ML20083K080

Multiple Subsequent Safety/Relief Valve Actuation Evaluation for Mark 1 Containment,Rev 1, Dtd Nov 1978.Tests of Torus Support Sys Were Run & Results Confirm Conclusions of Original 780727 Rept Re Safety

ML20083K080
Person / Time
Site:	Hatch
Issue date:	11/30/1978
From:	GEORGIA POWER CO.
To:
Shared Package
ML20083K077	List: ML20083K074 ML20083K080
References
TAC-08194, TAC-08195, TAC-8194, TAC-8195, NUDOCS 7901030219
Download: ML20083K080 (24)

Text

'

o' ,

J MULTIPLE SUBSEQUENT SAFETY / RELIEF VALVE ACTUATION EVALUATION FOR t

E. I. HATCH NUCLEAR PLANT UNIT 2 MARK I CONTAINMENT DOCKET NO. 50-366 PREPARED e .

FOR ,

t GEORGIA POWER-COMPANY 1[!

$00THERNCOMPANYSERVICES,INC.

5 3:

l .

i

~

I 790103C) W l, I -

.. l

' REVISION 1 nmic,.inen 1978 l

s . . .

O ..

e TABLE OF CONTENTS Page 1.0 INTRODUCTION 1-1 2.0 EVALUATION PROCEDURE , 2-1  ;

2.1 Consecutive Actuation Factors 2-2 2.2 SRV Setpoint Pressure to Reactor Pressure Multiplier 2-2 ,

2.3 Shell Stress Multiplier 2- 2 -

2.4 Structural Response Attenuation 2-3 2.5 Acceptance criteria 2-3 3.0 CALCULATIONS 3-1 3.1, Four Valve Case 3-1 3.1.1 Torus Shell 3-1 3.1.2 Columns 3-2

. 3.1.3 Column-to-Torus connection 3-4 3.2 Eleven Valve Case 3- 6 3.2.1 Torus Shell ,

3-6 3.2.2 Columns -

, 3-7 3.2.3 Column-to-Torus connection' 3-8 4.0 RESULTS AND CONCLUSIONS 4-1 -

5.0 REFERE!jCES 5-1 4

O O -

9-e e i 7

)'. ._ i y c

1.0 INTRODUCTION

In July 1978, the torus and support system for the Hatch Nuclear Plant Unit 2 were 'evalu'ated for effects of a multiple subsequent Safety-Relief valve (SRV) actuation due to isolation event (Refer- ,

)

ence 1) The evaluation was performed by extrapolating data .

1 from an SRV discharge test performed at the Peach Bottom Unit 2 plant. Subsequent to the evaluation, an SRV discharge test was performed at the Hatch Unit 2 plant in October, 1978. Thus the present report represents a" revision of Reference 1, wherein the evaluation of the torus for multiple subsequent SRV actuation has been repeated, based on the plant unique data taken in the Hatch Unit 2 SRV test. .

6 In the evaluation, the Hatch Unit 2 test data has been extra-polated to account fer differences between test and setpoint pressure, for the higher loading which would occur in the sub-scquent actuation condition, and for the superposition of mul-

~'

Eiple valves actuating simultaneously. Two cases have been considered in the evaluation: Subsequent simultaneous actua-tion of the -four lowest setpoint valves, and subsequent simul-taneous actuation of all eleven valves (see Figure 1). The extrapolation of data from the test conditions to the two cases evaluated has been carried out using,' procedures specified by General Electric in Reference 3. .

~

The evaluation included general membrane s'tresscs in the tores ~

~

shell, torus column loads, and stresses in the column.to torus welds. The calculated stresses were compared with acceptance criteria for the Mark I Short Term Program (STP). The evalua-tion procedure and results are described herein. .

O e

s e

' e 1

' ~ .- .

o .

0 .

M (1100 PSIC)

.[' *

(1110 PSIC)I' I - -

• 7

. G (1090 PSIC)

~

C (1090 PSIC) g g[* I (1100 PSIG ..

~ '

D (1100 PS10)

, y -

(1090 PSIC) ,

al (1110 PSIC) 3., '

(1110 PSIC , ,

B

- - (1090 PSIG)

A (1100 PSIC)

. ( ,

a - -

. 160 .

