ML20112D764

Program to Determine Capability of Millstone 3 Nuclear Power Plant to Withstand Seismic Excitation Above Design Sse

ML20112D764
Person / Time
Site:	Millstone
Issue date:	11/30/1984
From:	Kennedy R, Ravindra M, Sues R STRUCTURAL MECHANICS ASSOCIATES
To:
References
PROC-841130, NUDOCS 8501140478
Download: ML20112D764 (125)

Text

.

NTS/SMA 20601.01-R3 A PROGRAM TO DETERMINE THE CAPABILITY OF THE MILLSTONE 3 NUCLEAR POWER PLANT TO WITHSTAND

~.

SEISMIC EXCITATION AB0VE THE DESIGN SSE

/

4 prepared for -

NORTHEAST UTILITIES Berlin, Connecticut November, 1984 i

p g STRUCTURAL mECHAnlCS

"""""" ASSOCI W a cene.

ATES c....

. \ .

)'

5160 Birch 8 toot, Newport Beach, Cast. 92660 (714) 833-7552 8501140478 841130 PDR ADOCK 05000423 g A PDR 5

NTS/SMA 20601.01-R2 A PROGRAM TO DETERMINE THE CAPABILITY OF THE MILLSTONE 3 NUCLEAR POWER PLANT TO WITHSTAND SEISMIC EXCITATION AB0VE THE DESIGN SSE by M. K. Ravindra R. H. Sues R. P. Kennedy D. A. Wesley prepared for NORTHEAST UTILITIES Berlin, Connecticut November, 1984 l

g g STRUCTURAL mECHRnlCS

-W-RSSOCIATES a c.m. c....

5160 Birch Street, Newport Beach, Cahf. 92660 (714) 833-7552

TABLE OF CONTENTS Section Title Pa5gt 1 INTRODUCTION .................. 1-1 1.1 Background . . . . . . . . . . . . . . . . . 1-1 1.2 Objectives . . . . . . . . . . . . . . . . . 1-1 1.3 Report Outline . . . . . . . . . . . . . . . 1-2 2 OVERVIEW 0F MILLSTONE 3 SEISMIC PSS . . . . . . . 2-1 2.1 Seismic Hazard Analysis .......... 2-2 2.2 Seismic Fragility Evaluation . . . . . . . . 2-3 2.2.1 Fragility Model . . . . . . . . . . . 2-3 2.3 Plant Systems and Accident Sequence Analysis 2-5 2.4 Seismic Risk Estimates . . . . . . . . . . . 2-7 3 ANALYSIS OF SEISMIC CAPABILITY OF MILLSTONE 3 . . 3-1 3.1 Identification of Dominant Risk Contributors 3-1

3.1.1 Plant Damage States . . . . . . . . . 3-1 3.1.2 Structure and Equipment Failures .. 3-3 3.2 Structure and Equipment Fragilities / Margins. 3-5 3.3 Plant Damage State Fragilities / Margins . . . 3-7 3.4 Plant Damage State Frequencies . . . . . . . 3-9 3.5 Contribution to Seismic Risk from Different.

Acceleration Ranges ............ 3-11 4 SENSITIVITY STUDIES . . . . . . . . . . . . . . . 4-1 4.1 Overall Approach . . . . . . . . . . . . . . '4-1 4.2 Component Fragilities ........... 4-2 4.2.1 Fragili ty Cut-off . . . . . . . . . . 4-2 4.2.2 Interaction between Failure. Modes . . 4-4 4.2.3 Dependence between Component Failures 4-4 4.2.4 Variation of Fragility Parameters . . 4-7 4.2.4.1 Earthquake Magnitude Range . 4-8 4.2.4.2 Limiting Displacement for. .

Attached Piping ...... 4-9 i

_ _ . ~ . _ , . _ _ _ - . - --- . . -. .

TABLE OF CONTENTS (Continued)

Section Title Page 4.2.4.3 Coefficient of Sliding ...

Friction .......... 4-10 4.2.4.4 Buckling Capacity Estimates . 4-11 4.2.4.5 Ratio of Peak Ground Velocity to Peak Ground Acceleration . 4-11 4.2.5 Lognormal Distribution . . . . . . . . 4-12 4.3 Seismic Hazard ...............

4.3.1 Effect of Peak Acceleration Truncation 4-13 4.3.2 LLNL Hazard Curves . . . . . . . . . . 4-13 5

SUMMARY

AND CONCLUSIONS. . . . . . . . . . . . . . 5-1 5.1 Summary . . . . . . . . . . . . . . . . . . . 5-1 5.2 Conclusions . . . . . . . . . . . . . . . . . 5-2 REFERENCES APPENDICES A Component Fragility Plots B Plant Level Fragility Curve Tables I

4 ii

1. INTRODUCTION

1.1 BACKGROUND

Current industry codes and regulatory standards for the design of nuclear power plant structures and equipment impose significant levels of conservatism at a number of steps in the design process. Conservatism is introdaced at such stages as selection of design response spectra which are normally median plus one standard deviation values, use of low damping values, postulation of load combinations which include peak loads from LOCA and other extreme loads combined with seismic loads, use of code specified allowable values for strength, and limiting the design response to the elastic range. There is usually a compounding effect of the conservatism introduced at each stage of the design process such that the plant structurec and equipment have the ability to withstand levels of seismic excitation well above the design Safe. Shutdown Earthquake (SSE).

The median seismic ground acceleration capacities together with their variabilities, for the controlling structures and equipment items have recently been developed as part of the Probabilistic Safety Study (PSS)forMillstoneUnit3(NU,1983). The results from this study show that all the important structures and equipment items have median ground acceleration capacities well in excess of the 0.179 SSE for which Millstone Unit 3 was designed. In addition, the results of the PSS may be utilized, as shown in thic report, to demonstrate that there is a low frequency of seismic failure for Millstone Unit 3 at earthquake acceler-l ations well above the SSE acceleration. The motivation for the study l reported herein came from the NRC's concern that earthquakes larger than the SSE may potentially occu. in the Millstone site region.

1.2 OBJECTIVES Using the results of the Millstone Unit 3 PSS, the objective of this study is to investigate the capability of Millstone 3 to withstand 1-1

l seismic excitation above the design SSE. This will be accomplished using the procedure outlined below. l

1. Identify dominant contributors to seismic risk.

t

2. Evaluate the high confidence, low frequency of failure accelerations for critical structures and equipment.

3. Evaluate the high confidence, low frequency of i failure accelerations for the dominant plant damage states.

4. Evaluate the frequencies of occurrence of significant plant damage states from seismic events.

5. Evaluate the contributions of various acceler-ation ranges to the frequencies of occurrence of significant plant damage states.

6. Investigate the sensitivity of the results of the Millstone Unit 3 PSS to variances in the assump-tions or analytical models employed for the seis-mic fagility and hazard development.

1.3 REPORT [0fiLINE This report describes the approaches and results of two differ-

ent tasks. The fir:t task consists of addressing Objectives 1 through 5 outlined above. The'second task is to perform sensitivity studies of the seismic fragility and hazard models in order to address Objective 6 above.

This will allev in evaluation of the robustness of the conclusions derived from the results of task 1 and will address several concerns--

raised duriny'the NRC review of the Millstone PSS. Chapter 2 provides an

( overview of the seismic PSS of Millstone Unit 3, Chapter 3 describes the

( studies done using the results of the seismic PSS to evaluate the seismic j safety of Millstone Unit 3 structures, equipment and systems, Chapter 4 documents the results-.of several sensitivity studies performed on the component fragility and seismic hazard models and Chapter 5 includes a sumary. and the important conclusions of the study.

l l

1-2 i

2. OVERVIEW OF MILLSTONE 3 SEISMIC PSS The analysis of seismic risk as performed in the PSS can be divided into the sections as listed below:

1. Seismic hazard analysis - estimation of the fre-quency of various levels of seismic-induced ground acceleration occurring at the site.

2. Fragility evaluation - estimation of the condi-l tional frequencies of structural or equipment failure for given levels of ground acceleration.

1

3. Systems / accident sequence analysis - modeling of the various combinations of plant failures that could initiate and propagate a seismic core melt accident sequence.

4. Core melt quantification - assembly of the results of the seismic hazard, fragility and systems analyses to obtain estimates of the -

frequencies of core melt and the various plant damage states.

5. Public risk quantifications - assessment of the impact of seismic events on the containment and consequence analyses, and integration of these

results with the core melt analysis to obtain estimates of seismic risk in terms-of public health effects.

In the following, the methods used and the results obtained in l

each stage of the seismic risk analysis are briefly outlined. The

-results of the investigation described in this report are limited to

(- consideration of the first four sections listed above only. Thus, the i study concentrates on assessing the seismic capacity of-the plant as 1

described by ccre melt but_does not evaluate the consequences to the public as a result of a seismically induced core melt.

l 2-1 b

2.1 SEISMIC HAZARD ANALYSIS The seismic hazard curves - the curves showing the annual fre-

quencies of exceeding various peak ground acceleration levels - for the Millstone site were deveioped by Dames & Moore (D&M, 1983). Briefly, the procedura used to develop these curves consists of the following steps

o Identification of the sources of earthquakes such as fault zones and seismotectonic provinces.

. o Evaluation of the earthquake history of the region to assess the frequencies of occurrence of earthquakes of different magnitudes or epicentral intensities.

o Development of attenuation relationships to estimate the intensity of earthquake-induced ground motion  ;

t (e.g., peak ground acceleration) at the site.

I o Integration of all the above information to generate the frequencies with which different peak ground acceleration values would be exceeded.

The parameters that characterize this seismic hazard model are the seismic source zones, the activity rate (i.e., annual number of

. earthquakes occurring on a source), the relative frequencies of occur-rence of different sizes of earthquakes in a source, the attenuation of ground motion from a source to the site, and the upper-bound magnitude or epicentral intensity of a source. Uncertainties exist in all of these

! model parameters. In the reference, (D&M, 1983), the uncertainties of these model parameters are considered by postulating a set of hypotheses.

Each hypothesis consists of a specified configuration of the seismic

source zones, a value of the Richter slope parameter b (defining the f relativa frequwic;.: of different sizes of earthquakes), a value of the upperbound magnitude mb, max,and a selected attenuation equation. A

. probability value is assigned to each of these hypotheses, based on the analyst's degree of belief and expert opinion. A seismic hazard curve is generated for each hypothesis. The family of such curves, with associ-ated proabilities, represents the seismic hazard at the site. Figure 2-1,

shows the family of ten seismic hazard curves developed for the Millstone site.

! 2-2

, . a_. _. _ . . . . . ~ . _ . _ _ _ _ _ _ _ _ . _ _ _ . _ _ _ _ _ _ . _.

I 2.2 SEISMIC HAZARD EVALUATION The methodology for evaluating the seismic fragilities of struc-tures and equipment is well documented (Kennedy and Ravindra, 1984; ANS-IEEE-NRC, 1983; LLNL, 1982). Seismic fragility of a structure or equipment item is defined as the conditional frequency of its failure for a given value of the seismic input or response parameter (e.g., ground acceleration, stress, moment, spectral acceleration). Seismic fragilities are needed in a PRA to estimate the frequencies of occurence of initi-atin.g events (e.g., large LOCA, small LOCA, RPV rupture) and the failure frequencies of different mitigating systems (e.g., safety injection system, residual heat removal system, and containment spray system).

The objective of fragility evaluation is to estimate the peak ground motion acceleration value for which the seismic response of a given component (i.e., structural element or equipment) located at a specified point in the structure exceeds the component's capacity, resulting in its failure. Estimation of this ground acceleration value -

called the ground acceleration capacity of the component - is accomp-lished using information on the plant design bases, responses calculated at the design analysis stage, and as-built dimensions and material pro-perties. Because there are many sources of variability in the estimation of this ground acceleration capacity, the component fragility is i

described by means of a family of fragility curves; a probability value l 1s assigned to each curve to reflect the uncertainty in the fragility estimation (Figure 2-2). Note that the probability values (q, in Figure 2-2) sum to 1.0.

2.2.1 Fragility Model The entire fragility family for a component corresponding to a particular failure mode can be expressed in terms of the best estimate of l the median ground acceleration capacity 5 and two random variables.

! Thus, the ground acceleration capacity, A, is given by:

t A=4cR 'U (2-1) 2-3

in which cR and c u are random variables with unit medians. They represent, r?spectively, the inherent randomness about the median and the uncertainty in the median value. In this model, we assume that both cR and cu are lognotup1ly distributed with logarithmic standard deviations, S R and 8 , respectivelyl The formulation for fragility given by Equation 0

2-1 and the assumption of lognonnal distribution enable easy development of the family of fragility curves appropriately representing their uncertainty. For the quantification of fault trees in the plant system and sequence analysis, the uncertainty in fiagility needs to be expressed in terms of a range of failure frequencies for a given ground acceleration. This is achieved as explained below:

Withperfectknowledge(i.e.,onlyaccountingfortherandomvariability, cR), the conditional frequency of failure f gfor a given peak ground acceleration level, a, is given by:

f , 'n(a/I)'

E BR . (2-2) where 4 (-) is the standard Gaussian cumulative distribution function.

The relationship between fa and a is the median fragility curve such as is plotted in Figure 2-3 for a component with median ground acceleration capacity E = 0.90g and 8 R = 0.30. For the median frequency of failure range of 5% to 95%, the ground acceleration capacity would range from 0.55g to 1.48g for this particular component.

l When the modeling uncertaintyu c is included, the frequency of failure

becomes a random variable (uncertain). At each acceleration value, the l frequency of failure f can be represented by a probability density func-tion. The probability Q, of not exceeding a frequency of failure f' is related to f' by

~1n(a/5)+s V

• IO) f' = 4 8R ,

(2-3) 2-4

where Q = P[f < f'la], i.e., the probability that the conditional frequency of failure, f, is less than f' for a peak ground acceleration a

-1

, & (-) = the inverse of the standard Gaussian cumulative distribution function

- For exanple, the frequency of failure f' at acceleration 0.4a that has a 95% non-exceedence probability is obtained from Equation 2-3 as 0.22.

The 5% to 95% probability (confidence) interval on the failure frequency at 0.49 is~ 0 to 0.22. Subsequent computations can be made easier by discretizing the random variable frequency of failure f into different intervalsandassigningprobabilitygjintoeachinterval(Fiqure2-2).

Note that the sum ofq q assigned to all the intervals is unity. By this i process, a family of fragility curves, each with an associated prob-ability q, can be developed. This method is employed in this analysis.

In the PSS, the method of Monte Carlo simulation was used in calculating the distribution of failure frequencies.

ThemediangroundaccelerationcapacityEanditsvariabilityestimates gR and SUare evaluated by taking into account the safety factors inherent in the capacity predictions, response analysis, and equipment qualification. l l

In the Millstone 3 PSS, the values of d, B , and S were R U estimated for all safety-related structures and equipment. Table 2-1 shows these estimates for key structures and equipment that were included l in the event- and fault-trees of the PSS. i I

2.3 PLANT SYSTEMS AND ACCIDENT SEQUENCE ANALYSIS The modeling of accident sequences in the seismic risk analysis was accomplished with the use of fault trees expressing the plant safety systems logic. As discussed in the PSS, four seismically-induced accider.t sequence initiators were indentified: l I

\

]

2-5

loss of offsite power transient, large LOCA, small LOCA, and ATWS. The large LOCA seismic fault tree included the rupture of medium to laige f diameter (greater than 2 inches) Reactor Coolant System (RCS) piping, the failure of Nuclear Steam Supply System equipment, and the collapse of the 4

containment structure as possible large LOCA initiators. Similarly, the small LOCA seismic fault tree considered the possibility of this event being caused by either an uncontrolled ATWS scenario, the rupture of small diameter (2 inches) RCS piping, or failure of the reactor coolant

nmp seals. The ATWS scenario included the failure of the reactor scram

, system followed by the failures of support systems - i.e. electric power and service water. In constructing these fault trees, the indirect fail-ures of equipment caused by structure / system interactions (e.g., Demin-i eralized Water Storage Tank supply piping shear failure caused by sliding of the Engineered Safety Features (ESF) building, and failure of the aux-iliary feed water system as a result of ESF building collapse) were considered.

t 4

From the seismic fault trees, it may be ' determined that core melt occurs as one of three general types of consequences: transient with loss of cooling (MI), small LOCA with loss of safety injection or

) cooling (M2 ), or large LOCA with loss of safety injection (M3). A total of 19 plant damage states which describe the type of core melt accident, the timing of core melt, and the operational status of the l engineered safety features, initiated by seismic events were identified.

