Contractor needs advice

[dhearn]

The Project:
Building pad for a Reverse Osmosis and Brine Concentrator Water Treatment Plant.

Strip 18” of native soil.
Compact native soil subgrade to 95%
Place +/- 2’ non expansive fill @ 95%
Place wick drain blanket (12” drain rock between two layers of Mirafi 500X)
Install wicks 5’ on center
Place +/- 15’ non expansive fill @ 95%
Place 10’ surcharge @ 85% (native soil)

The Specifications:

Maximum Density: Is the density obtained in laboratory when tested in accordance with ASTM D 1557.
In place density: Is density determined in accordance with ASTM D 1557, or ASTM D 2922 and ASTM D 3017.

Non-Expansive Fill: Shall not contain rocks or lumps larger than 4 inches in the greatest dimension and contain no more than 15 percent larger than 2.5 inches. Non-expansive fill shall be predominately granular, have no environmental contaminants or debris. All structural fill shall consist of “non-expansive” fill. (No plasticity index spec.)

Compaction of Embankments and Roadway Fills:
1) Construct in layers of depths specified above.
2) Compact by rolling with power rollers, tamping rollers, vibrating rollers or pneumatic tire rollers.
3) Moisture Content: Bring each layer to between optimum moisture content and +3 percent of optimum moisture content for maximum density before compaction.

The Fill Material:
Overburden from a local sand mining operation. A silty sand with some clay. 100% passing a #4 sieve 48% passing a #200 sieve.
Multiple curves range between 120pcf @ 13% to 117pcf @ 14%
Moisture is +/- 20% in native state.

The Problem:
We have been unable to achieve more than 90 -92% compaction with material at optimum moisture.

What we’ve tried:
1. Place material in 6” lifts, wheel roll with a Cat 824 rubber tired dozer. 83-85% @ 18% moisture
2. Same as above, add 84” vibratory roller. 85-88%
3. Cat 815 sheep foot and 84” vibratory roller. 88-90% @ 13.5% moisture (different location)
4. Cat 825 sheep foot and 84” vibratory roller. 88-90% @ 20% moisture
5. Move all fill to mixing table and disc to dry.
6. Replace fill in lifts with following methods.
7. Cat 825 sheep foot and 84” vibratory roller. 88-90% @ 13.5% moisture
8. Cat 825 sheep foot and 84” vibratory pad drum roller. 88-90% @ 13.5% moisture
9. Cat 825 sheep foot and wheel roll with loaded Cat 623 scraper. 90-92% @ 13.5% moisture
10. Raygo Pactor and wheel roll with loaded Cat 623 scraper. 90-92% @ 13.5% moisture
11. Three different Geotechnical firms ran curves with consistent results.
12. Three different nuclear gauges with consistent results.

Where we stand:
Native soil compaction reached without difficulty.
The Design Engineer agreed to let the 90-92% pass for the fill under the wick drain blanket but still wants 95% for the rest of the fill. We did get one isolated test to get 95% @10.5% moisture and a couple of 93% tests but nothing consistent.
The dry density numbers are between 106pcf and 109pcf
Density numbers drop off rapidly when optimum moisture is exceeded.
The testing engineer has taken some pity on us and is using his lowest curve.
Between installing the wick drain blanket and installing the wicks we have about three weeks to solve this problem.
Alternate sources of material are not available in this area.
We have only placed about 6000 CY of the 139,000 CY required for this project.

Any advice would be appreciated.

 

[msucog]

95% modified Proctor (ASTM D1557) is roughly equal to 100% standard Proctor compaction in my part of the world. Is there any particular reason for such a high degree of compaction specified? if you can only get 90% with a CAT 815/825, something is wrong even if using the modified Proctor. If the subgrade is stable and you place fill in 6" layers with a 815/825, hit it with a smooth drum roller and run your test, their's no reason you should not be getting compaction or very near compaction with the moistures you describe. Is the subgrade stable under loaded dump trucks or loaded scrapers?
What part of the country are you in?

 

[dhearn]

The subgrade is stable and compacted to 95%. The site is located in Tracy, California

 

[dhearn]

Forgot to answer one of you questions. I can only assume that because of the installation of the wick drains and the surcharging of the pad that the engineer is concerned about settlement.

We've tried our best to eliminate all the variables. Astm 1557 is not that rare in this part of the country. It's beginning to look like this material performs better in the lab than it does in the field.

