Rebound of Sand after Surcharge - Storage Tanks

[motorbiketocrank]

I have a project for planned aboveground water storage tanks. Subsurface materials primarily consist of SP soils with decent blows (varies but typically N = 10 to 20 range). Due to the depth of the stress influence zone, anticipated settlements are several inches.

To mitigate this settlement my design called for construction of a fill to apply a surcharge that will equal the pressure of the planned tank mat (w/ overlying fluid). Surcharge will then be removed after settlements have occurred and prior to construction of the tank. My inherent related assumption is that there will be no significant rebound in these cohesionless soils after removal of the surcharge load and that the final tank load will cause no significant settlement.

I'm experiencing some pushback, however, by another engineer who feels there will be significant rebound in the sands after removal of the load. I grant that the "elastic" settlement in sands is part permanent movement in the particles and part true elastic compression which would rebound (I view it like loading up a big pile of marbles). But my feeling is that the truly elastic movement and related rebound component is small. I've done some searches to find literature discussing the mechanics of settlement in cohesionless soils and rebound potential but haven't found much.

Does anyone have thoughts/information on how much rebound (i.e. % of total settlement) is likely under this scenario? Thanks.

 

[Ron]

The purpose surcharging loose to medium dense sands is to take out the "reorientation" settlement, not the elastic compression. In short, rebound is not a big deal assuming reasonable time frame of construction. The elastic compression that is rebounded will be recompressed during construction, with a net of no significant long term settlement.

 

[motorbiketocrank]

I realize the surcharge will only take care of the "reorientation" part of the settlement and not true elastic behavior. But I can't seem to find any good information on how much of the calculated settlement (estimated using typical correlations of N value to modulus) is "reorientation" and how much is elastic movement that will rebound. If I calculate 5" of total settlement and 1/2" is pure elastic movement that will rebound then I don't have a problem. On the other hand if it will rebound 2" then I do have a problem. My instinct is that the rebound is small but just having problem finding documentation.

I really don't know the timeframe for tank placement after removal of the surcharge. I think, however, that any elastic rebound would happen almost immediately after surcharge removal and is therefore something I have to plan on regardless of timing.

I might be misinterpreting your intent but I generally don't agree that the rebound will be recompressed during construction resulting in no significant long term settlement. If there really were significant rebound (for example say 2"), it will not recompress until the water is placed into the completed storage tanks. This recompression would then be excessive settlement affecting my structure. Again I doubt there is much rebound.....but if I'm wrong, it will have an impact.

Thanks for the input.

 

[Ron]

I agree with you that most if not all of the "settlement" will be re-orientation. The elastic compression doesn't even happen until most of the re-orientation settlement is done....like trying to compress the soil below by pushing on a bed of corn flakes....the corn flakes will re-orient before allowing compression stress to be passed to the compressible soils.

Terzaghi, as with all things geotechnical, addresses this. Check his book as well as some of the other classical texts such as Sowers.

 

[fattdad]

read this article:

http://ascelibrary.org/gto/resource/1/jgendz/v113/i9/p967_s1?isAuthorized=no

f-d

¡papá gordo ain’t no madre flaca!

 

[BigH]

Is this a water tank founded directly on the soil (like a large oil storage tank) or is it an "up in the air" tank? If the former, I am not sure why you are really worried . . . Large diameter steel tanks can take a fair amount of settlement without problems - and 5" doesn't seem like it is overly large . . . Just use flexible connections of piping. it is always difficult to estimate the "rebound" of the sand. Can you carry out a large diameter consolidation test of the sand (at or near the original in situ density) and see what happens?

 

[motorbiketocrank]

It's an at grade tank (not up in the air). Can't do any additional testing at this point as the design was completed awhile ago and the concerns of the other engineer are now coming up during an oversight agency review just before the project is advertised.