Fault trees were constructed for each of these damage states and Table l

l 2-2, reproduced from the Millstone 3 PSS, shows the definitions of these l plant damage states. Analysis of the consequences of the plant damage i states (in the PSS) has shown that more than 90% of the contribution to core melt and almost all of the risk is from the following four plant damage states:

l

! V3 - LOCA with containment bypass AE - Large LOCA with failure of safety injection and containment quench sprays i

1

2-6 i

l

r I

i SE - Small LOCA or ATWS with failure of safety iniection and containment quench spra.ys TE - Transient (caused by loss of offsite power) with l failure of onsite emergency power or RCS heat t

removal l

In Chapter 3, these dominant plant damage states will be analyzed in

! depth.

2.4 SEISMIC RISK ESTIMATES l With the seismic fragility estimates of the key structures and l

equipment and using the plant damage state fault trees, the occurrence frequencies of the plant damage states at specified acceleration levels (conditional frequencies) are obtained by a simulation procedure. The annual frequency of occurrence of tae plant damage states are then obtained by convolving the conditional frequencies (again, the damage frequency given an earthquake of a particular peak ground acceleration occurs) with the earthquake occurrence frequencies (seismic hazard).

i l

l 1

l 2-7 L

TABLE 2-1 MILLSTONE 3 - FRAGILITIES OF KEY STRUCTURES AND EQUIPMENT i

i-SYMBOL DESCRIPTION A BR B U

l 1. LOSP Loss of Offsite Power (ceramic insulator 0.20 0.20 0.25 failure)

2. RECRHTEX Containment Recirculation Heat Exchangers 0.82 0.32 0.52

3. EGECLPSE Emergency Generator Enclosure Building (wall 0.88 0.20 0.46 footing failure)

4. RWST Refueling Water Storage Tank 0.88 0.30 0.36 l 5. EDGOILCL Emergency Diesel Generator (oil cooler anchor 0.91 0.24 0.43 bolt failure)

6. COREGE0M Reactor Vessel Core Geometry Distortion 0.99 0.31 0.33 i 7. DFCNTBLD ControlBuildingCollapse(diaphragm) 1.00 0.24 0.33 l

l 8. -CNTRLBLD Control Building Failure (sliding) 1.20 0.21 0.47 l

9. CRDS Control Rod Drive System (failure to SCRAM) 1.00 0.30 0.38

10. RPCWPUMP Component Cooli1g Water System Pumps 1.13 0.25 0.33

11. SWPIPE Service Water System Piping (due to 1.30 0.24 0.49 pumphouse sliding)

12. SWPHSLID Service Water Pumphouse Failu.e (sliding) 1.30 0.24 0.49 l

13. EGESLIDE Emergency Generator Enclosure Building 1.30 0.24 0.46 (sliding)

14. AUXBLDG Auxiliar 1.40 0.37 0.41 failure)y Building Collapse (shear wall

15. RCSPIPE Reactor Coolant System Piping (large LOCA) 1.59 0.48 0.51

16. RCSSMPIP Reactor Coolant System Piping (small LOCA)1 1.59 0.48 0.51 l

l 17. DWST Demineralized Water Storage Tank 1.60 0.25 0.43 l

18. TBDAFWP Turbine Driven Aux. F. W. Pump 1.60 0.30 0.26 Se , ice Water Pumphouse (shear wall failure) 1.60 0.33 0.34

( 19. SWPHCOLL l

2-8 ,

1 -. - . - - .- . . - - .-_- - _-_- - _ _ . _ _ _ _ __ _ _ _ _ _ . _

TABLE 2-1 (Continued)

MILLSTONE 3 - FRAGILITIES OF KEY STRUCTURES AND EQUIPMENT k

~

'I .

SYMBOL , - DESCRIPTION A BR B U

20. ESFBLDG Engineered Safeguard Features Building (base 1.70 0.23 0.43 mat shear wall failure)

21. RECPUMPS Containment Recirculation System Pumps 1.71 0.37 0.38

22. PORV Power-0perated Relief Valves (loss of 1.84 0.49 0.29 operability)

23. CVCSPIPE Chemical Volume Control System Piping2 2.17 0.42 0.52

24. RECRPIP Containment Recirculation System Piping 2 2.17 0.42 0.52

25. RPCWPIPE RCP Component Cooling Water System Piping 2 2.17 0.42 0.52

26. QSPIPE Quench Spray System Piping 2 2.17 0.42 0.52

27. CONTWALL Containment Crane Wall (collapse) 2.20 0.39 0.38

28. MCCFAIL 480V Motor Control Centers (trip) 2.21 0.28 0.57

29. RXVESSEL Reactor Vessel (support pads fail) 2.35 0.48 0.44

30. SWPUMPS Service Water System Pumps 2.40 0.31 0.53

31. RCPUMPS Reactor Coolant Pumps (large LOCA) 2.68 0.43 0.47

32. RCPSWHEX RCP Seal Water Heat Exchangers 2.68 0.46 0.40

33. RPCWHTEX RCP Component Cooling Water Heat Exchangers 2.68 0.46 0.40

34. CABTRAY Cable Trays 2.70 0.48 0.42

35. QSPUMPS Quench Spray System Pumps 2.93 0.23 0.68

36. QSHEADER Quench Spray System Header Piping 3.07 0.G2 0.37

37. STGEN Steam Generator Supports 4.00 0.44 0.'48 l

1 Conservatively. assumed to be the same as large LOCA Assumed to be the same as balance-of-plant i

2-9

TABLE 2-2 EXPLANATION OF PLANT DAMAGE STATES (PDS)

SEISMIC CONTAINMENT INDUCED ECCS ECCS IN3ECTION RECIRCULATION QUENCH SPRAY SPRAY RECIRC PDS INITIATOR Fails N/A Fails N/A AE Large LOCA Fails N/A Fails N/A

, SE Small LOCA or ATWS Fails N/A Fails N/A

}

TE LOSP Successful Fails Fails N/A l AL Large LOCA Successful Fails Fails N/A I . SLl* LOSP Successful Fails Fails N/A l SL2** Small LOCA or ATWS N/A Successful Successful AEC Large LOCA Fails j Successful Small LOCA or ATWS Fails N/A Successful SEC N/A Successful Successful 4 TEC LOSP Fails N/A Successful Fails

$ AEC Large LOCA Fails l Fails SEC Small LOCA or ATWS Fails N/A Successful j

N/A Successful Fails i TEC LOSP Fails Fails . Successful Successful i ALC Large LOCA Successful Fails Successiul Successiul

SLIC* LOSP Successiut Fails Successfit Successful SL2C** Small LOCA or ATWS Successful

} Fails Successful Fails Successful i ALC Large LOCA Fails Successful Fails SLIC* LOSP Successful f Successful Fails Successful Fails SL2C** Small LOCA or ATWS V3 LOCA w/ containment NOTE: All mitigating systems assumed to be failed for V3.

l l bypass

'PDS's including the combination "SLl" imply a consequential LOCA due to opening PORV's to perform feed and bleed, subsequent to f

a saismic-induced LOSP. * *"SL2" implies a late core melt following an S initiator.

14 2h C.S.8 9

=- - -

, 2 . .. . . .

m- -- m . .. .. .

M NhW,*wwA.::':: -

2,Rw W ,~ - ,s.**e

- * " "e' l ,,-. A w tumsx<

NN  %%V/ ,8'8

! ,,-4 k

- .=

b --

==- u s

x s a s s x

\\ \h Ik\'(

\bi \D6\ L\

to

-a _ L

's '.. A ' s '. ' , ',

1 xm iit x s T, \\M \ \\ \l i

\N tk brYL 33.

808

....... . WWO.. ._ .._ , .

tese SG ACCELSRAfen(eaWeseI l I

a t

l FIGURE 2-1. SEISMIC HAZARD CURVES FOR MILLSTONE l

l l 2-11

l l

l 1.9 1

l 5 ...

6

0.4 -

q3 q2 43 9

, 0.4

• q

! I E ..: .

3 .

[ { 0.4 0.4 1.2 1.8 2.0 l

SSE Peak Grwnd Accelerettee. 3 FIGURE 2-2. FAMILY OF FRAGILITY CURVES FOR A COMPONENT 2-12

1 l

i l

1.0 95%

1 j 0.8 "Nonexceed.

Probability 4 Curve Median

% Fragility 0.6 -

Curve m 0.4 -

, 5%

! Nonexceed.

0. -

Probability

- Curve f J. ..-nu. .

, .4 ... i.. 1..

SSE Peak Ground Accelerstica, 3 l

l .

l e?

FIGURE 2-3. MEDIAN, 5% NONEXCEEDANCE, AND I 95% NONEXCEEDANCE FRAGILITY CURVES l FOR A COMPONENT l

l l

2-13 .

1

3. ANALYSIS OF SEISMIC CAPABILITY OF MILLSTONE 3 3.1 IDENTIFICATION OF DOMINANT RISK CONTRIBUTORS The Millstone PSS (NU,1983) evaluated seismic fragilities for critical structures and equipment items whose failures could potentially lead to plant damage. Of these, 37 were identified as key components, based on the capacities evaluated. These results were sumarized in the previous section (Table 2-1).

i In the following, the components that represent the dominant contributors to seismic risk are identified. This is accomplished by an analysis of the plant damage state Boolean expressions (derived from the fault trees given in the Millstone PSS), such that those components that perform critical safety functions and have seismic capacities compar-atively lower than other critical components (and, therefore, dominate the seismic risk) can be singled out.

3.1.1 Plant Damage States The plant damage states for which the seismic initiated accident sequences have been shown to be important contributors as discussed ear-lier are: TE-Transient (causedbylossofoffsitepower)withfailure of onsite emergency power or RCS heat removal; AE - large LOCA with failure of safety injection and containment quench sprays; SE - small LOCA or seismic ATWS with failure of safety injection and containment quench sprays; and V3 - LOCA with containment bypass. The Boolean expressions for these damage states are given in Equations 3-1 through 3-4 below,

. . . t r .

TE = 1 ^ j 3 v 5 v 7 v 12 v 20 v 30 v 34 v R19 v (17 v R5)^4 h(3-1)

J i . - .. .

AE = 3 v 4 v 5 v 7 v 12 V 20 v 30 v 34 v R19 ^ 15 'v 29 v 31 (3-2) 3-1

SE = 3 v 4 v 5 V 7 v 12 v 20 v 30 v 34 v R19 ^ 6v 9 v 16 (3-3)

V3 = 27 v 37 (3-4) where the numbers correspond to the component numbers given in Table 2-1.

Here, the notation, v, is used to designate an "or" operation, thus, for example, damage state V3 occurs if Component 27 or Component 37 fails; the symbol A designates an "and" operation, thus, for example, damage state AE occurs if any of the components in the first set of brackets fails and any one of the components in the second set of brackets fails. Descriptions of the components and their fragility parameters have been presented in Table 2-1.

Note that these expressions also include the non-seismically induced random failures, R5 and R19, random failure of all three auxiliary feedwater trains and random failure of both emergency AC systems, respec-tively. The respective median frequencies for these two systems to fail are approximately, 5x10-5 and 2x10-4 per demand.

It should also be pointed out that if capacities for more than one failure mode for a particular structure have been evaluated (e.g.,

items 7 & 8 in Table 2-1), only the dominant (lowest capacity) failure mode has been considered. This simplification, which amounts to an assumption of perfect dependence between failure modes, has an insignif-icant influence on the results, as discussed in Section 4.2.2.

Finally, the Boolean expressions for damage states AE & SE have been developed with the assumption that loss of offsite power (component 1)isconceded,thatis,itisassumedtooccurunderany level of earthquake excitation. This assumption simplifies the deriva-tion of the Boolean expression from the fault trees and adds only a very minor conservatism.

3-2

3.1.2 Structure and Equipment Failures With the plant safety systems logic described (the Boolean expressions) and the fragility evaluation, those structures and equipment l which dominate the seismic risk can be identified. Also, the significance of random equipment failures during a seismic initiated accident sequence I can be evaluated.

For damage state TE, the components which dominate the seismic risk can be seen to be Components 3, 5, 7, 12 and 20. These are,

1. 3 - emergency generator enclosure building wall footing -

wall footing failure is conservatively taken to fail all attached equipment thereby failing the emergency power system,

2. 5 - diesel generator oil cooler - anchor bolt failure of this peripheral is taken to cause failure of the diesel generators,

3. 7 - control building diaphragm - loss of the diaphragm is taken to cause loss of actions from the control room.

4. 12- service water pumphouse sliding - fails connected piping causing loss of service water and loss of diesel generator cooling, and

5. 20 - engineered safeguard features building - failure of

, the shear wall near the base mat is taken as loss of l engineered safeguard features These components are the dominant contributors since they are the lower capacity components in Equation 3-1, whose failure can lead to damage i state TE, given a loss of offsite power (Component 1). Note that although the refueling water storagae tank (Component 4) has a compara-l tively low capacity, failure of the demineralized water storage tank (Component 17), which has a relatively high capacity, must occur concur-rently to result in damage state TE. Thus, the failure of the RWST does not significantly contribute to the occurrence of state TE. The remain-j ing components in Equation 3-1, 30 and 34, have very large capacities, l such that seismic failures are not credible.

i 1

3-3

The random failures, R19 (emer$,ency power system) and R5 (auxil-iary feed water system), must also be considered. For the Millstone Unit -

3 emergency power system, the random failure frequency of 2x10-4 per l demand is relatively low such that the frequency of any of Components 3, 5, 7, 12 or 20 failing due to seismic loads dominates. Also, the auxil-iary feedwater system random failure must occur concurrently with seismic failure of the RWST to lead to damage state TE. It is clear that since the random failure frequency is 5x10-5 this accident sequence does not contribute to risk.

Finally, it should be pointed out that the consequences of the component failures considered here are conservative. For example, dis-tortions of structural members have been assumed to result in failure of an entire safety system. Hence, the significance of the structural fail-ures 1.1ay be somewhat overstated.

l From Equation 3-2, it is seen that one of the components in the first set of brackets and one of the components is the second set of brackets must fail to result in damage state AE. The component failures in the second bracket represent the occurrence of a large LOCA. Since i

the failure accelerations of Components 29 and 31 are very high, failure

! ofCor;onent15(reactorcoolantsystempiping)istheonlycredible l source of a seismic-induced large LOCA (this fragility is believed to be l conservative). The components in the first set of brackets represent l loss of emergency cooling (caused mainly by loss of offsite and onsite AC power, i.e., station blackout) and have much lower capacities than the large LOCA. Thus, damage state AE is governed by failure of component 15.

For damage state SE, as in damage state AE, the components in the first set of brackets represent failure of the emergency cooling system. The second set of brackets represents the seismic-induced small LOCA or seismic-induced failure to scram. Since the capacity of the I

3-4 l -.- -- -. _ - . - - - _ _ _ - _ _ - - _ _ . . _

scram system (Components 6 or 9) is significantly lower than the small LOCA occurrence (Component 16), damage state SE will be governed by the failure of the scram system. Thus, Components 6 and 9, described below, are the dominant contributors to damage state SE.

1. 6 - reactor vessel core geometry - distortion of reactor vessel core prevents insertion of control rods

2. 9 - control rod drive system - failure of the scram system Damage state V3 will occur if either Components 27 or 37 fail.

]

Since the capacity of Component 37 (steam generator support failure) is significantly larger than Component 27, Component 37 does not contribute significantly to the risk. Thus, failure of Component 27, the containment crane wall is the dominant contributor to damage state V3. Although the crane wall failure occurs at only relatively large accelerations, its failure is significant from a radiological release point of view since it is assumed to cause a containment bypass. .

