ML20116M423

Rev 0 to Final Rept Point Beach Refueling Water Storage Tank Seismic Capacity

ML20116M423
Person / Time
Site:	Point Beach
Issue date:	06/30/1995
From:	STEVENSON & ASSOCIATES
To:
Shared Package
ML20116M418	List: ML20116M411 ML20116M419 ML20116M423 ML20116M430 ML20116M446 ML20116M477 ML20116M493 ML20116M497 ML20116M572 ML20266H889 ML20266H893 ML20266H897 ML20266H900 ML20266H904 ML20266H908 ML20266H912 ML20266H915 ML20266H919
References
REF-GTECI-A-46, REF-GTECI-SC, TASK-A-46, TASK-OR NUDOCS 9608200150
Download: ML20116M423 (26)

Text

91C2696-DR-004 j

l l

i i

f Point Beach Refueling Water Storage Tank i

Seismic Capacity I

i l

Final Report i

i June 30,1995 s;t?

e Prepared for I

Wisconsin Electric Power Company l

4 i

Prepared by Stevenson & Associates l

10 State Street, Woburn Massachusetts 01801 l

i O

I jeRS"28sa888"ajg6 I

t i

0025A-41

!d i

FIGURE 2.6 i

STEVENSON & ASSOCIATES DOCUMENT APPROVAL ShlET i

I DOCUMENT Poin + Bene 4 KM T So's-a capa e. G

\\

TITLE:

i.

l 00CUMENT. TYPE: Criteria

-Interface Report b Specification O Other O Orawin, O PROJECT NAME:

Posn+ 8ea.J usz A - K, J08 NO.:

9,2 c,24 94 CLIENT:

Niss=*.r/n E/ccNe fo ues-Camp e This document has been prepared in accordance with the S & A Ouality Assurance

_ Manual _ and project requirements Initial issue (Rev. 0):

b 7f A-b W

Oate Prepared by:

//![

Oate f

/

Reviewed by:

Date Y/

Approved by:

REVISION RECORO:

Revision

, Prepared Reviewed

. Approved by/

Description of Revision No.

by/Date by/Date Date i

m /w

>hhi alud,s Fm a km

• wr w3 i

'l O

S&A Project 91C2696 Final Repon Page 1/23 O

4 Point Beach l

Refueling Water Storage Tank Seismic Capacity 1

l Final Report O

June 30,1995 Prepared for Wisconsin Electric Power Company Prepared by Stevenson & Associates 10 State Street Woburn, Massachusetts 01801 O

l

S&A Project 91C2696 Final Report Page 2/23

/

\\s TABLE OF CONTENTS

1. Introduction..............................................................................................................3

2. S u m m ary o f Re s u l ts............................................................................................

.......3 3. E n g i n e e ri n g D ata........................................................................................................... 3

~

4. Ana lys i s C rit e ria...................................................................................... -................

.3 4.1. F a i l ure M od e s.......................................................................................................... 4 4.2. M a te ri a l S tre n gth.................................................................................................... 4 5. C o n fi gu rat i o n................................................................................................................ 5

6. B ase Anchorage Capacity Determination............................................................................. 6 6.1.AnchorBolt................................................................................................................6 I

6.2. Bolt C hair and B ase P late................................................................................................ 6 6.3. Local Tank Shell Bending at Bolt Chair...................................................................... 9

7. S e i s m ic Re s pon se............................................................................................................. I 3 8. Ca pac ity v s. De m and........................................................................................................ 1 3

9. Conc l us io n and D iscuss ions............................................................................................... 15

10. References.......................................................................................................................16 1 1. Appendix - Seism ic Response Analysis......................................................................... 17 1 1.1. S im plifled Dynam ic Model........................................................................................ 17 11.2. Impulsive Response Calculation Procedure................................................................. 22 1 1.3. Vertical Frequency and Demand................................................................................ 23 O

11.4. Sloshing Frequency and Demand...............................................

................23 V

i 1.5. SEWS Form...............................

......................................................................23 i

1 O

i' l

e

l S&A Project 91C2696 Final Repon Page 3/23 O

~

1.

Introduction I

This report summarizes the seismic capacity of the Point Beach Nuclear Unit i Refueling Water Storage Tank (RWST), equipment ID IT13 (1-3]. The results are also valid for the Unit 2 RWST (equipment ID 2T13). He seismic demand is based on the site design basis spectrum as defined in the Point Beach FSAR [8].

j.

A design basis analysis was performed on the tank circa 1970 [4]; however, recent and more accurate analysis techniques rely on a substantially different model of behavior. The RWST is re-assessed based l

on modern techniques. The anafjsis basically follows the Conservative Deterministic Failure Margin (CDFM) approach as defined in EPRI NP-6041 [5] Appendix H.

2.