~

)( - = 1.Ok'EST SET POINT ~

- VALVES

, FIGURC,1: SArtTY f(El.1017 Vid VC l'A: !.1' RAD 1.0 CAT 10:iS 5::ti SCT l'0!!!TS FOR llATCil U:llT 2 ,

9 ,

~ '

A l

o, f O.

7. . ,

,:2.0 EVALUATION PROCEDURE Based on the information available from the . ongoing Mark I pro-gram, the !!atch 2 torus support system would experience the ,

highest loads following the actuation of the safety relief valve D. Thus, the evaluation was based on the preliminary l- information measured for a single actuation of , safety relief .

valve D in the cold condition. The measured data on maximum shell stress and outside column load formed the baseline for the evaluation. The approach used in the evaluation is con- Y sistent with Reference 3. ,

The baseline data was extrapolated to predict response for the multiple valve second actuation cases. Consecutive valve actua-tion factors obtained from Reierence 3 were applied to account for the " Hot Pop" effect. Since the Hatch Unit'2'SRV test data was obtained at' reactor pressures slightly lower than the SRV set-point pressures, a' multiplier was applied to account for the "pressu're difference" effect.

The procedures in Reference 3 were used to determine the effect of simultaneous actuations of multiple valves. Attenuation curves were applied to determine shell stress and column load in the test bay based on position of each valve in the multiple ' actuation. The attenuation curves were derived from results of testing at l Monticello(Reference 3) and Peach Bottom (Reference 2), which in-l- cluded actuations of various salves outside the test bay. Peach .

l Bottom was selected as a reference since.it has eleven (11) SRV's

! in 4-4-3 setpoint grouping, SRV ramsheads located on torus ring gir'ders, and full saddle supports for the torus, all similar to I Hatch 2.

l The evaluation was performed for general membrane stresses in the l torus shell, column loads, and stresses in the column-to-torus weld connections. The results were c'ompered against acceptance l criteria for the Mark I Short Term Pr'ogram. The procedure is described below.

2-1

'w

1 G

l 2.1 Consecutive actuation Factors .

Consecutive actuation factors of 2.37 for torus shell stress and 1.96 for column load, as recommended by General Electric in Reference 3, were applied;to account for the higher discharge loading in the hot condition.

2.2 SRV Setpoint Pressure to Reactor Pressure Multiolier The test data for the Hatch Unit 2 was obtained for a cold actuation of valve D, with reactor pressure at approximately 973 psig. The SRV pressure setpoints are at 1090-1110 psig.

Therefore, the data must be corrected ~to setpoint pressure.

The reactor pressure to setpoint pressure multipliers were calculated by applying the analytical methods described in Reference 5. The multipliers were found to-be 1.09 for the four valve case (extrapolation to lowest setpoint pressure) .

and 1.10 for the eleven valve case (extrapolation to highest setpoint pressurs). ,

2.3 Shell Stress Multiplier .

In the Hatch Unit 2, all shell stresses were measured using

, , strain gages on the outside of.the torus shell. The measured

~

. stresses thus reflect membrane an.d bending components, while the shell structural integrity is7 judged relating to the membrane component. In accordance with the General Electric'recommenda-tion i*n Reference 3, the measu ed shell stresses were divided

~

by a factor of 1.17 to obtain the shell membrane stresses.

i .

l. '

l

4 i

9; 9 , .

. . I 2.4 Structural Response Attenuation

, Figure 2 and 3.show the^ attenuation curves for,the torus-shell stress and column load. The shell stress attenuation curve in' Figure 2 was obtained from the results of Peach Bottom SRV-discharge test..(Reference 2) . The column load attenuation curve in Figure 3 was given by General Electric in Reference 3.

The Peach Bottom data is indicated on Figure 3,for comparison.

Multipliers were taken from the curves for each valve-included in the combination, based on location in the torus.

2.5 Acceptance criteria Dend weight, hydrostatic pressure and seismic effects were added to shell stresses and column loads calculated as above. In

. accordance with the NRC requirements for the evaluation- ,

~

(Reference 6), the results were compared.with the STP acceptance criteria. The STP criteria'were defined in_ terms of-strength ratio. The strength ratio for an element is'the

-stress (or strain ~) in the element for a given applied load divided' by the lower bound of the value ~ of stress (or strain) for which one would predict failure of the element. The STP criteria ~, described in Reference 7,' require that a strength -

ratio of 0.5 may'not-be exceeded for a'ny_ component.- -

e O 6 6 9

. Y e

4 9

I i

. ..m . . .. ;

s .

