Compaction level reduces with moisture content?

[EireChch]

Hi all,

I am working on a project where there is a debate as to how the material was QA/QC'd to 90% compaction (standard proctor), however after settlement of some infrastructural, the compaction level is ~80%. The material is well graded sand with trace fine gravel.

One theory being floated around is that the material has a very bell shaped OMC/MDD curve. Someone has proposed that the material was compacted to 90% however as there were rainfall events / drying events, the moisture content has changed. As the moisture content changes, the density also changes. Refer to below sketch, we have an OMC/MDD line similar to line 1 which is sensitive to changes in MC.

I dont buy it to be honest. I argued that the purpose to the OMC/MDD curve is on moisture content during construction, ideally you are at OMC +/-2% (or whatever your spec says), if you are too wet or too dry you will likely find it difficult to achieve 90%MDD.

Infrastructure is supported on granular material and doesn't routinely settle due to changes in MC, does it?! Yes if there is changes to water level as a reduction in groundwater level can increase in effictive stress in soil causing additional settlement, but not if the moisture content changes from 14% to 8% after placement?

 

[GeoEnvGuy]

90% spmdd for infrastructure seems low. If you found 80% the qc/qa could have read the bulk density as dry density when doing nuke shots. You can back calculate the dry density using the moisture content and bulk density to see if that's what you have now.

Is the material filter compatible to prevent loss with the below material or was cloth used?

If the well graded material internally stable or has finer sand washed out?

 

[Mad Mike]

GeoEnvGuy makes an excellent point about where the discrepancy may lie- I've had 2 jobs now where the chaps taking nuclear density readings related the compaction to bulk density instead of dry density- both of them resulted in court cases when the material turned out to be horribly under-compacted.

Agree also that 90% Standard Proctor is very low...

All the best,
Mike

 

[fattdad]

crazy talk!

If the theorists intend on making any sense, they'd have to explain embankment dams! Heck, they'd all fall over!

f-d

ípapß gordo ainÆt no madre flaca!

 

[maxim22]

Think well graded sand should have a pretty flat curve and not sensitive to normal mc changes. NMC 5 to 8%?

Not sure you could place and roll at 14%.

Cheap to do a sand replacement density test in this material.

 

[Okiryu]

I think that if the fill was properly placed and compacted, changes in moisture will not affect the grade of compaction. One exception could be expansive soils but you have granular fill.

Was the original 90% and post 80% compaction determined by the same testing company? Perhaps the 90% compaction results were not accurate?

Did you have heavy rains between fill lifts placement? Water ponding? Did you have trucks and/or other construction equipment traveling on top of the fill during construction?

 

[EireChch]

Thank you all.

We are due to receive the QA/QC test results and also the new field density test results soon. They have undertaken sand cone tests. No nukes involved thank god. 90% was set for back fill in trenches. It is in a non trafficked area so may not need to be 95%. Anyways, this was set a number of years ago by another consultant. We are acting as forensic consulted.

GeoEnv - no geotextiles placed. The natural ground is similar material which we understand is well graded granular material. Slightly cemented so there is some apparent cohesion. My gut says that fines wash out into the natural material below should be limited, its not as if its a sand material sitting upon a coarse gravel. In saying that, open to a recommendation of how to confirm its washout potential is low?

Maxim22 - wouldnt a well graded material have a bell shaped OMC/MDD curve? Refer to screen shot below from Nowak - Guide to Earthworks. I havent seen our OMC/MDD curves yet, my sketch above (and reference to 14% and %) was theoretical to illustrate that a reduction or increase in moisture content of a material that has already been placed should not change its current density. When I have OMC/MDD curves i will post them.

 

[EireChch]

Okiryu - thats my thoughts too. It is the earthworks contractor who is floating the theory, trying to defend their work to avoid the remediation costs. In saying that, their work was QA/QC'd by the clients on site engineer too so it is difficult to say how the material has reduced in compaction.

A new company has done the current FDT. As far as i know it was placed in sunny conditions, no rain. Not sure re construction equipment.

One thing is certain is that the infrastructure has suffered settlement due to the loose material below. DCP test (ASTM D6951 DCP), show loose (less than 3 DCP blow per 100mm) to up to 2m below infrastructure.

