This action might not be possible to undo. Are you sure you want to continue?

# PRESENTED AT THE SESSION ON ADVANCES IN l)TORAGE TANK DESIGN API, REFINING 43RD MIDYEAR MEETING SHERATON CENTRE TORONTO

, ONTARIO, CANADA TUESDAY, MAY 9,1978

BASIS ~DFSEISMICDESI'GN

P·R9VIS·IONS

FOR WELDED STEEL OIL STORAGE TANKS

**SPONSORED BY SIC ON PRESSURE VESSELS AND TANKS MANUFACTURERS SIC ON TANKS AND VESSELS
**

R.S. Wozniak Chicago Bridge- & lron Company Oak Brook, II11nol8 and W.W. Mitchell Standard 011Company of California san Francisco, Callfol1!1la

CBT-5359

IIT

f

I

L.... n

:j

.-

BASIS OF SEISMIC DESIGNPROVISIONS FOR WELDED STEEL OIL STORAGE ANKS T R. S. WOzniakl and W. W. Mitchell2

r , "

.,

-r

"

q .'

.

'.

ABSTRACT Recommended design provisions are described fGr the seismi,c' design of flat bottom storage tanks which are propos,ed to b~ included in API Standard 650. The basis for establishing design loads is presented including seismic zone coeffi,cients and the essential facilities factor.' The design procedure is based on the approximate method of Professor Housne:r: except that amplification of ground motion is recognized in determining the impulsive response. The derivation of curves is presented which simplify the calcula·tion of the weight:s of the effective masses of tank contants, their centerlS of gravity, and the period of vibration of the sloshirlg mode. The basis of the design lateral force coeffil:ients is given. Resistance to overturning for unanchored tanks is providl:!d by the tank shell and a portion of the tank contents which depends on the width of bottom annular ring which may lift off the foundation. The basis for determining this width is presented. A curve is included in the provisions for calculating the maximum longitudinal compression force in the sh4i!11 for unanchored tanks. The derivation of this curve is pre~:;ented and an approximate formula for the curve given. The fOl~mulas for maximum longitudinal compression force in the sh4i!11 for anchored tanks and required anchorage resistance aze explained. The basis is given for es.tablishing the maximurllallowable shell compression which takes into account the efjEect of int.ernal pressure due to the liquid contents.

I

·1

!

Supplemental information is presented for calculating the he:LI~ht of sloshing of the liquid contents, for designing roof S\l~port columns to resist forces caused by sloshing, and to calc~late the increase in hoop tension in the shell due to seismi<: forces.

. ***

1 ChiCcl'~o Br idge , I ron Company, Oak Brook, I L 2 Standard Oil Company of California, San Francisco,

CA

1

~,::: -~:~:

...

.~ -._" ...... ....

_..

~ ..

INTRODOC: TION

'.

Fteports of damage from major earthquakes within the last fe~r decades cite cases of damage to flat bottom welded steel storage tanks [1, 2, 3, 4]3. Damageto the tanks falls in four general categories: 1. Buc:k:1ingof the bottom of the tank shell due to longitudinal compressive stresses resulting from overturning forc~es. This buckling is most frequently in the torm of an outward bulge in the bottom foot or two of the tank shell extending partly or completely around the tank terlD.edan -elephant's foot bulge.Damageof this type has generally been limited to unanchored tanks ranging between 10 and 100 feet in diameter. Loss of contents has resulted in some of the more severely buckled tanks. Damiilgeo the roof and upper shell of the tank and to t int4!rnal roof support columns due to sloshing of the tank conl::en • ts Damcilge piping and other appurtenances to tan}~due to movement of the tank. connected to a

2.

3.

4.

Damclge resul ting from failure of the supporting ground, notcl):,ly from liquefaction, washout due to broken piping, an~ .slope failure due to high edge loads.

~~hedamage reports have led 1;0 increasing interest in the seismic design of tanks to be located in seismically active axeas , Tank builders, owners, and, in some instances, regulatcIll:'Yagencies have developed their own criteria for seismic design. To provide uniform guidelines, recommended design i='J::'ovisions have been prepared which are proposed to be included as an appendix (Appendix P) in API Standard 650, Welded S:f~eelTanks for Oil Storage. Similar provisions are being de·veloped by the American Water Works Association for water s tc)rage tanks.

I

SCOPE OFD£SIGNPROVISIONS . The proposed Appendix P to API Standard 650 covering seismic clesign of storage tanks is included in Appendix 1 to this papE!r. Detailed requirements are included to assure

3 Refere:n.ces are listed

at the end of the text.

2

..-

- ."~ .....

. .. •••..

'........

_o_."

._ •... ~, .... ', ... :

. ".' ...

_.-

... -. -

.- ..

.' . life safety or to the . ':'_--- -- • ~'. A requirement is included to provide suitable flexibilit~r in piping attached to the shell or bottom of the tank.. it is important that soil conditions at prospective tank sites in seismically active areas be investigated for potential instability including liquefaction during an earthquake. ::::::::-'::'. (J"-. the pr.-- . --- -_.~ stability of the tank shell against overturning and to preclude buckling of the tank shell due to longitudinal coepress Icn for the level of earthquake ground motion which has a :r:easonablelikelihood of not being exceeded during the life c)E the tank in the region in which the tank will be Loca eed .. - . As noted in the Introduction to Appendix P. These latteI' items normally do not pose a risk to. but may be considered o from ~le standpoint of economic risk to the tank itself..Emsion takes this ductility into account. Additi. . Housner [5] and included in Chapter 6 and Appendix F of ERDA TID 70~:4 [6] with modifications as suggested by Professor A.. S. . W. The she Ll. Current practic:e for the seismic design of welded steel tanks for hoop t. -~ --:-. .. _ I . -._.-.sClLEetyf surrounding facilities. DESIGN LOADING The design procedure presented in Appendix P is based on the simplified procedure developed by Professor G.~- .. i'lhen this is done. The response of tanKs to earthquake ground motion also Inc ludes an increase in hoop tension in the shell... have substantial ductility in hoop tension and can absorb energy resulting from earthquake ground motion through yielding.~-.. increa~ied hoop tension should be investigated... Guidelines for this are included later in this paper..f !f. VelEttsos [7]. Guidelines are included later in this paper for the design to accomplish these objectives. ~'lhen tanks are designE!d for higher levels of earthquake ground motion. ..c1cedureconsiders two response modes of the tank and 3 ~ . however. This has led to rupture of the shell in the past for riveted tanks...- -.·A... The proposed Appendix P does not address soil stability since this does not affect the design of the tank. !~.. !Seismic response in hoop tension does not govern the design of the shell for the maximum level of earthquake ground motion proposed for API Standard 650. no provisions are included for hoop tension. Consequently.. However..s of welded tanks..c)nalitems which the tank purchaser may wish to ccns Idar to minimize or avoid overflow and damage to the roof and upper shell and to roof support columns are noted.

