Field Density Test Results Don't Match Observed Soil Behavior 1

[theCorkster]

Hopefully someone can help shed some light on a dilema I'm encountered on a fill construction problem.

We're involved as the geotechnical engineer on a project that involves construction of some minor fills (2 to 4 feet)for light weight, single-story structures. During construction the on-site laboratory for the owner was calculating failing relative compaction (RC) results. So with some side-by-side testing with the on-site laboratory we determined that soil moisture content corrections from oven-dried moistures (nuclear gauge indicating 2 to 3 percent higher than oven-drived); this is something we've seen with volcanic origin soils in the area. Using these corrections 90 percent relative compaction was achieved.

Further fill construction with a sandy clay with gravel/ clayey sand with gravel material excavated from about 3 to 5 feet in a detention basin has generated relative compaction ranging from about 82 to 89 percent RC. During this time our geologist has been on-site performing more side by side tests and monitoring the contractor's operations - soil 2 to 5 percent above optimum, loose lifts of 8 inches, and compaction with a Caterpillar 815 sheepsfoor compactor. Test pads cut 3 to 5 inches into the compacted fill cut through weathered gravel clasts and leaves a sheen on the surface where the clay content is greater.

With additional compaction, including loaded 635 scrapers and loaded water trucks, there is no visible deflection of the fill, nor is there an increase in the inplace density! Increasing the soil moisture content to 7 to 8 percent above optimum, followed by the same compaction process, yields similar numbers to those within 2 to 3 percentage points of optimum. Again, no deflection under fully loaded scrapers and water trucks.

Laboratory curves are being run by both labs and corrected for pluse 3/4 gravel when appropriate; regardless, the field compaction with a lot of effort falls in the 85 to 90 percent range. Both laboratories are generating curves and optimum moisture contents within inter-laboratory precisions (ASTM). And sand cone density tests in a small test pad compare nicely with the nuclear gauge densities.

I'm considering shifting to a method spec for compacting this material.

Any similar experience out there that might illuminate what I'm not seeing so I can get some sleep?

Thanks

 

[fattdad]

Have you checked that they are using method C for the proctor curves? Is this standard or modified? Have you tried optimum compaction moisture content and vibratory compaction? I mean you should be able to get 95 percent compaction or better.

f-d

¡papá gordo ain’t no madre flaca!

 

[theCorkster]

Fattdat;

We're screening all samples on the No. 4, 3/8 and 3/4 sieves and correcting for soil/gravel moisture prior to selecting whether the curve is A, B or C method (modified, ASTM D1557).

We've varied the moisture content from optimum based on the curves (12 to 15%) up to about 22%. The densities increase by about 2 pcf; even at 22% the fill does not pump or show signs of instability.

We've considered using a vibratory compactor, but haven't tried it since the fines content in the fill runs about 20 to 30 percent. I'll explore that idea and try it to see if it will make a difference.

Thanks,

 

[BigH]

Just a thought - have you "dug out" say 6 inches of the placed fill and done the compaction test on the lower layer? While I agree that the material's fines makes it unlikely that it behaves like a clean sand (where I would test the lower layer rather than the placed layer) - might consider checking this just to rule out something fishy in the upper zone under the sheepsfoot. (Have you tried elephant feet instead?? - see an article in the Canadian Geotechnical Journal many many years ago where someone was pulling the wool over geotechs eyes talking about this . . .)

 

[Ron]

I find it a bit odd that at 7% over optimum in a clayey material you don't have any pumping or instability.

It sounds like you are following reasonable procedures to figure out the issue, but something doesn't smell right.

I agree with f-d that you should be able to get 95% of a modified proctor with proper procedures, particulary in the upper foot or so. BigH makes a good point that if the sheepsfoot is not "walking out of the material" then you will have disturbance in the upper 6 inches or so.

Run a loaded "wobble wheel" compactor over it and see if there is any instability or rutting.

 

[theCorkster]

Ron, Big H, Fattdad;

On some isolated occasions we've observed similar disconnect between the visually observed behavior of fill and the density test results.

The sheepsfoot compactor is observed walking out of 8 inch loose lifts of fill after compaction.

Test pads are cut with the blade; the surface "sheens," weathered gravels are cut smooth, and durable gravel/cobble rip out of the pad.

While we don't have a wobble wheel compactor on site, we've run over the fill with both fully loaded 623 scrapers and fully loaded water truck without deflection.

I'm feeling like Alice in Wonderland on this one. I haven't figured out the disconnect. Thanks for the ideas and reality checks!

 

[muuddfun]

Have you tried compacting dry of optimum, maybe 2 percent down? Have you tried to take larger sand cones, or ring tests? How sharp or flat is you curve?

What is your EI? Is something weird going on with the expansiveness?

