Settlement of Water Tank 1

[moe333]

I'm trying to estimate the settlement of a steel water tank to be constructed over medium dense sany alluvium, water table at about 13 feet. The tank will about 55 foot diameter and 23 feet high, and will be set on a concrete ringwall foundation. I have a CPT profile to 50 feet. I don't think settlements will be excessive for these conditions, but want to get an estimate to confirm this.

I am considering using the Schmertman (70, 78) method of estimating footing settlement on sand with the strain influence approach and elastic theory, but this method is for a rigid footing, whereas the tank should be flexible.

Another simplified method I'm considering is by Myerhoff (65), modified by Duncan and Buchignani (76). This is a simple formula which uses blow counts and bearing pressure.

Any suggestions are appreciated.

 

[SlideRuleEra]

moe333 - Go to my (non-commercial) website and download the classic paper "Oil Tank Foundation - Boberg" for addtional insight in to ringwall action.

http://www.angelfire.com/mech/slideruleera

Best Wishes

 

[GeoPaveTraffic]

I would be very carefull using any method for a rigid foundation for a steel water tank. I don't have a good reference for another method but will take a look. Also, unless your CPT went to rock or another hard pan, your boring is not deep enough.

I would expect the settlements under the center of the tank to be several inches. You should also consider the effects of other groundwater conditions when comunicating with the owner.

 

[Ron]

Yes, the tank is flexible and the ring wall is rigid....but because of the confinement of the ring wall and the soil below the tank, it acts as a unit with reasonable rigidity to act as a foundation unit. I would use Schmertman's approach as it is well researched on sands.

 

[JoeTank]

After many years of designing and inspecting water and oil storage tanks I have come to the point that I question the "confinement" capability of a concrete ringwall. Most of the tanks that I have observed after many years of service often show local settlement of the soil adjacent to the inside wall of the ringwall. This seems to be due to the fact that most contractors cannot properly compact the soil immediately next to the wall or they just do a poor job of it. It would seem that this condition would invalidate the lateral confinement assumption. Perhaps someone would like to comment on this problem. I'm just a dumb structural guy, not a geotech dude.

Steve Braune
Tank Industry Consultants

http://www.tankindustry.com

 

[SlideRuleEra]

SteveBraune - I can't speak about the geotech issues either, but I have been involved in the design / construction management of a few ringwalls. Contractors not "being able" to properly compact the soil next to the wall is not the issue. Contractors not "being required" to compact this area is highly probable. Using powered hand tampers, adequate compaction can be obtained all the way to the face of the wall - this level of detailed construction management is often ignored.

A specific example: During construction of a new power plant in 1981, the contactor constructed the ringwall and properly compacted the fill, both inside & outside the wall. The industrial waste tank was erected. During hydro, the tank/ringwall started leaning to one side. After about a week (with the hydro water still in the tank) the lean stopped, with both the tank & ringwall about 15 degrees out of plumb. We had the tank dismantled - under the tank bottom, the properly compacted fill had been contained by the ringwall and had also "tilted" 15 degrees.

The tank/ringwall had been constructed over an area at the edge of a (backfilled) 15 ft. deep excavation for the plant foundation. This was were the excavation had been sloped - one side of the ringwall was on 15 ft. of 6 month old fill (that had not given a high degree of compaction). The opposite side of the ring wall was on undisturbed soil.

My point is that the ringwall "took" the interior confined, compacted soil "with it" during this differential settlement "experience".

 

[BigH]

My question still remains - if you read my posts in other similar threads - why are you using a ring wall? I don't think it is necessary - I've been involved with more than 20 large tanks (up to 150 ft in diameter) and none of them had a concrete ring wall - all were normal steel plate tanks resting directly on a granular "tank pad" on the natural soils - and these ranged from firm clays to till; and they all behaved very well even for the ones on less than stellar foundation soils (Hamilton Harbour Area in Ontario). As for settlement, try out three or four reasonable methods - obtain the estimated settlements and get the average of the various methods - this will likely be in the right ball-park (as per a method proposed years ago in Ground Engineering).

