Steel Tank Partially on Fill 1

[JedClampett]

We have a large steel tank (>100 ft. diameter) that's planned for the side of a hill. The hydraulics and property line issues dictate that the best location is partially on excavated hillside and the rest on filled hillside. Some of the fill is twenty feet deep. The tank will have a ringwall. My question is:
Does anyone have a concern about this, in particular due to settlement? We'll put in flexible pipe connections. In my opinion, steel tanks are not as susceptible to differential settlement as other types of tanks or, for that matter, other types of structures. Other senior people in our company are worried. We'll ask for very controlled fill compaction, but contractors never can be trusted with earthwork. So I'm not under any illusions that we can eliminate settlement. But I think that the tank is flexible and as long as the settlement is reasonable, it should be OK.

 

[BigH]

What's the cross section look like "downhill" of the tank? and of the tank relative to the original and proposed slopes?

I believe that you might have some worries about edge failure if it is very near a slope crest heading downhill. 20 ft of fill is a lot - and while it might be compacted well, you might want to consider building the fill as a reinforced fill (see some past experiences with steep reinforced slopes for railway embankments - one that comes to mind is near Hamilton Ontario). Also how are you going to blend the new fill with the existing ground onto which it is to be placed (see Ohio's DOT tech memo on sidehill fills).

While not exactly the same, the paper by Bjerum in the 1957 London ISSMFE conference would be worth a look - in it they had an edge failure of a tank on clay.

 

[FixedEarth]

Your assumption on flexible tank is valid for the interior and not for beneath the edges (ringwall foundation). As mentioned above 20 ft of Fill is substantial. The fill settlement will not be completed at the conslusion of earthwork, so some settlement will take place for a period after the grading operations.

You may consider installing settlement plates on the fill side and see when the settlement levels off. This tells us when the fill settlement under its own weight has completed, however, we will still have more settlement under the 20 ft fill than under the cut section. In this case, try to take consolidation tests, say 15 ft below the tank floor on the cut and on the Fill side.

From the consolidation test results, your geotechnical firm will be able to tell you if differential settlement is a concern, if shallow ring foundations are still feasible, or if you need deep foundation support below the ring wall foundation, such as drilled piers. If you decide to support the Fill side with piers, you will need to check differential settlement between the pier supported and the shallow foundation on the cut side.

Then there is daylight distance requirement to safe guard against the edge failure, BigH mentioned. I would agree with the concerns of the senior people.

http://www.FoundEng.com

 

[JedClampett]

One of our problems is that we can't get much value out of our Geotechnical Engineer. He wasn't selected by us. We're not really sure if he comprehends the issue.

 

[cvg]

hire your own qualified geotechnical engineer, one that has successfully designed more than one tank. Ask him to do a peer review of the design. Shouldn't cost a lot to just review another engineers report but will certainly help you sleep at night.

 

[BigH]

If you do hire an independent geotechnical engineer, make sure he and you advise the other geotechnical engineer and client that the geo report is being reviewed (Code of Ethics stuff). I would still like to see a cross section of the hill with the tank and fill superimposed and the nature of the soil(s) encountered. Interesting problem . . .

 

[racookpe1978]

Not a geo or structural engineer, so this is an open question:

Would the problem - the threat to the tank and its foundation by movement of the fill - be reduced by any kind of retaining wall around the fill, instead of a simple fill and natural slope of the fill earth? Obviously, any retaining wall is both more expensive to build than an open un-restrained slope, but is the advantage of "no slump" greater than the higher pressure of the wall's weight over a smaller area of fill? Could that ringwall be incorporated into the retaining wall design?

Also, is a threat of the higher part of the hill coming down onto the tank over time? Seems like that happens frequently on highway cuts when the rains return.

 

[chicopee]

You can always drive sheet piles on the downside of the hill.

 

[cvg]

sheet piles and retaining walls will not prevent the differential settlement.

