Storage Tank Foundation Settlement 1

[FTank]

I would appreciate hearing from anyone that has had experience in storage tank settlements, both uniformly & differentially.API 650 doesn't address this very much that I can find other thanar. 5.10.4.7 d. "The columns shall be effectively....in the event of vertical movement of the columns relative to tank bottom of up to 3 in.".Par. 7.3.6.6 (hydro test)"...differential settlement greater than 1/2" per 32 ft. of circumference or uniform settlement of 2 in."......Neither of these paragraphs address differential settlement across the tank bottom. Other paragraphs refer to "excessive" settlement without defining what is meant by "excessive".Case in point - I am designing a tank that is 100' Dia. x 64' Shell Height, 1% bottom crown up from center to outer shell. Single center column-supported roof with rafters.Geotech report says I can expect settlement of 8.4" at edge and 15" at center of tank which equates to differential settlement of 6.6" Question: do I take the 3" vertical movement of "up to 3" mentioned in Par. 5.10.4.7 d, as the max. differential allowable or am I missing something here?Any thoughts, suggestions, past experience, etc. would be really appreciated.I realize that API 653 has some guidelines but this is a new tank and thus API 653 does not apply. Any other industry standards that would apply?

 

[fattdad]

J. M. Duncan has a paper in the ASCE Geotechnical Journal.

f-d

¡papá gordo ain’t no madre flaca!

 

[Ron]

The settlement estimates are excessive and will give you problems, particularly with connections. For settlements of this magnitude, I would consider some ground improvement techniques or a pile supported slab.

Settlements of this magnitude can cause shell buckling and other issues.

 

[BigHarvey]

Codes are written mainly for teh structural aspects and not the geotechnical ones. A good reference is " Criteria for settlements of tanks" by Marr, Ramos and Lambe, Journal of the Geotechnical Engineering Division, ASCE, vol 108, N°GT8, august 1982.
I recently had the problem for 31 m diameter double wall steel circular tanks ( 8800 m3 capacity )founded on rigid inclusions. We had to propose total and differential settlement criteria because they were not contmplated in the codes.
We proposed the following :
total settlement center : 12 cm under 150kPa
differential edge/center : 1/250
tilt : 10 cm
differential along edge : 1/450
We obtained settlements during the hydro test 1/3 of calculated ( FEM analysis ).
Hope it will help !

 

[ishvaaag]

Also, ascertain if the given settlement by the geotechnical party is some worst to be expected or the likely to be expected, mid to long term. I have seen too much a number of excessive settlement predictions incompatible with everything doable and done; scarcely compatible with reality, in other words. In some cases we have insisted (and got) a more proper statement of what expected (not making anyone to state what do not think, but making them amenable to FACTS and even common criteria far far less conservative than those initially proposed. Remember, this is engineering, and as one of your US colleagues made me aware of, "what to some is just barely enough, to others is a barbarity".

 

[fattdad]

If you have a 100 ft diameter storage tank and you consider it fully loaded there will be stress increases with potential to mobilize settlements to depths of over 100 ft. You mention that the geotechnical engineer has forecasted settlements of 8 inches at the edge and 15 inches a the center. We don't know, however what the geotechnical engineer used to make this forecast. Is there boring data that shows the variation of the soil layering to dephts of 100 ft? What will more than likely mobilze settlements in the soil layer is the change in stress reacting to the soil modulus in the various soil layers. The approximation of soil modulus can be a difficult parameter to estimate accuratly.

Many geotechnical engineers use correlation to blow count to estimate soil modulus. Let's say we consider a value of Es=11N where N is the SPT N-value and Es is the soil modulus [tsf]. If you have a soil modulus of 150 tsf (i.e., N=14) for a layer between the depths of 40 and 50 ft and the stress is increasing 2,000 psf, you may conclude that there will be 0.8 in of settlement just from that interval. However, this simplified calculation really doesn't consider the complete story.

Soil modulus can vary with confining stress. This is one of the reasons that Professor Mike Duncan developed the hyperbolic model to model soil compression. Very few projects can accuratly predict hyperbolic parameters to gauge the stress-dependent nature of soil modulus. So, it's rarely used in practice.

If you are dealing with a deal breaker, I'd consider the use of a dilatometer. This tool will give you a direct measurement of soil modulus with depth and in a spreadsheet form can then be used to integrate the change in load and how it'll result in compression.

Just as an aside, I highly recommend the paper I referenced above. It is a great reference. Many ground supported tanks settle in excess of 6 inches at the center.

Good luck.

f-d

¡papá gordo ain’t no madre flaca!