3.2 STRUCTURE AND EQUIPMENT FRAGILITIES / MARGINS In the previous section, those components which significantly contribute to the seismically-induced frequency of damage states TE, AE, SE and V3 were identified It is the capacities and seismic margins of these components which must be addressed. Table 2-1 includes the results of the seismic fragility analysis for each of these components. As can be seen, the median ground acceleration capacities (which r'epresent best estimates of the component's capacity) are all more than 5 times the 0.179plantdesign(SSE) level. In fact, of all the 37 key structures and equipment for Millstone Unit 3, only the offsite power has a median ground acceleration capacity only slightly above the SSE level.

This failure, however, is not a contributor to risk due to the avail-ability of the backup emergency power system (which has a median ground acceleration capacity significantly larger than the SSE, as stated above).

3-5

_.y _

As there is uncertainty in the fragility evaluation, the 5%-95%

confidence bounds on the median are of interest (Table 3-1). That is, we have 90% confidence that the median capacity falls within these bounds.

Alsc, we have 95% confidence that the median capacity is above the lower bound, f.05, and there is a 5% probability that it is above the upper bound, A ,95 These values are obtained, based on the fragility model formulation (Section 2.2.1), as

. . 4 -I(.05)xd U . -1.65 s U A.05

= Ae = Ae (3-5) 4 -1(.95)xs 1.65 8 U A.95

=

Ae u=Ae .

(3-6)

From the table, it is seen that of all the components which are dominant contributors to the seismic risk, the lowest 95% confidence value on the median capacity is 0.419 . It may be noted that failure of Component 2,

, the containment recirculation heat exchangers (which has a relatively low capacity), does not contribute to the four important plant damage states under consideration here. Thus, it is seen that even the lower bouno estimates on the median capacities are all more than twice the SSE level.

Finally, the high confidence, low frequency of failure levels for each of the dominant components may be examined. This acceleration value, also shown in Table 3-1, considers both the uncertainty and random-ness variabilities, discussed in Section 2.2, and is the acceleration value for which we have 95% confidence that the frequency of failure is less than 5%. That is, it is an acceleration value for the component for which we are highly confident that given this level of ground acceleration

, there is only a small chance of failure. This value is obtained, based on the fragility model formulation (Section 2.2.1), as A

high conf.,

low freq. of failure s.- .

-1.65 ( su + 8R}

= Ae (3-7) 3-6

. . _ = . . _ , . . _.. _ - . _ _ _ - . _. --

Of all the dominant risk contributors, none have high confidence, low frequency of failure values below 0.3g.

( From these results, it may be concluded that the critical struc-

, tures and equipment at M111stene 3 all have seismic safety margins well

( above the plant design level of 0.17 9 , and, in fact, there is high confi-

l. dence that there is only a low frequency of failure for any risk contri-buting component for earthquakes below 0.39 Appendix A contains plots of the fragility curves for each of the 37 key structures and equipment at Millstone Unit 3. Shown on the

+

figures are the 5%, 50% and 95% confidence fragility curves. These curves represent the range of uncertainty in the fragility values. They are interpreted, for example, as follows: if the~ failure frequency at a

particular acceleration level is read from the 95% confidence curve.

l there is 95% confidence that the frequency of failure (at that accelera-tion level) is less than the value read.

I i

^

3.3 PLANT DAMAGE STATE FRAGILITIES / MARGINS From the plant damage state fault trees, the plant damage state Boolean expressions were derived. The Booleans,'as mentioned earlier, sunnarize those components or combinations of components that must fail in order to result in plant damage. Following the rules of Boolean algebra, the individual component fragilities ~ were combined, two at a time, using the Discrete Probability Distribution (DPD} spproach of l Kaplan (1981) to fonn plant damage state fragility curves. With the DPD l approach, the individual component fragility curves are first discretized l into a family of fragility curves, each with a probabilistic weighting, representing the uncertainty (characterized by the08 value)inthe

( fragility evaluation. Each one of the, say n, curves of one component is then combined, according to the rules of Boolean algebra, with each one l-l of the n curves of the second component. The resulting nxn curves are

! then condensed back to n curves, which are then combined with the n curves of the third component. This process is continued for all the L

i I . 3-7

. - + , . - , + ,-.,.wv.---..- --,...m-e--e--.- ,.--e., ---,---,-4,,--~%-,, --.-n-.-,-,.--, ,,,-------.-,.--,+e,----.m.--.w,---.-,.-.----

components in the Boolean expression, such that the end result is n plant damage state fragility curves. These plant damage state fragility curves, '

like a component fragility curve, represent the frequency with which the occurrence of the plant damage state (failure) is expected to occur, given the occurrence of an earthquake of a particular peak ground

! acceleration. Appendix B lists the complete fragility curves for the four plant damage states. Note that the fragility for each damage state is represented by a family of 11 curves, each with an assigned probabil-

!, ity(orweight)asdiscussed. Figures 3-1 through 3-4 show plots of the I

damage state fragility curves, wherein, the f amily of 11 curves has been j reduced to the 5%, 50% and 95% confidence fragility curve fomat previ-ously described.

i The results of the plant damage state fragility analysis are sumarized in Table 3-2. As for the individual component fragilities, the plant damage states all have significant margins above the SSE level.

The lowest capacity damage state, damage state TE (which is the dominant contributor to seismically induced late fatalities, NU, 1983), has a l median capacity of 0.61g with 95% confidence that this median is above O.399 and a high confidence, low frequency of occurrence level of 0.26g.

3 Note also that damage state V3, which is the dominant contributor to i seismically induced early fatalities (NU,1983), has a high confidence, j low frequency of failure of 0.6g.

i Notice that for damage states TE and V3, the acceleration capac-l ities are below those of the lowest capacity component in the Boolean expressions (excluding loss of offsite power). This result is expected because for these two damage states single component failures (conceding loss of offsite power) can lead to the damage state.

l

. 0f course, when there is more than one failure that must occur in order to lead to plant damage or when there are redundant safety systems, the reverse .is true. This is the case for damage states AE and

SE for which there must be a seismic-induced LOCA or ATWS and a concurrent J

3-8 f

4 p ,7- - - -,,m-~,. - -

-.,.w-.,.-.,.-.--.m,v-,m,..,g_w..m,,-,m%,- -

c+,-,,- ---m,wcw,,ww---,,-.s,. , , , - -

seismic failure of the mitigating systems. Here, the capacity is at least as great as the capacity of the quantity enclosed by the brackets on either side of the "and" (A) symbol (Eqs. 3-2, 3-3). For damage state AE, the capacity is governed by the failure of the reactor coolant system piping (large LOCA) and for damage state SE, by the core geometry distor-tion preventing insertion of the control rods (note that the capacities are slightly lower than the capacities of the controlling components since for AE, the large LOCA can actually result from any one of three component failures; 15, 29 or 31 and for SE, the small LOCA or ATWS can result from the failure of components 6 or 9 or 16).

Finally, it is pointed out that since these four damage states represent the major contributors to seismic-induced core melt (NU, 1983),

the Boolean expressions can be combined (according to Boolean algebra) to obtain a Boolean expression for core melt. Due to the appearance of common components in the separate damage state expressions $nd considering the relative capacities of the components, it was found that 3a core melt expression is essentially that given by the damage state TE express-ion. Thus, the conclusions derived for damage state TE are appli-cable for core melt and it is, therefore, concluded that the plant has a significant seismic safety margin against core melt.

3.4 PLANT DAMAGE STATE FREQUENCIES In order to obtain the annual occurrence frequencies of the plant damage states, the family of plant damage state fragility curves (Section 3.3) are convolved with the family of seismic hazard curves (Fig.2-1). In this way a set of doublets for the plant damage state frequency are obtained.

hPj'f i ij (3-8) where f g is the seismically-induced severe core damage frequency and p93 i is the discrete probability of this frequency, given by 3-9

[

i

. Pj i 91 p j (3-9) 3 f

ij fi (a) , da (3-10) l Here Hj represents the jt.!! hazard curve, fj the iS plant damage I stata fragility curve, qi is the probability associated with the 1S

! fragility curve and pj is the probability associated with the j S l hazard curve.

t l The above equations state that the convolution (Eq. 3-10) between the seismic hazard and the plant damage state fragility is carried out by selecting a hazard curve j and a fragility curve 1; the probability assigned to the plant damage frequency resulting from the convolution is the product of the probabilities p j and q) assigned to these two curves. The convolution operation given by Equation 3-10 consists of multiplying the frequency of occurrence of an earthquake peak ground acceleration between a and a+da (obtained as the derivative of Hj with respect to a) with the conditional frequency of the plant damage state, and integrating such products over the entire range of peak

• ground accelerations 0 tom.

-In this manner a probabilistic distribu' tion on the frequency of plant damage state can be obtained. Table 3-3 sumarizes these results for the four plant damage states at Millstone for which the seismic event j is an important contributor _ The table shows the median', mean and 5%-95%

l confidence bounds on the annual occurrence frequency. Note, again, that this statistical description'of the occurrence frequency results from i

uncertainty in the hazard (represented by 10 subjectively weighted curves, fig. 2-1) and uncertainty in the damage state fragility (represented by 11 weighted' curves).

i l From Table 3-3, it is seen that even the 95% confidence occur-rence frequencies are all very low. .The sensitivity of these results to variations in the fragility and seismic hazard modeling is explored in Chapter 4.

3-10 E

Again, as was mentioned in Section 3.3, the conclusions for damage state TE are applicable for the core melt accident. Hence, it may be concluded, that the occurrence frequency of core melt is very low and is significantly below that due to internal events (NU, 1983).

3.5 CONTPI8tQ01 TO SEISMIC RISK FROM DIFFERENT ACCELERATION RANGES l

I The annual occurrence frequencies of the plant damage states are obtained, as mentioned earlier, by convolving (i.e., integrating over the entire range of acceleration values) the seismic hazard curves with the l plant damage state fragility curves. The question has arisen as to what ranges of acceleration contribute most significantly to the overall frequency of occurrence of the damage state. For small accelerations, the frequency of occurrence of the earthquake is large, while the frequency of plant damage is small. For large accelerations, the fre-quency of occurrence of the earthquake is small while the frequency of plant damage is relatively large.

~

This question can be addressed by performing the convolution (integration) over small acceleration ranges and comparing the occurrence l

frequency obtained with that obtained by integrating over the entire range of accelerations. Figure 3-5 illustrates the result of this analysis.

Note that for the larger capacity damage states most of the contribution l

comes from the higher acceleration ranges as would be expected. Note

j. also, that more of the contribution to the 95% confidence frequency comes from lower acceleration ranges than for the median frequency, since the l

95% confidence frequency is dominated by the upper hazard and fragility curves. Finally, what is perhaps the most important observation is that the contribution from the acceleration ranges below 0.39 is very small.

This indicates that very large earthquakes must occur in order for any significant damage to be done to the plant and that the plant is well equippeJ to resist earthquakes significantly larger than the SSE. For damage state TE, (and, therefore, core melt) the major contribution to the median frequency comes from earthquakes with peak ground acceler-ation between 0.45 and 0.85g.

3-11

TABLE 3-1A 1

FRAGILITIES OF KEY STRUCTU'tES IN MILLSTONE 3 5%-95%

Confidence High Conficence, f v Bounds on Low Frequency A 8 6 Median Failure level No. Symbol Structure / Failure Motle (g) R U (g) (g)

) 3 - EGECLPSE Emergency Generator 0.88 0.20 0.46 0.41-1.87 0.30 Enclosure Building 4 RWST Refueling Water Stora'ge 0.88 0.30 0.36 0.49-1.59 0.30 i Tank (Buckling)

I 7 DFCNTBLD

• Control Building 1.00 0.24 0.43 0.58-1.72 0.39 (Diaphragm Failure)

{

8 CNTRLBLD Control Building (Sliding) 1.20 0.21 0.47 0.56-2.59 0.39 12 SWPHSLID Service Water Pumphouse 1.30 0.24 0.49 0.58-2.90 0.39

) (Sliding)

, 13 EGESLIDE Emergency Generator 1.30 0.24 0.46 0.61-2.76 0.41 i Enclosure Building (Sliding)

14 AUXBLDG 1.40 0.37 0.41 0.71-2.74 0.39 Auxiliary (Shear WallBuilding Failure)

! 19 SWPHCOLL Service Water Pumphouse 1.60 0.33 0.34 0.92-2.79 0.53

(Shear Wall Failure) i 20 ESFBLDG Engineered Safeguard 1.70 0.23 0.43 0.84-3.44 0.58

{ Features Building (Basemat/

Shear Wall Failure) 27 CONTWALL' Containment Crane Wall 2.20 0.39 0.38 1.18-4.10 0.62 i

TABLE 3-1B FRAGILITIES OF KEY EQUIPMENT IN MILLSTONE 3 51-951 Confidence High Confidence.

. Sounds on Low I w uenc 8

ho. Symbol Equipment /Fallure Mode (g) 8 R 0 f g) 1 LOSP Loss of Offsite Powe 0.20 0.20 0.25 0.13-0.30 0.10 f.Cere itc Insulator Failure) 2 RECRHTER Containment Rectrculation 0.82 0.32 0.52 0.35-1.92 0.21 Heat Enchangers 5 EDG0ILCL Emergency Otesel Generator 0.91 0.24 0.43 0.45 1.84 0.30 (Oil cooler anchor bolt fatlure) 6 COREGEOM Reactor vessel Core 0.99 0.21 0.33 0.58 1.70 0.35 Geometry Distortion 9 CROS Control Rod Ortve System 1.00 11. 3 0 0.38 0.54-1.86 0.33 (Failure to SCRAM) 10 4PCWPUMP Camponent Cooling 1.13 0.25 0.33 0.66-1.94 0.44 Water System Punes 11 SWPIFE Service Water System 1.30 0.24 0.49 0.58-2.90 0.39 Ptping (due to pumphouse sildtag) 15 RCSPIPE Reactor Coolant System 1.59 0.48 0.51 0.69-3.67 0.31 Piping (Large LOCA) 16 RC35MPIP Reactor Coolant Systee 1.59 0.48 0.51 0.69 3.67 0.31 Ptping (Small LOCA) 17 OW5T Deatneralized Water Storage 1.60 0.25 0.43 0.79-3.24 0.52 Tank 18 TSOAFWP Turbine Driven Aux F.W. 1.60 0.30 0.26 1.04-2.45 0.64 Pua, 21 RECPUMP5 Containment Retirculation 1.71 0.37 0.38 0.92 3.19 0.50 System Pumps 22 PORY Pouer Operated Relief Valves 1.84 0.49' O.29 1.14-2.96 0.51

. (loss of operability) 23 CVCSPIPE Cheetcal Volume Control 2.17 0.42 ' O.52 0.92-5.09 0.46 System Piping 1

/

3-13 P

-- vs- ---w- ,- ,

l l

l I

i TABLE 3-1B FRAGILITIES OF KEY EQUIPMENT IN MILLSTONE 3 (Continued) t l

l 5 95#

Confidence High Confidence.