866 -* (2a) '''!here the ratio of filling height to radius is greater than 1._ .~ll resu1 ting from these forces is given by the following formula in Section P.333. WI' Housner presents the following formula for tanks where the ratio of filling height to radius is less than I... The forces associated with these modes are normally termed the impulsive force and the convective force. -- .-. --- ...!> (D/H greater than 1. _ . The design overturning momentat the bottom of the sh.._'-_.. _. ri__ its COlrltents: the response of the tank shell and roof togeth4er with a portion of the contents which moves in unison with the shell. N..masses 4)f the tank contents.~.:. to maximum D filling height. . c·· -_.enes .. respec1~ively..•._ - _-. Housner's procedure considers the 1iq.. are glven In Figure P-2 of Appendix P for various ratios of tank diametE~l~...5 times the radius to respond -as a rigid body as far as impulsive forces are concerned.!. - . These curves are based on the formulas developed by Housner and presented in TID 70:2:4. "'4o . and the fundamental sloshing mode of the ccnt..86sfl0. This leads to the formula: (2b) .1: (1) .uid contents in the lower part of the tank below a depth e'gual to 1. _... Cl and C2 are the respective lateral force c~)efficients for the impulsive and convective forces and W141ndW2are the corresponding weights of the effective _. .-_ .... The effective weight 4:lf the uppeLportion of the contents is determined from fOlt'lIlula(2a) using DIH • 1.5 (D/H less than 1. H.3.":. Curves for determining WI and W~as 21 ratio to the total weight of tank contents..333).. _-- _ .__ ...In this formula.-_ .333): tanh O..~re force. :- . The total effective weight ·is determined by adding the full weight of the lower portion of the contents to the effective weight of the upper portion. For the weight of the liquid contributing to the impulsi.

315 (4a) .l.. (D/H less than 1.333). ~ '. as ratios to the maximum filling heignt. these are based on the work of Housner•. which is based on Housne:I:" corrected version of TID 7024.-. _.- --.333): height to radius is greater H x.' ..67 sinh (J. . the formula for the height 5 to the centroid of the impulsive force is: / . applieq to the foundation can be determined by substlt~lltlng Xl and X2 from the following formulas for Xl and X2' res:pectively. is as follows: s (3) The heights Xl and X2 fro.0- il. (D/H greater than 1.67} COSh \i5iH -1.: .' · -C . For tarLks where the ratio of filling height to radius is less ~an 1 •.0 the centroids of tne lateral seismic f~rces applied to WI ClndW2' respectively.. Where the ratio of filling than 1. /" JJ II . .2 = H X 1.: of the . - .( . .c. . The total overturni~91 mC?ment. H- Xl _ 0.09LQ H (4b) The formula for the height to the centroid convect. in formula (1): 5 ~--:... _ 0.3.61) OIH (5) ~. the bottom of the tank shell 1. the overturningr momentcalculated in accordance with formula (1) is that applied to the bottom of the shell.5. The formula for the weight of the effective contents used ee determine the convective force. As noted at the end of paragraph P. Again.I ve force is.. are given in Figure P-3 of Appendix P for various D/H rat~ios..:. H.500- o. ~ ~ ' -.0 'O/H 3.

In reality. . . ..67 sinh ( 3.. The value of 0.. concll~des that the impulsive force can be reasonably well estim'!IIted from the solutions derived for a rigid tank except repla1cing the maximum ground acceleration with the spectral value of the pseudo-acceleration corresponding to the fundamental natural frequency of the tank-fluid system.. . ..0) a (6a) I ~ H - 1. Housner considered the tank to be infinitely rigid so . ~> H . Velet:sos.10 to 0. X: H = 0." ! . :' _...0 +1. . .1.'1 • .25 seconds.937 3. Since the calculation of the fundamental period is complex for tillnks hich do not experience uplift and unknown for w those that do.. .0) exclucUng any soil factor. in this_study of thin wall flexible tanks. .. a constant value is proposed in Appendix P for Cl which represents the maximum amplified ground motion... The soil factor does not appear appropriate for structures with a very low natural period of vibra1~ion. In his work.866* i -] 1.0 - cosh D/H 3.333 l ( 0.67 ) ( D/H (7 ) ( .333: -. " ~ z .: \ r .-.... 6 <.67 D/H )-1. . tanks are not infinitely rigid. Tanks on the ground are inherently rigid.24 is consistent with the Uniform Building Code maximu:1D value for structures other than buildings (It • 2.hemotion of the tank shell and roof together wi th that por t Icn of the contents that moves in unison with the shell coincides with ground motion. The high value of Cl in comparison with buildings ~-7f ~ C :. I ~~ :. ~"':-....375 f. . that t. Storage tanks typically have natural periolds of vibration in the range of 0..t.866 tanh 0.