 

[BigH]

I, in my old age, is becoming somewhat disillusioned by the reliance of "designers" on codes and not on good engineering principals to their own problem and their long experience. I must presume that borings were done and the geotechnical engineer followed reasoned and rational methods of determining that the centre will settle 15 inches and the edges some 8.5 inches. This constitutes large settlements but, in my view, are not "destructive" or excessive that a well designed tank cannot handle. I remember of a case history (Dames and Moore) where the tank was put on the ground (not on piles) and settled some 40 inches. The tank was still useable. Of course, with settlements of this nature, you must use flexible connections between the piping and the tank. I do not think that you need to put the tank on piles. It was indicated that the centre of the tank "coned" from the edges by one inch. Maybe in the final configuration - but what you should do is to camber up from the edges some 7 or 8 inches. then when the tank settles, the edges go down their 8, the centre down it's 15 - but the net centre to edge is down 1 or so inches. - of course, this is based on settlement computations that, if one is lucky, is within 30% of the actual values. You might wish to use some wick drains and load the tank in stages. But in the end, the tank is doable - in practice. You should review the paper by Bjerrum (1956 ISSMFE Conf in London) that talks about edge failures - you might have to consider that. That is my view.

 

[oldestguy]

What is the tank to hold? Let's say it is some volatile stuff, not water, well then you would be more conservative.

Normally these types (in my experience) first had a water test to check that they are OK. If not, repairs can be made before use, in some cases, but more difficult for a double wall job.

 

[FrCivEng]

Refer to the answer posted on API Forum on 20-Dec-09

Geotech report highlights a predictible settlement of 8.4" (21 cm) at the edge and 15" (38 cm) at the centre of the tank ;
Do never consider a geotech report as conservative or not, as it is a factual report which results must be respected by each party involved ( civil contractor, customer, tank contractor)

The difference between the edge and the centre (6.6") does not correspond to what is usually considered as the differential settlement ; for differential settlement definition, refer to API 653 Appendix B

API 650 11th Art 5.10.4.7 does not apply in the present case, it refers only to bottom plate local deflection to guarantee supports legs stability ; it addresses locally bottom plate deflection and earth backfill deformation underneath but without giving any acceptance criteria for earth foundation deformation

API 650 Art 7.5.5 is the criteria for foundation checking before tank hydrotest only, when the tank is empty.

API 650 11th Art 7.5.2 is an additional criteria to be checked before tank hydrotest only, when the tank is empty, after tank foundation checking according to Art 7.5.5 ; Contractor or Customer shall perform a topographical survey to measure shell verticality and roundness ; note that when tank foundation is OK, out of roundness / verticality defects result from tank construction only.

If your tank is cone up type, you can easily manage the difference between centre and edge ; the min requested slope at long term for a cone up is 1 %, so you have to built the earth backfill foundation underneath the bottom plate with 1,6 % in order to anticipate the backfill deformation under loading/offloading cycles during 30 years ; with D = 100 m , the centre will be at 1,6 % x 50 m = 80 cm > 38 cm predictible which is still less than 1 % min ie 50 cm

More touchy to manage is the edge settlement ; API 650 11th Art 7.3.6.6 gives now acceptance criteria for settlement at hydrotest but they seem to be very stringent and may be unrealistic with no real feed back from construction sites and civil contractors; I mean that a lot a tank foundations will probably be refused even in presence of good soil & foundation conditions ; my personal understanding is that this is to protect tank builders responsibility

On the other hand, API 653 Appendix B acceptance criteria, based on tank shell allowable deflection is very less stringent, applicable to existing tanks

However you have to perform your design according to API 650 11th criteria ; assuming you have an external floating roof on such a tank ;
For a 100 m D tank, the max distance being 32 ft (9.75 m), you should consider 36 measurement points 8,73 m distant and check the measured differential is less than 1,3 mm/m x 8.75 = 11,4 mm;
For new tanks, it is very usual to consider Pr De Beer criteria; still referred by Ph. Myers (AST book, 1997) which is

?Si <= L/450.

With :

- ?Si : edge foundation differential deflection calculated at each point of tank circumference, from settlement measurements and analysis,
- L : distance between 2 points considered on tank circumference ; in your case, L = 8,75 m
- ?Si = Si – (S i-1 + S i+1) / 2
- Si : variable settlement calculated from total settlement measured (levelling measurements on site) reduced of uniform settlement (definition as per API 653 - App B)

A 21 cm predictible settlement at the edge (corresponding to the total settlement value) is not a big deal for a 100 m D tank ;
The cheapest recommended solution is the preloading ; lets assume you need about 50.000 m3 earth to build the bund walls around, this is equivalent to 5 m x 1,8 = 9 t/m2 preloading on tank footprint which is 60 % of the total long term load for crude oil ; preloading time is depending on soil nature ; this is an additional geotech study ;
More expensive are stone columns and for really bad soils and they have to be dropped at the right depth to be efficient
It is better and cheaper to improve the soil at foundation construction, tank lifting by any mean is very expensive and the cost of business interruption must be added ....