. Sounds on Low Frequency No. Syeol Equipnent/ Failure Mode (g) f a  % g}

24 RECRPIP Containmeni Recirculation 2.17 0.42 0.52 0.92-5.09 0.46 System Piping 25 RPCWPIPE RCP Component Cooling Water 2.17 0.42 0.52 0.92 5.09 0.56 System Piping l

26 QSPIPE Quench Spray System Piping 2.17 0.42 0.52 0.92-5.09 0.46 28 MCCFAIL 480V Motor Control Centers 2.21 0.28 0.57 0.87-5.63 0.55 (trip) 29 RIVESSEL Reactor vessel (support 2.35 0.48 0.44 1.14 4.84 0.52 pads fati) 30 SWPUMP5 Service Water Systen Pupps 2.40 0.31 0.53 1.01-5.72 0.61 31 RCPUMP5 Reactor Coolant Pumps 2.68 0.43 0.47 1.24-5.79 0.61 (largeLOCA) l 32 RCP5WHEC RCP 5eal Water Heat 2.68 0.46 0.40 1.39-5.16 0.65 l

Inchanters 33 RPCWMTEX RCP Component Cooling 2.68 0.46 0.40 1.39-5.16 0.65

, Water Heat Exchangers 34 CASTRAY Cable Trays 2.70 0.48 0.42 1.36-5.3B 0.62 35 05PLMP5 Quench Spray System Pumps 2.93 0.23 0.68 0.96-8.94 0.66 36 QSMEADER Quench Spray system Neader 3.07 0.62 0.37 1.67-5.63 0.51 Piping

! 37 STEEN 5 team Generator 4.00 0.44 0.48 1.82-8.79 0.08 Support Failure l

I 9

3-14 i

TABLE 3-2 l

FRAGILITIES OF DIFFERENT PLANT DAMAGE STATES i

5%-95%

. Confidence High Confidence Plant Damage State A Bounds on Low Frequency (g's) Migjg ofFagygLevel V3 LOCA w/ containment bypass 2.05 1.16 - 3.45 0.60 AE Large LOCA with Early 1.22 3.75 - 1.91 0.45 Core Melt SE Small LOCA or ATWS with 0.77 0.58 - 1.04 0.40 Early Core Melt TE Transient (loss of 0.61 0.39 - 0.84 0.26 offsite power) with Early Core Melt 3-15

TABLE 3-3 SEISMICALLY-INDUCED ANNUAL FREQUENCIES OF PLANT DAMAGE STATES nnual Fmquency Plant Damage 5% - 95%

State Median Mean Confidence 8ounds V3 2 x 10-9 1 x 10-7 0 - 7 x 10~7 AE 8 x 10-8 7 x 10-7 1 x 10 3 x 10-6

' SE 4 x 10-7 2 x 10-6 2 x 10 8 x 10-6 TE 2 x 10-6 6 x 10-6 2 x 10 2 x 10-5 4

s l

l I

3-16 l

l

4m-e a h %_ah--.i.-.a 6_ .w __MA..&JA*4-.4. h-E e. -aan.- . :e.e .*..e_+- am---. a-

. LI~C CONDITIONAL FREQUENCY OF FAILURE 1.00 g.co o,.20 0,.40 0,.60 0,.80 ,

O 1 0 P1 D

X a ?-

xn GO C

' Z 8 O A -

w D P-

,'. Oe nO rT1 o r-- e

$ m a, a 2 m Ir_

$ du M -O O

" Z m -

a i m

O_ '

R -

Z

• a W O

O

=

G N i e i I O

O 9

9

l l

l l

0 e

o r

O O

\ i ' i N

3

) 5

\

nc g

-s s

- -r o d w r S

Z B O m OW wA $

• • m O -T D" E-  %.

1 OO e on

- a e

w

\ .

a -

\ Z

• ."k

\ \

t

• C c x-Q m

\

\

-? a

!- X 33 170 0 7 1

00.pc l00 1 oe.o os.o o4.o o2.o ERULIRF FO YCNEUQERF LANOITIDNOC c-ts -

l i -. . .__ -. . .

a .6 *Ame. a,u . J ,s. . m-.. 4 m-- e, a _%* .J .Aa * .4 ma aA.-. - 4 _ _

61'-E CONDITIONAL FREQUENCY OF FAILURE J.00 0.20 0.40 0.60 0.80 1.00 l O j 3 0 i

M

~D X

O P- -

Zm

.n -OO e i 5 C $  ;

sii m

z i O

Y m

." I P- R -

n eo a OO

1. M F

m m =

4 r._ _

m m m wO I

n G  !

E Z S

C

! a-

\

a cn O

E Di ro . . , , i O

O S

m. m_a - - ..w..%..as,,es - w=e-e- a.-w*.A=-

OZ-E l

CONDITIONRL FREQUENCY OF FAILURE 0,.20 0.60 0,.80 1.00 J. 00 0,.140 O ~

"D U r I

X O.~

7) .c .

0 0 m C m E Z O kO )

t'* O c._ -

n cn u, 00 "

o TT1 h m a u' m

y 'D . . -

M --% \

4 "O

m G m Z E

a h O. - ~ -

m 9 O O

tv ' I f g O

O O

e e e

l I

L

(3-Sa) PLANT DAMAGE STATE TE l

l

. n i n

'50 -

50 -

Median Frequency 95% Confidence Frequency e*

I ~

, .m -

• 25 -

25 - ._

"O- 4

=

1 2 -

M e nm w-d 1 ,-

dr  % _

- i i

i i 1

0.5 1.0 1.5 0.5 1.0 1.5 Peak Ground Acceleration (g) Peak Ground Acceleration (g) l l

(3-5b) PLANT DAMAGE STATE V3 t n j db 50 -

l 50 -

Median Frequency 951 Confidence Frequency 25 25 -

~ * - -

~_ _ _ _

i F i

. h  :

F~ 1 I

T1 I 0.5 1.0 1.5 O.5 1.0 1.5 PeakGroundAcceleration(g) Peak Ground Acceleration (g)

FIGURE 3-5., . PERCENT CONTRIBUTION OF DIFFERENT ACCELERATION RANGES l

l 3-21

PERui.NT CONTRIBUTION OF DIFFERENT ACCELERATION RANGES (3-Sc) PLANT DAMAGE STATE AE s,

50" - 50 -

Median Frequency 95% Confidence Frequency 25 - 25 -

m- "?

• M O C. - -

m.N .e m

W

b. M .

i i 1 i i i i l O.25 0.5 1.0 1.5 0.25 0.5 1.0 1.5 PeakGroundAcceleration(g) PeakGroundAcceleration(g)

(3-5d) PLANT DAMAGE STATE SE 50 - 50 -

Median Frequency 95% Confidence Frequency

' 25 - 25 -

. W

~~

L' -

4_

i .-

l w si no% ~

l o[ dd i

h_i i~

4 I i i ' i l

0.25 0.5 1.0 1.5 0.25 0.5 1.0 1.5 l

PeakGroundAcceleration(g) PeakGroundAcceleration(g) l I

l .e l FIGURE 3-5. PERCENT CONTRIBUTION OF DIFFERENT ACCELERATION RANGES (Co'itinued) l 3-22 b

l

4. SENSITIVITY STUDIES 4.1 OVERALL APPROACH l- During the review of the PSS by the NRC staff and their consultants (USNRC,1984), questions regarding the sensitivity of seismic I

risk estimates to different modeling assumptions were raised. Some of these sensitivity issues have been addressed on a generic basis (Ravindra, et al, 1984). In this chapter, sensitivity studies performed using plant-specific information on Millstone Unit 3 are described.

The overall approach to evaluating the sensitivity of seismic risk estimates is as follows. The plant damage state fragilities and

unconditional frequencies of occurrence due to seismic events discussed

{ in Chapter 3 were calculated using fragility and hazard models judged to be realistic 1.e., " base case." If a specific modeling parameter value or assumption is considered to be different, its impact is assessed by recomputing the plant damage state fragilities and frequencies of occur-l rence. Note that each modeling parameter is treated separately 1.e.,

several parameters are not assumed to be different simultaneously from the base case values.

7 The significance of changes in the plant damage state occurrence

~

l frequencies will be addressed on a case by case basis. Whether or not a <

partit.ular change is to be taken as significant must consider not only the percentage change in the occurrence frequency but also the actual level of the occurrence frequency. For example, if the occurrence frequency of a plant damage state is of the order of 10-8, increases of less than an order of magnitude are insignificant. This is based on a consideration of both the relative accuracy of risk estimates at low occurrence frequencies and NRC priority rankings for effecting fixes to l reducerisk(USNRC,1983).

l l

4-1 l

l

. . _ . _ . - - _ _ . _ _ . . . _ , _ - - . . - - - _ . - _ _ _ _ . -~-_

l l

4.2 COMPONENT FRAGILITIES The influences on the plant damage state frequencies of different assumptiens made in the development of component fragilities are examined in this section. Effects concerning fragility lower tail truncation l

l (cut-off), dependence between component failurc.s, variations in fragility i parameters, and the assumption of a lognormal distribution model are discussed.

4.2.1 Fragility Cut-off Thb lognormal model used for fragility description admits very low capacities of components as being possible. However, experience has shown that properly designed and built components rarely, if ever, fail below their design levels. Hence, a cut-off on the lower tails of fragility curves is thought to be appropriate. The annual plant damage state occurrence frequencies reported in Table 3-3 were calculated using the following cut-off model:

The cut-off on the lower tail of the median (50 percentile) fragility curve was taken as:

k exp(-28C ) (4-1)

C0 =

l l whereECO is the cut-off on the median curve, 5 is the median peak ground acceleration for failure and BC is the composite logarithmic standard deviation, i.e., SC * / Bg 243U The cut-off for the lower tail of other fragility curves was taken as:

l A C0 exp(-x 8C /1.65) (4-2)

CO "

l t

4-2 l

1 l

i-1

l l

l where x is the ratio of the deviation from the median divided by the standard deviation. For instance, x=0 for the median fragility curve; for the 95 percentile curve and above, x=1.65 and for the 5 percentile curve and below, x=-1.65. Figure 4-1 shows the fragility cut-off model.

In order to assess the sensitivity of the plant damage state frequencies to the fragility cut-off model used, the plant damage state frequencies I

given in Table 3-3 were recalculated using fragility curves without any lower tail cut-off. Table 4-1 shows a comparison between the plant damage state frequencies calculated for these two fragility models. It is observed that the fragility cut-off had hardly any effect at all on the plant damage state frequencies for Millstone Unit 3 except for the ,

median occurrence frequency of state V3. The reason for the lack of sensitivity is that the contribution to the plant damage state frequencies from acceleration ranges below the fragility cut-off level is generally negligible. This is so even when the fragility curves are -

4 cutoff well above the plant SSE of 0.17g. For example, the emergency generator enclosure building (wall footir.g failure) fragility curves were cut-off at the following values:

kCO = 0.32g and 95 percentile AC0 = 0.19g i It is only in the case of damage state V3, for which the e ound acceleration capacity is very large, that the lower tail of the frilility curve contributes to the risk. This is because the hazard curve vabes are extremely low at the median acceleration capacity of damage state

V3. It is only such cases in which the cut-off model employed would show a noticeable effect. However, these cases involve damage states with l

high capacities and low occurrence frequencies. In light of this and the l

l discussion is Section 4.1, the change 'of occurrence frequency exhibited for damage state V3 is considered insignificant. Therefore, it is

( concluded that the fragility cut-off model utilized in this study has not significantly biased the plant damage state frequency estimates.

l l

4-3 i

l .

. _ _ _ _ _ . ~ - - _ _ _ _ . _ . _ _ _ . _ . _ ___.___ _ _ _ . . _ _ _ _ . . , ___ _. _ _.__...._ _,__.._..._.__ _ _ _,.

4.2.2 Interaction between Failure Modes For some of the Millstone Unit 3 structures more than one potential failure mod? was indentified. In the PSS these failure modes were assumed to be incependent, that is, knowledge of the occurrence of one mode does not influence the occurrence probability of another mode.

Thus, for example, the diaphragm failure and the sliding failure of the control building were modeled as two separate failures. This assumption is conservative. For the best estimate " base case" results presented in Chapter 3 of this study, different failure modes for the same structure were assumed to be perfectly dependent. Thus, the capacity of the structure is governed by the lowest capacity failure mode. Clearly, the correct approach lies somewhere in between these two extremes.

In order to assess the sensitivity of the plant damage state frequencies to the extreme assumptions of perfect dependence and independence, the plant damage state frequencies were recalculated, herein, assuming that the failure modes are perfectly independent.

Table 4-2 shows the results of this study. As can be seen, there is essentially no difference no matter which assumption is used. This results from the fact that, in the perfectly independent case, failure modes with capacities significantly larger than the capacitis of components and failure modes already considered are being added. Thus, the total failure probabilities are not significant1v increased.

4.2.3 Dependence between Component Failures Since all components are simultaneously being excited by a single earthquake some dependence is likely to exist between earthquake-induced failures. For the best estimate " base case" study the Boolean expression for the different plant damage states were reviewed:

the Booleans were found to consist of highly dissimilar items of structures and equipment (i.e., buildings, tanks, electrical equipment, control rod drive system, RCS piping, and service water system pumps),

their locations in the plant and within the structures were found to be different and their dynamic characteristics were also judged to be 4-4

different. Hence, correlation in the seismic responses and seismic capacities of these components was assumed to be minimal. As such, the component failures in these Boolean expressions were treated to be j statistically independent. Note that in the Millstone PSS independence j between component failures was also assumed.

l In order to evaluate the significance of this assumption, a separate case study was performed wherein the component failures were ,

treated as perfectly dependent, i.e., both the uncertainty and randomness parameters utilized to describe the component fragilities were assumed to be perfectly correlated. Assumptions of perfect dependence in the uncer-tainties of different component fragilities means that the median ground acceleration capacities of all components are known if the median ground acceleration capacity of one component is given. Since the uncertainty arises from the insufficient understanding of structural material properties, approximate modeling of the structure and inaccuracies in the mass and stiffness representations, and the use of engineering judgment in lieu of plant-specific data, it is expected that all the components will be affected to some degree by these uncertainties. Therefore, some probabilistic dependence between component median capacities may be 1

expected. Perfect dependence in uncertainty is an extreme case.

t

, Dependence in the randomness arises from a consnon earthquake generating the responses in different components and consnon structural /

material properties. Assumption of perfect dependence in the randomness means that if the fragility of a component for a given peak ground accel-eration is known, the fragility of the other components is modified by l

! that knowledge. For example, the fragility of the " union" of the component failures A and B is given by:

! ' f@g f@k~f@ikI@ 2 f O

ik i'

k (4-3) l [ [

l Perfect Independence Perfect dependence l

For the " intersection" of component failures-i 4-5

_. _ ._ _, _ _ , _ _ _______._.____._a_____.._____

t I@i f@ k ,

ik t "I" ,@i'I k B

,(4-4) l T T i Perfect Independence Perfect dependence

! These two assumptions of perfect dependence and independ-ence are expected to yield realistic upper and lower bounds on the plant damage state frequencies. Note that an assumption of perfect independence )

l between failures of components in the union operation (i.e., those within  !

the parenthesis of the Boolean expressions of AE, SE and TE plant damage ,

states) and of perfect dependence between failures of components in the

" intersection" operation would give an absolute upper bound on the plant damage state frequencies. However, this is clearly an unrealistic l situation and, in any case, would yield occurrence frequencies only l slightly more extreme than those obtained with the bounding assumptions

! used.

Table 4-3 gives the comparison of plant damage state frequencies for the two cases of perfect statistical independence and perfect statis-

tical dependence between component failures. Note that the effects of l the two assumptions are most evident in damage states V3 and TE. For these damage states, occurrence is primari1k,the result of failure of any one. component. For this case, that is, whenA p Boolean expression is the union of component failures (note that for damage state TE, once

offsite power is lost the Boolean is essentially reduced to the union of components) the assumption of statistical indep2ndence is always conservative. Note that the PSS assumed statistical independence between component failures and will bias the results toward the conservative side f for these damage states. Note also, that since damage state TE is the dominant contributor to core melt the independence assumption will bias l

'the core melt frequencies reported in the PSS toward the conservative

(

side. Damage states AE and SE do not show large changes and the effect l of correlation between component failure is not considered significant in light of the discussions in Section 4.1. Finally, it can be seen from Table 4-3, that with either of the bounding assumptions employed, the i

l 4-6

h.igh confidence, low frequency of failure acceleration values are all significantly above the 0.17g SSE level.

4.2.4 Variation of Fragility Parameters The fragility parameters for different dominant structures and equipment identified in Chapter 2 were derived using plant-specific data and realistic models of failure modes and capacities. However, during the NRC review, some alternative models for failure modes and capacities were suggested. In the following, the sensitivity of the PDS frequencies

to these revised fragility models is discussed.

In order to allow for a tractable sensitivity study, a reduced form of the Boolean expressions given in Equations 3-1 through 3-4 was used. That is, those components with relatively high capacities appearing in the union operations were found not to contribute to the risk and were removed. These so-called " reduced" Booleans are listed below for the four damage states:

=

TE 1 ^ (3 v. 5 v 7 v 12 v 20 v R19) (4-5)

AE =

(3 v 4 v 5 v 7 v 12 v 20 v R19) ^ (15) (4-6)

SE =

(3 V 4 v 5 v 7 v 12 v 20 v R19) ^ (6 v 9) (4-7)

V3 = 27 -

(4-8) i l where the notations and symbols have been defined in Section 3.1.1. Also 5 curves rather than 11 curves were used to represent the fragility uncertainty for each component. That is, not quite as fine a discretization on uncertainty was used. The first column of results in Table 4-4 shows the annual occurrence frequeicies obtained when the l

reduced Boolean and the 5 curves discretization is used. Note that only

! minor differences are observed between these results and Sose of the base case (Table 3-3). Since this simplified model is used ve carry l

l 4-7 l

out the fragility parameter sensitivity study, the results in column 1 of Table 4-4 represent a benchmark against which the sensitivity cases can be compared.