 

[GeoPaveTraffic]

It appears that you are doing everything correctly. If i read the post correctly, you are having trouble with the silty sand with some clay material. 117 to 120 pcf seems high, even for a modified Proctor on that type of material, but it could be that high. I assume, since so many tests have been ran, that percent of clay in the fill and in the Proctor samples has been compared and that they are about the same?

When the material was compacted with the scaper, was it pumping any at all?

How large an area is being filled? Is the area square or is it long and narrow? If it is long an narrow, the problem may be related to the sides of the fill not being contained and the energy going to the side.

If all of the lifts you have tried have been 6" thick or so, even though it is counter intuitive; you may want to try a 12" lift and see what happens.

 

[civilperson]

How is the density testing being done? Insist on a sand cone or water balloon duplicate test if using nuclear gage and run a one point proctor to insure correct curves used. Thin lifts and testing at least one lift below the surface is recommended.

 

[BigHarvey]

I would not comment on the compaction but check with the wick drain contractor if he can go through the compacted layer and through the drainage blanket. To me the grain size distribution of your drainage blanket is not adapted to wick drains rigs. You might have 3 weeks to solve this problem before it starts to be one.

 

[dhearn]

The Testing Engineer did do one sand cone test to check against the nuclear gauge. The tests came out the same.

The pad is +/- 60,000sf and roughly square.

The material does not pump under the loaded scraper when the moisture is close to optimum.

Another thing that's puzzling me is that tests with moisture ranging between 6% over optimum and 3% under only vary a couple of percent in density.

I’m thinking about having the Geotech run another curve using the standard proctor and recalculate the tests to see if the difference varies more than would be expected.

 

[VAD]

Tend to agree with Big Harvey. You may wish to install wicks prior to using rock. Leave extra length of wicks. Regarding compaction of the silty sand. Have you undertaken a Standard Proctor test just to see what the values are. It is often prudent to do so to have a feel for how the material is likely to behave. I would guess that you need to have your field moisture at around 9 or 10 % moisture. At 13 % you are too wet especially if the curve drops off dramatically - typical of silty materials. In that case I would be at least 4 points below the optimum moisture.

Very often we expect compaction to be achieved but there is such a thing as determining the compactibility of materials. Unfortunately this is not part of specifications.

 

[TDAA]

I am curious as to what the Liquid Limit and Plastic Index are for the soil. If the material is non-plastic, you will likely have issues compacting in the field. I have had a site with this similar situation, where even near optimum, there was not enough apparent cohesion or friction to hold the newly compacted material together.

If the curve is found to be correct, one thing to try is to do a few lifts, and cut back down (ie 12-18 inches) to test. The over burden may help hold the material together, beneath the compactors. This helped on our site, but eventually, the contractor decided to use a different material for a portion of it.

 

[msucog]

i deal with non-plastic silty sands/sandy silts most of the time and this scenario reminds me of one geologic area in particular. the soil is about 95pcf @ 15% on the standard Proctor and looks like most other soils around here. but the material is a nightmare to deal with at pavement subgrades and even at footing subgrades. pretty much, the material has to be within 1+ to 2- of it optimum moisture and has to be overbuilt by a lift or so to obtain its required density. when the pad was built, the contractor ran 2 cat 825's over the area. since we run a lot of drive ring densities (since the material has a fair amount of mica in it), we have to dig down 6-8" below the footmarks prior to driving the ring because the surface materials "fluff up" so much. we tested the building pad area at 98% std Proctor and when the footings were excavated, there is always 4-6" of loose, disturbed material. the contractors are not over digging from what i can tell....but the material is so easily disturbed. the stuff has a CBR value of about 3.
surely the local geotechs have enough experience in the area to dig down a few inches in such materials when present, but i would ask them to dig down about a foot and very carefully run a drive ring method density just for comparison's sake. i personally prefer drive ring whenever possible.

 

[BigH]

msucog makes a very valid point. If you look at the original D'Appolonia information, you see where the compactive effort is applied compared to ground surface. (was used in determining best fill thicknesses). I always have felt that the traction, etc. on the surface of some soils renders the absolute surface unreliable as far as compaction goes. Of course, I would do my test at the surface, but if it were slightly under, I would dig down about 1/2 layer and redo the test - usually with good (passing) results. In simplistic terms, for some soils it is best to specify the level of compaction of the preceeding layer rather than the layer being placed (increase in compaction due to confinement, etc.). If in road design, the upper fill below pavement would be additionally compacted when placing the first base course. The pavement layers are a different animal on compaction mainly due to the fact that in almost all cases, they are not natural soils/aggs but crushed stone or crushed gravel.

