Settlement behind retaining wall 7

[eric1037]

I have a bit of a conundrum. We have a project that is a new school building. Part of the school has a lower level and part is slab-on-grade. At the interface between the two levels, there is a below-grade retaining wall. The wall retains about 13 to 14 feet of soil. The subgrade materials at the bottom of footing consist of loose to medium dense sands with a maximum allowable net bearing pressure of 2,000 psf. The design bearing pressure was verified prior to construction of retaining wall foundations.

During some of the utility installation, the plumbers discovered a significant void (4 to 5 inches) under the floor slab of the slab-on-grade portion adjacent to the retaining wall.

Our company observed the backfilling behind the retaining wall. The fill consisted of a relatively clean sand. The fill was placed in lifts of about 8 to 12 inches and compacted to 95% of the modified proctor. The compaction was done using a large vibratory roller and a walk-behind plate compactor close to the wall to avoid additional stress on the wall.

We observed the outlet of the retaining wall drainage system and did not see any evidence of sand at the outfall. We did not see any evidence of wall movement or settlement. No cracks - nothing.

Our approach right now is to perform several soil borings through the slab-on-grade behind the retaining wall. We will determine the relative density of the fill as it is right now. In addition, we are verifying the Maximum Dry Density by performing a new Proctor and performing a sieve analysis.

We have also recommended performing a ground penetrating radar scan of the slab to determine if the void is widespread or localized.

Lastly, we have recommended video survey of any utilities in the vicinity of the observed voids.

I guess I find it relatively hard to believe that the sand behind the wall could settle 4 to 5 inches under its own weight. Especially since we observed the fill placement and tested the fill as it was placed. We have somewhat ruled out settlement of the subgrade since we have not seen any movement or settlement of the retaining wall.

Have any of you experienced anything of this nature? Am I missing anything? Should I be performing any other tests or investigations?

Thanks in advance!!

 

[SlideRuleEra]

Sinkholes and subsurface solution voids come in all sizes, miniature to gigantic, could collapse of one of these be a cause?

http://www.SlideRuleEra.net

 

[eric1037]

We are in west Michigan. We don't really have that phenomenon here as far I know. We do have such things due to gypsum in the rock, but rock is at about 200 feet at this location and the gypsum deposits are more to the north near Grand Rapids.

It is suspicious because I haven't seen any settlement of foundations or movement of the retaing wall. I wouldn't expect the sand fill to settle that much.

 

[SlideRuleEra]

Another thought - Your company observed placement and compaction of fill behind the wall and under the slab on grade. How about the subgrade under, or around the lower-level floor? If this material was not placed properly, the void that you are seeing now could have resulted from raveling all the way down to there - I have seen this happen (a tank on a shallow ringwall foundation, on well compacted soil, started tipping over during hydro-testing because of inadequate backfill of an adjacent deep structural slab).

http://www.SlideRuleEra.net

 

[eric1037]

We did observe the subgrade under and around the lower level. Everything seemed OK.

I just received a report back from the field and it appears the voids are less towards the ends of the retaining wall. We are doing more cores to try to identify the areas with the largest voids.

I am beginning to suspect that this has something to do with the utilities behind the wall. I am not sure if one of the storm or sanitary drains has broken or if there was insufficient compaction of soils after installlation of utilities. (Our company didn't test the backfill of the interior utilities.)

 

[eric1037]

Has anyone seen settlements of 4 to 5 inches in compacted sand fill with basically no load on it? Granted, there is 14 feet of fill. Even if it wasn't compacted to 95% of the Modified Proctor, I can't imagine that magnitude of settlement.

 

[GeoPaveTraffic]

I suspect that you are loosing material into one of the sewer lines. It is unlikely to get that much settlement of a sandy material only 14 feet thick.

 

[MRM]

I wouldn't expect to see more than about 0.3 to 0.5% of the fill height in settlement on good compacted sand backfill (about 3/4" for 14 feet). Of course, most of that movement would take place during fill placement too. Some secondary compression of sand does occur, but again, it would be small in that situation. I think geopave's theory is a good one.

 

[fndn]

If your compaction was done during Unusually DRY weather period, any you ALSO had an intense period of heavy rains, your 14 foot sand fill would have been jetted resulting in the observed settlement. Similarly this could happen if the underground contractor was testing the hydrants and the water was diverted to the backfill zone.

 

[DRC1]

One other possibility is an abandoned sewer. Especially if it is an older school that has been modified in the past. Many municipality records are far from perfect. This could explain the void, yet the lack of sand in the known utilties.

 

[GeoPaveTraffic]

fndn,

I have to disagree with you. Jetting of a well compacted sand fill will not induce additional settlement or compaction of the fill.

 

[Geo69]

eric,

We had a project several years ago where a house settled. Approximately 17 feet of free-draining gravel was placed below the house and supposedly compacted. During our investigation, several inches of settlement were observed between the bottom of the floor slab and the top of the fill. We determined that the fill was not properly compacted.