4.2.4.1 Earthquake Magnitude Range The fragility analysis in the Millstone PSS assumed that the ma,iority of seismic risk resulted from the earthquakes that have magnitudes between 5.3 and 6.3. This was reflected in the choice of a magnitude 5.8 spectra as the median spectra representative for the site (Dames & Moore, 1983). Further analysis by Dames & Moore has indicated that events in this magnitude range dominate the hazard even for ground motions as large as 0.6g. Therefore, it is appropriate to use a single spectral shape to represent ground motion for calculation of structural and equipment fragilities. The choice of a magnitude 5.8 spectrum is further justified since it represents the stated range.

In its review, the NRC has suggested that the contributing earth-

~

quakes would be in the magnituda range of 5.8 to 6.8. This results in a reduction of the duration factor CD (see Millstone PSS for details) from 1.4 to 1.0. Therefore, the effective ductility of some structures and equipment which fail in the ductile mode may be lower than assumed in the base case. The components affected are the Emergency Generator Enclnsure Building (wall focting failure, EGECLPSE), Control Building diaohragm failure (DFCtfrBLD), Engineered Safeguard Features Building failure (ESFBLDG), Containment Crane Wall (CONTWALL), and Control Rod Drive System (CRDS). With the revised fragility parameters, the high confidence, low frequency of failure acceleration level of the control

[

rod drive system (CRDS) changed from 0.33g to 0.30g. The other components had minor changes in the high confidence, low frequency of failure acceleration value. Table 4-4 shows the small change to PDS frequencies as a result of this assumption of different magnitude range

[

-(Case 1).

i I

l l

l 4-8

4.2.4.2 Limiting Displacement for Attached Piping Some of the Millstone 3 structures (e.g., Service Water Pumphouse, and Emergency Generator Enclosure Building) may begin to slide on their foundations if sufficient seismic response occurs. The initiation of sliding does not constitute structural or equipment failure. Damage to safety-related equiment as a consequence of structure sliding can occur only if the resulting sliding displacements are of sufficient magnitude to cause impact with adjacent structures or failure of attached piping. In the PSS, an approach recomended by Newmark (1975) was used to conservatively determine the peak ground acceleration corresponding to estimated sliding failure-induced displacements.

Failure of buried piping entering a structure was conservatively assumed to occur when the sliding displacement transverse to the pipe axis causes buckling of the pipe. A total median sliding displacement of

approximately four inches was found to be necessary to develop pipe buckling. Typically, two inches of clearance exists between the pipe and

the wall penetration sleeve. Significant pipe forces cannot be developed ,

until the gap between the pipe and the wall penetration is closed. A sliding displacement of.two inches therefore constitutes a lower bound on the displacement necessary to cause pipe buckling. In the PSS, a median structure sliding displacement of four inches was assumed to be necessary to develop pipe buckling and the ground acceleration capacity of the

! structure was estimated correspondingly.

As a result of the review questions by the NRC, the median displacement needed for pipe buckling was assumed to be 2 inches (Case 2) and 6 inches (Case 3) separately. Only the service water pumphouse

!- slidng failure mode was affected. The revised median acceleration capacity of the service water pumphouse was calculated as 1.19 (Case 2) l 'and 1.4g (Case 3), respectively. The median acceleration capacity of l this structure was estimated in the base case as 1.39 The high confidence, low frequency of failure acceleration capacity of the structure was calculated to be 0.349 (Case 2), 0.41g (Case 3) and 0.399

(

i l

4-9 i

i (Base Case). Table 4-4 shows that the PDS frequencies are not significantly changed by the different assumptions in the limiting displacement values.

4.2.4.3 Coefficient of Sliding Friction The service water pumphouse base mat bears upon excavated rock or fill concrete poured on intact rock. A large percentage of the base

!- mat plan area bears upon the excavated rock surface based upon available

excavation drawings. The surface of the fill concrete was raked during construction.

1 Development of median coefficients of friction for the determini-nation of structure sliding resistances was based upon known test results for other interface conditions and engineering judgment. The results of

testing conducted by Gaston and Kriz to determine the coefficient of friction for formed concrete interfaces are reported by Walker (1969).

Individual concrete blocks (not masonry) were cast using steel or plastic-coated plywood forms. Based upon testing of these blocks, a coefficient of friction of about 0.8 was found for these relatively smooth concrete surfaces. The raked surface of the fill concrete below the pumphouse base mat is rougher than the concrete block specimens tested by Gaston and Kriz. A median coefficient of friction of 1.0 was c therefore estimated for the base mat fill concrete interface. The surface of the excavated rock is very uneven and is therefore expected to have higher friction than the raked surface of the fill concrete. Conse-quently, a median coefficient of friction of 1.2 was estimated for the base mat / excavated rock interface. A riedian coefficient of friction of 1.1 was used in determining the sliding resistance of the pumphouse.

This value represents an average of the coefficients of friction for the two different foundation interface conditions. It is considered to be somewhat conservative since a large persentage of the base mat plan bears on excavated rock which has the higher coefficient. -

4 l

i 4-10 t

. . .v.. -..v . - . . . . , _ , _ _ _ _ - . . . . - - . _ , , . , . . . , , , - . _ - - , _ . . - . _ . . , _ - - . _ , _ . _ . . . . . _ _ , . , _ , . . . ~ , . ~ .

I l

NAVFAC DM-7 (Dept. of the Navy, 1971) recommends the use of a l coefficient of friction of 0.7 for mass concrete on clean sound rock when specific test data is unavailable. This is a conservative value for the design of structures and is not appropriate for use as a median value in j fragilities evaluations.

l The sensitivity of PDS frequencies to the asst:ned median coef-

! ficient of friction was studied by assuming the median value was 0.7 as l

contrasted to the value of 1.1 used in the PSS. The median ground accel-

eration capacity reduced from 1.3g to 1.1g. The high confidence, low frequency of failure acceleration of the service water pumphouse was I reduced to 0.34g (Case 4) from 0.39g (Base ~ Case). However, the calculated PDS frequencies did not show any change (Table 4-4).

4.2.4.4 Buckling Capacity Estimates The refueling water storage tank (RWST) fragility was estimated I

for the potential failure mode of tank wall buckling. The capacity of the tank was evaluated using the median buckling stress formula from NASA SP-8007(1965). In the design of such tanks, a lower bound buckling stress coefficient is generally used. The sensitivity of PDS frequencies to the buckling stress formulas used was evaluated. The median acceleration capacity using the lower bound buckling stress coefficient

! was estimated to be 0.73g (Case 5) compared with 0.88g (Base Case). The high confidence, low frequency of failure acceleration capacity value reduced from 0.30g to 0.28g. However, Table 4-4 shows that the PDS frequencies are not changed by this different model for tank capacity.

4.2.4.5 Ratio of Peak Ground Velocity to Peak Ground Acceleration l

! Determination of fragility parameters for structure sliding-induced failure is, in part, dependent on the ratio of the peak ground velocity to the peak effective ground acceleration. A median velocity to L acceleration ratio (v/a ratio) of 28 in/sec/g was selected for use in the Millstone 3 structural fragility evaluation based upon the recommendation .

4-11 l

I, of Newmark (1973) for rock sites. A value of 36 in/sec/g has also been recomended (Newmark & Hall,1978). However, the 36 in/sec/g v/a ratio is not appropriate for the fragility evaluation since it is a

1. conservative value more appropriate for design.

The sensitivity of PDS frequencies to the assumed values of v/a ratio was studied by using two values of v/a 1.e., 17 in/sec/g (Case 6) i and 36 in/sec/g (Case 7). The median acceleration capacity of the l

affected structure (i.e., service water pumphouse) in the two cases were i

estimated to be 1.6g and 1.1g, respectively. The base case capacity was

! 1. 3g'. The high confidence, low frequency of failure acceleration values were 0.50g (Case 6), 0.379 (Case 7) and 0.39g (Base Case). Table 4-4 shows.that the PDS frequencies were not affected by the different assump-tions on v/a ratio.

4.2.5 Lognormal Distribution As explained in Chapter 2, the random variables reflecting the inherent randomness (c ) and the uncertainty (c ) are m deled by lognonnal R U probability distributions in the fragility analysis. The advantages of i this formulation are:

i

1. The entire fragility curve and its uncertainty can be l expressed by three parameters - A,Sp and gg. With the very limited available data on fragility, it is much easier to only estimate three parameters rather than the entire shape of the fragility curve and its uncertainty.

l 2. The formulation given in Equation 2-1 and the lognormal l distribution are mathematically tractable.

l The lognormal distribution can be justified as'a reasonable distribution since the statistical variation of many material properties (Freudenthal, et al,1966) and seismic response variables may reasonably f be represented by this distribution (Newmark,1973; Bohn, et al,1984).

lL Figures 4-2 to 4-4 reproduced from Bohn, et al, (1984) show that the j simulated response data of various structures and equipment, supports this observation.

i 4-12 L

L

W In addition, the central limit theorem states that a distribution consisting of products and quotients of distributions of several variables tends to be lognormal even if the individual distributions are not l

lognomal . The overall factor of safety, which is modeled as the product l of a number of factors, can, therefore, be considered to be approximately lognomally distributed.

l 4.3 Seismic Hazard In the Millstone 3 PSS, the results of the seismic hazard analysis for the site performed by Dames & Moore (1983) were utilized to compute the frequencies of dominant plant damage states. In this section, the sensitivity of PDS frequencies to peak ground acceleration truncation and to different seismic hazard modeling is discussed.

4.3.1 Effect of Peak Acceleration Truncation In the base case study described in Chapter 2, some of the seismic hazard curves were truncated at specified, maximum peak ground acceleration values of 0.6g, 0.8g and 1.0g. The basis for this truncation i is given in Dames & Moore (1983). In order to study the sensitivity of peak acceleration truncation, a modified set of hazard curves was l

developed wherein all the truncated curves were extrapolated on a loglinear scale. The plant damage state frequencies were calculated using this modified set of seismic hazard curves. As seen in Table 4-5, the effect of peak acceleration truncation on PDS frequencies is minor especially at hig5 confidence levels, since the significant risk contribution is from earth pake levels below the cutoffs.

4.3.2 LLNL Hazard Curves l Recently, the Lawrence Livermore National Laboratory has published a preliminary report on the seismic hazard studies it conducted on Eastern United States sites (LLNL, 1984). Figure 4-5 shows the seismic hazard curves developed by the LLNL study for the Millstone 4-13 i

site. It can be observed that the LLNL study predicts higher frequencies of exceedence compared to the Dames & Moore (1983) hazard curves used in 4

the PSS (see Figure 2-1) over the entire range of acceleration values. A critical examination and comparison of the modeling assumptions made in both the LLNL and Dames & Moore (D&M) studies is given in Dames &

Moore (1984). It was identified in this reference that excess conservatism in the LLNL hazard study results comes from (1) the inclusion of all earthquakes with body wave magnitude mb 2 3.75 in the data base as contrasted with the D&M approach of including earthquakes with m p 4.5 and (2) use of attenuation models wherein no limits on the peak ground acceleration were placed at any given magnitude and distance from the source. Two alternative sets of hazard curves have been provided (Dames & Moore, 1984). Set I was based on inclusion of only earthquakes of m b 2 4.5 in the data base and maintaining all other features of LLNL hazard model and Set 2 was developed using acceleration truncation in the attenuation equations along with the inclusion of earthquakes of magnitudes larger than 4.5 only.

The plant damage state frequencies were recalculated for these three sets of seismic hazard curves - the original LLNL hazard model and revised LLNL hazard model Set 1 and Set 2. Table 4-6 shows these l

frequenies. It is seen that the revised LLNL hazard model (Set 2) wherein I

the excess conservatisms have been removed, gives results comparable to those obtained in the base case (D&M hazard model).

l Although the plant damage state frequencies give valuable informatio1 on the seismic safety of the plant, they are sensitive to the seismic hazard curves used in the analysis. However, the contributions to the frequencies of occurrence of significant plant damage states by earthquakes in the range of 0.259 to 0.35g are small even when the original LLNL hazard curves are used. Figure 4-6 shows these results in the form of histograms and compares similar histograms obtained using the Dames & Moore hazard curves. This reinforces the earlier conclusion that very large earthquakes must occur in order for significant damage to be done to the plant and that the plant is well equipped to resist earth-quakes significantly larger than the SSE.

4-14

TABLE 4-1 l

1 l

EFFECT OF FRAGILITY CUTOFF ON ANNUAL FREQUENCY OF PDS Median 95% Confidence Plant Damage without with without with State cutoff cutoff cutoff cutoff V3 8 x 10-9 2 x 10-9 7 x 10-7 7 x 10-7 ,

AE 8 x 10-8 8 x 10-8 3 x 10-6 3 x 10-6 SE 4 x 10-7 4 x 10-7 8 x 10-6 8 x 10-6 TE 2 x 10-6 2 x 10-6 2 x 10-5 2 x 10-5 4-15

i TABLE 4-2 EFFECT OF CORRELATION BETWEEN FAILURE MODES ON ANNUAL FREQUENCIES OF PLANT DAMAGE STATES Annual Frequency High Confidence -

95% Confidence

• F 9 Median p 1u e erationfol Plant Perfectly Perfectly Perfectly Perfectly ' Perfectly Perfectly

• Independent Dependent Dependent Independent Dependent Independent, Failure Failure Failure Failure F.aflure Failure Modes Modes Modes Modes Modes Modes V3 2x10-9 2x10-9 7x10-7 7x10-7 0.60 0.60 3x10

-6 3x10-6 0.45 0.45 AE 8x10-8 8x10-8 SE 4x10-7 4x10-7 8x10-f 9x10-6 0.40 0.40 TE 2x10-6 2x10-6 2x10-5 ( 2x)f 5 0.26 0.26 l

4-16

1 l

I l

TABLE 4-3 I

l 1

l EFFECT OF DEPENDEf'CE BETWEEN COMP 0NENT FAILURES l

BOUNDING ASSUMPTIONS l

l Annual Frequency High Confidence -

Low Frequency of Failure Median 95% Confidence Acceleration (a)

Plant Perfectly Perfectly Perfectly Perfectly Perfectly Perfectly Damage Independent Dependent Independent Dependent Independent Dependent State Failures Failures Failures Failures Failures Failures V3 2x10-9 8x10-10 7x1077 6x10-7 0.60 0.62

-8 8x10-8 7x10-6 0.45 0.31 l AE 8x10 3x10-6 SE 4x10-7 2x10-7 8x10-6 9x10-6 0.40 0.33

-6 -5 2x10-5 0.26 0.30  ;

l TE 2x10 4x10-7 2x10 l

l l

l 4-17 l

_FEAGILITY SENSITIVITY CASES

, VARIATION OF ANNUAL PDS FREQUENCIES h

Case 1- Case 2 Case 3 Case 4 Case 5 Case 6 Case 7

. Plant Reduced

{

Damage Boolean h[pjf h[p$f Concrete: Rock B ck ng ST yj, , yj, ,

State 5 Curves EQ Mag. Capacity Capacity y = 0.7 17 in/s/g 36 in/s/g t

-2 =6 NA NA NA Median 3 x 10-9 4 x 10-9 NA NA NA

! V3 NA NA NA j 95% 7 x 10-7 8 x 10-7 NA NA NA i

5 x 10-8 5 x 10

-8 5 x 10

-8 5 x 10-8 5 x 10-8 5 x 10-8 4 x 10

-8 5 x 10 -8 I Median 2 x 10

-6 2 x 10-6 2 x 10-6 2 x 10-6 2 x 10-6 2 x 10-6 2 x 10-6 2 x 10-6

7 95%

I E

-7 3 x 10-7 3 x 10-7 3 x 10-7 3 x 10-7 3 x 10-7 Median 3 x 10 4 x 10-7 3 x 10 7 x 10-6 6 x 10-6 3 x 10-6 6 x 10-6 6 x 10-6 6 x 10-6 6 x 10-6