 

[dhearn]

Thank you all for your advice.

We are looking for other sources of material, but we are hampered by the agency insisting that sources must be covered by the SMARA act of 1975 (strip mines). There are only so many in the area and many need their overburden for reclamation purposes.

We will have a curve run using the standard proctor and compare the results.

I’ll update the thread when we have more information.

 

[msucog]

what happens if the standard Proctor comes up a little lighter? wouldn't that blow your mind...i've seen it happen and screws up the scheme of things when you try to rationalize the information to people that don't think about the statistics of it all. the good thing is that the vast majority of time, the modified will be heavier.
i've got a question for you: is the material being placed as fill material that was excavated from deeper than say "near the ground surface"? is there a possibility of the material causing erroneous results with the nuke gauges? try a drive ring or sand cone density test if it's a possibility...

 

[dhearn]

We would expect the standard proctor to come up lighter (3-5%)because of the smaller hammer, lower height, and fewer layers.
But if it exceeds 5% and the test were in the 97-99% range we might be able to make an argument to change the spec.

We have had one sand cone test done that confirmed the gauge, but I must admit I've never heard of a drive ring test before.

By the way the we finished the wick blanket yesterday, and today the GC sent us a letter demanding we submit a different material within 48 hrs.(even though we have two samples already submitted and are awaiting results) So much for the three week breathing room we thought we had.

 

[msucog]

maybe the drive ring method is more local...i've never worked in other parts of the country. in my opinion, it's the easiest, quickest, most straight forward way to test in-place density. it will not provide reliable results if the material tested is extremely granular or rocky. but it works great for sandy silts/silty sands. sand cone and drive ring are equivalent in my mind...so if you've got sand cone results to correlate the nukes, then i'd say you've probably got a reliable test.

you've got a very interesting problem...i'm stumped if the subgrade is stable, the tests are being run below the surficially disturbed layer, and they're packing the heck out of the thin lifts. only thing that crosses my mind is that maybe it's somehow related to the local geology. please let us know how it turns out. sorry i can't be more help.

 

[Stego1]

Sorry if this has been covered above, but has coarse material been removed from the sample prior to doing the proctor test? This may reduce the field compactability in relation to the lab compactability.

Similarly, when the material destined for the lab was sampled from source, was the sample representative?

It may be something that simple!!

 

[Shadenem3]

I'm probably a little late jumping on the bus here but....

In my days as a contractor, I ran into, on several occasions, where compaction was not attainable because of the underlying geotextile. The problem was not enough granulars over the fabric. This usually occured when there was less than 18" of granular over the fabric. When the roller drum contacted the surface, it would create a movement (not deflection) of the surrounding granulars as a result of the fabric distributing the load of the roller (as it should). I'm thinking the "drain rock sandwich" would behave in the same manner.

 

[DRC1]

The material has 48% fines and the fines appear to be sitly weak clay. Low strength and limited drainage will limit the amount of compaction the material will be able to acheive. In the lab, the material is compacted in a small mold which provides confinement, increasing the apparent strength of the material.

I am not a big fan of rigid adherence to compaction specs. (I hear a few groans as some members say here he goes again.) 95% compaction does not tell you anything quantitatively about a soils ability to support loads. We just know that other soils on other projects have been able to support loads at this compaction.

1.) The material as is may not be able to achieve desired compaction. This is probably due to the high percentage of fines. Yes proctor testing does work for fine grain soils, but never as smoothly as with sands. I would suggest adding fairly clean (not nessearily washed) sand maybe 10-15% and see if that improves things. Note this will require a new proctor.
2.)I always suspect the testing. With three tests, that would seem to eliminate the question, but the 48% fines need adequate time to air dry and if all 3 labs rushed to do the proctor, the material may not have adequate time to dry.
3.) Try compacting slightly dry of optimum. Water content does not effect the quality of the fill (except where low permiability is important). Optimum water content is only an indication of least compactive effort. Obviously the lab results and the field are different. Going a little dry may allow water in the fines to bleed out.
4.) Check you base. It appears it would be okay from your information - which by the way, you did a great job giving us data to work with. However, if the base is failing, you will not get compaction of the overlying material.

Good Luck and keep us posted.