 

[JoeTank]

BigH,
Several years ago I chaired a task group charged with the general revision of the foundation design appendix of a nationally recognized tank standard. There were about 10-15 members, about evenly split between major oil companies and tank contractors. This group had extensive tank foundation experience. It was interesting to note that there was a "use a ringwall" crowd and a "don't use a ringwall" crowd. Both had their reasons, motives and opinions... and there was no way to get anyone to repent of their approach. Needless to say, all types of tank foundations are addressed. Neither side should prevail as it would be foolish to adopt a position to always use a ringwall or never use a ringwall. Personally, I'm convinced that a ringwall is the better way to go, not for geotech purposes, but for maintenace purposes. Earthen foundations seem to have too many edge settlement problems, often related to site drainage.

Steve Braune
Tank Industry Consultants

http://www.tankindustry.com

 

[Ron]

SteveBraune....I've seen the condition you describe inside the ringwall and yes,it is usually due to poor compaction, whatever the reason. Once that "settlement" is complete, which is usually fairly quick, then the confined section acts as a unit.

Ring wall or no ring wall?? Either works. A high modulus pad with sufficient extension beyond the walls works as well as a ring wall. One reason for using a ringwall is to minimize the undermining potential from erosion.

 

[moe333]

Thanks everyone for the insight. The civil/structural engineer for the project has made the decision to use the ringwall foundation. Based on the insight from this forum, and from review of literature, I plan on estimating the tank settlement from several methods assumming the tank induces a unifrom contact pressure on the subsurface soils.

 

[BigH]

Further clarification on "tank pad" design. This consisted of well graded sand and gravel - often, as in Ontario equivalent to MTC Granular B Type I (crushed) or Granular A. It extended at least 5 ft (1.5m) beyond the edge of the tank. Tanks are typically within a bund for spillage control so there is little erosion problem - of course, if there may be, you can treat the materials to minimize/obviate it.
Thanks for the comments SteveBraune. It seems that in the various threads in the various forii here, I fell like the only one who questions the need for a ring wall.
For moe333 - don't forget that the centre of the tank will settle more than the edges so the tank should be constructed initially with a slight upwards camber towards the centre of the tank.

 

[geoleng]

I would use caution in assuming that the tank area has uniform loading. Usually the loading is centralized; with a larger load within the center or 1/3D(not sure about that number)of the tank area.

 

[JoeTank]

If the tank has a flexible steel bottom and the product is a liquid, the load is uniform. Can't be anything else. Yes, there are load concentrations at roof support columns and at the shell, but other than that the load has to be uniform. The uniform load due to the liquid is equal to the unit weoght of the product multiplied times its depth.

Steve Braune
Tank Industry Consultants

http://www.tankindustry.com

 

[MRM]

Maybe there was some confusion between load and settlement? While the load is uniform, the resulting settlement is likely to be greater in the center of the tank...

 

[geoleng]

Sorry, I obviously didn't fully understand the problem. The tank is flexible and will have a uniform load.

 

[dmoler]

I would like to add my 1/2 cent worth. Just because there is uniform loading, soil settlement characteristics do not allow soil to settle uniformly. The center of gravity of the loading would influence settlement characteristics of the structure. While the loading would be uniform on the bottom of the ring structure, the center of the ring should experience more settlement with the edges of the ring experiencing less (1/2?) without any footing. The differential settlements are, of course, the killer in any project, rather than total unit settlement.
The differential settlement due to settlement in the center and the edge must not exceed an allowable value. Slideruleera's case actually sounds like a shear failure of the soil rather than excessive settlements.

 

[tincan]

At 23-feet of water, the soil loading will be around 1,500 psf. This normally in its self would not require a ring wall, and no anchor bolts are required for overturning.

The bottom plate is normally 1/4" plate and the plate is extended 1+ inch beyond the face of the tank wall. This 1+ inch extensiion + the wall thickness + an equuivalent length of 1+ inch on the inside of the tank is used to carry and transfer the shell loading to the subbase structure. The structure should weigh around 85,000# and have a perimeter loading of 85000/55pi = 175 plf = 14.5 #/in resulting in a loding of approximately 7 psi.