95% 6 x 10-6 i

i 1 x 10-6 2 x 10-6 j Median 2 x 10-6 2 x 10-6 2 x 10-6 2 x 10-6 2 x 10-6 NA i TE 2 x 10-5 2 x 10-5 95% 2 x 10-5 2 x 10-5 2 x 10-5 2 x 10-5 2 x 10-5 NA

}

• i i

i t

i

TABLE 4-5 EFFECT OF PEAK ACCELERATION TRUNCATION Median Mean 95% Confidence Plant Damage Base No Base No Base No State Case Truncation Case Truncation Case Truncation V3 2x10-9 7x10-9 1x10-7 1x10-7 7x10-7 7x10-7

-AE 8x10-8 9x10-0 7x10-7 7x10-7 3x10-6 3x10-6 SE 4x10-7 4x10-7 2x10-6 2x10-0 8x10-6 8x10-6 TE 2x10-6 2x10-0 6x10-6 6x10-6 ,

2x10-5 2x10-5 I

i-4-19 e

er e- sv' w m-e -eT-=*e *Tvt*t s>PN'-=v?-Nw-"T"P+==P----r- -m'-9't+977 i *'ef'F*-e*T--'e r""W"?9--T-'F n-'e t--** y-Tmwe-**pm-e4 v-mv 7- Te-

TABLE 4-6 l

l EFFECT OF DIFFERENT SEISMIC HAZARD MODEL ANNUAL FREQUENCY OF PLANT DAMAGE STATE l

l Revised LLNL Hazard Model Plant D&M LLNL ("b, min = 4.5)

Damage Hazard Hazard State Model Model No Acceleration Acceleration Truncation Truncation Set 1 Set 2 Median V3 2 x 10-9 5 x 10-7 2 x 10-7 2 x 10-8 AE 8 x 10-8 4 x 10-6 2 x 10-6 2 x 10-7 SE 4 x 10~7 1 x 10-5 5 x 10-6 8 x 10-7 TE 2 x 10-6 3 x 10-5 1 x 10-5 2 x 10-6 95% Confidence V3 7 x 10-7 3 x 10-5 2 x 10-5 2 x 10-6 i

AE 3 x 10-6 1 x 10-4 6 x 10-5 1 x 10-5

, SE 8 x 10-6 3 x 10~4 1 x 10~4 3 x 10-5 TE 2 x 10-5 7 x 10~4 3 x 10-4 8 x 10-5 i Mean V3 1 x 10-7 9 x 10-6 4 x 10-6 5 x 10-7 AE 7 x 10-7 4 x 10-5 2 x 10-5 2 x 10-6 SE 2 x 10-6 8 x 10-5 3 x 10-5 6 x 10-6 l TE 6 x 10-0 2 x 10-4 8 x 10-5 2 x 10-5 l

l 4-20

(

l l

1-95% 50% 5%

Conditional Frequency og 0.5 - - - - - - ----

Failure ,

I C

1 I

I O

a h

a 5 A Peak Ground Acceleration dexp(-3sC)I d exp(-2sC )

Iexp(-sC) i l

FIGURE 4-1. FRAGILITY CUT 0FF MODEL 4-21

1.0  ;

, . . g . , ,

,. . , . g . ,

-(b)

-(c)

. - (a) 1 . . .

, 0.5 -

Median: 19.59 ft/s2- - 2 Modeen: 3.01 ft/s .

Sets: 0.33 - -

. Median: 19.90 ft/s - a -

Beta: 0.4 -

",. Sota: 0.40 5 0 ' I I ' I ' '* ' I ' l ' l ' l ' l ' l '

l g

l 10 20 30 40 10 20 30 40 2 4 6 g 1.0

i . . i ;i. .,, , . .

- (d) .

-(e)

-(f) ,

0.5 -

Median: s.90 tils z

,,,;,, 3,73 ,,f,2 --

Beta: 0.33 Beta: 0.27

. . - Median: s.31 ft/s2 _ _ _

_ . . . Gets: 0.38 - - -

0 I''''I '* ' I ' ' ' ' ' ' ' I ' ' ' I 5 10 15 5 10 2 4 6 8 Peak acceleration (ft/s2 )

i (a,b,c) X, Y, Z components for the top of containment shell.

l (d.e,f) X, Y, Z components for the internal structure.

NOTE: The X direction is west, Y is south, and z is vertical; calculated response is shown dotted and the fitted lognormal distribution is by a solid line.

l l

l FIGURE 4-2. CUMULATIVE DISTRIBTUION OF COMPUTED PEAK STRUCTURAL ACCELERATIONS l

l l

I 4-22

- _ . _ - . . . ~ , _ _ _ _ . _ _ _ _ _ _ - _ . . ~ - _ . _ _ . _ _ _ _ _ _ _ _ _ _ . _ _ _ . _ _ _ _ .

i 1.0 1 . . p i..i. . ,

i *

-(s) .

- (h) -

05  ? -

2- Medien: 9.74 ft/s2-_

l ,

Medien: 10.21 ft/s , ,

Osts: 0.34 y Sete: 0.38 .

l , . ,

5 I ' I ' I I'I

0 10 20 l 10 20 30 o ^

g. 10 i

.,., , , ,. , , , , , , , , i

i y , i

!= - (i) -

- (j) <

- (k) 0.5 -

2 ,

• Medien: 2J7 ft/s2,

. Modeen: 5.87 ft/s , , ,

. Sees: 0.30 . . . . Bete: 0.42 .

! . . .  ; Medien: 5.03 ft/s 2 , ,

. Sete: 0.42 , , ,

'I ' ' ' ' I ' ' " ' ' I ' ' ' ' I ' I ' I ' I ' I 0

. 5 10 5 to 1 2 3 4 5 Peak accelerstion (ft/s ; 2

! (g,h)X,Ycomponentsforthesteamgenerator.

1 l

(i,j,k) X, Y, Z components for a point near the control room in the AFT complex.

! NOTE: The X direction is west Y is south, and Z is vertical; calculated l response is shown dotted and the fitted lognormal distribution is by a solid line.

l i

l l

FIGURE 4 3. CUMULATIVE DISTRIBUTION OF COMPUTED PEAK STRUCTURAL ACCELERATIONS t

4-23

1.0 i g i i i . 1 i

i i i

, 1-

. (a) ,

• - - (b) .

. . a Median: 20.11 ft/s2 . Median: 780.5 ft-4b 0.5 - ~ ~

• ~~

, Seta:0.48 Sete: 0.50 Mpe sias:

Mpe slae: -

E , OD = 3.5 in. CD = 3.5 in.

Wall thickness = 0.438 in. - Wall thickness = 0.438 in.

0 'l I I I I I I ' ' ' I ' I 10 20 30 40 50 80 70 1000 2000 3000 2

Peak accelermion (ft/s ) f%sk monent (ft-lb)

(a) auxiliary feedwater system, inside containment--check valve.

(b) auxiliary feedwater systems, inside containment--nozzle.

NOTE: The X direction is west, Y is south, and 2 is vertical; calculated response is shown dotted and the fitted lognormal distribution is by a solid line.

FIGURE 4-4. CUMULATIVE DISTRIBUTION FOR PEAK RESULTANT RESPONSES r

l l

l l

l l

l l

4-24

_- ,.._-..,n-,., ..-.,-.,~.-.-w,-. . , , - - - - - - . , . . - - - _ _ , , , . - - - - - - - - - , - e c - - . - - + - -

i

-1

10 . . . . . . . . .

1

-2 10 - -

>- -3 l

C 10 -

d ,

85 < -

I" s

a. a

\ x % Confidence -

w to - m\ N  % ,

g Median 8  % '*%s, u

l 5 -e 10 m  % - .

5% Confidence ~

m%

f -4 10 m

! x N

-7

• =

10 ' ' ' ' ' '

j ga w

a

• ACCELERAT10N ChVSCC"2

= ~ =

l l

l FIGURE 4-5. SEISMIC HAZARD CURVES FOR MILLSTONE (LAWRENCELIVERMORENATIONALLABORATORY)

I 4-25 -

(

l DAMES & MOORE HAZARD CURVES l

l J L 50 -

l 50 -

! Median Frequency 951 Confidence Frequency N

.m -

25 -

25 -

9~- 4 0 - =

~ * -

~ .

n e' ,2 o f"~ h _

d 1- .

i i 1 i i 0.5 1.0 1.5 0.5 1.0 1.5 PeakGroundAcceleration(g) Peak Ground Acceleration (g)

LLNL HAZARD CURVES 1

I d(

50 '

50 L

i Median Frequency 95% Confidence Frequency 25 -

9 25 -

e e -e m

-=-

o -

NN

~

9N N __-

l 0

I i I

.~

I

} I h

I O.5 1.0 1.5 0.5 1.0 1.5 PeakGroundAcceleration(g) PeakGroundAcceleration(g)

FIGURE 4-6. PERCENTCONTRIBUTIONOFDIFFERENTNCCELERATIONRANGES-PLANTDAMAGESTATETE 4-26 l

L

DAMES & MOORE HAZARD CURVES I

db I

50 - 50 -

Median Frequency 95% Confidence Frequency 25 - 25

~

F ,

. h ,-

. 2fT , ,

_T1 .-

0.5 1.0 1.5 0.5 1.0 1.5 PeakGroundAcceleration(g) Peak Ground Acceleration (g) l t

LLNL HAZARD CURVES a t a tSO -

50 l Median Frequency 95% Confidence Frequency 25 -

25 i -

m- _

ww -

l X

I I i X

i

~ ~

i D. I' l 0.5 1.0 1.5 0.5 1.0 1.5 PeakGroundAcceleration(g) PeakGroundAcceleration(g)

FIGURE 4-7. PERCENT CONTRIBUTION OF DIFFERENT ACCELERATION RANGES - PLANT DAMAGE STATE V3 4-27

I DAMES & MOORE HAZARD CURVES a

50 -

Median Frequency 95% Confidence Frequency 25 -

, _ m, d "

q 2 q

. m

,q* -

i r-i i hm, l

U. i i h _

0.25 0.5 1.0 1.5 0.25 0.5 1.0 1.5 PeakGroundAcceleration(g) PeakGroundAcceleration(g)

LLNL HAZARD CURVES o n 20 50 Median Frequency 95% Confidence Frequency l

85 25

_ 9 _

w R- ~_ m w2 h i i l .

rN i

Ti i l

0.5 1.0 1.5 0.5 1.0 1.5 PeakGroundAcceleration(g) PeakGroundAcceleration(g)

DIGURE4-8. PERCENT CONTRIBUTION OF DIFFERENT ACCELERATION RANGES - PLANT DAMAGE STATE AE 4-28

i

. DAMES & M0 ORE HAZARD CURVES l

i 50 -

Median Frequency 95% Confidence Frequency 25 -

d - R _

,C - og~ -

wd ma A -

N _

8 I I I ~ i i i i' O.25 0.5 1.0 1.5 0.25 0.5 1.0 1.5 Peak Ground Acceleration (g) PeakGroundAcceleration(g)

LLNL HAZARD CURVES o n 50 -

50 Median Frequency 95% Confidence Frequency 25 - 25 ud

, b~-- ~

d h .

N I

~[

i I' I i O.5 1.0 1.5 0.5 1.0 1.5 PeakGroundAcceleration(g) PeakGroundAcceleration(g)

FIGURE 4-9. PERCENT CONTRIBUTION OF DIFFERENT ACCELERATION RANGES - PLANT DAMAGE STATE SE 4-29 '

_ .-. , , , - - - - -.-,---,,_.- -.-- ,yy,- . _ , . . _ - - - - . - , - , - . - - - - - , .

. - . , , - , . - - - , - - - .-.m,. - -

i

5.

SUMMARY

AND CONCLUSIONS d

k

5.1

SUMMARY

j. In this study, the results of the Millstone Unit 3 PSS were i utilized to foces on the issue of seismic capability of Millstone 3 l structures and equipment beyond the SSE acceleration level. This was accomplished by evaluating the high confidence, low frequency of failure 2 acceleration values for the dominant risk contributing structures and equipment and for the significant plant damage state <. In addition, the contribution of various ranges of acceleration to the overall annual frequency of seismically induced plant damage was evaluated. This allowed determination of the level of earthquake accelerations that contribute most significantly to risk. Finally, in order to investigate concerns raised during the NRC review of the Millstone PSS and in order to explore the robustness of the PSS results, an ,in-depth sensitivity analysis was perforced. In this analysis, the effects of varying the fragility model, assumptions used in the component fragility development, and the hazard model were evaluated.

The following section sumarizes the results and conclusions of this study.

5.2 CONCLUSION

S From this study the following conclusions were drawn.

I 1. The study indentified the dominant contributors to seismic risk and demonstrated that these structures and equipment have high confidence, low frequency of failure accelerations of at least 0.30g l~ (Tables 3-la and 3-lb). Also the high confidence, low frequency of i failure accelerations for the dominant plant damage states were shown to l be greater than 0.25g (Table 3-2). It is concluded, therefore, that the

Millstone 3 structures and equipment exhibit seismic capability l significantly beyond the 0.179 SSE acceleration level.

5-1 1

-- ~ wn , men,----e-------n,-,,--~~mnrw--mm,-~.-----~r-,--,re-en nr.~ -, , , - -~~-,,n,-n--- ---n<,~,-r.rns~r~.e -

--m----

i

2. A review of the plant damage state fragilities indicates 4

that the plant damage state TE, which is the dominant contributor to

seismic induced core melt, has a high confidence, low frequency of l failure acceleration value of 0.269 Thus, there is considerable safety margin above the SSE against core melt. Also the plant damage state V3, which is the dominant contributor to seismic-induced early fatalities, i has a high confidence, low frequency of failure acceleration value of 0.60g. Therefore, there is very high seismic capacity with respect to l seismic induced accidents leading to early fatalities.

3. By convolving the plant damage state fragilities with the

! site seismic hazard curves, the annual occurrence frequencies of the

! dominant plant damage states were obtained. Table 3-3 shows these results and demonstrates that the annual occurrence frequency of seismically-induced plant damage at Millstone Unit 3 is very low. It may be concluded, therefore, that seismic risk at Millstone 3 is indeed very low and is less significant than risk due to internal events.

j 4. An analysis of the contributions of different peak ground acceleration ranges to the plant damage state frequencies demonstrated i that the contribution of earthquakes of up to 0.30g is not significant.

This supports the conclusion that very large earthquakes must occur in order for any significant damage to be done to the plant and that the plant is well equipped to resist earthquakes significantly larger than the SSE.

5. The influences on the plant damage state fragilities and j occurrence frequencies of different assumptions made in the fragility modeling was investigated. Effects concerning the fragility lower tail cut-off, correlation between failure modes and correlation between component failures are sumarized in Tables 4-1 through 4-3,

! respectively. From this and the discussions in Sections 4.2.1 through

! 4.2.3 it is concluaad that the results are f airly insensitive to these variations.

! 6. In order to address several specific concerns raised during i the NRC review of the Millstone PSS with regard to assumptions made in

! the component fragility evaluations, the fragility parameters of components affected by these assumptions were varied. The specific areas of concern were, the earthquake magnitude range that dominates the seismic risk, the limiting displacement (capacity) for attached piping, the coefficient of sliding friction, tank wall buckling capacity, and the ratio of peak ground velocity to peak ground acceleration. Each of the areas was addressed individually by varying the values (e.g., coefficient offriction)usedinthefragilityevaluation,recalculatingthe fragilities of all affected components and then re-evaluating the plant damage state frequencies. Table 4-4 shows the results. Based on these results, it can be concluded that varying the assumptions used in the fragility evaluation does not have a significant effect. This lack of sensitivity is discussed further is Section 4.2,4.

f 5-2

l I

i 1

7. The lognormal fragility model was re-examined, and was j concluded to reasonably represent component fragilities. '

i

8. The plant damage state occurrence frequencies were obtained by using a family of seismic hazard curves, some of which were truncated at specified maximum peak ground acceleration values of 0.6g, 0.8g and 1.0g. In order to study the sensitivity to peak acceleration truncation, a modified set of hazard curves, wherein all truncated curves were extrapolated on a loglinear scale, were used. Table 4-5 shows the results of this analysis, from which it may be concluded that acceleration truncation does not significantly affect the plant damage state occurrence frequencies. This results from the fact that the

significant risk contribution is from earthquake levels below the cutoffs.