Summary of Results The analysis concludes that the tank capacity is limited by the anchor bolt yielding capacity. He CDFM capacity of the tank is 0.127g according the conclusions of this study [1]. His capacity exceeds the site Safe Shutdown Earthquake (SSE), which has a Peak Ground Acceleration (PGA) of 0.12g. The median

]

fragility is estimated to be 0.27g.

j 3.

Engineering Data The following data were employed in this analysis. Dese data are considered reasonable and conservative:

i RWST water level is 69 ft, 4

Soil shear wave velocity based on soil property data from Reference [6],

e Anchor bolts are A307 or better material, I

e Tank meets USI A-46 Generic Implementation Procedure caveats for attached piping requirements 4

j

[7] based on Point Beach A-46 SQUG walkdown.

The A-46 seismic evaluation form is attached at the end of the report.

j e

1 1

I 4.

Analysis Criteria The tank was analyzed for a water level of 69 feet, equivalent to about 294,000 gallons of water. This is j

slightly above the overflow level of about 290,000 gallons as reported in the Point Beach Tank Level Book [1].

I The seismic demand is defined by the site design basis ground response spectrum in the FSAR [8,9].

i The tank is analyzed for seismic loading conditions beyond the SSE. Consistent with the CDFM methodology, inelastic material deformation and large-displacement effects are allowed. Because of uncertainty in post-buckling or post-sliding behavior, failure is assumed to occur at first buckling or i

sliding, or the initiating of extensive displacement.

i I

e J

t 4

S&A Projrct 91C2696 Final Report Page 4/23 r~

N 4.1.

Failure Modes The following failure modes are considered:

1. Failure of anchor bolt, i

2.

Failure of bolt chair and connection to tank,

3. Local tank shell bending near the bolt chair,

4. Global shell buckling due to overturning moment,

5. Sliding of the tank,

6. Tank wall failure due to increased hoop stress.

Stoshing of water against the tank dome will also be addressed. Note that piping and failure of nozzle at attachments are not addressed here. These issues were addressed as part of the USI A-4 46 walkdown effort and were found acceptable.

4.2.

MaterialStrength The material strengths are listed in the table below. Ai'owable stresses and loads are based on the guidance in [5].

i item Strength Basis Tank Shell Sm = 20 ksi ASTM A240 Type 304 fy = 30 ksi

[10]

T fba = 2.4 Sm = 48 ksi

'd fu = 75 ksi i

/va = 0.42fu = 31.5 ksi Tank Base & Bolt Chair Sm = 12.6 ksi ASTM A283 Grade C fy = 30 ksi

/ba = 2.4 S = 30 ksi m

I fu = 55 ksi fva = 0.42fu = 23.1 ksi

~

P = 41.7 kips Assume A307 or A36 1.25" Anchor Bolt b

material, allowable load ref. [ll]

Concrete fe'= 3 ksi

[4]

where Sm = allowable stress intensity, fy = nominal yield stress, fba = allowable bending stress, fu = ultimate stress, fva = allowable stress, P = allowable bolt load, b

O andfe' = concrete strength.

O i

6

S&A Project 91C2696 Final Report Page 5/23 5.

Configuration The RWST contains water and is about 70 feet tall and 27 feet in diameter. He tank shell material is ASTM Type 304 stainless steel. It is founded at grade on a large 3.5-foot thick R/C slab and anchored by 271.25" dianieter cast in-place bolts. He slab also supports a 20-foot high R/C pipeway that is roughly 17 feet square in plan. Details of the tank can be found in Refs.12 and 17. A sketch of the tank is shown below.

Shell-Type 304SS 1 Oi Chair-A283 Grade C

]25' typical 4

27 - 1.25" Anchor Bolts in 28 bolt pattern Liquid - water at atmospheric pressure 35* to 95'F 70' 27.0' O.32' near base 0.25*

{

f Mat c

9.

~

3.5 Slab 1.5 4

r --

o I

Soil 1

1~

12*

d 2"

u Z_F 1.25" "0.75*

/ast C

Place 1

The slab is founded on deep layers of soil. Recent geotechnical surveys and analysis have provided soil 4

property data for use in engineering analysis [6]. The best estimate soil shear wave velocity is about 900 ft/sec for the upper layers of the soil profile.

4 Tank Base Weight W, = n(167)2(0.25)(0.283) = 6,200 lbs = 6.2 kips Tank Roof Weight e

e

S&A Project 91C2696 Final Report Page 6/23 O

He thickness of the roof steel is 3/16" and it is stiffened. The weight is estimate based on base weight including an additional 5% for slope and the adddd weight of 14 channels. An additional 10% is included on channel weight for miscellaneous weight.

