In situ sampling methods for new embankment fill?

[jdonville]

Here's the situation:

A new embankment to support a single-lane ramp is proposed. The embankment crest is up to 68 feet above one toe and about 35 feet above the other due to the existing ground surface. Otherwise the embankment is symmetrical with 2H:1V sideslopes. According to Janbu, the unit cohesion needs to be on the order of about 1500 psf in order to have an undrained global safety factor of 1.3 or greater. The embankment fill will likely be comprised of A-4 and A-6 borrow. NavFac 7.2 Table 1 indicates that 1500 psf is feasible with these kinds of soils.

I would appreciate ideas of what kinds of tests to perform to verify that the 1500 psf criterion is being met.

I would think that rig-mounted CPT would be too expensive to use as a verification tool. How about DCP? or a geoprobe-mounted tool?

What about the frequency of testing? It is certain that nuclear density and moisture testing would be performed on the same material also. Are there any open references to determination of a relationship between %compaction, moisture and cu for new embankment in order to permit traditional sampling and lab determination of qu much less frequently?

McCoy - any thoughts or references on probabilistic verification of the strengths, with respect to method, frequency/spacing of tests?

Thanks in advance for your responses.

Jeff


Jeffrey T. Donville, PE
TTL Associates, Inc.

http://www.ttlassoc.com

The views or opinions expressed by me are my own and do not necessarily reflect the views or opinions of my employer.

 

[BigH]

How about field vanes? - or you could obtain tube samples and test in UCC. I know that they used to have small portable UCC compression machines back in the middle east for "site" testing. If I were to use "driving" methods - I'd rather go with SPT than DCPT. You can do the SPT by hand - heavy but I've done it before.

 

[jdonville]

BigH,

The field vane idea satisfies the following important criteria:

1) can obtain data over the entire lift thickness.
2) the equipment can be relatively portable to permit many tests over a large lift area (like a nuc, I think you'd want to find the problem areas on the current lift so they can be fixed before placing the next)

I am concerned about damage to the vane if gravel sizes are present in the embankment fill materials. Plus, field vane needs to be done slowly by a patient technician. I am not sure how long a test will take, but the ASTM spec for field vane shear specifies a max rotation of 0.2 degrees per second (1 revolution per 30 minutes), so we'd need to use a gearbox to control the rotational speed.

I don't think that we'd be primarily interested in determining the remolded strength (except as a research project), so this will cut down on the time required to perform an individual test. Also, if we reach the target strength (torque) without yielding the material being tested, then we can mark it as a pass and move on.

Again, we would probably want to develop a correlation between %compaction and field-determined su, so we would need to look at performing vane and nuc tests close by each other, probably by the same tech.

Jeffrey T. Donville, PE
TTL Associates, Inc.

http://www.ttlassoc.com

The views or opinions expressed by me are my own and do not necessarily reflect the views or opinions of my employer.

 

[BigH]

Yes - if you have "stones" in the fill, it might be problematic. But you can make a pretty robust vane. As far as the ASTM methodology . . . When I first started out, we put a t-bar across the top of the A-rods - then we used fish scales hooked onto rings of the t-bar at 12 inch or 18 inch from the centre each side. We had two pull the fish scales uniformly while one "read" out the poundage. When it failed, we would write 20#ES@12 (twenty pounds each side at 12 inches - or 40 ft-pounds. We did not fail the sample over a long time - but usually 20 seconds or so. Not as precise as the Sprague and Henwood vane table with the gears which I have also used - but, say, "good enuf!" I would probably work up something like this - of course you could use a torque wrench - if it is good enough for structural bolting, it would be good enough for you! - and yes, there is no need to overdo it if you have reached your "minimum". Guess you might want to put a bit of the Bjerrum correction factor on it - but if you could do a correlation programme of vanes vs UCC, then you would be able to correlate directly.

 

[jdonville]

BigH,

UCC?

Jeff

 

[Panars]

UCC=UnConfined Compression

 

[DRC1]

I would not be concerned about the cost to verify the embankment. Ask yourself what is the cost of the desired testing program, what is the cost of the embankment in place and what is the cost if the embankment is not suitable? I think in that light the additional cost to obtain high quality information that you can rely on vs. The cost to get a comprimise program, you will see that in the big picture, the cost of the additional quality is minimal. It never ceases to amaze me how owners cut engineering costs only to spend it ten times over on field problems.

 

[Dirtguy4587]

I agree with DRC1. Given that, I would think that the CPT would be a good tool to use. Tests are typically quite quick to perform, and the value in getting continuous readings throughout the embankment could be very valuable.

