammer.

Anyway, we used a company out of VA that is probably one of the DMT varsity. They have a rig that can do CPT with in-line PMT (!) and also flat plate DMT. If interested, contact In-Situ Soil Testing L.C. at insitusoilATprodigy.net (Roger Failmezger).

Jeff

 

[bvn]

VAD,
The embankment is project to be 3 to 11 feet high. Settlement would not be a large concern if not for shallow underlying utilities and concerned utility owners. My concern is that the desired predicted settlement accuracy by the utility companies is unattainable using any published method, complex or simple.