 

[BigHarvey]

Your comment about contractors who can't be trusted is not welcome, particularly when you look for advice from people who can be contractors like myself.
I you are worried about differential settlement you can use rigid inclusions for the filled area in order to reduce the total settlement in that area and therefore eliminate differential settlement . You will have to consider a 1 to 1.5 m thick compacted distribution layer over the whole area to transfer the loads to the rigid inclusions network. From a construction point of view, rigid inclusions are piles ( steel driven or ACIP, or any kind of piles ) and from a design point of view, they are designed to settle with controlled settlement values. Design is generally carried out using FEM.

 

[BigH]

If you hire a competent contractor - as with BigHarvey - you would have no problem with fill placement to achieve desired relative compaction. The various suggestions about retaining walls, sheet piles - seems, to me at least, overkill and the tank becomes a very expensive proposition. Would be better to take the tank down very low with perhaps soil nails holding up a nearer vertical uphill wall and put the structure almost all, if not all, on the natural soil - but then I still am not sure of what the hill is made of. Further to as what BigHarvey suggests, you might add 2% cement plus a % of fly ash to the granular fill to provide a cemented fill - and differential settlement of the fill will be minimal due to the rigidity of the fill.

 

[77JQX]

I work for a water agency that has over 40 tanks on hillsides. We require new tanks to be constructed either entirely on fill, or entirely on a cut. We don't allow construction on part fill/part cut. What's the design life of your tank? 50 years? More? What's the consequence of differential settlement? That's a lot of time for settlement to appear, and with a wide, heavy structure like a 100-ft diameter tank, your geotech needs to consider settlement in much deeper layers than with an ordinary structure. Is there room on the site to build a replacement structure if this one shows signs of differential settlement? I doubt it, since you're already having property line issues. If this is the only hydraulically acceptible site, then you've already answered your question, but if it was me, I'd look hard for another solution.

 

[mudman54]

i think BigHarvey and BigH are on the right track to do some ground improvement alternative to make the entire foundation consistent unless, as the last commenter stated, a different site for the tank is available.

 

[molerat2210]

Hire a geotech with steel tank experience. Some steel tanks can be leveled after filling with water and allowing a settlement period.

 

[muuddfun]

What you should do is blow out the back half of the excavation, and then you will have 20 feet of fill under the entire tank.
You are talking about 3,000 yards or so additional grading, that is not very much money to fix your problem.

 

[CEUS]

Without a raft foundation: the scenario is as follows, big deferential settlement is expected to occur because water weight is supported by the ring beam and the bottom plates over the the fill material. as consequence, water leak will be a frequent problem due to weld cracks of the bottom plates of the tank.

I highly recommend a raft foundation system as effective solution to the problem.

 

[BigH]

Basically @CEUS, if you use a cement treated aggregate base as the fill and place and compact, as we have suggested earlier, to 98% MDD modified, you have, in effect, produced a raft foundation. Tank doesn't need a concrete ringwall.

 

[cvg]

correction to above, water weight is not supported by the ring wall. the ring wall supports the outer shell of the tank plus part of the roof load. the water load bears on the bottom plate and is directly supported by the fill. while cement treated base could work, I have never seen a tank partially supported on any sort of fill, only on cut

 

[apsix]

We have used piles in the fill for this situation, but it was a concrete tank.

 

[BigH]

I think that this thread shows the old adage - give the same problem to three geotechnical engineers (and toss in a structural to boot) and you'll get 5 different but plausible scenarios.

This is a steel tank, a cement treated aggregate base will exhibit a compressive strength of more than 5 MPa at 7 days (at 4 to 5% cement - and well graded crushed stone) and it will likely increase to greater than 10 MPa at 28d - clearly this is pretty close to the rigidity of the rock and should not, for a steel tank, be of a concern - a few mm, maybe 10mm, of settlement? As for the ring wall, many a large steel tank have been built without them to "support the tank wall and roof. I've been involved with several - one of a 150ft diameter and >50 ft high. No problems