Frequency of OMC/MDD insufficient?

[charnott]

As Program Manager on a large scheme we are concerned by in-situ density test results which regularly show compaction >100% of sampled MDD. The Engineering Design Consultant's specification to which the Contractor is working states that Lab OMC/MDD should be tested 'once a week'. With the contractor currently producing 10,000m3/week and planned production to recah 42,000m3/week I am concerned that the specified frequency is not appropriate and am planning to raise an RFI to the designer asking for his review of this issue.

COuld there be any other reason compactions >100% are being reported?

For information: the material is sand, <15% fines, 100% passing 75mm sieve.

Thanks in advance

 

[jgailla]

charnott,

Yes, you should send an RFI.

I assume you are in Britain?
In the U.S. the Standard or Modified Proctor (ASTM D698 and D1557) is usually used to determine those parameters.

A compaction 1 or 2% over 100% is possible and can happen frequently, depending on the material type and compactive effort.
For instance, in a sandy soil with the Standard Proctor lab test (which usually wouldn't be done), and 5 passes with a 15-ton vibratory roller, I would expect to get over 100% on every test if the soil was close to optimum moisture.

Consistently getting over maybe 102% indicates that the test method is not the right one, the field test procedure is off, or the site contractor is working harder than he needs to on the compaction. You should be proofing your field densities, which I assume you're using a nuke for, with occasional sand cone testing.

10,000 m^3 is about 1,000 dump trucks, no?
If you have a good consistent soil source, once every 1,000 trucks may not be too bad, although it seems like too many trucks per test.

Your technicians on site should be trained to recognize when the soil has changed enough to require another test.

Specifications should relate the testing, assuming a consistent soil source, to the amount of fill placed, not a time element. If you only have 10 trucks a week, the testing is excessive. If you have hundreds of trucks a week, the testing is not frequent enough.

 

[cvg]

not sure what "regularly show compaction >100% of sampled MDD" means. Does that mean the you have some tests greater than 100 and many less than that each day? or that the majority of tests are over 100?

I would not be surprised with this much volume that the material tested in the lab may not always be representative of the material tested in the field. One lab test for every 50,000 cubic meters (soon to be 210,000) is not sufficient in my opinion.

 

[motorbiketocrank]

Yes, you should be doing more tests. At least several per day for this type of quantity. As noted you should be doing sand cones once in awhile to confirm nuke gauge. And you should be doing regular proctor tests as MDD and OMC will vary. If the material is very consistent you could do a test program to determine the number of passes necessary and have that as your criteria with lesser frequency densities (still would test every day even under this situation). Proctor tests might also not be completely "accurate" for that low of fines content and you could consider relative density tests.

Having a spec that determines # of tests based on a timeframe isn't good practice. It should be per volume or per area for each lift.

 

[Ron]

Agree with others. It is ludicrous to specify a test frequency base on time when the test has nothing to do with time. If you are consistently getting >100% compaction, then you are likely outside the range of the MD relationship...at the least, run some check points on the MD relationship to see if the material still falls on the curve.

 

[charnott]

thanks for the comments.

jgailla - im in Qatar, designer is British (as am I), we the PM are american with a spec that references both BS and ASTM standards within it. the spec calls for sand replacement for measuring in-situ density relative to the OMC/MDD from Proctor (to BS1377). Field density tests (again to BS1377) are specified at 3 tests per 1000m2 of each layer however the lab MDD/OMC is only specified as once a week, hence my concern.

cvg - of the 20 in-situ density results we have so far (work has only recently commenced) 9 of them are >100% with 4 >105%. i wouldnt be too concerned if a few tests were marginally >100% however 25% > 105%MDD is of concern and the obvious potential cause is the very light requirement of OMC/MDD testing when compared to the intensive field density testing specified (see above comment).

geobdg - i would have prefreed nuclear gauge with sand replacement for corroboration however the Contractor proposed sand replacement only and the Designer has accepted that. A compaction trial was undertaken using varied layer thickness, passes etc to optimise the methodology hence my belief it is the unrepresentative lab OMC/MDD results causing the >100% in-situ density values.

Ron - in addition to the RFI i have commissioned the projects independant material testing consultant to collect samples for OMC/MDD and undertake their own in-situ density..........we await their results!

THANKS to one and all, i was looking for re-assurance that this was a genuine concern that warranted an RFI to the designer and i think i got that. Excellent website and I'll be spending a lot more time here me thinks!!!

 

[BigH]

I may presume you are working on one of the new highway ringroad projects.

I am still not certain if the specifications call for the light tamping (2.5 kg hammer) or heavy tamping (4.5 kg hammer). Of course, if the specs called for the light tamping, it is very conceivable that you will have compaction test results consistently above 100% - remember this is only a comparison of a field dry density to a laboratory determined maximum dry density using a specific energy level - and the heavy rollers impact more energy than the light tamping method.

The Indian Highway specification basically state the following (third revision - of which there will be a later one that I don't have):

Granular: one density of compacted layer per 500 m2
(Borrow Material for embankment and subgrade) - 2 tests per 3000m3

Now what constitutes a "test" - or as they imply a "set": they say that a field density "test" is a minimum of 6 determinations (sand cone) or 12 determinations (nuclear density) - then they apply a statistical function to determine if the compactions passes muster . . .

RC(average) > RC(specified) + (1.65 - (1.65/N^.5))x stdev where RC is the relative compaction and N is the number of individual tests run in a set.

This makes sense to me . . .

 

[charnott]

BigH - thanks for your comments, Im not on a highway project FYI. Does the highway specifiation you refer to provide recommended frequencies of testing of lab determined OMC/MDD or only the in-situ density?

 

[Drumchaser]

Agree with cvg

"I would not be surprised with this much volume that the material tested in the lab may not always be representative of the material tested in the field"

...also...it has been my experience "sand" is tough to get required compaction on without flooding/rolling with loaded scraper and maybe letting it dry back....shooting thru top lift to test previous lift...unless the sides are confined which is rare here.

 

[BigH]

@Drumchaser - you have reinforced my contention that any "sand" fill be density checked on the previous layer . . . When this wasn't the case and the "surface" test was low, I always dug down 3 to 4 inches and redid the test - better results.

@charnott - see the second point - 2 tests per 3000m3 (borrow area - so this is the MDD/OMC requirement). The project I am working on presently (earth dam) requires one MDD/OMC per day but at least 1 per 10 field density tests - field density tests as 1 (set) per 500 m3 but at least 2 per layer (nuke) and 1 per 5000m3 (sand cone). Obviously the sand cone is a correlation check on the nuke and of course you should have sufficient correlations at the start of the work to confirm the variations.

 

[charnott]

BigH - thanks for the clarification, your specification sounds considerably more sensible than the one I'm forced to abide by and I'll bear that approach in mind when i recieve the designers response to the RFI.