 

[FixedEarth]

Loaded scrapers are not causing much deflection, so your soils are competent. Sand cone and nuke dry densities are similar so field testing procedures are correct.

Just wondering, is the "reported" procter value, the lab maximum dry density and not by error maximum wet density?

 

[theCorkster]

FixedEarth;

Two different laboratories are measuring about the same dry density results for maximum dry density. We've had two different staff perform ours with consistent results.

The last effort was to use two additional different compactors with varying passes to see if there was any discernable difference in the field inplace results. Using 3, 4 and 5 passes with a Rex Pactor and a Cat vibratory padsfoots, it's apparent that the Rex compares with the Cat 815 compactor, while the vibratory padsfoot compactor was not as effecient.

We're going with a method specification for this material using the Cat 815 since it has a better production rating than the Rex. While I liked the way the Rex would breakdown the material and I think does a better job of mixing the soil than the 815, the inplace density results from the 815 and Rex were comparable.

Thanks for the thought!

 

[oldestguy]

As I read the original question and cements, I question the use of words that do not fit with my usage.

Is your term "relative compaction" = "percent compaction" or "relative density" at your job? The two acceptance methods are very different.

Is there a job specification, such as "95 percent of Modified Proctor" ASTM D-1557 or something like "Relative density at 70 percent or higher"

If these are light buildings, then why so high a compaction requirement as 95 percent comaction? Are you assuming you will need design footing pressures on the order of or over 8,000psf?

With your situation, perhaps changing to something like "Fully loaded scraper number XX leaves ruts no deeper than 2 inches" might save all the grief.

 

[theCorkster]

oldestguy;

I've used the terminology "relative compaction" for a long time. I started using it since it refers to the ratio of field compaction to laboratory compaction, regardless of which laboratory compaction test is run.

I seldom use the term relative density as many contractors are not familiar with the term and concept. And when it is necessary to run both the vibratory max and impact compaction to determine the maximum density, I don't care to waste time running a minimum density test. If I did, I'd have to correlate which provides the desired result, 90 percent compaction or 75 percent relative density.

On occasion I've run into soil in this area that will not result in 90 or 95 percent of ASTM 1557, as I'm experiencing here. In the past I've then resorted to observation/proof rolling, and provided the moisture contents are optimum or slightly above, accepted the construction. This project owner's engineer wants proof that 1) the constructed product will perform as intended, and 2) the contractor has essentially demonstrated that regardless of effort used, the inplace test results do not achieve 90 percent of the laboratory method.

And so we have had to provide effort and justification to support using a method specification.

Regardless, I'd like to understand where in the testing process the dissidence comes from.

 

[oldestguy]

Mr. Corkster:

In your comment above, for items 1 as well as 2, how do you normally "prove that the result will perform as intended?

In my experience, we have done it the lazy way, perhaps to cover our rear (when we know we have far more compaction than needed). We have said essentially "with this effort, or with this degree of compaction, our past experience tells us this site is good for xxx PSF bearing pressure". Of course we could go on and run laboratory and field tests to prove it. I even recall running plate load tests, extrapolated up to the job loads, but that is very rare.

On one job I had a loaded scraper sit on a footing over night to prove it.

On a few jobs for a big national chain, I cut it close, to keep costs down, but made sure that the owner's representative knew, in writing, there is some more risk than normal as to settlements expected. No problems resulted.

So in your experience, is not your past experience sufficient to answer the question? If they doubt that, then who do they trust?

 

[theCorkster]

Oldestguy;

Great questions.

So here goes:

Fill soil parameters (compressibility, strength, permeability) are functions of soil type and density. And for many projects the degree of compaction is set as the most easily measured bar to exceed to "prove" that the minimum acceptable magnitudes of strength, compressibility, permeability, based on project requirements are reached. And clearly from the design side, the degree of compaction is conservatively selected based on experience, empirical relationships, and unfortunately too infrequently on specific analysis.

I realize that the analytical methods (field and laboratory density) were created to provide another quantifiable measurement besides visual monitoring, I suspect primarily to assist in dispute resolution. Now, since much of the geotechnical industry has placed so much of its criteria for accepting soil compaction solely on test results, visual monitoring and experience regarding actual soil response to construction equipment loads has to be explained and/or defended. Numbers without knowledge.

And yes, my past experience is sufficient to know and to opine that the proposed construction will perform as designed. However, regardless of my professional registrations and years of experience, the project owner and their consultant (with years of experience and multiple professional registrations), who was performing the compaction control tests, needed to be satisfied that I was truly addressing this issue while keeping the best interests of the owner and their contractor in mind.