. of spectral acceleration falls in the region of maximum spectral velocity or disp11Icement•.amping inherent for stora~le tanks. it may be desirable to calculate the fundaalental period and use a lower spectral acceleration value.2 ft/sec2. This c~mes from the formula: T = 21T0 ..578 (10 ) . 3. S.1 to 1.679 (3./( _ ~. For very rigid tanks which are anchored.nonstructural load bearing elemetkts..5 for soft ee medium stiff soils.67) tanh D/H (9) Substituting 9 • 32. The period can be determined from the ell:pression: T • kD~ (8) where k is obtained from Figure P-4 for various DIH ratios. is determined from Table P-2 and varies from 1. k is then: k = Vtanh(~~) 0. depending on soil type.65 feet.The formula presented for C2.~C) - is appropriate because of the low d. The site amplification factor.3 ft/sec and a maximum spectz'a1 displacement of 1. The period of the first sloshing mode is relatively long lind the corresponding value. taking Cl as ·the maximum amplified grouncl motion may be overc:onservative. These amplifications factors correspond to those recommended in the Final Review Draft of 7 . S. The calculation of c~ requires the determination of the naltural period of the flrst sloshing mode and the site amplification factor. For some tanks. and the lack of ductility of the tank shell in longitudinal compression. is based on a maximt~ spectral velocity of 1.0 for rock-like soils to 1. the lack of .5 to 2.

Cl and C2' are appliccible for the areas of highest seisluici ty and for tanks which are not required to be functional for emergency post earthqt:lake operations.-. For oil storage4. Z. The value of the zone coefficient. e .5. Z._.ilhich are required to be functional for emergency operations after an earthquake. o 0 _0·· .Recomm~mdedComprehensive Seismic Design Provisions for Buildings (B} prepared by the Applied Technology Council study ~Project ATC-3) sponsored by the National Science Fqundation and the National Bureau of Standards.our s of approximately equal seismic risk and is cons Ide red to be an improvement over the zone map included in the Uniform Building Code which is based on historic earthquake damage levels. are included in for. corresponds to the Occupancy Importance Factor specified in the Uniform Building Code.05 o 2 1 ~~heessential facilities factor. The values of Z correspond to those .__ .4 0. The relationships between the zones shown on the Claps included in Appendix Pand the contour ranges shown on the ATC-3 maps are as follows: Appendix P Seismic Zone 4 3 ATC-3 Map Contour Ranges OVer 0. The zone maps are based on peak ground motion acceleration contour maps Lnc luded in the Final Review Draft of the ATC-3 Project. The zone coefficient.1 Under 0. For the 48 contiguous states. this should apply to tanks such as those storing 8 ..2 0. in a The lateral force coefficients.1 to 0. The ATC-3 map is for use in establishing ultimate design loads and the acceleration values depicted should be reduced for application in working stress design procedures.05 to 0.defined in Figure P-l. the essenti.4 0. The ATC-3 map depicts ccnt.al facilities factor. ~~heUBCrequires that "structures or buildings which must be: safe and usable for emergency ourposes after an earthquake in order to preserve the health and safety of the general public" be designed for a factor of 1. the map used is that where the accelerations are a measure of effective peak velocity so to be appropriate for the long period convective force as well as the short period impulsive force. I.uula (1) for the! design overturning moment to provide an adjustment for tarliks located in less seismically active areas and for tanks '11. is obtained from Table P-l for the various zones . I.2 to 0..--.-~: .specified in the Uniform Building Code. and.

fuel fc)r power generating facilities emergency postearthquake operations. ... RESIS'I'JWCE TO OVERTUIU~ING which are essential for - The factors which may contribute to resistance of the overtulmin9 momentare noted in para9raph P.. ..~ . _. elemen1~alstrip of the bottom plate perpendicular to the shell '''hich can be lifted off the ground.... ._ o· . .1..... .••_ ...... Pt· . The wei9ht of the COil tents which may be utilized to resist overturning is based on the calculated reaction at the tank shell of an._. deflection and momentdiagram are shown below:: • . .._-.. ' p .4..... ..-"'...• "-:. ..... .____... one at the junction to the shell and p the other at some distance inward from the shell.'.~ p... •. :ow. 9 _'::-" ....... .•. The assumE!dloading.. The calculation is based on small deflection theory and assumes the development of tWCI lastic hinges...-" . p.J ...

.-_. to 6 to 7 ipercent of the tank radius [9)."1lumongitudinal l ccnpreas Icn force.._- t-···· -~'. b.s of uplift occur. w 1.----. i~ • .. w • (12) substituting and W = 62. (13) " .ng table model tests of tanks [lO) show significant chanqas in the response characteristics which.::=:::. and -_ L = 0. ••• -..25 ~:.. 10 ' .._ .---_ _-.---. ~.--. The limitation of ilL to 1. _ -_.- --. at the bottom of the tank shell are \0 \..-'. are not accounted for by current design procedures when greater amount:s of uplift occur.I.Methods for determinin.HO limits L to about 6.1.9\ of the radius.:"-..cltionships: solution leads to the following (11) >.-.. .> ) . and L= " ~ :' ..101_!:. - .-_ The equilibrium rel... The above procedure to establish the maximum resist:ance of the liquid contents to overturning of the tank is conservative since it doe~ not take into account membrane stres:ses which will de-velop in the bottom upon uplift. ••-:. the maxi. L. . Further studies need to be undertaken to better determine the uplift resistance and to account for the changes in response when large amount.0274 (14) Practice has been to limit the uplift length.~.••• --.-- . t '.__ '. SHELL COMPRESSION .GH. ..:. Recent shaki.