How the material has got there is the question, these leads to question the QA/QC test results or the new FDTs which are showing less than 90% compaction.

 

[Okiryu]

With that good fill material (well graded sand with trace fine gravel), I just can think that the contractor did not do a good job placing that fill....

 

[maxim22]

Too sleepy in this office, its my mistake the bell shape curve is correct. Geologist confused well graded with well sorted.

 

[Ron]

The moisture content at compaction is just that. It's what helps to keep us on the Proctor curve. Once compaction is done, the moisture content can go up or down and should not affect compaction in place (the exception is some clayey materials subject to moisture expansion/contraction). For granular materials it makes no difference at all.


A Great Place For Engineers to Help Engineers

Follow me there.....

 

[MTNClimber]

Isn't loose granular fill dumped out of the back of a truck around 80% SPD? That's at least what I remember when I was an extremely bored field engineer prior to smartphones helping ease the pain of waiting for something to happen.

I want to agree with the theory that the tester was a severally confused individual but that's hard to prove. Is there a chance with (random) deeper settlement of natural material would cause the fill to loosen up?

 

[GeoEnvGuy]

You may consider determining the erosion boundaries using the method by foster and fell.

Another mechanism to consider is internal erosion in the more permeable trench with steep gradients. Any evidence of pressurized liquids being released?

Any notes on if spring line compaction was completed for larger diameter pipes. How wide was the trench and would they be able to do spring line compaction.

 

[EireChch]

The moisture content at compaction is just that. It's what helps to keep us on the Proctor curve. Once compaction is done, the moisture content can go up or down and should not affect compaction in place (the exception is some clayey materials subject to moisture expansion/contraction). For granular materials it makes no difference at all.

Thats was my argument too, its only for use in the compaction process.

I have also been teasing apart field density calculations, we are always comparing dry density to MDD. When we receive the new FDT, the insitu moisture content is used in accordance with the wet density to determine the DD. (DD = Wet density / (1 +mc)). Therefore, if we check DD at the same location on a dry day and after a rainfall event say.

After rainfall - the wet density and moisture content will be both be increased, and therefore we should still arrive at the same DD as when it was checked on a dry day. Does that sound logical?

Now if we start tracking a machine over compacted material that has been subject to an increase in moisture content from rainfall or what ever or we start to place new material over a wet layer, this can reduce the density of the underlying wet layer but this is down to disturbance form machinery or compaction effort.

Isn't loose granular fill dumped out of the back of a truck around 80% SPD? That's at least what I remember when I was an extremely bored field engineer prior to smartphones helping ease the pain of waiting for something to happen.

I want to agree with the theory that the tester was a severally confused individual but that's hard to prove. Is there a chance with (random) deeper settlement of natural material would cause the fill to loosen up?

Ive heard that but never actually confirmed it or thought it through. Havent had to. Natural ground is medium dense SPT N 30-40 from 2 to 4m below base of trench then 50+ below 4m. I dont think that could be a reason.

You may consider determining the erosion boundaries using the method by foster and fell.

Another mechanism to consider is internal erosion in the more permeable trench with steep gradients. Any evidence of pressurized liquids being released?

Any notes on if spring line compaction was completed for larger diameter pipes. How wide was the trench and would they be able to do spring line compaction.

No pressurized liquids, it would be nice if it was as it would give a good causation of settlement. I am at a loose end though as to how it was QA/QC'd ok and has since changed. I will have more information soon that i will share.

 

[Ron]

EireChch.....a couple of additional thoughts...

With compaction that low and it being a granular material, the disturbance of the surface from further construction activity can reduce the compaction as you noted.

The contractor is probably confused between compaction and stability. You can get compaction fairly easily in granular materials, but compaction and stability are two different things, obviously. Had the material been stabilized and compacted, you probably wouldn't be having the argument; however, stabilization costs money. A smart contractor will sometimes eat the cost because his life is much easier! He also often has the contractual obligation to use means and methods consistent with accepted practice to control his own work and the protection of such.


A Great Place For Engineers to Help Engineers

Follow me there.....

 

[geomane]

Make sure you have a proctor test run on a representative sample. I see contractors all the time trying to find the ‘sandiest’ sample they can get for their proctor test because it will return a lower maximum dry density and make it easier to achieve relative compaction.