1.0 i *

. 0.9 -

e 0.8 - -

0.7 -

h 0.6 -

d e

U 0.5 -

d u y .

E " 0.4 -

8 s 5 .

C 0.3 -

Q -

, 0.2 -

0.1 -

0 I I I I I I I I 1/4 3/4 1 2h 34 4 Sk 6h 7

, BAYS FROM RAMSilEAD CENTERLINE FIGURE 2 TORUS SHELL STRESS ATTENUATION RING GIRDER DISCHARGES (FROM PEACH BOTTOM TEST DATA IN REFERENCE 2')

-.Y- . _ _ _ _ - _ - - - - -

+ g .

e .

~

e O

S 1.0

. O MONTICELLO MIDBAY

'h HATCH I MIDBAY h HATCH I RING GIRDER - .

0.8 -

8 PEACH BOTTOM RING GIRDER ,

h ~

O

- < .70 . -

S.

O

$ 0.6 - ,

s2 3 -

o .

O O '

!", 0.4 -

s<

2 a: .

O Z

0.2 -

.15 .15 g o .12 ,

g h 4~~ , ? _ _ _*

G i i i i i i i 3 4 5 6 7 8 0 1 2 BAYS FROM RAMSHEAD CENTERI.INE FIGURE 3 TORUS COLUMN LOAD ATTENUATION (FROM GENERAL ELECTRIC REFERENCE 3 REPORT)

~ ..

e

'3 0

. CALCULATIONS The actual calculatfions-to determine strength ratios for the torus shell, columns, and column-to-torus connections are given below.

Calculations are presented for a four valve case and an eleven .

valve case. ,

3.1 Four Valve Case In the first case, the four lowest setpoint valves were assumed to undergo subsequent simultaneous actuation. Referring to Figure 1, the lowest setpoint valves are'G, C, B and F. 'The evaluation follows.

3.1.1 Torus Shell The highest measured shell stress for single valve cold actuation occurred at torus quarter span, along the. axis of the ramshead dis-Based on preliminary, data, this measured value was esti-charge.

mated to be 7.7 ksi (Reference 9). The calculations for multiple second actuation follow:

Maximum shell membrane stress for single valve. cold actuation j = 7.7/1.17 = 6.6~ksi

.. Factor for multiple ac,tuation:

e 4

O 9,

'. 3-1

9 ,

C" From Fiqure 2 Multiplier for' Valve C = 1.0 Multiplier for Valve G = 0.64 Mult,iplier for Valve B = 0.11 ,

Multiplier for Valve F = 0.10 Net multiplier

= 1.2 (1.0)2 + (0.64)2 ^+ (0.11) + (0.10) = 1.44 Consecutive actuation factor = 2.37 SRV setpoint pressure to reactor pressure multiplier = 1.09 Resultant shell stress due to SRV discharge

= (1.44)(2.37)(1.'09)(6.6) = 24.6 ksi Shell stress due to dead weight, hydrostatic pressure and earthquake was taken from Reference 8, = 1. 4 ksi .

Total shell stess = 24.6 + 1.4 = 26.0 ksi The yield stress for the SA 516, Grade 70 torus material is 38 ksi. Since the calculated stress is within yield and the strength ratio is about 0.34, it. satisfies the Short Term Program Acceptance .

Criteria. .

3.1.2 columns.

The maximum column load measured a t llatch tini.L 2 for a ningic valvo col,d actuation was 84 kips (Reference 9). This value was measured near the top of the outside column. In'a torus with saddle supports, column load is_ greater at the bottom than at ,the top, due to load transfer by shcar from the web O

e

. . .l l of the saddic. In the Peach Bottom SRV tou t (Reference 2),

~

load measured at the top of the column was approximately 80% of that at the bottom. This ratio was applied for Hatch

~

Unit 2,,since the saddle design is similar to Peach Bottom, ,

resulting in a baseline outside column load of 84/.80 = 105 kips.'