~_ • ~~~.--:-- 4._ __ ~_ .. For tanks which experience uplift. _ -_. . ..2... _- - - --_..USN Ci. • . ~. • 0. The curve in Figure P-5 is derived from the following assWl!ed load distribution around the shell of the tank: .C '. b can be determined from the value of the compressive force parameter obtained from Figure P-5 as a function of the overturning moment parameter.273M ~ 02 (15) • which assumes that the force varies directly with ·the dist4lnce from the centerline of the tank in the direction of the lateral loading.. ~. .l and P. ._....~..5.*.--_ __ •• ~ ••• _.4~ ~~""~/. .. For unanchored tanks wherEI there is no uplift and for anchored tanks the compr'ession force can be determined readily from the formula: 1.S._.~--~~~~~~~~~~ ~~~~-3II--r: wL ~~~~---l- • a""'dA. f giverllin Paragraphs P..A/A ! I 11 '--''''1'...

. The relationship may be approximated with 19'ood accuracy up to a value of the moment parameter of 1. the relationship between the two paramElters on the lefthand side of the expressions is obtairled..3. and R the shell radius. -:--~ .... -_·_.:..·-7~·.125 Bt/R where E is the modulus of elalsticity. _.:. _. the critical stress for very thin wall ahellli..·~ ."_.5...'.sin28 sin18-8cos8 M _ .._4·~'_"_".::-_--::-=:: _ . . was established as 0. These were established to provide a safety factor agains't buckling of about 1.6-2-6-2--0..~~>" ._ -- _ _ .::._:.. Osing£..t 1 I~ i>- _ Prom the summation of vertical forces and overturning . this leads to an allowable stress of: (1 . :..'. This J... ..· : __ .5..~~M~-~::-r2171 (18) ~h+~ Formulas for determining the maximum allowable 10ngit:1udinal compussion in the shell are given in paragraph :P.000 psi and a safety factor' of about 1. -'." 12 ':.'_ -.000.formula: t = -O-.: .:::_"::-=. Excluding the effect of intenl41ll pressure. Iron Company.... __ -' ... Miller (ll) of Chicago Bridge' .IS based on the work of C.'.'-'... i . D. the following expressions are obtained: 1.~-::--.. . '.~_ .5.54 11i1i'i th the following..P"='".:.--' . t the shell thickness.-'-86-6-7~F. 29.."'.. This relationship applies for values of the moment parametter from 1f/4 where uplift commencesto 71'/2 where the shell becomes unstable...- ":--:- .. 1T -u- (17) By substituting values of ~ from 0 to 71'radians in these expressions.. moments. sin8-Scos8 71' cosB) (1- (16) I28.'.

------.. Crate and Schwartz [12] determined through theoretical analysis and experimental tests that the cri~ical buckl~~g stress in compression for thin wall cylinders incre~n:Beswith internal pressure. ': .'.400. where 11: in inches and D is the diameter in feet • is Lo.r-..-'-. the allowable longitudinal coepreas Ive stress for thin wall tanks for values of GHo2/t2 greater than 200..--".1028. -:-... .. . .- 80Q._ .-.. . . ~..000.tablished as: 400.6 Et/R.. ..... However..000 was established as: --Eo A._-------_ . · o. This value is reached when the value of GHo2/t2 is about 200.""_ -. --. Thus. only limited tests ltlave been made to date. .200.•.~ _ .'-: .. ~_... rather than by hoop stress... for values of GHI)2/t2 less than . As the thickness of the shell in proportion to the diame11:erof the tank becomes relatively large.OOOt o + 2GHO t (21) Formula (21) will normally apply only to very small tanks where the shell thickness is established by minimum valuel. with sufficient internal pressure the critical buckli~g stress will reach the classical limit of 0. Crate and Schwartz showed a doubling of the critical buckling streS!i as the nondimensional parameter -. -..._. Based en this. _....000 the allowable compressive stress was e!. Theoretically.. "' .000l o (20) I The tests of Lo. · ( ~. _ -0.._. ".. The tests made by Lo. Crate and Schwartz showed a nearly linear increase in critical buckling stress with internal pressllre up to the limits of their tests. _ . formulas (20) 13 . r~rrn~reased from :i~ero(no internal pressure) to a value of 0.000t o (19) .

f Fa • 0.." ff ) (22) ROOF SUE-PORTIUG COLUfolNS ~~en it is desired to design roof supporting columns to resis:t the forces caused by the sloshing of the liquid ~ \..f thickness to diameter ratio. The presence of internal pressure and the radial restrain provided by the bottom leads to a different form of buckling than has been experienced in most experimental work on the buckling of c:ylind.f 14 """. _.ersunder axial and bending loads.istance provided by the weight of the tank shell is not considered. tanks need not be anchored when the requirl!d resistance to overturning can be provided by the tank shell and internal contents without exceeding the maximum value permitted for wJ...eight of the sloshing wave ma. The h. _.y apply only to small diameter tanks (under 15 ft diamete!r) • ( • The longitudinal compressive buckling stress of tank shells is a subject whiCh needs further study. careful attention should be g1ven to the attachment of the anchors to the shell to avoid the possibility of tearing the shell.y be determined from the followi'ng formula based on Housnet'"s corrected version of TID 7024: d = 1.5 Ft is established in Appendix P to maintain an adequate safity factor throughout the intermediate range c. When anchorage is required...124ZIC2 r2 lanl\ (..-.4~GE OF TAUKS Generally. 4· . . ANCHOR.. Miller presents formulas for crltical buckling of shells without internal pressure for intermEtdiate and thick shells leading to a maximum value of criticcLI buckling stress equal to the yield stress. ~ SLOSHING WAVE HEIGHT In some cases it may be desirable to provide freeboard in the 'tank above the maximum filling height to minimize or avoid o'... This limit will normall. _.erflow and damage to the roof and upper shell due to sloshil1l9of the li_g_uidcontents. The specified anchorage resistance given in paragraph P. -:-. 6 fOl~ anchored tanks provides a factor of safety in that the re!.) are no longer applicable. and (2J... The limit c.