 

[aeoliantexan]

If the sand is clean, relative density may be more appropriate, and considering it may provide some insight. For the materials for which I have used relative density, the difference between maximum and minimum densities was often about 20 to 25 pcf. Then 80% of maximum density would be close to 0% relative density! Even modest compaction would give something higher. Also, any good inspector would use a second method to evaluate compaction, such as a probe rod, shovel handle, stability under foot, or how it digs. Sand may densify when it gets wet, but it will not reduce in density. If it is really at 80% MDD now, it was not compacted.

How it escaped the judgement of the inspector is another issue.

 

[EireChch]

An update.

The settlement (100mm)is within pipes that connect to the main line as shown below. Some cracking has occurred and was noticed during filling, the pipes were never operational. There would have been some areas where water was leaking which would have worsened the situation. However, there are areas where pipes have not cracked / leaked and settlement has still occurred.

The degree of compaction required was 90% SPD, however the new OMC/DD test results are for modified so 80% MPD is a little closer to 90%SPD. Regardless of this there is still the issue as to why the pipes have settled.

From reviewing additional information I now understand that the connecting pipes are supported on backfilled material which was placed as bedding and backfill for two underlying large diameter pipes. The backfill material is site won Sand with silt. Coefficient of uniformity (Cu) is between 3-4, so the material is uniformly/poorly graded, so not good backfill material.

We also have some irregular OMC/DD curves (kind of like 1 and 1/2 peaks). See below image. MDD=1.96g/cm[sup]3[/sup] OMC 10%. I have been doing some research and i suspect that the irregularity may be due to the fact that the lab has started with a very low moisture content of less than 2%. I would have gone 6,8,10,12 and 14% MC.

There is some negative pore pressure that occurs due to the low moisture content at the start of the curve. Negative pore pressure starts to hold the particles together which increases the density, as more moisture is added the negative pore pressure is lost and particles become less dense. I have seen only a small number of irregular curves this bad so open to other theories? Crappy lab?

There has been some minor loss of fines it appears, fines content was 14% at QA/QC, from our recent testing fines is slightly lower 10 to 12%, which could be due to wash out from leaking pipes. However the error could be in test accuracy also. It is a possibility but with less guts than id like.

The connection pipes were installed prior to backfilling, these are rigid pipes that are spanning unsupported from the excavation cut (left hand side) over the adjacent pipes and into the main lines. As they were in place during backfilling of the pipes beneath, compaction beneath the connection pipes would have been very difficult due to the access limitations. Also, compaction may have been reduced/limited over the underlying pipes as a precautionary measure to reduce the risk of damage. I bet the QA/QC field density testing was undertaken at the areas which received the best compaction i.e. between the main lines or adjacent to the edges where there are no pipes beneath. No QA/QC testing would have shown the loose material directly beneath the connection pipes.

Thoughts....?

 

[BigH]

As pointed out - for sand, the relative density test is the most appropriate. Ron hit it on the head. You place a material at the optimum moisture content (or plus minus as the spec says). it is compacted. The test isn't done for 5 days - and the air has dried out the moisture that was there at the time of compaction. Unless there is a volume change the dry density determination would be the same if measured at the time of compaction or 5 days later when dried. If it rains and is measured 5 days later - the dry density would still be the same UNLESS there is a volume change. Volume change as Ron pointed would be by disturbance of the surface due to traffic - even people walking on it.

I have said this before, but for sands, it would be prudent to be testing the previous lift - not the lift you are placing. This is because the surface 50 mm or so might be disturbed due to the traction of the compactor. Also, the surface is not well confined in sand. I used to test the surface and if it was low, I would then dig out 50 mm or so and do another test. But, preferably the spec would say to do the density check on the previous lift/layer.

That is my opinion.

 

[Ron]

Agree with BigH about testing the compaction below the existing lift. This confinement helps to keep from disturbing the sandy soils during the test.

After rainfall - the wet density and moisture content will be both be increased, and therefore we should still arrive at the same DD as when it was checked on a dry day. Does that sound logical?

That is exactly correct and why the moisture content after compaction doesn't matter except for stability...not compaction.




A Great Place For Engineers to Help Engineers

Follow me there.....