Obviously the owner had been "cowboyed" before and wanted to know that this wasn't the case with this issue. And I respect their geo consultant's response as well. Turned out we'd had some similar educational background, which I believe was helpful in developing their response to their client, while, as you say, covered their interests.

The project is now moving forward, and the owner has expressed appreciation of the efforts to resolve this, and I believe has learned something new about one nuance of earth construction.

Again, thanks for the input and questions.

 

[dgillette]

I've been watching this thread all along without saying anything, expecting someone would provide a clear answer as to what's going on. (Cue theme music from "Twilight Zone" and intro by Rod Serling.) I'm still curious about a couple of things:

What are the actual values of w-opt, gamma-d-max, PI, and % gravel, and do you have a comparison between D1557 and D698? Have you run specific gravity? Was the Proctor run on oven-dried material that is being wetted back? I'm told that can change the mineralogy of some clayey volcanic materials (affecting Atterbergs and perhaps the Proctor as well), but have no experience with it myself.

Any chance the gravel correction is fouling things up? You mention "cut through weathered gravel clasts." If those are numerous and are breaking down in the screening process or are getting broken up in the lab compaction more than in the fill, that might help to explain it.

The thing that really jumps out is getting what appears to be good compaction and trafficability at 7% wet of optimum. One reason I ask about SpGr is to see where the field measurements of density and %w fall w.r.t. the zero-air-voids curve. It might also be close to w-opt from D698. What does 7% wet mean for liquidity index? For clays, w-opt from D698 is usually in the ballpark of plastic limit.

On the other hand, if the fills are no more than 4 feet thick, the settlement due to consolidation of compacted fill under light buildings can't be too awfully large.

Regards,
DRG

“Nature has no contract with mathematics – she has even less of an obligation to laboratory test procedures and results,” according to one Karl Terzaghi.

 

[hokie66]

theCorkster,
As a structural engineer who would have been on the other side of this debate, I applaud your approach and wish for a geotech on all my jobs who exhibits common sense and diligence.

 

[oldestguy]

dgillette hit on something that I have found explains why you can't always reach the desired percent compaction (I still go with this wording, since it is commonly accepted wording for many).

Running the usual impact compaction tests, even with a fresh sample for each point, will break down the stuff into a different gradation than the original, usually easily compacted to more density.

Run some gradation tests on the before and after soils and see if there is a difference.

It does not happen a lot, but when it does, it needs an explanation as to why.

 

[dgillette]

oldestguy - I usually say "percent compaction" because "relative compaction," while a legitimate term, is too easily confused with "relative density." You can't distinguish the meaning from the words - it's like dirt bike vs mountain bike, and road bike vs street bike.

 

[theCorkster]

dgillette & oldestgyuy-

The D1557 max density and optimum for 6 samples have ranged from 114 to 121 with optimum from 12 to 17 percent. We haven't run D698 since it isn't specified for the project.

Specific gravity of gravel ranges from 2.5 to 2.7 with absorptions ranging from about 1.5 to 5 percent. Clasts of weathered volcanics (tuff fragments) drive the lower specific gravity and higher absorptions. These clasts are not as durable as some of the basalt/andesite clasts.

With regards to the gravel correction - lab samples are processed over No. 4, 3/8 and 3/4 screens manually. Some of the less durable clasts do degrade, producing more finer material, which tends to drive the laboratory max density upwards.

The catch with this process is that in the field the plus 3/4 inch fraction of the samples comprises on the order of 0 to 20 percent of the sample, but comprises about 50 percent of the Plus 4 material. The fine gravel in the sample is typically durable, and we haven't seen breakdown during processing. At the onset of the problem, we believed that since the plus 3/4 gravel is removed, resulting in proportionately more of the minus 3/4 gravel, the compacted density would be higher. Conversely, the inclusion of the lower specific gravity coarse gravel in the fill would tend to lower the measured inplace density.

To evaluate this notion we processed two samples by including all of the plus 3/4 gravel after manually breaking the gravel to pass the 3/4 screen. The results of two of these tests fell in the 119 and 121 pcf range.

As suggested, we plan on running gradings on future samples to see what degree of breakdown may be occuring.

Thanks for the perspectives and ideas!

Corkster

 

[oldestguy]

Another option for compaction tests with plus 3/4" material would be to use a larger mold and proportionately more blows, etc. with an equal work effort per unit of volume and all the gravel is used.

US Bureau of Reclamation used to have a standard for this; number E-38. My searching the WEB site for them did not find it now. It used a hammer weight of 185.7#, at 18" drop and a mold of 20" diameter (outside diameter), 16" high.

However, the degradation is still possible with any impact compaction.

 

[dgillette]

E-38 is now called 5515-89. You might be able to locate it under that name. If you want, I can scan it and post it here. (No copyright issue since it's a federal government document.)