W, = 1.05W, +1.10(14 x 6.7 x 12.3) = 7,800 lbs = 7.8 kips The roof c.g. is estimated to be H,a 71.0' Tank Shell Weiaht W = 2n(162)(70 x 12)(0.26)(0.283) = 62,912 lbs = 62.9 kips 3

Water Weight W, = x(13.5)2(69)(62.4) = 2,465,000 lbs = 2,465 kips 6.

Base Anchorage Capacity Determination The capacity of the tank anchor chair is determined in terms of equivalent bolt capacities.

6.1.

Anchor Bolt ne anchor bolts are 1.25" diameter cast-in-place bolts. He embedment length is (17]

I, = 48" - 4" - 18" - 3/4" = 25-1/4" l

According to GIP, Appendix C (7), the nominal bolt capacity based on 34 ksi allowable stress is i

P,=41.7 kips i

i The required embedment is 12.5" forf,'= 3.5 ksi concrete. Since I, >> 12.5", no reduction is l

applied for the 3.0 ksi concrete.

i j

He required spacing is 15-3/4". The actual spacing is approximately 36". There is no reduction i

for the spacing effect.

1 The required edge distance is 11", the actual edge distance is very large, greater than 5'.

i Therefore no reduction is applied due to the edge distance effect.

I 6.2.

Bolt Chairand Base Plate e

Following the yield line analysis solution recommended in Ref. [2], the capacity of the bolt chair i

base plate is calculated in (3].

5 O

e a

"4 S&A Project 91C2696

{

Final Report Page 7/?3

$O V.

The A283 steel has well defined yield point, o, = 30 ksi 2

2 M' = o ' tE- = 30 75 = 4.22 k -in / in 0

4 4

e Moment Canacity of 0.2'" Base Plate - Tyne 304SS The 304SS material does not have a well defined yield point; use o, = 30 ksi, o, = 75 ksi, j

a i

Gu 4

i a ' + 2 0,' t-- = 54 25 = 0.844 k-in / in oy i

f 2

O c

y 1

M, = 0.9

(

3 s4 4

l The 0.9 factor is applied to introduce slight conservatism.

Uu

]

Moment Canacity of 0.32" Shell Wall - Tvne 304SS

'o " + 2a "' t 032 2

2 i

M' = 0.9

- =54

= 1382 k-in / in 4

4 i

3 s

s Yield I ine Annivsis b = 6" l

I 4

l R

i c

e=

m I4 mc l 2.69" c

a=

d

=

Re 4.69" e

mc b

c 13 l

c i

Hole Radius R, = 0.75" Point of Contact of Nut c = 2.69"-0.94"= 1.75" d = 3.0"-0.94"= 2.06"

]

1 = 4.69"-2.69"-0.75"= 1.25" 3

t l

a

S&A Project 91C2696 Final Report Page 8/23 fi G

e - R, cos0 2.69 - 0.75 cos0 tan =

=

b 3.0 - 0.75 sin 0

- R, sine 2

I, sin p = e - R, cos0 = 2.69 - 0.75 cos0 m = m, + m, = 0.844 + 1382 = 2.226 k - in / in i

c' = d tan, which is an estimate of the bolt deformation.

Pc c' = m b + 2m tan p(a + 1 )+ 2m l,(tan p sin p + cos p) i c

3 c

= 2226 x 6 + 2 x 422 x tan p + 2 x 422 x I, sin p(tan + cot p)

The bolt load associated with various trial angles e is tabulated in the following:

Trial 0 tanp 14sinp c'

Pc 0

0.65 1.94 1.33 61.3 10 0.68 1.95 1.40 59.2 20 0.72 1.99 1.49 57.0 30 0.78 2.04 1.60 54.9 40 0.84 2.12 1.73 53.0 50 0.91 2.21 1.88 51.4 60 0.98 2.32 2.03 50.2 70 1.06 2.43 2.18 49.3 80 1.13 2.56 2.33 48.7 90 1.20 2.69 2.46 48.5 100 1.25 2.82 2.57 48.5 110 1.28 2.95 2.64 48.8 CDFM Capacity = 0.9 x 48.5 = 43.7 kips The 0.9 factor is applied to introduce slight conservatism.

The lowest equivalent bolt load for the base plate is 43.7 kips which is higher than the anchor bolt capacity. The failure of the anchorage will be governed by the bolt itself.

Punchincr Shear at the base otate The shear stress may be expressed as, Y= * ' '

O

(

i S&A Project 91C2696 l

Final Report Page 9/23 where D,f s the effective diameter. Based on AISC for 1.25" hex nut, the distance across flats is i

1-7/8". D,f s estimated to be 1.5". The allowable stress is assumed to be 0.42F,, or 23.1 ksi i

[10].

l 41.7 V=

= 11.8 < 23.1 ksi l

n(1.5)0.75 Side Plates According to the requirements in [18],j must be greater than 1/2" and 0.04(h-c), andjk must be greater than P/25. For the bolt chairj = 0.75 j>1/2" j > 0.04 (11.25) = 0.45" 2

jk = 0.75(3.71) = 2.78 > 41.7 / 25 = 1.67 in Side plates are checked. Note that the side plates are in tension, so buckling is not a concern.