Inside column loads were conservatively taken equal to the outside column loads. The calculations for multiple valve second actuation follow:

4 Factor for multiple actuation:

From Figure 3, reading abscissa for ring girder discharge ,

Multiplier for Valve C = 1.0 ,

Multiplier for' Valve G = 0.70

~

Multiplier for Valve B = 0.10 Multiplier for Valve F = 0.08 Net multiplier = 1.0 + 0.70 + 0.10 + 0.08 = 1.88

. Consecutive actuation factor u 1.96 .

~

SRV setpoint pressure to reactor pressure multiplier = 1.09 Resultant compressive column load Outside column = (1.88)(1.96)(1.09)(105) = 422 kips Inside colum

= (1.88)(1.96)(1.09)(105) = 422 kips

- Column loads due to dead weight, hydrostatic pressure and earthquake were taken from Reference 8. .

Outside column = 92 kips ,

Inside column

= 74 kips

6_

~

g s

Total column load .

Outside column = 422 +-92 = 514 kips Insi e column = 422 + 74 = 496' kips To calculate strength ratio for the columns, capacity was taken as 1.6 times CRC static capacity (Reference 7).

On this basis

. Outside column SR =

3 20

~ '

496 = 0. 25 < 0. 5 Inside column SR =

g 3.1.3 Column-to-Torus connection The maximum load measured near the top of the outsido column, for a single valve cold actuation was 84 kips (Reference 9).

The load on the inside column was conservatively assumed equal to that on the outside column. Calculations for multiple valve second actuation fo11ow: ,

Factor for multiple actuation: .

From Figure 3, reading absicissa for ring-girder discharge Multiplier ,for Valve C = 1.0 Multiplier for Valve G = 0.70 - .

Multiplier for Valve B = 0.10 Multiplier for Valve F = 0.08 Net multiplier = 1.0 + 0.70 + 0.10 + 0.08 = 1.88 Consecutive actuation factor = 1.96 e

(St .

CP SRV'setpoint pressure to reactor pressure multiplier = 1.09 Resultant column-to-toru's connection force:

Outside column connection .

= (1.88)(1.9C)(1.09)(84) = 337 kips Inside column connection

= (1.88)(1.96)(1.09)(84) = 337 kips Column-to-torus connection forces due to dead weight , hydrostatic pressure and earthquake were taken from Reference 8.

Outside column connection = 74 kips Inside column connection = 58 kips Total column-to-torus connection loads:

Outside column connection = 337 + 74 = 411 kips -

Inside column connection = 337 + 58 = 395 kips The capacity of the connection was calculated in accordance with Reference 7, Appendix A. On this' basis:

. Strength ratios are:

Outside column connection 2810' 411 - 0.15 < 0.5 Inside column connection = y_3_9 5 = - 0 .19 <0.5

\

. ,e D% I

,. ' 3*2 El,9ven V_ h Case h.'

l In the second case, all cloven valve.. were assumed to undergo subsequent sinulta6cous, actuation. The ovaluation follows.

3 . ?, .1 orus Shell -

As described above, the highest measured'shell s' tress was 7.7 ksi at quarter span. The procedure is as before.

~

-Maximum shell membrane stress for single valve cold acteation

= 7,7/1.17 = 6.6 ksi Factor for multiple actuation:

From Figure 2 .

Multiplier for Valve C = 1.0 Multiplier for Valve G = 0.64 Multiplier for Valve D = 0.20 Multiplier for Valve H =.0.13 Multiplier for Valve M = 0.12 Multiplier for Valve L = 0.11

^

Multiplier for valve A = 0'11 .

Multiplier for Valve K = 0.11 Multiplier for-Valve B '

or Valve E . = 0.11 Multiplier for Valve F = 0.10 Net multiplier

= 1. 2 j[(1. 0) 2+ (0. 64 ) 2+ (0. 20) 2+ (0.13) 2+ (0.12) 2+ (5) (0.11) *+ (0.10) 2

= 1.49 Consecutive actuation factor = 2.37 SRV setp.oint pressure to reactor. pressure multiplier = 1.10-Resultant shell stress due to SRV discharge e 0.49)(2.37)(1.10)(6.6) = 12 5 . 6 . k s'i

g g Shellstressduetodeadweihht, hydrostatic pressure and carthquake was taken from Reference 8, = 1.4 ksi ,

Totcl shell tress = 25.'6 + 1.4 = 27.0 ksi The yield stress for the SA 516 Grade 70 torus material is 38 I l kri. Since the calculated stress is within yield, and the strength ratio is about 0.36, it satisfies Short Term Program Acceptance Criteria.  ;

3.2.2 Columns

~

From Section 3.1.2, the maximum column loads for.a single valve cold actuation,were:

Outside' column = 105 kips Inside column = 105 kips

~

Factor for multiple actuation:

• From Figure 3, reading abscissa for ring girder discharge Multiplier for Valve C = 1.0 .