PI is zero at the s11rface and maximum at the bottom (Y • B). c~)V ( / t. 750: PI = 2011ZICIG02 [0. - (24) where Y is the distance in feet from the liquid surface to the ~)int under consideration. PI may be determined as follows: Y<0 . / r/:' • contents. As can be seen. Pl may be deterllll. P2 may be determined-from the fo'Jllowing formula: 15 _ .\.ined the following formula: from .333.ueo the convective force. SOOP ~l'ENSION - When it is desired to analyze the tank shell for increC!Lsed hoop tension due to earthquake ground motion.. t For tanks where O/S is great~er than 1.T [ y 1 (25a) (25b) The convective hoop tension. these forces may be determined as described in Appendix 2 of this paper. Wfiere O/H is less 1than 1.333. the increaLsed hoop tension Ps per inch of shell height can be obtailLed from the followlng expression: • (23) where Pl is the tension due to the impulsive force and P2 is that °ld.150 .

.~.'....-. \ ..'~ = ...-.... ~ ... Pe' should be dhrided by a ductility factor of 2.~. are recommended... .\ OJ ~) (26) The increased hoop tension du~ to earthquake-ground motion should be added to the hoop tension due to hydrostatic pressure.' ' .... -: . ...'~:~ / \ ..hoop tension due to earthquake ground motion for use when it is desired to take these factors: into consideration in the seismic design • It has been seen that the design provisions are based on the simplified procedure developed by Housner for rigid tanks except that the maximum ground acceleration is replaced with the spectral value of the pseudo-acceleration corresponding to the fundamental natural frequency of the tankfluid sys cem as suggested by Veletsos. ~ c"J.. further study of these effect!:. CONCLWSION ..68 cosh (3.- .signat normal allowable design tensile stresses. ::-::._ ~.---- .68 H-Y....~ ~.. Provisions are includt~d to insure stability of the tank shell against overtul:'ning and to preclude buckling of the tank shell due to longitudinal compre-ssioni however.. Supplemental informcltion has been presented to determine the sloshing wave height. ~) 16 l .!*'pw.. ___ ': .:. ..._..cosh ( 3.:___..- The basis has been presented for the seismic design prOV1~il:LOnS oil storage tanks which have been proposed as for an appttndix (Appendix P) to API Standard 650. the dE!..1:. the forces on roof supporting columns caused by s Iosh Inq and the increased ."'--·-0'-- "~~:: .4 .40". The formulas have been given for the curves included in the proposed revisic)n to facilitate design calculations..0 for applicatlon in .. The hydrodynamic portion of tbe stress..

. respectively. psi increased hoop tension in tank shell due to impulsive.a~ximum longitudinal shell compression force. lbs/in. psi E • Fa • I~ximum allowable longitudinal compressive stress in 1~ankshe 11. lihellcircumference lateral earthquake coefficients for impulsive and convective forces. in. ft t~ankdiameter.( b • NOMENCLATURE .0 for water) IDaximum filling height of tank. and total earthquake force. ft overturning_moment applied to bottom of tank shell. in. • internal pressure. (sec2If t)~ bottom uplift length.site amplification factor 17 . ft essential facilities factor :~arameter for calculating T.-lbs/in. Ibs/ft of . --9 G B • 4!lccelerationdue to gravity = 32. respectively. s . ft I~dulus of elasticity. Bt I k L M Hp· p R • tank radius.2 ft/sec2 lsp'ecific gravity (1. • d D • - beight of sloshing wave above mean depth. ps i • minimum specified yield strength of bottom annular ring and tank shell. convective. ft-lbs plastic bending moment in bottom annular ring.. ft total height of tank shell.. respectively • • . psi i 't -~.

/' -- -' -- \-- . thickness of bottom annular ring. for computing total overturning moment on foundation. in. ft height from bottom of tank shell to centroids of _~ulsive and convective lateral earthquake forces. lbs total weight of tank contents. lbs/ft of shell circumference any . respectively. '~eight of tank shell. sec a tb· T • w • • unit weight on tank bottom. . respectively. lbs ~ weight of effective masses of tank contents for determining impulsive and convective lateral earthquake forces. for computing H.. 18 '"_ _.__ . in.A s-: I t• thickness of cylindrical shell.. radians. ft • • height from bottom of tank shell to centroids of impulsive and convective lateral earthquake forces.here shell uplift cocunences. ft Y • 'rertical distance from liquid surface to point on shell being analyzed for hoop tension. . ft Z• B • sleismic zone coefficient c:entral angle between axis of tank in the direction of etarthquake ground motion and point on circumference ". lbs/sq ft maximum weight of tank contents which may be utilized to resist shell overturning moment. total weight of-tank roof plus -portion of snow load. if lbs total weight of tank shell.. lbs/ft of shell circumference. sloshing wave period. . lbs r \ Xs - height from bottom of tank shell to center of gravity e)f shell. . When used in formulas for tank design applies to thickness of bottom shell course excluding corrosion allowance.

ASC:E. 8.Bulletin of the Seismological Society of ~~rica. Inc. Applied~·· Tec:hnology Council. 3. E.S. J.isions for Buildings.. Palo AI to. Espinosa and J. No. Volume 47. .s Second Annual E En~Jlneering Mechanlcs Dlvision Special ty Conference. R. G.. C.tainers. 19 . ATC·-3-05. E Tec:ii.· Bulletin of the seismol2?ical societ..· The prince William Sound. 1'. Department 0 Commerce. Clough. 1974: DamageDistribution. NUI:lear Reactors and Earthquakes. June 1911. Alaska. csi.. of -Damage of Storage Tanks. -Earthquake Response of Lic~uid-Storage Tanks. . Earth uake of 151154. I.· Advances in Civil Engineering Th]~OU9hngineerini Mechanics. P •. 7. s. 9. -Experimental Evaluation of Seismic Design Methods for Broad Cylindrical Tanks. 1441-147 .~rGeodetic Survey. n~'I.'· The GZ:'I~at Alaska Earthquake of 1964. 1-24-r May 1977. Iron Company. A. D. Coast Vo ir. 1967• R. Wozniak.· University of CalIfornia Ear·thquake Engineering Research Center Report Number UC9/EERC-77/l0. S. pp. of America.. ·Lateral Seismic Loads on Flat Bottomed Talll]~s. Volume ii. 5. -RecommendedComprehensive Seismic Design Prc>'. 183-195 and 367-390. Veletsos and J.. ·D. October 19 7. ume II-A.REFERENCES 1.Final Review Draft. Lockheed Aircraft Corporation and Holmes & Narver. January 1977. Rinne. pp.·Chicago aridge . w. Housner. Alaska Earthquake of lSIIS4. Proceedin. August 1963. U. Nov'mnber1971. I • 2. Y. February 9. Washington.arthquake Engineering Research Laboratory. Yang. National Ac:ndemy Sciences. California. pp. F.· Engineering Fecltures of the San Fernando Earthquake. C. Engineering. 6. 4. January 1957. ·Dynamic Pressures on Accelerated Fluid Co:n. The Water Tower. Husid. 15-35. -Oil Storage Tanks. May1977. 10. Chapter 6 and Appendix P. D. de las Casas. pp. R. -The Lima Eal~thquake of October 3._. Hanson.-EROA TID 7024. ·Sehavior of Liquid Storage Tanks. Jennings. . 1913. A.