Bolt Chair Welds The bolt chair to tank wall welding is checked in this section. Using Ref. [18], Equation 7-5 to 7-7, peak line load Won the 1/4" fillet weld is, P

41.7

139 k / in Wy

=

a + 2h 6 + 2(12)

Pc

_ 41.7(2.69) = 1.17 k / m.

W=

- 0.667(12)2 y

2 0.667h W = ]Wy + Wj = 1.82 k /in 2

1.82

= 103 ksi < 30.6 ksi a y = 0.707(025)

The bolt chair welding is adequate.

6.3.

Local Tank Shell Bending at Bolt Chair Nor.linet finite element analyses are performed to assess the capacity of the local tank shell bending nea-the bolt chair [3]. The COSMOS /M finite element program [19] is selected for the analysis. COSMOS /M is capable of performing linear and nonlinear, both elastic-plastic and large displacement, types of analyses. The analyses are run on an IBM 80486-based personal computer. The program is independently verified by S&A [20].

FEM Model i

\\

l l

S&A Project 91C2696 Final Report Page 10/23 A section of the tank near the anchor chair is modeled in (3]. The tank shell, the bolt chair, and the anchor bolt are modeled. To simplify the model, the tank is assumed to be lifted up axisymmetrically. Thus only a slice of tank between anchor bolts is modeled. The behavior of this slice represents the most stressed tension side for a tank under overturning moment.

He height of the tank modeled is equal to the width of the model, or 36.4 inches, which is the distance between anchor bolts. The tank wall is modeled with a 24x24 grid of shell elements.

The FEM mesh is illustrated in the following picture.

_:~~__-

~_-

~--

--H

-\\--

' --b

~

L Q

--~

-m :::$

-m V

---m

.-m

-1 p-

_~_

ah The tank shell and the bolt chair are modeled by the SHELL4T element, which is a 4-noded nonlinear quadrilateral thick shell element, and SHELL3T, which is a 3-noded nonlinear quadrilateral thick shell element.

The anchor bolt is modeled by a nonlinear 3-D truss element, TRUSS 3D.

Sunnorts and Constraints

1. Base of the anchor bolt fully fixed.

2. Base of the tank shell and base of the anchor chair are constrained from horizontal motion due to tank base plate in-plane restraint.

3. The boundary condition at the sides of the model, i.e., at the centerline between anchor bolts, are assumed to be symmetrical. He circumferential displacement, the rotation about the radial O

axis, and the rotation about the vertical axis are constrained.

~

S&A Project 91C2696 Final Report Page 11/23 4.' The rotational constraint at the tank base due to the tank base plate is considered by a series of torsional spring elements.

Material Pronerties COSMOS /N incorporates a bilinear elasto-plastic material for nonlinear elements. The stainless steel is known not to exhibit a pronounced yield point. From the stress-strain diagram for 304SS, the plastic modulus is calculated by

]

E, = (o, -o,)/ e, = (75 -30)/ 020 = 225 ksi limited to a maximum strain of 20 percent. For the modeling of the anchor bolt, half of the recommended E, value (21],450 / 2 = 225 ksi, is used due to lack of consideration of the initial perfect plastic region. 'Ihe base plate is made of low carbon steel, the same plastic modulus as the anchor bolt is assumed.

The plastic modulus values used are summarized in the following table:

Component Yield Stress Ultimate Stress Plastic Modulus Tank Wall 30 ksi 75 ksi 225 ksi Bolt Chair 30 ksi 55 ksi 225 ksi

\\

Anchor bolt 36 ksi 58 ksi 225 ksi The failure will be defined by either exceeding the limiting strain or the ultimate stress.

LoadStenn The loads are applied in 15 steps to 1.5 times the reference load. Load factor 1.0 corresponds to the load capacity of the anchor bolt.

O

S&A Project 91C2696 Final Report Page 12/23 s

Step Load Factor 1

0.25 2

0.5 3

0.6 4

0.7 5

0.8 6

0.9 7-1 8

1.1 9

1.2 10 1.25 i

i1 1.3 12 1.35 13 1.4 14 1.45 1

15 1.5 i

Annivsis Results i

The first nonlinear analysis (PBRWSTNI) failed to converge at load step 11, or load factor 1.3.