As noted by Steve above, there are ringwall people and there are pad people. Reguardless, the center of the tank will settle more than the edge. We grade the granular backfill of the ringwall sloping from the center to the edge to account for the deflection. As a rule of thumb, we use a 1-inch in 12-foot grade. For your 55-footer, grade the center 2 1/4-inch higher than the edge.

Best Tincan

 

[GAFenton]

SteveBraune: don't put yourself down. Your "dumb structural guy" thoughts agree completely with elasticity theory. Wish that we could all be so dumb. Elasticity theory says that a rigid foundation will have infinite stresses at its edges. We all know that isn't going to really happen, but the theory nevertheless suggests that some (or a significant amount) of the settlement which would normally occur towards the inside of the tank will occur at the edge once the edge is strengthened, which, of course, makes sense. Clearly, additional compaction will help, but against infinity I'm not sure if I would entirely blame the contractors. Probably it would be better to go with BigH's suggestion to not bother with a ringwall unless you want to reduce the interior settlement for some reason (e.g. to reduce the shear stresses on the steel membrane near the tank edges? There might be excellent reasons to do this if one is at all concerned about protecting the environment against leakage from an oil tank. However, this is just a gut feeling; I have no idea how well these tanks are reinforced against failure at the bottom edges. It might be worth looking into this before deciding to discard the ring walls. My gut feeling is to use ringwalls for oil tanks, don't worry about it so much for water tanks and/or reinforce the tank bottom edges).

Regarding settlement equations, it seems to me that Harry Poulos compiled a book on theoretical settlement solutions. Here it is;

H.G. Poulos, and E.H. Davis, "Elastic Solutions for
Soil and Rock Mechanics", John Wiley & Sons, 1974.

This might help you, moe333. Hopefully, your library has a copy since it has settlement solutions for flexible circles, circular rings, rigid circles, etc. I don't have the book, but it might have even include flexible circles with rigid rings. However, be aware that while theoretical solutions are not too bad for flexible loads, they have this infinite stress "problem" at the edges for rigid foundations. The latter leads to a (possible) over-prediction of settlement for rigid foundations (or, at least, I've read complaints about the predictions, which may only mean that they are not bad). For semi-rigid foundations, the overall settlement will be somewhere between the max flexible settlement and zero. I would assume the same total displaced volume, as in a boat.

 

[am]

hello every one

I have been reading with interest this valuable post and noted all interesting recommendations.

I have a problem of a different nature. I do not have any experience in tank design but this is definitely a very interesting subject. We have large process water tank, 26m diameter and 7m high, resting on 300mm thick ring wall. I do not have the design calculation of the structure and foundation and can not comment on the adequacy of the structural components, but assuming that they have adequate strength - being done by suitably qualified and experienced people.

The tank has three large water inlet pipes. The site people reported that they are experiencing significant. vibration in the tank when those inlet pipes, pumps are running. Those pipe and pump foundations are isolated from the main tank foundation. I am wondering if any of you have experienced such a situation and what type of action can be taken to remedy such a situation. Is it posssible that this may not be structural engineering related problem also!

Any input will be greatly appreciated.

Kind regards

AM

 

[BigH]

GAFenton - if I remember right, you are a Canuck? My experience with the oil tanks on pad foundations and not on ring walls are with two very large Lake Ontario refineries, a storage site for a large pipeline company in southern Ontario, tanks in the Hamilton area and just about every major oil storage tank in Sarnia - at least up through the 1980s. To my knowledge, none ever had concrete ringwalls; they have all stood up without problems so it goes, then, as to why? Sure, the edge plates are beefed up a bit - If I was back home, I could probably find this out - but I'm a bit far away (10k klicks or so). Sorry.

For am - what kind of soils are you on? You say that the pumps, etc, are isolated (i.e. separate) from the main tank foundation - still, the pump pad isn't big enough and is not damping sufficiently. The vibrations may not, then, be isolated from the main storage tank foundation. You may wish, then to truly isolate your pumps via trenching or other "cut-off" measures (See Richart Hall and Woods' book on Soil Dynamics).