I agree there's a chance the settlements may not be a problem. I think my calculated totals are pretty conservative and will be less in reality (wish I had some dialtometer testing to better estimate). But my design was performed very early in the project when many specifics of the tank weren't finalized. Structural designer wasn't sure but thought it likely that a concrete ringwall would be used. My main concern was for differential settlement at the edges that could cause problems for the ringwall and connections between the bottom and sides of the tank.

The subsurface materials are free draining and therefore surcharge settlement will pretty much be immediate (so no significant construction schedule impacts). Plus they are bringing material on site for use as fill in other areas (so no significant extra material import costs). Basically my thought was "since there's not a significant cost or schedule impact why not temporarily stockpile the imported fill material over the tank location and help eliminate the potential for settlement problems". And especially given the uncertainties over the tank type, etc at the time of design this seemed to be a no brainer.

I did some more looking on the rebound issue and found a good discussion in Fang about stress history (he indicates if a granular material is loaded, unloaded and then reloaded, the reloading modulus can easily be 5 to 30 times that of the original loading.....which to me is the same as saying that there is little rebound). Mitchell also had some information discussing the topic which gave me the general impression the true elastic behavior is mostly at small strains and the larger settlements are primarily plastic type behavior.

Thanks again for the feedback.

 

[BigH]

You might want to do a search of the Eng-Tips site for discussions on tank foundations. There are several good ones. Many prefer ring walls; others, myself included, don't see a real need for them in most instances. Our company designed many tanks - some large ones on clayey strata with Su values only in the range of 50 kPa (1000 psf) and although there was settlement, the tanks behaved perfectly fine. Many years ago in, I believe, ENR there was an article about a large tank that Dames&Moore designed - they said settlement under the tank would be something like 24 inches. It turned out that it was almost a 39 or so - the client removed the tank and put it on piles. A lawsuit, of course, resulted - D&M contention was that even though the settlements were larger than anticipated (and if you estimate the consolidation settlement to within 30%, you've typically had a good day), the integrity of the tank was not compromised - it still behaved and acted as intended. Don't know the outcome of the lawsuit, though.

 

[BigHarvey]

You should observe a settlement of 1/3 of the initial settlement when you reload your sand layer ( at least it's what is considered in pressuremeter theories )

 

[fattdad]

Please consider reading the D'Orazio and Duncan paper that I linked in thread above.

Surcharging the site will (may) increase the modulus value, but it won't make it so elastic compression will become moot. If current soil conditions and imposed stresses calculate out to show 10 inches of settlement a preload that results in 10 inches of settlement will not justify zero settlement after the preload is removed.

Get the ASCE publication and learn about settlements below big tanks.

f-d

¡papá gordo ain’t no madre flaca!

 

[aeoliantexan]

For a sewage treatment plant in the Missouri river floodplain we removed surface clays and silts down to silty sand which was over 100 feet thick and contained some thin clay layers. We built a permanent fill to 5 to 20 feet above original grade plus a rolling surcharge to simulate the large treatment structures plus 10%. Several inches of settlement occurred. When we removed the surcharge, the rate of rebound per foot of fill was roughly one-fifth the original settlement rate.

 

[motorbiketocrank]

Fattdad - I did read the paper (thanks for the link). I plan to keep it on file for future use but am not sure how applicable it is for my current site since it primarily discusses behavior on cohesive soils. I also don't have a defined compressible layer of limited thickness that it presents will result in the worst distortions.

I'm really not too concerned that I'll have problems at my site. My calculated settlements (5"-6" range) are relatively small overall for a tank and likely I could get away without the surcharge at all. And I think these settlements are conservative (on other recent projects using dilatometer has cut our calculated settlements in half as compared to SPT correlations). But, considering the very small cost and time to surcharge, I think it's good insurance, helps limit potential impact of various uncertainties in the analysis and limits potential problems with utilities etc. So I was mostly just trying to get a handle on the amount of rebound so I could intelligently discuss with the reviewer.

Thanks also to others for their input.