Even though the analysis could continue by using smaller load increments or through a I

displacement control process, it is approaching a stage where the system h'as softened p

]Q significantly due mostly to the yielding of the anchor bolt. He peak von Mises stress in the tank i

shell is 37.6 ksi(element 203), and the maximum strain for e, is 0.0055. He maximum strain is far less than the ultimate strain,0.20, defined previously, i

The 1.25 load factor is not sufficient to demonstrate that the capacity of the tank shell is higher than 41.7 kips since a 0.75 factor will have to be applied on the capacity value when a nonlinear 4

ultimate analysis approach is used [2].

Since the purpose of this investigation is to obtain the capacity of the tank shell and the system failure due to the anchor bolt yielding, the capacity of the anchor bolt and the chair is increased by a factor of 4/3. A second nonlinear run (PBRWSTN2) is made to investigate the stress / strain distributions in the tank shell.

The second analysis completed till the load factor reaches 1.5 without divergence. The failure is still due to the anchor bolt. He maximum stress in the tank shell is 37.5 ksi and the maximum strain for c,is 0.0077.

It is confirmed that the tank shell capacity is higher than either the anchor bolt and the chair.

Displacement Histories The displacement histories of the two nonlinear analyses are shown in the following graph.

l S&A Project 91C2696 l

Final Report l

Page 13/23

!\\

i 1'5 NF e.n qJq 47: qn$ y 7

y ww3:3 ' -sy, y y.f.

-.rF

.M ay

.. A -Q pes. ;..

. r

, :.r..,, _

'wv y - -

1

..hr--. J - ~ ~ ~ E ur-l y

3-1

.ev i

n 0.00 0.10 0.20 0.30 O.40 l

Displacement (in) l Note that the deformation is nearly linear up to the bolt capacity. He initial yielding in the tank j

wall does not result in significant system softening.

7.

Seismic Response The seismic response of the tank impulsive mode could be obtained from the SSE ground response j

spectrum. However, the massive foundation and soft soil result in two concerns:

4 l

1.

Shift in fixed base impulsive frequency (F,) due to base flexibility j

2.

Participation of the foundation mass in the response i

l The concern on frequency could be dismissed on the basis of conservatism. The frequency of the fixed.

l base impulsive mode lies to the left of the peak of the ground response spectrum. Adding flexibility will only shift the frequency to a lower spectral demand. To account for the second concern, a simplified Soil-Structural Interaction (SSI) model is developed to estimate the response of the impulsive mode.

He procedure for the dynamic response analysis is included in the Appendix.

8.

Capacity vs. Demand The procedure described in Appendix H of [5] is used to compare the capacity to demand. A computer program has been developed to perform most of the calculations. He program, called "VETSA1" has been verified for project specific use (13,14]. He input and output files are documented in (1]. There

i S&A Proj ct 91C2696 Final Report Page 14/23 i

are some differences between the VETSAl procedure and the referenced procedure. These differences

)

are listed below:

)

i i

Change Comments l

Instead of one bounding case for the vertical earthquake Recommended by Dr. R. P. Kennedy combination, VETSAl checks two cases: vertical in (15].

response upward and vertical recponse downward.

3 The program calculates fluid hold-down using small Recommended by Dr. R. P. Kennedy displacement theory all the way until fully plastic in (15].

moment is reached, rather than linear extrapolation after some point.

Fluid force moment contribution is summed More accurate and easier to numerically, therefore no best fit linear approximation implement.

is needed (i.e. actual fluid force based on above is used)

Among other output, VETSAl prints the moment demand, shear demand, moment capacity, and shear capacity. Since the tank capacity depends on the demand, VETSAl accepts as input a scale factor on input spectral acceleration. This scale factor is adjusted until the demand approximate equals the i

capacity. Internally, VETSAl checks for the global anchorage failure and shell buckling due to overturning moment, and the sliding of the tank with respect to the base mat.

Details of the input data for VETSA1 is available in Ref. [1,13]. The VETSAl analysis results are shown in the table below.

Item Value Comments Scale Factor on applied spectral acceleration 1.1 Capacity = Demand at this value Demand Moment Ms 16,656 kip-ft Applied base moment Demand Shear Vs 517 kips Applied base shear Q = Angle to the neutral axis 113' See (5,13] for definition Allowable compressive line load 3.68 kip /in See (5,13] for definition Maximum fluid hold down line load 0.287 kip /in See (5,13] for definition Total anchor bolt forces 658 kips Allowable moment M, 16.821 kip-ft Allowable shear V, 2,137 kips 1

CDFM Canacity To account for the effect of missing anchor bolt, the missing bolt is assumed to locate at the worst direction, M, = M, - P,1,

= ( + nn(1

- 90*

I8 8 ft l

S&A Project 91C2696 Final Report Page 15/23 The adjusted allowable moment is then Afj = 16,821 - 41.7(18.78)=16,037 kip-ft Scale the input to obtain the CDFM PGA capacity PGA, = (Scale Factor) PGA,,r A// / Afs = 0.127 g where PGAref = reference PGA = 0.12 g.