9. The sensitivity of the. plant damage state frequencies to seismic hazard modeling was addressed by comparing results obtained using hazard curves developed by the Lawrence Livermore National Laboratory (LLNL) with the results obtained using the hazard curves of the PSS, developed by Dames & Moore (D&M). It was shown that when the conservatisms in the LLNL curves were removed, results comparable to those obtained using the D&M curves were obtained. This leads to the conclusion that, while gross differences is hazard models can significantly affect the plant damage state frequencies, the results obtained in the PSS are fairly robust since the hazard model employed has already considered a wide uncertainty range. In addition it was found that even when the original, unmodified LLNL hazard curves are used the contribution to the plant damage state frequencies from earthquakes below 0.30g is small, reinforcing conclusion 4.

In sumary, based on the Millstone Unit 3 PSS and the further confirmatory studies performed herein, it is concluded that the Millstone Unit 3 power plant has the capability to withstand seismic excitation above the design SSE with a high degree of confidence.

l l

l l

s l 5-3 i

I REFERENCES l

1. ANS-IEEE-NRC, PRA Procedures Guide, NUREG/CR-2300, January,1983.

l 2. Bohn, M. P., et al, " Application of the SSMRP l Methodology to the Seismic Risk at the Zion Nuclear l Power Plant", NUREG/CR-3428, Lawrence Livermore

National Laboratory, January,1984.

( 3. Dames & Moore, " Seismic Hazard and Design Spectra at

! Millstone Nuclear Power Plant Unit 3." Report prepared

! for the Northeast Utilities, October 26, 1983.

t

4. Dames & Moore, " Sensitivity of Seismic Hazard Results at Millstone to LLNL Study Assumptions on Attenuation and Seismicity" Final Report to Northeast Utilities, June 15,1984.

5. Department of the Navy, " Design Manual - Soil Mechanics, Foundations, and Earth Structures", NAVFAC l DM-7, March, 1971.

6. Freudenthal, A. M., Garrelts, J. M., and Shinozuka, M., "The Analysis of Structural Safety", Journal of

the Structural Division, ASCE, ST1, pp. 267-325,

t February, 1966.

7. Kaplan, S., "On the Method of Discrete Probability Distributions in Risk and Reliability Calculations",

j Risk Analysis, Vol. 1, 1981.

8. Kennedy, R. P. and Ravindra, M. K. " Seismic i Fragilities for Nuclear Power Plant Risk Studies",

Nuclear Engineering and Design, Vol. 79 No.1. May 1, j 1984, pp 47-68.

9. Lawrence Livermore National Laboratory, Seismic Safety

-Margins Research Program, Phase I Reports, 1982.

10. Lawrence Livermore National Laboratory, " Seismic L Hazard Characterization of the Eastern United States:

Methodology and Preliminary Result for Ten Sites",

Preliminary Report dated February. 1984.

p l 11. National Aeronautics and Space Administration, L " Buckling of Thin-Walled Circular Cylinders", NASA j-SP-8007, 1965.

R-1

4 REFERENCES

12. Newmark, N. M., "A Study of Vertical and Horizontal Earthquake Spectra", WASH 1255, Nathan M. Newmark Consulting Engineering Services, prepared for USAEC, April, 1973.

13. Newmark, N. M., Letter correspondence to A. J.

Bingaman, et al,

Subject:

Factor of Safety Against Sliding, June 10, 1975.

14. Newmark, N. M., and Hall, W. J., " Development of Criteria for Seismic Review of Selected Nuclear Power Plants", NUREG/CR-0098, May,1978.

15. Northeast Utilities, Millstone Unit 3 Probabilistic Safety Study, 1983.

16. Ravindra, M. K., Banon, H., Sues, R. H., and Thrasher, R. D., " Sensitivity Studies of Seismic Risk Models", EPRI NP-3562, Project 2170-5, Report prepared for the Electric Power Research Institute, June, 1984.

17. Structural Mechanics Associates, Inc., Seismic Fragilities of Structures and Components at the Millstone 3 Nuclear Power Station, Report No. SMA 20601.01-R1-0, prepared for the Northeast Utilities, March,1984.

18. U. S. Nuclear Regulatory Commission, "A Prioritization of Generic Safety Issue", NUREG-0933,1983.

19. U. S. Nuclear Regulatory Commission, "A Review of the M111 stone-3 Probabilistic Safety Study", Docket No.

50-423, May, 1984.

20. Walker, H. C., et al, " Summary of Basic Information on c Precast Concrete Connections", PCI Journal, December, 1969.

l l

l R-2

APPENDIX A COMPONENT FRAGILITY PLOTS This appendix contains fragility curve plots for the components listed below. The plots show the 5%, 50% and 95% confidence level curves.

These may be interpreted as follows: if, for example, the failure frequency of a particular acceleration is read from the uppermost curve (the 95% confidence curve) then there is 95% confidence that the frequency of failure (at that particular acceleration level) is less than the value read.

Component No.

, and Symbol Description i

2 RECRHTEX Containment Recirculation Heat Exchangers 3 EGECLPSE Emergency Generator Enclosure Building (wall footingfailure) 4 RWST Refueling Water Storage Tank 5 EDG0!LCL Emergency Diesel Generator (oil cooler anchorboltfailure) 6 COREGE0M Reactor Vessel Core Geometry Distortion 7 DFCNTBLD Control Building Collapse (diaphragm) 8 CNTRLBLD Control Building Failure (sliding) 9 CRDS Control Rod Drive System (failure to SCRAM) 10 RPCWPUMP Component Cooling Water System Pumps A-1

l Component No.

and Symbol Description 11 SWPIPE Service Water System Piping (due to pumphouse slidng) 12 SWPHSLID Service Water Pumphouse Failure (sliding) 13 EGESLIDE Emergency Generator Enclosure Building (sliding) 14 AUXBLDG Auxiliary Building Collapse (shear wall failure) 15 RCSPIPE Reactor Coolant System Piping (large LOCA) 16 RCSSMPIP Reactor Coolant System Piping (small LOCA) 17 DWST Demineralized Water Storage Tank 18 TBDAFWP Turbine Driven Aux. F. W. Pump 19 SWPHCOLL Service Water Pumphouse (shear wall failure) 20 ESFELDG EngineeredSafeguardFeaturesBuildiig(base mat shear wall failure) 21 RECPUMPS Containment Recirculation System Pumps 22 PORV Power-Operated Relief Valves (loss of operability) 23 CVCSPIPE Chemical Volume Control System Piping 24 RECRPIP Containment Recirculation System Piping 25 RPCWPIPE RCP Component Cooling Water System Piping 26 QSPIPE Quench Spray System Piping 27 CONTWALL Containment Crane Wall (collapse) 28 MCCFAIL 480VMotorControlCenters(trip)

( 29 RXVESSEL Reactor Vessel (support pads fail) f 30 SWPUMPS Service Water System Pumps 31 RCPUMPS ReactorCoolantPumps(largeLOCA) 1 A-2 e

Component No.

and Symbol Description

, 32 RCPSWHEX RCP Seal Water Heat Exchangers 33 RPCWHTEX RCP Component Corling Water Heat Exchangers 34 CABTRAY Cable Trays 35 QSPUMPS Quench Spray System Pumps 36 OSHEADER Quench Spray System Header Piping 37 STGEN Steam Generator Supports s

A-3

g _ _ - _ _ . _. . ._ -- . _. .. -

o .

.I I I LLI

% CONF.

go ,,

8 m 0.050

- g- 0.500 -

R 0.950  ;

g LL.

LL.

00-

~* ei a

p a-U Z .

LLI E

, D .

2 Oo R i

LLI:r -

T *-

u_

J T

Zo Om

~

w f-w Q .

Z O o Ua

k. 00 0'.40 0'.00 l.20 l'.60 2.00 PERK CROUND ACCELERATION C RECRHTEX MILLSTONE III SENSITIVITY STUDY , COMPONENT FRACILITY

-o pg "I I I to E CONF. ,

a g m 0.050

- f- 0.500. -

R 0.950  ;

g LL '

d a

LL .

oo* M i

. o

> o-a

> z .

e, LLI $ ,

D s i ao R LLI:8' E a trd- .

LL

__2 T

, zo om

)

m f-1 m

j o' . .

z i O o ao i I I I I i .00 0.110 0.80 1.20 1.60 2.00 i PERK CROUND ACCELERATION G ECECLPSE MILLSTONE 111 SENSITIVITY STUDY , COMPONENT FRRCILITY

o

~

~ 1 . I La E CONF.

go ,

u a2 0.050 df-g 0.500 0.950 N

LL .

LL

• Oc .

to .

p f-E U

Z .

? LLI E

= a  ;-

Go R LL]2 E (Ef-Lt.

J C

Zo ON -

Mf-

+ .

M .

-o .

Z C o Uo I I I I

.00 0.40 0.80 1.20 1.60 2.00 PERK CROUND RCCELERATION G RUST MILLSTONE 111 SENSITIVITY STUDY , COMPONENT FRRCILITY

o o

~ i  : -

to boo CONF.

0.050 u

df-g 0.500 0.950 N

ts_

LL Oc g

~

a -

h o~

U z "

> tu 4 D L Go R Ld2 E .

Odd-u_

a 1

Zo em -

e-e g-W m

)

i a .

Z C o Uo I I I 1.60 I

2.00

.00 0.140 0.80 1.20 PERK CRGUND RCCELERATION G EDCGILCL MIL LSTONE !!! SENSITIVITY STUDY , COMPONENT FRACILITY

o v i i tt!

CONF.

No to 0.050 u

df-g 0.500 0.950 N

. LL.

l ".

Do =.

LD p g- -

E o

Z .

? LLI $

" a s -

Go R LLIT E grd-tt-J T

Zo ON -

m r f-m -

O~ .

Z '

Do Uo

.00 0'.110 0.80 l'.20 l.60 2.00 .

PERK CROUND ACCELERATION G .

CORECEOM  ;

MILLSTONE III SENSITIVITY STUDY , COMPONENT FRACILITY  !

.. i

o I I I I to a CONF. ,

$m 0.050- u dd-g 0.500 0.950 N

LL. ,

LL 00

• 4 o

p g- -

U

> Z .

a tu 8 a s Oo R Ld 2 E K '-

LL

-s T -

Zo ON s g- -

H Q' .

.Z Co Uo I I l I

.00 0.140 O.80 1.20 1.60 2.00 PERK CROUND RCCELERATION C DFCNTBLD MILLSTONE III SENSITIVITY STUDY , COMPONENT FRRCILITY

l 01-V 1

4

.. i CONDITIONAL FREQUENCY OF FRILURE-0.60 0.80 1.00l J.00-0.20 I

0.40 I i i 1 i

'o  !

7 o ooon -i

..o ,  ;

TT1 m sic z '

p ui o u1,1 coc*

I 2n

?5 O P- -

C;# xc ,

y$ GC I zo C

Z O O m T.- -

Q O cn m no O rn a r--

rn 4 x w ~D 7 -

$ H rU o

so O

z ,

n a a. '

A o 5 O 2

m O

C u . . . .

I a

!

• o 84/08/14. 17.56.30. RT70MHS o

l l

o

._, e i . i I

Lu

• C CONF.

-ga m u

0.050

.d f- 0.500 N 0.950  ;

j g i tt.

t N ooLD a .

1 a

x d-

o l' Z .

, .tu

.'. a

~

co R

, tu =r -

=

Ef-tt.

4 a

C zo om s g- -

i i o .

i z

\ D o

oo -

4 , , , ,

, 3 .00 0. LIO 0.80 1.20 1.60 2.00 i PERK CROUND ACCELERATION C CRDS MittSTONE III SENSITIVITY STUDY , COMPONENT FRRCit!TY

o o

m I I g I

, Lt!

% CONF.

go. ,

g m 0.050

- g- 0.500 -

R g 0.950 E LL LL 00 =

1 S a ,

. ~

p o~

U

? Z .

G- Los a D 2 Go R tt! -2' E

, x f-tt.

_.J E.

Za j ON

m g- -

I--

>=-4 -

O Z~

O a Uo i i i

.00 0.40 0.00 1.20 1.60 2.00 PERK CROUND RCCELERATION C RPCUPUMP MILLSTONE III SENSITIVITY STUDY , COMPONENT FRACILITY

l l

1 l

o GDDfiti *SE*97'60 '60/90/h8 o >.

I "

I I I N G

a:

a:

L o E

• u

-J o o r

a u

z ,

O >.

o e-e o N& B m

_ -; a-or >-

w. C

_a E W C 00 m ou

-g a m l O C Z m D =

l co wr wS*

l

-d o kd mz

.c o o

u. tn o tn C l

z o in m D .**

w og t

uooo a

1 O*I 09'O 09'O Oh"O 02'O '

00 *f 380lIBJ JO ADN30038d 18NDI110NO3 i

-A-13

a d I i i

LLI E CONF.

ga ,

o 0.050 j,mf- 0.500 -

N g 0.950 E LL .

LL =

ooLD ".

y f-8 U

Z .

l LLI $

Y U U 4

g Oo R LLI" . E Ef-LL

__J G

Zo om ~

M 1- f-m O

• Z I I I I

.i .00

, 0.40 0.00 1.20 1.G0 2.00 PERK CROUND ACCELERATION C SWPHSL10 MILLSTONE III SENSITIVITY STUDY , COMPONENT FRRCILITY

o I i i  :

.m LLI

$am CONF.

0.050 u

' df- N 0.500  ;

. c 0.950 tt.

LL 00

• os o

p a-i O Z .

LLI E 3 -

id

,'. I m Oo R LLl2 E u

f-J T

Zo ON s g-r M .

Q -

' I I I

%. 00 0.40 0.00 1.20 1.60 -

2.00

PEAK CROUND RCCELERATION G

) ECESLIDE i MILLSTONE 111 SENSITIVITY STUDY , COMPONENT FRACILITY

o o

s, a

~ r.

i tu

$am CONF.

0.050 o

df-g

-0.500 0.950 N

.u_

u- $

00 i m -

8

>_ f -

o

, z a s tu  ;,-

=

a e 00 tu =r =

x u_

d-a C

zo em s g-r a -

4 z

C o Ua

.00 0'.LIO O'.80 l.20 l.GO 2.00

PERK CROUND ACCELERATION C i RUXBLDC MILLSTONE III SENSITIVITY STUDY , COMPONENT FRACILITY

o a

I I i -

tu E CONF. ,

gam g 0.050 0.500.

R

- f- E g 0.950 LL.

~:

LL

• Do ei o

w o-U Z .

E

? L1J k

O D Go R LLl2 E E

tu d-

-J E

Zo ON s g- -

s s '

O Z-

-D a Uo .

.00 0'. llo O'.00 l.20 l'.60 2.00 PEAK CROUND RCCELERATION C RCSPIPE MILLSTONE 111 SENSITIVITY STUDY , COMPONENT FRRCILITY

o-o 5 i i i LLs

$a g oo CONF.

0.050 g

~ f- 0.500 -

R 0.950 ;E c

11-LL.

Do .

of p ?-g a

U Z .

Y LLI $

l E D L

.i Oo R LL12  %

M a*-

LL.

J C

Zo ON *- -

H i H ,

O -

Z O o Ua i i i

.00 0. llo 0.00 1.20 1.G0 2.00 PEAK CROUND ACCELERATION C

~RCSSMPIP MILLSTONE III SENSIT]VITY STUDY , COMPONENT FRRC]LITY

.g. . .

o d s 1 i Ltl

% CONF. ,

go u .

co 0.050 df-CI-0.500 D.950 N

E LL

• e-LL ".

O C3 (D .

p g- -

8 .

U Z .

U k i O .