Paper 12823.. ·Suckling of Thinl-Walled Cylinders Under Axial Compression and Intltrnal pressure.o.. . No.. Miller. C.· Journal of the Structural Division.... pp.!!". 51'3. 695-721.." ..• ~ •• ~:".... Marich 1977.- ••.11. J[. Crate and E. Schwartz. ASCI. Proc.· NACA 2021. 1950• TN 12. . VollWlle 103. B. F~· ~ 20 . 8. -Buckling of Axially Compressed Cylinders. • ( \ r ' ~ t . 10. B.

P. No pro'lIisionsare . contents: (1) the relatively high frequency amplified responHe to lateral ground motion of the tank shell and roof togethur with a portion of the liquid contents which moves in unison with the shell.ntents in the fundamE!Rtal sloshing mode..Juireluents herein must be by agreement between purchaser and :Il2L1'lufacturer..-P.i9nprocedure considers two response modes of the tank and it. These requirements represent accept.l. 2 l]~TRODUCTION lb- . ----.l lSCOPE SEISMtC DESIGN OF STORAGE TANKS This clppendix establishes reco~mended minimum b~sic requirements for the design of storage tanks subjec~ed to seismic load as specified by purchaser. . tensio~n. .APPENDIX 1 PROPOSED APPENDIX P TO API STANDARD 650 . and (2) the relatively low frequency amplifj. The design requires the determination of the hydrodynamic mass associated with each mode and the lateral force and overturning fIlomentapplied to the shell resulting from the response of the masses to lateral ground motion. in hoop ..ed response of a portion of the liquid cO.:.. 'Any deviation from the rE!I..tto..l P. HowevEH'. Overturning Moment The overturning moment due to sei~ic forces applied to the bOttom of the shell shall be determined as follows: 21 -~_ . Provisions are included to assure stability of the tank shell against overturning and to preclude buckling of the tank sbell due to longitudinal compression.included regarding the increase. '" . jt is recognized that other procedures and applicable fact~I~S or additional requirements may be specified by the purchaser or jurisd ictional author ities.l DI~SIGN LOADING .ess for the lateral force coefficients soecified herein taking into account generally accepted increased allowable stress and ductility ratios. P.ed practice for application to flat bO.due to seismic forces since this does not affect shell thickn.JI tanks. The de~l.

J . Nl • Weight in pounds of effective mass of tank contents which moves in unison with tank shell. Ws • Xs • Total weight in pounds of tank shell.3.3.2(a). .} . Essential facilities ~actor. I· I. of tank shell to centroid of lateral seismic force applied to WI' determined per para3raph P. • ~ J Where: M • Overturning moment in foot pounds applied to bOttom of tank shell • • • . The tank foundation is subjected to an additional overturning moment due to lateral displacement of the tank contents which may need to be considered in the design of some foundations such as pile supported concrete mats. Cl and C2 • Lateral earthquake forcoecoefficients determined per paragraph P. Xl· Height in feet from botto. Wr • Total weight in pounds of tank roof plus portion of snow load.Safor tanks which must be functional for emergency post earthquake operations and 1. I I Zone coefficient from Figure P~l and Table P-l. as specified by purchaser.2(a). determined per paragraph P.3..2(b)~ Note: The overturning moment deter. determined per paragraph P.0 for all other tanks.3.( "r . Height in feet from botto. r Bt • Total height in feet of tank shell. W2· Weight in pounds of effective mass of first mode sloshing contents of tank.2(b).3. of tank shell to centroid of lateral seismic force applied to w2' determined per paragraph P. f 22 . if any. .3. Height in feet from bottom of tank shell to center of gravity of shell.ined per this paragraph is that applied to the bottom of the shell only.

3. . c.p. 23 . Nuclear Reactors and Earth~luakes. = 1'he heights from the bottom of the tank shell to the centroids of the lateral seismic forces applied to WI and W2' Xl and X2' may be determined by mutiplying H. by the ratios XI/H and X2/H. * Techn:Lcal Information Document 7024.3 a. Narver. Inc. Ef[ective Ma~s of Tank Contents eff\.24. The curves in Figures P-2 and P-3 are based on a modification of the equations presented in ERDA Technical Information Document 7024*. S. obtained from Figure P-3 for the ratio of 0/8.purchaser). WI' W2' Xl and X2 may be determined by other analytical procedures based on the dynamic characteristics of the tank.. Atomic Energy Commi£l:sion. obtained from Figure P-2 fc.August 1963. P.3. L~teral Force CoefficIents The lateral-force coefficient CI shall be taken as 0. prepared by Lockheed Aircraft Corporation. 14aximum filling height of tank in feet from bottom of shell to top of top angle or overflow which limits filling height. The • Where: • TOta' weight in pounds of tank contents (product specific gravity specified by.2 a. u b. Alternatively. for the -U. respectively. and Bo1me~. D = Tank diameter in feet..~ctivcmass lIl' Clnd the effective mass H2' may be determined by mul tiplying '"IT' by the rutioG ''11/'-11' and U2/'tlT' respect ively. .r the ratio D/U.