Check Hoon Stress in the Tank Total pressure is the sum of hydrostatic (P.), impulsive (P ), and the vertical response (P,)

f P = P, + P, + P, From VETSAl output at base P = 29.9 + 1.27 + 4.21 = 35.4 psi The hoop stress near the base, where t = 0.32"

/~N l

f,, = fd = 0.0354(162) 032 = 17.9 ksi < 2.0S, = 40 ksi f

4 It is clear this is not the controlling failure mode.

9.

Conclusion and Discussions From a combination of hand calculation and computer analysis including nonlinear finite element models, it is concluded that the base anchorage capacity is governed by the yielding of the anchor bolt which has a capacity of 41.7 kips. This capacity is based on a conservative assumption that the bolt i

material is A36 steel. If the bolt capacity is higher, then the governing failure may shift from the bolt to 1

the bolt chair or bolt chair / tank wall interface.

A simple soil structural-interaction model has been developed to estimate the seismic demand on the 4

tank considering the lower bound, the best estimate, and the upper bound soil properties. The maximum demand is determined to be 0.202g.

Based on the calculated anchorage capacity and the seismic demand, the tank is analyzed based on the CDFM approach developed in the EPRI-NP6041 report. A computer program VETSA1 is developed to implement this procedure. The analysis concludes the CDFM capacity of the tank is 0.13g (0.127g)

(3 according the conclusions of this study. This capacity exceeds the site Safe Shutdown Earthquake (SSE),

> Q which has a Peak Ground Acceleration (PGA) of 0.12g.

O e

d

)

S&A Projrct 91C2696 Final Report i

Page 16/23 i O The median fragility of the tank can be estimated from the CDFM. Due to the conservatism built-in the CDFM approach, the CDFM is assumed to equal HCLPFu. The ratio between HCLPFu and HCLPF o is 5

typically exponential of 0.29, which is the peak and valley factor for the Eastern US sites.

[

Median Fragility

= HCLPFa

""8 *'d e

3

-_ ' HCLPFu',uyp,.pa

(

, o 29

'CDFM' up, a

e

\\ 134 >

For p, = 0.46, Median Fragility = 2.15

• CDFM = 0.27g.

10.

References 1

(

1. Stevenson & Associates, " Point Beach RWST Seismic Capacity," S&A 92C2696-C003, Rev. O, June 1994.

4

2. Structural Mechanics Consulting, Inc.," Point Beach RWST Seismic Analysis,",. Letter report to S&A, O

March 22,1994.

i l

3. Stevenson & Associates," Point Beach RWST Tank Shell Local Bending Nonlinear Analysis," S&A 92C2696-C004, Rev. O, May 1994.

l

4. Bechtel Corp.," Seismic Analysis of Refueling Water Storage Tank," June 1970, Job 6118.

i i

5. Reed, J. W. et al., "A Methodology for Assessment of Nuclear Power Plant Seismic Margin," Revision l

1, EPRI-NP-6041-SL, August 1991.

l

6. GEI Consultants, " Point Beach Nuclear Plant IPEEE" Draft Report, Project 93109, March 30,1994.

S&A 92C2696-DC-056a.

j

7. SQUG," Generic Implementation Procedure (GIP) for Seismic Verification of Nuclear Plant Equipment," Revision 2, Corrected February 14,1992.

8. Foint Beach Nuclear Plant," Final Safety Analysis Report," Unit Nos. I and 2.

f

9. Point Beach Response Spectra Summary, NPM 93-0547, S. R. S'. Amour, St

• ember 3,1993.

10. ASME Boiler and Pressure Vessel Code,Section III, Division 1,1980.

11. URS/ John A. Blume & Associates," Seismic Verification of Nuclear Plant Equipment Anchorage -

j Volume 4: Guidelines on Tanks and Heat Exchangers," Rev.1, EPRI, June 1991 EPRI NP-5228-SL.

(

l Y

S&A Project 91C2696 Final Report Page 17/23

12. Wisconsin Electric, Graver Drawing L-24015,"lT13 RWST Assembly & Detail," Rev,5, 12/18/1992.

13. S&A Calculation 92C2746-C011," Verification of Utility Program VETSA for Seismic Analysis of Vertical Tanks," October 1993.

14. S&A Calculatiod 92C2746-C009," Median Seismic Capacity of the CY Condensate Storage Tank,"

October 1993.

15. Reed, J. W. and Kennedy, R. P.," Add-on Seismic IPE Training Course," sponsored by EPRI, August 1993, Boston, MA.

16. American Society of Civil Engineers," Seismic Analysis of Safety-Related Nuclear Structures and Commentary," September 1986, ASCE 4-86,

17. Bechtel Job 6118, Drawing C-142, " Auxiliary Building South Wing & Facade," Rev.18, 12/15/1992. Drawing C-149, " Auxiliary Building - South Wing Reinforcing," Rev.11,12/15/1992.