 

[Mccoy]

I'm sorry for the late answer Jidonville,
I just saw this thread.
FVT has a pretty good reported reliability, this would cut on the uncertainties, it beats the heck out of SPT for sure.
You may carry out a probabilistic analysis in different ways, the best may be by inputting variables directly into your stability equation, but in this case you'd need a specific software. If you are using a spreadsheet I might try a probabilistic analysis, once a few potential failure surfaces have been singled out. Otherwise, you might execute a treshold analysis, that is probability for the Cu to be Lesser then the 1500 psf needed for reasonable safety, that is probability of reasonably safety (not probability of failure, that would be by the rigorous method). Also, a capacity-demand analysis could be carried out building a distribution on "safe Cu" by expert judgment and convoluting it to the observed Cu.
distribution.
There are many sampling strategies, but if you don't have problems of locating weak targets, it all gets down to sampling along a regular 3-D grid which be representative enough. I'm going to look deeper into that. A good idea would be to thicken the sampling around the area where the stabilty models (or your engineering judgment) predict a higher likelyhood of occurrence of a failure surface. And try to spot resistance patterns into the fills, that is, spatial populations of different strenght.
Let me know when you have the Cu data I can do a preliminary assessment on it. Just remember, as you sure know, quantity is good, quality is better.

 

[jdonville]

All,

Thanks for your input and perspectives. The embankment is still in the design stage, and the strength requirement and testing regime will need to pass review with the DOT's engineering brass.

Dirtguy & DRC1, given that this embankment is plenty high and will be paid for by the State, I have no fear of an inadequate testing program. I simply don't want to recommend an unwarrantedly expensive or ineffective testing program for construction.

To make more clear than perhaps I have stated above, I would ideally like to find a relationship between the degree of compaction / in-place dry density / in-place moisture content and the in-place peak shear strength of the embankment fills.

*[ I have yet to find any papers addressing this relationship (or lack of relationship) with respect to cohesive fills. Any pointers or references to the literature would still be appreciated. ]*

This relation ship would be helpful because:

1) Degree of compaction (as measured through % of Proctor dry density) is a commonly performed field test and will, no doubt, be performed on any embankment compaction job anywhere. I have never seen field vanes performed on newly placed embankment, and I don't imagine that it is that common a practice - at least in the USA.

2) The in-situ degree of compaction determination can be performed rapidly with minimal interference with the embankment construction processes. Again, I have never seen the use of field vanes used for construction and don't know how rapidly (or reliably) they might be performed on lift thicknesses of less than 1 foot.

3) It seems that there would likely be efficiencies to performing a majority of only one type of field test (easier data reduction, improved repeatability through technicians performing more tests of one type).

4) The embankment fills will almost certainly be excavated from local borrow. The local soils do contain gravel sizes. I am concerned about damage to field vanes, although criteria could be established for technicians for the amount of effort to expend on inserting them into the compacted soil.

5) The goal would be to identify and remedy areas with an average shear strength deficiency before they get buried by the next lift. CPT might be used at the end of the job as a final verification that the embankment mass meets the strength criteria, but would be too expensive to keep handy to test just the current lift (operator + equipment).

Mccoy,

Your suggestion about concentrating testing on the theoretical critical failure planes is well made.

Thanks to all for the responses. Please share any additional comments you may have.

Jeff

 

[BigH]

Jeff - why don't you establish a programme in the lab for it. Do proctors and then trim to get Su values, etc. In ORN 31 there is a chart of CBR vs proctor curves for clayey soils (I am almost sure). Also, I remember way back in university that we saw some correlations of proctor values to unconfined compressive strengths. Again, while the vane might not be the best, you could obtain "core" samples (there is an ASTM spec on this I believe) to establish compaction - then using a portable compression machine (I have seen them in a pick-up truck bed before used in middle east) - you could determine the UCC value. McCoy is correct in that the most critical strength would be along the critical surface - and this is likely at the interface of the embankment and natural soil. I may have actually sent you the ORN on the first disk.
Ciao.

 

[jdonville]

BigH,

Sorry, not on the first disk (I just rechecked). Is there a web source for these? I have a relatively fat pipe.

Jeff

 

[BigH]

From ONR 31 on its way - CBR vs MDD curve.

 

[jdonville]

BigH,

I'm not sure if the charts you sent are entirely helpful in this instance, but your comments and references are always appreciated. Thanks. Please forward any additional material you think might be helpful as you think of it.

Jeff