The lateral force coefficient C2 shall be deter~ined as a function of the natural period of the first mode sloshing. T.ed to resist overturning is r 24 t . T may be determined from the following expression: T • kD~ k• c. D/H.5: f • C2 • 0. The spectrum for C~ should correspond to the spectrum for Cl except modlfied for a damping coefficient of 0. and the soil conditions at the tank site. P.30 S T iihenT is greater than 4.5: C2 Where: • 1. Re'!. the portion of the corltents which may be utiliz. When T is less than 4. For unanchored tanks. T = Natural period in seconds of first mode sloshing.5' of critical.4 Il~CSISTANCE O OVERTUJUoJING T a.24 times the acceleration of gravity.35 S T2 s • Site amplification factor fro~ Table P-2. Cl and C~ may be deter~ined from response spectra establlshed for th~ specific site of the tank and for the dynamic characteristics of the tank.istance the overturning moment at the bottol1l f the to o shnll may be provided by the weight of the tank shell and by' the weight of a portion of the tank contents adjacent to the shell for unanchored tanks or by anchorage of the tarlkshell.b. The spectrum for Cl should be established for a damping coefficient of 2' of critical and scal~~ to a maximum am~ified acceleration of 0. Factor obtained from Figure P-4 for the ratio Alternatively.

d4!pendent on the width of the bottom annular ring which lifts off the foundation and may be d-etermined as follows: • elt. shall not exceed the thickness of the bottom shell course. ThE~ thickness of the bottom annular ring. wtlll!re: • • Maximum weight of tank contents in pounds per foot of shell circumference whicb may be utilized to resist the shell overturning ~oment. tb'.S. the width of .785: b • 25 .- Design specific gravity of contents as specified by purchaser. thle annular ring in feet shall be equal to or greater th.c:ept that wL shall not exceed 1.a. 25GHD. or % inc:h.0274 Gii wL p •5 SI!ELL COMPRESS ION P. b. Where the bottom annular ring is thicker than the remainder of the bottom. G a .n: 0.l • ThE! Unanchored Tanks maximum longitudinal compression force at the bottom of the sbell may be determined as follows: M NhE!J'l 02 (Wt+vL) is equal to or less than 0. • Thickness of bottom annular ring in inches. • Minimum specified yield strength in pounds per square inch of bottom annular ring. whichever is greater.

400. shall not exceed the maximum allowable stress. Whlere: b • o 2 is greater than 0.000: t.5 26 to . Weight of tank shell in pounds per foot of shell circumference. ' -. b. .S. D2 P.'_.785: (WtflfL) M Maximum longitudinal shell compression force in pounds per foot of shell circumference. Fa' determined as follows: l~en the value of GHD2 is greater than 200. P.000: ITt tT 0I '~en the value Ilty• of GaD2 - 800.2 ~nchored Tanks The maximum: longitud inal compression force at the bottom of the shell may 'be determined as follows: b • Wt + 1..273 M .3 Maximum Allowable Shell Compression The maximum longitudinal compressive stress in the shell.( • When b+WL b 'aay be computed from the value of the parameter --~Wt+¥L obtained froID Pigure P-S.000 D t tT D is less than 200.' ..000 t + 2 GHD Itxcept that in no case shall the value of Fa exceed 0.5.

the columns shall be desi:lgned to resist the forces caused by the sloshing of the! liquid contents. • ThiCkness in inches.ilgeof tanks shall be designed to provide a minimum anchor. p •6 • -Maximum allowable longitudinal compressive strels in the Ihell in pounds per square inch.7 P:[PING Provis:lons for suitable flexibility in all piping attached to the sh4!11or bottom of the tank shall be considered.iIIgeresistance in pounds per foot of shell circumference of: 1.Where: t '. a ThE! base of the roof supporting col~s shall be rel!I'~rained to prevent lateral movementduring earthquakes. of the shell in AJ~CBORAGEOF TANKS Anchor. On unanchc)red tanks subject to bottom uplift. Minimum specified yield strength pounds per square inch. 'rile above formulas for '. P. P. Ttut purchaser shall specify any freeboard desirea to airalmize or avoid overflow and damage to the roof and UP5)ier bell due to sloshing of ~e liquid contents. excluding corrosion allowance. b.273 M D2 r If' The stl~esses due to anchor forces in the tank shell at the points of attachment of the anchors shall be investigated. -- a. of the bottom shell course. . piping connected to the bol~tomshall be free to lift with the bottom or shall be locatec! so that the horizontal distance Ileasured from the shell 1~o the edge of the connecting reinforcement shall be the width e)f the bottOm hold down as calculate~ in paragraph P. 4 (b) :plus 12 inches. rty. take into account the effect of internal pressure due to the liquid contents. WhE!ll specified by the purchaser.8 AI~OITIONAL CONSIDERATIONS . 27 .

f • AL.A.. •• . ~-. MA~ ( ... PIClU .KA . •••• MIC ZaN. r . .

2.0 6.2 .. .._. 0 ..2 0 0 H .0 D/H FIGURE P-2 ~ ...-£ 0... 0 1.0 8...... .6 :a" I~ '0.0 D/R nCiORE P-3 - ... Xl . . NI'" :II • ..0 8.s " 0....4 1::>.. 0 0. .4 ID.. ..0 2.0 " 3.0 4.0 3.0 4....O ~ "' (•• 8 ~I= B ~'= k 0 e. .0 :~'" 5. .0 .0 5.. ' Jl..0 7...8 :II .0 7.0 6.. 1.