18. American Iron and Steel Institute, " Steel Plate Engineering Data Volume 2 - Useful Information on Design of Plate Structures," December 1992.

19 SRAC," COSMOS /M User Guide", Version 1.61, August 1990.

20. S&A," COSMOS /M Verification," Rev.1,1994.

21. Joint Committee of WRC and ASCE," Plastic Design in Steel, A Guide and Commentary," 1971.

4 11.

Appendix - Seismic Response Analysis 11.1. SimpilRed Dynamic Model The simplified model schematic is shown in the following picture.

O

S&A Projtet 91C2696 l

Final Report Page 18/23 Y

h NE L

o I

El i

a u

p

..y

... e, gH l

M M

I n e Ky g

a I

We

= Effective impulsive total weight include impulsive weight and tank steel Hg

= Effective height for Wg, includes effect of variable base pressure

= Total foundation weight and rotary inertia l

Fr If j

tg

= Foundation slab thickness = 3.5' l

EI

= Equivalent beam properties to produce fixed base impulsive frequency F if f

slab is restrained Ku Ky = Soil springs is

= Distance between soil spring, set to provide proper rotational stiffness of slab The model is designed toward estimating the equivalent peak acceleration response of the tank impulsive mass. This value will be used in lieu of the spectral acceleration value in the referenced procedure. Calculation of the model parameters follow the schematic. Three separate models were created as indicated below.

ModelID Description 2T13LB Lower bound soil shear modulus, upper bound impulsive mode frequency estimate 2T13BE Best estimate soil shear modulus, best estimate impulsive mode frequency estimate 2T13UB Upper bound soil shear modulus, upper bound impulsive mode frequency estimate Based on judgment, the lower bound soil / upper bound F, was considered to be the worst combination for mass participation of the foundation (rigid soil modes are higher than F,). The upper bound soil / upper bound F,is the worst case for RS spectral acceleration. A best estimate case was also run. A 5% damping value is used in all cases [5].

Horizontal Imnulsive Mode (F,1 -Fixed Base t=

+&

-4

+

S&A Project 91C2696 Final Report Page 19/23

~

The frequency F,is interpolated from Table 2.2 of[1l]. For H/R = 69 /13.5 = 5.1, t/R = 0.25 /

162 = 0.0015, and R = 13.5, F = 2.61 Hz f

Adjust for stainless steel vs. carbon steel, F, = 2.61 '

= 2.51 Hz Y30000 l

Accuracy on fixed-base frequency estimated to be t 15% [5], so the upper bound on F,is 2.9 Hz, lower bound is 2.2 Hz.

a imnulsive Weight W,and Height X, Based on Ref. [5], equations H 5 and H-6, W, = W,(1.0- 0.436R / H) = 2254 kips X = N(0.50-0.188R/ H)= 31.96 ft Effective Imnulsive Weight (Wy Ob Assume tank steel acts with impulsive water weight [5]

W, = W, + W, + W, + W, = 2331 kips Base Moment Due to Imnulsive Pressure (My Based on [5], page H-10, the peak impulsive pressure on the base is, ksf P, = 136RH' = 0.824Sn For linear pressure distribution over circular area, use beam stress analogy to calculate the base

moment, M,, =

P, = 1592S, k - ft 4

Effective Imnulsive Weight Height (Hy The composite c.g. ofincluded weights is, O

S&A Project 91C2696 Final Report 1

Page 20/23

,X, +,(0)+ W,(71)+ W (35) y H,, =

= 32.09 ft Ws including the moment due to impulsive base pressure My the effective height is, H = (H,,W, + M,)/ W, = 32.8 f1 i

Foundation Weight and Rotarv inertia (WM)

Including the 3.5 ft slab and pipeway in rigid foundation mass, the pipeway fixed-base frequency is approximately 13 Hz [4], well above impulsive mode frequency. Therefore, the pipeway mass is assumed to be rigid.

Based nn Ref. [4), pipeway mass Fr = 1069 kips, and slab Ws= 850 kips, Wp - W + Ws = 1919 kips f

For rotation about the base, an effective circular slab radius Rg of 20.5 ft, pipeway height Hp of 40 ft, and 25 ft between tank centerline and pipeway centerline L, of 25 ft are assumed.

I = 1,Ry' +1,H,2, y,7

= 900,000 k-f1 2 W

W f

o 4

12 The above values apply for rocking about axis normal to line connecting the tank and the pipeway centerlines. It is judged to be the worst case.