Multiplier for ' Valve G or D = 0.70

. Multiplier for Valve H = 0.15 .

Multiplier for Valve M = 0.15 .

Multiplier kor Valve L = 0.11 Multiplier for Valve A = 0.12

~

Multiplier for Valve K or E = 0.09 Multiplier for Valve B = 0.10 Multiplier for Valve F = 0.08 ,

Net multiplier = 1.0 + (2)(0.70) + ( 2 ) (0 .15 ) + 0.12 + 0.11 + 0.10

+ (2)(0.09) + 0.08 = 3.29

-e 9 eme .

• q A =

5 -

Q.

s Conse.cutive actuation f actor a 1.96 m .

SRV setpoint pressure to reactor pressure multiplicr = 1.10 l

Resultant compressive column load Outside column = (3.29)(1.96)(1.10)(105) = 745 kips Inside column = ( 3. 29 ) ( 1. 9 6 ) ( 1.10 )'(105 ) = 745 kips Column loads due to dead weight,' hydrostatic pressure and earth-quake were taken from Reference 8.

Outside column = 92 kips Inside column = 74 kips Total column load -

Outside column = 745 + 92 = 837 kips

, Insideco'humn = 745 + 74 = 819 kips

,. To calculate strength ratio for the columns, capacity was taken as 1.6 times CRC static capacity (Reference 7). On this-basis:

837 Outside column SR = 3120 = 0.27 < 0.5 , .

819 Inside column SR =

1970 = 0.42 < 0.5 3.2.3 column-to-Torus connection .

The maximum loads measured near the tops of the columns, for a single valve cold actuation, were:

Outside column connection = 84 kips Inside column connection = 84 kips e

e

G

• * . \

Factor for multiple actuation:

From Figure 3, reading abscisca for ring girder discharge Multiplier for Valve C = 1.0 l =

Mult-iplier for Valve G or D = 0.70 .

Multiplier for Valve H = 0.15 Multiplier for Valve M = 0.15 Multiplier for Valve L = 0.11 4

Multiplier for Valve A = 0.12 Multiplier for Valve K or E = 0.09 Multiplier for Valve B = 0.10 Multiplier for Valve F = 0.08 Net nultiplier

= 1. 0 + (2)( 0.70 ) + (2)(0.15) + 0.12 + 0.11 + 0.10 + (2)(0.09)

+ 0. 08 = 3. 29

~

Consecutive actuation f actor = 1.96

'SRV setpoint pressure to reactor pressure multiplier = 1.10 Resultant column-to-torus connection force

. Outside column connection ,

= (3.29)~(1.96)(1.10)(84) = 596 kips Inside column connection

= (3.29)(1.96)(1.10)(84) = 596 kips Column-to-torus connection forces due _to-dead weight and carth-quake werb taken from Reference 8 Outside column connection = 74 kips .

Inside' column connection =-58 kips

.. m_. ______ -

e G

- r .

Total' column conneption load Outside column cotmeetion = 596 + 74 = 670 kips Inside column.cennection = 596 + 58 = 654 kips- '

Thb capacity of the connection was calculated in accordance

- with Reference 7, Appendix A. On this basis:

Strength ratios are: ,

670 Outside column connection = 2810 = 0. 24 < 0. 5 -

654 Inside column connection =

2119

= 0 . 31 < 0 . 5 -

)

e 4

+

O e

O e e

j e

4

• l e

t e

e e

_. l

~

- y- .- - - < y - -s p- , v r-m-, , , , , - - ,

1

. (y v ,

4.0 RESULTS AND_ CONCLUSIONS The results of the evaluation for the four valve case are summarized in Table 1. .The maximum primary membrane stress intensity calculated for the torus shell was about 26.0 ksi, which is within yield strength of the material, and thus, well within the recuired margin of two against failure. The ,

worst case strength ratio for support system components was found to be that for the inside column. The calculated strength ratio for the inside column was 0.25, which satisfies the STP acceptance criteria of 0.5.