..---.6 0.. = ..':' .t FIGURE P-4 -""""T"' ..8 1.!-::-:r.-r. 1::::=:-:::+1-----1 _ !. .=T----..0 .-._.. .too.=_:..o:. -:::.' :•.-__ _ =..-:....'.-I-·-'-----.._... ---t-..~ .490 ='f~-r = .l=. .. wt + 1.' . o _ ~~ ·... . .:::.-- --f· 1.273 M ~ u.- 1. -._ =-..__..-------~ -~i~-.._ ._j-_ ..~' ~.___ _ _ -__~ ~ -+--'-1-'-- I---~~.. '1'- ~-.-:.'1'- ~j...---!- ++ .:- .... _ -~ =_. <0....--....t~~....--.!- - --.•0_ 1---1---..--i::=t .. --..__.o... --+=-_. _--....·-."t::::.~._.--t ..-_~--11.=t= When b c: _ ':.'.6 1: 0.._~ -=-~:..+ 1-.. When M =t-- > -to 1---. . --.785. . -: ._.. - r F = _1_ .#-- __ ·-1 ~F -_- ·t· -._.0 1.f ... _..:__ 1..:-:: __.f-: ~"1: 3::== ~ ±----.... -~ -.=r-=: T-- ..::: _ .• .- ~~.5.-t: 'i: . _ T~===r. :"':':':~-"::'1:._..':. ~----...~ 1.::::t=·.. •• _ -1.__f:t.::E ~ . .2 M 30 .1-' .~ ~~ ~ iiL- -. .J. -_-_-~_·t_::··:_-·---_~ I- M rf!(Vt+-L') .::t: .

.Ef soil conditions wbere the soil deptb is less than 20CI feet and the soil types over lyin. including sites where the soil depth exceeds 200 feet and the soil types overlying rock are stable deposit!5 of sands. Such material 1Ilaybe characterized by . either shale-like or crystalline in nature. gravels. or stiff clays. gravels. rock are stable def~sits of sands.reater than 2.0 1. or Stl.2 1. SOIL Pll~()FILEYPE8 is a profile with deep cohesionless or T stiff c::Layconditions. p. TABLE P-2 31 . Z 0.1875 No earthquake 0.ZONE COEFFICIENT Seismic Zone Per Figure P-l 1 2 __!_ . .5 C SOIL PROFILE TYPEA is a profile 1.375 0. SOIL PROFILE COEFFICIENT Soil Profile A S B Type 1. In locations where the soil profile type is not known in sufficilent detail to determine the soil profile type. SOIL PRC. with: Rock of any characteristic. characterized by 30 feet or aore of soft-to-'m~dium-stiff clay with or without intervening layers of sand or other cobensionless soils. SOil Profile C shall be assumed.75 1.I!l shear wave velocity .FILE TYPEC is a profile" with soft-to-lDedium-stiff clays arId sands.500 feet per sec:1)nd.0 design required" for Zone TABLE p-l o. or stiff clays~ 2.

and the acceleration force of an effecl~ive . and acceleration. cosh Tr (0 cos 2fT Tcos 1T 1f T' ( 1) coShfTt u= . u.column of water. H-Y) X " u= 2dgT o. The acceleration force of the ' column and its effective water lIass are functions of the seism:Lc'factor.APPE~~DIX 2 HORIZONTAL FORCES ON COLUMNS CAUSED BY SLOSHING OF FLUID IN CYLINDRICAL TANKS ~he follOwing presentation is considered a reasonable approximation for the determination of seisaic induced loads on co:lumns. The drag and inertial forces are functions of th. acceleration force of the column mass.! fluid velocity. u. ' a !be total horizontal force acting per foot of column length includes the drag force. inertial fo~ce. 4fTdg 0- cosh n{H-Y) 1r sin 2fT Tcos fT T 1f X cOshTrt (2) The average force per foot of column is: -2 ul~1 • u 32 .

Por circular columns. Tbe drag jE~ctor is corrected for rectangular column to account for tJle additional resistance to flow. For simplicity.:... An iteration with time sloshing wave is necessary to search load.on of the column over t~be period of the out tll'e maximum colWln Equation (4) is a function of time and In the tank... a FIG 1 Substituting Equation (1) and (2) into Equation (3) and i~lte9rating yields the following average force per foot of column: . -_ ..• .~ t7 y cos21TT'tcos1T~lcos 21T~ cos7rXI f.a 1J1_ \"'" U __ -. It is recommended that AISCprimary column allowclll)les be used in the beam-column design since secondary 33 . _--=.. COSh _(1T*) ~~I_ .." Equation (3) may be applied to circular and rectalilgular shaped interior columns. the design of the column for combined beam-c:~)lumn action is made assuming the sei_ie load acts unifol~IDlyover the full height of the column rather than the fluid beight.. Dc' j~:s the maximWD dimension of the membercross-section as show~1 Pigure 1. The analysis for rectangular columns is in based on an equivalent circular column with diameter Dc.. 21Trr+21T H W H\ The solution of loca ti.

ft/sec. A value of 1. time from beginning of wave cycle. ft.i . Ib-sec2/ft. fluid particle velocity. lb/ft.2 horizontal distance in direction of earthquake force from center of tank to center of column.6 for wide flange structural shapes. . weight of effective column of water. 'd and'i mc u H • average total force on column. fluid particle acceleration.- columrl allowables have safety factors too low to allow an additi'onal increase for the seismic load.0 is recommended. sec. ft/sec. lb/ft. • drag coefficient.. 34 . • • column weight. "4 ~ • • P ~ • mass of fl~id. CM • mass coefficient. varies I'nw = mw • u x 1fDc· _. • A value of 2. lb/ft. ~ from 0 to T.ner terms I see the Nomenclature following the main body of thE! paper. drag and inertia force on colUmn. For ot.0 is recommended for round columns and 1. lb/ft.. Dc • maximum cross-section dimension of column.4 • . ft. NOME~CLATURE rOR APPENDIX 2 • ~e following defines terms used in Appendix 2 only..