Eauivalent Beam Pronetties (EI)

Assume the fixed-base impulsive tank mode responded with weightless cantilever beam 1

' 3EIg 5N XlW, therefore, El = (2. 1.)* XlF ', I 3g For the three soil cases, the values are Case FI El Lower Bound 2.2 Hz 1.627 x 10' kip-ft' Best Estimate 2.5 Hz 2.101 x 10* kip-ft' Upper Bound 2.9 Hz 2.827 x 10' kip-ft' Soil Soring (K @ and la y

S&A Project 91C2696 Final Report Page 21/23 O

The foundation is 3.5 ft thick with an irregular shape. Based onjudgment, an equivalent circular footing with effective radius Rr is used to' represent footing stiffness. Rr = 20.5 ft is scaled from the design drawing [1].

Based on Ref. [16], Table 3300-1 A,y = 32(1-v)GR, 7-8v 14GRy Ky = 2 1 -v (two springs)

Rocking stiffness K* = 8GRf 3(1-v)

For this model K~ = Kyl' l 2 O

I, = )2K, l Ky The effective shear modulus is calculated as G = F,a s'Ps V

where F,a is the reduction factor due to strong motion strain, V, is the soil shear wave velocity, and p, is the soil density. 'Ihe reduction factor F,a f 0.7,0.85, and 1.0 are used for lower o

bound, best, and upper bound estimates respectively. The calculated properties are summarized in the following table.

OG

4 S&A Project 91C2696 l

Final Report i

Page 22/23 iO

,U Parameter Lower Bound Best Estimate Upper Bound

]

V, 650 ft/sec 900 1,350 p,

0.13 kip /ft' O.13 0.13 i

v 0.49 0.49 0.49 i

G,,,,,,,a 1,706 ksf 3,270 7,358 s

0.7 0.85 1.0 F,a G

1,194 ksf 2,780 7,358 Ky 1.297E+5 kip /ft 3.019E+5 7.992E+5 j

Kr 9.599E+4 kip /ft 2.235E+5 5.915E+5 l

K, 5.373E+7 kip-ft/ rad 1.251E+8 3.31IE+8 l

1 33.46 ft 33.46 33.46 3

1 11.2. Impulsive Response Csiculation Procedure I-The COSMOS /M VI.61 finite element program (19,20] was used to calculate the natural

)

frequencies and mode shapes of the three models. He models were then subjected to the SSE earthquake input using S&A's EDASP for Windows (EDASPw) program's PSD analysis capability. He peak response of the impulsive mass was obtained for each of the three models.

}

ne procedure is described below.

1 STEP DESCRIPTION COMMENTS 1

Generate modal model of system COSMOS /M 1.61 finite element

{

package is used to calculate the natural i

frequencies and mode shapes i

2 Translate modal data to EDASPw Utility program TRANSLAT3 used to format translate COSMOS /M data 3

Generate PSD from site SSE ground RS Use S&A's SPEUIKA program to j

calculate the PSD l

4 Calculate response of impulsive mass USE S&A's EDASPw program to calculate response to base excitation l

Details of the analysis can be found in [1]. The peak acceleration of the impulsive mass results are summarized in the following table Model Peak Acceleration (g)

Comments l

Lower Bound Soil 0.201 ff = 2.23 Hz l

Upper Bound Ff l

Best Estimate Soil 0.201 ff = 2.36 Hz Best Estimate Ff l

Upper Bound Soil 0.202 ff = 2.69 Hz i

Upper Bound Er Fixed-Base 0.187 Spectralacceleration at 2.9 Hz j

Upper Bound Ff Eventually, a demand of 0232g is used which is the envelope of all models.

i i

1 I

i

S&A Project 91C2696 Final Report Page 23/23 i

i 11.3. VerticalFrequency and Demand From (16), Equation 3500-13 4

39 eg 3

+-

= 4.5 Hz f' = 4H _p tE Ki_

s where K= bulk modulus of the fluid.

1 1

'62.4 '

27 1

+

= 43 Hz f' = 4(69) j 32.2 s0.0216(4 x 10') 4.5 x 10's The vertical SSE spectral acceleration equals 2/3 of the horizontal. For 5% damping curve, the 2

peak below 4.5 Hz is 1

4 S,=2 0.20 = 013 g 3

Because the participation of the slab may increase demand, include an increase of 25%,

]r3 S, = 0.16 g

- ()

}

11.4. Sloshing hnquency and Demand i

From Table 2.1 of[l1], for H/R = 5.0, R = 13.5' F,= 0.34 Hz The SSE spectral acceleration for 0.5% damping is Su = 0.065 g i

The sloshing height from [5] is 1

h, = 0.837RS, / g = 8.8" Since the freeboard is greater than 12", the allowable PGA is scaled to 0.16g.

i

}

11.5. SEWS Form V

"---m---

l l

e

\\

I 4

e l

J l

d 4

n 4

4 f

CST 1

4 4

1 i

t 4

i f

i A

i h

i i

I J

J f

i

?

J, i

4 b