4 The results of the evaluation for the eleven valve case are cummarized in ' Table 2. The maximum primary membrane stress intensity calculated for the torus shell was about 27.0 ksi, which is also within yield strength of the material. The worst .

case strength ratio for support system components was again

. found to be that for the inside column. The calculated strength .

. ratio for the inside column was 0.42, which satisfies the STP

~

acceptance criteria of 0.5.

9 4

m o

a 4

O e O

9 4

_a

i . . . .

g g

- s TADLE l

• SUt@iARY OF RCSULTS FOR HATCH UNIT 2 SUBSEQUENT $IMULTANEOUS ACTUATIOh1 OF FOUR LOWEST SETPOINT VALVES:

EVALUATION AGAINST STP CRITERIA -

1 STRESS CATEGORY i EXTRAPOLATED LOCATION OR l FOUR VALVE STRENGTH RATIO LOAD HOT POP i

DISCHARGE t SHELL 'O.34

• 26.0 ksi AT SPAN , Pm i

OUTER  !* AXIAL FORCE .

514 kips 0.16 COLUMN I AT BOTTOM -

OUTER COLUMN- AXIAL FORCE TORUS 'AT TOP

• 411' kips 0.15 CONNECTION .

~

INNER AXIAL FORCE '

• COLUI24 AT BOTTOM 496 kips 0.25 INNER COLUMN- AXIAL FORCE TORUS . AT TOP 395 kips 0.19 CONNECTION O

9

.9 1

1 1

1 l

e e TABLE,2 SUMf;iARY OF RESULTS .

FOR HATCH UNIT 2 SUBSEOUENT SIMULTANEOUS ACTUATION OF ELEVEN VALVES .;

EVALUATION AGAINST STP CRITERIA STRESS CATEGORY EXTRAPOLATED LOCATION OR ELEVEN VALVE STRENGTH RATIO LOAD HOT POP DISCHARGE t . SHELL

• AT k SPAN Pm 27.0 ksi 0.36 OUTER AXIAL FORCE COLUMN AT BOTTOM 837 kips 0.27 OUTER COLUMN- AXIAL FORCE TORUS AT TOP .

670 kips 0.24 CONNECTION

• INNER AXIAL FORCE -

8 COLUMN AT BOTTOM 819 kips 0.42 ,

INNER COLUMN- AXIAL FORCE TORUS- AT TOP -

654 kips 0.31 CONNECTION .

On e

e .

6

O

5.0 REFERENCES

1. Multiple . Subsequent Safety / Relief Valve Actuaticn Tor E. I. Hatch Unit 2. Prepared by Bechtel

~

Power. Corp. for G'eorgia Power Co. and Southern Services, Inc. Revision 0 dated July 1, 1978.

- 2. Peach Bot, tom Atomic Power Station, T.est Report

" Response of Suppression Chamber to Main Steam Relief Valve Discharge", Prepared by Philadelphia Electric Co. and Bechtel Power Corp., Rev.1 dated March, 1978.

3. Mark I Containment Program - Multiple Consecutive S/RV Actuation Evaluation - Task 7.1.3, Generic Structural Multiplier Development Procedure, dated June 16, 1978', by General Electric Company.

4. Final Report In-Plant Safety-Relief Valve Discharge Load Test - Monticello Plant, August, 1977, by General Electric Company.

5. " Safety Relief Valve Discharge Analytical Models", .

General Electric Report NEDE-20942-P, by P. Valandini, dated May 1975. .

6. USNRC Letter John F. .Stolz to Charles F. Whitmer,

" Multiple-Subsequent Actuations of Safety / Relief .

. Valves Following an Isolation Ev~ent", dated March 22, 1978.

7. Description of Short Term Program Plant Unique Torus Support Systcms and' Attached Piping Analysis, Prepared for Mark I Containment Owners Group by Nutech, June 1976. .

8. Structural Evaluation of Pressure Suppression System

. for-E.I. Hatch Nuclear Plant Unit 2, July 1977 by Bechtel Power Corporation.- ,

e n p -

e -

9. Results of S/RV Discharge Test for E. I. Ilatch Unit 2, Given by Telecon from T. T. Robin to R. Broman, betober 17, 1978. ,

O 4

e e

O 9

e e

O O t

4