I would recommend checking the compaction and the moisture content of the fill. Sometimes fill compacted dry of the optimum moisture will experience settlement when it becomes wet.

Some other questions/considerations:

Was the fill compacted to 95% of a modified proctor?

Did the person testing the fill choose the right proctor?

How much time elapsed bewteen the fill being placed and the floor slab being placed? Even compacted fill will move or settle into place shortly after placement.

What type of sand was placed? We have seen contractors have difficulty with certain types of rounded, free-draining sands in the past. It's like trying to compact sand at the beach. Even after much compaction effort, you can still walk across it and leave footprints.

Hope this helps.

 

[eric1037]

A bit more in the ongoing story:

We performed hand augers and DCP testing and found that there were some relatively loose zones.

We collected samples and performed natural moisture contents and a verification Modified Proctor.

Turns out that the wrong Proctor may have been used on at least part of the fill.

When we went back and recalculated the percent compaction, there were some areas where it didn't meet the spec.

Now, you all know as well as I that it's difficult to go back and second guess the test reports. However, it looks like some of the material may have only been compacted to 90% of the modified proctor.

I know its difficult to estimate settlement based on percent compaction, but I am wondering if even being off by 5% would be enough to cause 4 to 5 inches of settlement. I would think to get that much settlement, almost no compactive effort would be required.

Our next phase is to perform ground penetrating radar on the slab to identify the areas of void and to attempt to determine the magnitude of the void space.

We are also recommending video survey of the utilities in the area.

Thanks for all of your input so far. Anything that I may be missing?

 

[MRM]

Eric,
You shouldn't be advertising the fact that you used a DCP to investigate this problem. Don't you know how unreliable and meaningless the results are?

Of course I'm kidding!! Using a DCP was a great idea for this application; particularly since you're using it in "clean sand" fill.

If you don't mind me asking, what DCP instrument do you use, what kinds of blow counts were you getting at which depths, and what is the actual gradation, mineral type, and particle shape of the sand?

I've got a few correlations I could share with you if you're interested. I'm north of you, in the Traverse City area so maybe there's a chance the "clean sand" fill you're testing is similar to what I have calibrated my instrument in.

 

[Geo69]

eric,

you may consider remolding samples of the sand to say 85%, 90% and 95% compaction into a consolidation ring and measuring their compressibility. You could use this to estimate the amount of settlement that has and/or could occur.

 

[BigH]

I find it disbelievable that you would get 4 to 5 inches of settlement under self-weight of 13.5 ft of backfill (average). You have indicated that you may have as low as 90% MDD - modified; well, approximately, this would be equivalent to 95% standard MDD which is a typical specification for highway fill. Highways will not tolerate 4 to 5 inches of settlement (after placement) for such a low embankment.

Let's look at it another way - suppose that the 90% modified MDD can be very conservatively taken as a relative density of 40%. This implies (admittedly through correlations) that the "N" value is in the order of 10. If you go to a chart of N vs allowable bearing pressures for 1 inch of settlement, you would be permitted to put on as much as 3000 psf (150 kPa). 13.5 ft of fill applies only some 1500 to 1600 psf to the foundation stratum. If you take the half-way height of say 7 ft - you get approximately 800 psf acting on the whole of the bottom 7 ft - this is about 0.3% of 1 inch - or expecting 0.3 inches of settlement due to overlying weight on the lower 1/2. Taking this up - one would expect self weight compaction to be less than 1 inch for sure.

There are other reasons that might cause additional settlement - saturation (i.e., puddling) but this, in my view, is doubtful for 4 to 5 inches - and as indicated earlier, it is likely the loss of material by washing out or washing into services, etc.

 

[BigH]

oops - not "0.3% of 1 inch" - but ""30% of 1 inch"".

 

[eric1037]

A quick update on this project:

We video surveyed the utilities in the area and could not find any evidence of soil or damage. We even went so far as to video the perforated drain behind the retaining wall.

We could just not believe that consolidation of sand soils could result in the amount of void that we observed.

So, in the interest of time and client satisfaction, we have volunteered to pay for the grouting of the void space. We made it clear that we do not believe we are entirely responsible for the void. In addition, we have requested an unconditional waiver on further responsibility for the problem.

I'm sure we could have made a research project out of this and denied responsibility, but we most likely would have lost a good client and would have ended up in court. That would have most likely have ended up costing as much or more than we paid to have it fixed with no guarantee of winning.

Lesson learned: Make sure you perform several proctor tests during significant backfill.

Live and learn!!

Comments? Criticism?

 

[eric1037]

BTW:

I gave you all a star for helping me think through this problem.

This place is great!

 

[SlideRuleEra]

eric1037 - Good business compromise!

http://www.SlideRuleEra.net