G Go R LLI:r E M g-tu

-s T

Zo ON -

s

~ g- ,

e-e C =

Z C o UC ,

.s

i  : I

'h.00 0.40 0.80 1.20 1.60 2.00 PEAK CROUND ACCELERATION C i

DWST MILLSTONE III' SENSITIVITY STUDY , COMPONENT FRACILITY t

o

.M I I I I W

$o CONF.

0.050 u

'mf- O.500 -

N a- 0.950 Lt.

LL. -

Do "

e m

.~ -

O U

Z .

> W 8 i '3 ?d io Oo R LLl2 E Cr a*-

. u_

._J C

i Zo ,

i ON -

M w f-i H Q . .

Z

] O o Uo s e e u

.00 0.40 0.00 1.20 1.60 2.00 PERK CROUND ACCELERATION C T

l TBDAFWP l MILLSTONE Ill SENSITIVITY STUDY , COMPONENT FRRC1LITY

sanone st sz sc so/so/ns 8 .

a 1 I i i I g G

G tt:

6 o E 9 E o

- - a E

u

~z .

O >.

om o Nw  ?

n

- g z >-

W C

-s t W t au m ou

- -g c n o

Z w

-> z ao AB.

=r x M*

-d o kn mz 000 1 u.nn a nn W

8 oooo

. *. 9 an o .

00*I 08'O 09 0 Oh*O 02 0 i

00 f' 3807I65 30 AJN30038d lBNOIIIONO3 A-21

o

- i i i i to

$om CONF.

0.050 g

g- 0.500 -

g 0.950  ;;

.c ,

LL.

to.

00 i

m ~.

>-g- -

a o

z .

,. tu =

a a  ;;

" R Go w =r  :

K f-u-

-- I c ,

j zo on s g- -

n o

4 z i O o uo -

4 i i i i

%. 00 0.40 0.00 1.20 1.60 2.00 i- PERK CROUND RCCELERAT. ION G 4

ESFBLOC l MILLSTONE 111 SENSIT]VITY STUDY . COMPONENT FRACILITY

~

o o

i i i i tu 00 CONF. =

30 0.050 -

3 N

8

._ 'O.500 E c 0.950 m .

E L1- 6 N

oo

> o-U Z d LLI R P D 8 5 Oo 3

=

to:r -

or d-u_

_I E

Zo ON -

m g' -

r s

O -

Z Co l UO .

k.00 0.40 0.00 l.20 l.60 2.00 PEAK CROUND ACCELERATION C RECPUMPS MittSTONE III SENSITIVITY STUDY , COMPONENT FRACILITY

o i e i I Ltl

% CONF. m 30 0.050 0 dg-0.500 O g 0.950

• LL- .

E LL. i eo N

> o-U

? Z- ~

d

'g LLI R D 8 Go Lt)::r

)=

x d-to.

_J T / -

Zo ON s g-n--

Q .

Z C o Ua k,00 '

0. LI O O.00 l.20 l.GO 2.00 PEAK CROUND ACCELERATION C PORV MILLSTONE Ill SENSITIVITY STUDY , COMPONENT FRRCILITY

o o j a

I I e I g

w 1 E CONF. .

Ua 8 "m

~0.050 0.500 R

0.950 E g

LL.

I d

LL. . .

Do 2 ]'

D ~

& a

>o~

l U a

]

I  ? Z d N W i a -

E 4

Go l

W2 ~

E  !'

l E o*- ]

LL.

__J C

Zo ON _  :

g ._

1U s

i a .

4 Z

O a i Uo  ;

a u u k.00 e

1.60 0.40 0.80 1.20 2.00 i PEAK GROUND ACCELERATION G l CVCSPIPE  :

MILLSTONE III SENSITIVITY STUDY , COMPONENT FRACIL]TY j I

o a

i s I i to E CONF. =

30 '0.050 0 df-O.500 6 a- 0.950 -

LL .

E L' d OO

> o-U

> Z d 4 tLI R m a a Go )

Ld d* _

Ed-to.

._J T

Zo .

ON -

.-i g -

i-

.--e C

• Z O o Uo I 8 '

.00 0', LI O 0.00 1.20 l.60 2.00 PEAK CROUND RCCELERATION G RECRPIP HILLSTONE Ill SENSITIVITY STUDY , COMPONENT FRACILITY l

1.

o o

i s i i LLI E CONF. m Uam .0.050 8 df-0.500 $

e g 0.950 l' .

3 L' 6 m

Oo g m -

a C

U Z g

? W -

R ti a g Oo )

LLIC ~

l M d-1 LL-J l T l Zo ON ~

1 m

w f-H .

C Z

Oo

. J n u o '

3 .00 0 . 11 0 0.80 1.20 l.60 2.00 PERK CROUND ACCELERATION C l

l RPCUPIPE l MILLSTONE !!! SENSITIVITY STUDY . COMPONENT FRACIL}TY

NIN NN'l' o

o i I i tu C CONF. m Dao 0.050 S dg- ~0.500 O

e a- 0.950 L^- .

L1-Oo N O as o

p g-O Z -

? LLI E M a i Oo Lt3-r

)

a

~

E a* -

LL.

-.J C

Zo ON -

e- e g-I--

e-e Q -

Z O a Uo i # '

k.00 0. Lj 0 0.60 I.20 l.60 2.00 PEAK GROUND RCCELERATION C QSPIPE MILLSTONE Ill SENSITIV1TY STUDY , ;0MPONENT FRACILITY

e . 5 ~ - -

~

0

=8 $= 34~do gRg)m 0 Y

. T I

- - - 2 L I

C A

1. 0 6

R F

T N

E N

i O

lC P M

O C

N ,

O Y 0I D 2 U

.T T S

i

'lA R Y T

E I V

L I E T I

0C S N

0C E i

R S 0

I I

D I N E U LN LO 0 O AT US 4R TL NL

'0.C i

OI CM 000 K F505 A N059 E O ...

C0O0 '

0 P 0

oo- 0g oLD .~ o orr & u. oN g aa UZLDGtTt

. i I I LLE3dgL L L

^O L J_TZOs

_ i H ZOU

>'c*

o

o o

~ i I i tu M LONF.

a .

o .0.050 S df-g 0.500 0.950 R

E Lt_

J LL 9 eoe a g

~

l M o' ~

U g Z g o LLI p

~D 5 l Go R l Ld:2' E

~

l Z o'-

i ti-l l

__J T

Zo ON s -

I- g -

M .

O Z

O o Oc I i i i 3 .00 0.40 0.80 1.20 1.60 2.00 PERK GROUND RCCELERATION G MCCFRIL MILLSTONE III SENSITIVITY STUDY , COMPONENT FRACILITY

o a

i I I i

tu E CONF. S 3a 0.050 g

'a -

0.500 0

g 0.950 u- a a

b "

Do l o _

8

> o-o Z N y tu A S D R co o La =r _

e g-u_

-s C

Zo ON _

sf-r e--e O

Z Ua -

' 2.00 0.80 l.20 l.60

.00 0 . 14 0 PEAK CROUND ACCELERATION .C RXVESSEL MILLSTONE 111 SENSITIVITY STUDY . COMPONENT FRAC]LITY l l

o o

i i i i LLj M ONF 3 0.05b $

>0 -

0.500 R

-o -

e g 0.950 LL .

N a

b m Oo ~

m m a

>o_

2 U w' Z m N LLJ e

=

D a Oo 3 l tt;:r -

Ea-LL

-s C

Zo -

ON -

m . -

I- o s

O -

Z O o Uo i i a i k.00 0. LI O 0.80 1.20 1.GO 2.00 PEAK CROUND ACCELERATION G SUPUMPS MILLSTONE 111 SENSITIVITY STUDY , COMPONENT FRAC]LITY

. ' -i% ,;&.2.9. ; Q.,;.':.. .n . . . j:i '

. l ~:' .v *' L , '. - -r yl..- ' 9/ ' ?, i . i.. i_ i-2:1.),."'s. .Q. ::.Q se.;"V %.W:%.;y J.y, Q/ ;

. ,_ y .. . .., ;..f'.* ~ 6.,,. ..a.. .. g ::, -:;; ~;~;; ;_.

t i 'c' ;'. , . ;:- .. ..s,p,4

_, ~

y , , . . , . '~y .-;.-.

?

- , .: , '.g;,, . + e~..:,'.

. . .: . .. . - , - 7;,,

u,- +-

- s; ,

, ,. . .:., ; p f. _ . ? . .x*. ,

-, .s.>

?.

A' . .

.- ,. 7. ,

.s. ..,.: . , , .

pr, :k*

" . [.

' 'a

- -'* ..- . . ,; - ,: ', E .

• k .- ,- .

~

t . .h . _ l b _

.c s_ . . b ' *

• L '. , . . , .

i: .g . . R_) - J . -f- ,. ,'

0 h pe ,9j~ao den ;;:e 0 Y

_ . T

_ I

_ - ~ ~ 2 L I

C A

R F

T O . N G E N

'. O I

lG P

_ M O

C

_ N ,

O Y 0I D U

_ 2 u .T T S

1A R Y E

T I

V L I E T I

0C S N

8C E S

0 .A u

i I

I D I N E U SO N

0O PT MS 4R UL i i0 .G PL CI RM b00 K

05 A 59 E

00 0 P 0

o* Og om of. f oN g oo L

L gL Lo L L uZILD L e W g u_ J_EZOm

_ i mOZOo

$w

o a

i i I I La M CONF. =

30 0.050 $

df-0.500 O g O.950 LL .

U LL 6 Oc .

8

> o-U Z d

= LLI R O D E Go )

w :r _

x LL d-J E

l Zo l QN _

l m

l-f- .

H Q -

Z O o Ua -

k. 00 0.40 0.00 l'.20 l.GO 2.00 PEAK GROUND RCCELERATION C RCPSUHEX MILLSTONE III SENSITIVITY STUDY , COMPONENT FRRCILITY

o o

s a a I to E CONF. n 30' 0.050 -

'd N

0.500 E

[ 0.950 L1- .

n aoO ,

d o

u Z g LLI e.

> U E l

0

=

Go to:r

)

=

l M b-LL-J T

Zo ON -

s r f-s C

Z C o Uo

.00 0.40 0 0 l.20 l.60 2.00 PEAK CROUND ACCELERATION G RPCUHTEX MILLSTONE !!! SENSITIVITY STUDY , COMPONENT FRACILITY

0 j 3 R -

.1 A".% do d e *)= 0 Y .

T h..,

I

- - - - 2 L I

C A

R F

jf .

O T

N 4,

G E

N O

h; i

'1C P .

N M

O C [9.

O

[?

j:-

Y 0I D 2 U 5

.T T S E i

1A R Y T

h.a.,

E I V

L I E T I

+

0C S N

8C i

8 0 .A E

S I

I Ne f

D I - y N E j f,

U YO N [- .

0 O 4R RT RE TL [ .

0 .C i

s BL RI 000 F505 N059 O ..

C000 K

A E

P CM

[

f[ <

0 0 -

oo*, am o

oN Q*

0ru "o _ s_

_oC.

u tMU#- - gL. t uO Uzu30LML t

t C Z O m b s a Z C Uoa k rM

%c. .

- l llli1ill

[.

)

a m s I I tti

- M LONF.

a ,

m 0.050 l3 df-g 0.500 0.950 R

0 Lt_

ti k Oo  %

• d

.- a pg -

U Z .

> LLI h L, 3 ai Oc R tt!-r E m f-La_

_J T

Zo ON -

O .

- Z O

US

.00 0'.40 0.00 l.20 l.60 2.00 PEAK CROUND ACCELERATION G QSPUMPS MILLSTONE III SENSITIVITY STUDY , COMPONENT FRRCILITY

x -- , , _

.o.

i i e i LLI i a CONF.' ,

.$ca '0.050 9 df-g O.500 O.950 N

'4 LL.

LL oa* =

4 a

y g-U ,

Z .

l LLI E

> S f, 3

= aa R LLIT E g g-u_

__J l

E Zo '

om s g-l r

r.

O -

Z O

UO -

, .. e i 3 .00 0. LIO 0.80 1.20 1.GO 2.00 PEAK GROUND RCCELERATION G QSHERDER MILLSTONE III SENSITIVITY STUDY . COMPONENT FRRCILITY

u- - - - - -- -

o a I

i i

, i LLI

% LONF. ,

ga g 0.050

'og- 0.500 N

l; g 0.950 Lt.

E'.

Lt .,

Do os a

p f-o .

P Z 8 a Ltl  ;

a R oo  ?

LLlo' -

gg-Lt.

1 Za ON -

m g-I-

s t

O .

Z uS , ,

1

%. 00 0. LIO 0.00 1.20 1.GO 2.00 PEAK GROUND ACCELERATION C l STCEN MILLSTONE Ill SENSITIVITY STUDY , COMPONENT FRf1CIL ITY a ,

APPENDIX B PLANT LEVEL FRAGILITY CURVE TABLES 3 This appendix contains fragility curve tables for the four plant damage states for which the seismic event is an important contributor: V3,

[=

AE, SE and TE. At each acceleration level listed, 11 frequencies of occur-rence for the plant damage state are given. These 11 values represent uncertainty in the frequency of occurrence. The probabilistic weightings for each frequency given in the table are, respectively, 0.0122, 0.0278, 0.0656, 0.1210, 0.1747, 0.1974, 0.1747, 0.1210, 0.0656, 0.0278, and 0.0122.

I, I

s H~

b l B-1 1

l urtinig l gie i

1V913 9-L Wtitstoua nuts g gm6e stesa AE 3tmilK oJ dttuS 14AaL JJe6gL62las i

yssatsavatov 3oupittoust Jaabnaust og ossnJauop 6 -

• 05" 1*0oni-10 T'0003-20 t*0003-20 1*0003-70 t*0003-20 1'0003-20 t

• 0003-20 t'00C!-iO T*C007-?0 1*0007-20 t'0003-20 l

• IN T'Ma -/6 1*0rcl-7n 1*r003-70 1*0003-70 1*OC03-20 1*0003-FC t'?er4-FO T*0003-20 t'C003-2C 1*0003-20 1*0003-20

• 140 t amc!-po t'noc3-re T'0003-70 t*0003-20 T*0003-F0 t'0003-F0 1 000> -re i .are3-ic t*C003-20 t*0003-20 t'0003-20

'70n T*00P9-ic t'0003-20 1*0003-70 1*0003-70 T*0003-70 t*0003=l0 i 0002-re t*0003-20 t'0003-20 1*0003-20 1*OC03-2A

• 760 T*00C3-TO T'0003-70 1*f003-20 t'0003-70 1*0003-20 T'0003-F0 1*00C3-2C t*0003-20 t'C003-20 t'0003-20 t'0003-2A.

j

• 'C" T*0 ort-2c 1*0002-70 1*0003-70 t*C003-rO t

• 00 03- 1*0003-20 1*C0C3-20 1*0002-20 1*0003-20 1*0003-20 1'0003-.70 20

• I58 1*00C3-FO T'C0el-?0 1*0003-20 T'0003-FO t*0003-F0 1'0003-70 I*0D03-20 1'0002-20 t'0003-20 1*0003-20 t'0003-20

• • T '000 3 -7 f T'0003=fO 1*0003-20 T'0003-TO t 0003-?0 t*0003-F0 t'00C3-20 t'0003-20 t'0003-20 t*0003-20 t'*t013-02

• • 'O T

• 0009 -TO t*0003-20 T'0003-20 t'eoP3-iD T*0003-20 1*0003-20 t*00C3-20 1*000J-20 2*A923-01 t'1993-07 9*0013-02

'C0n . taoCca-gO 1 0003-70 f003-20 I*0003-70 t*0003-7 1*0003-20 J* 0012-2A-- A'?aA43-ah T.*2281-82 I* . 2'4021-02 1'0Atase.0 J 1*0001-FO T*00C3=fO , t*0003-20 T*0003-20 t'0003-20 t'0003=te

' 6_5 0_ ,.nnem_ra ,esnen.ns reivu_na _.E *A&2AraZ _ 2?5 02J*42 -

• • 00 1*00CN-20 t*00C3-20 t'C0C3-20 m yg3,34 T

• 0 00 3-70 _. t '0 003-20 ._t '.000_3_-2,0_.

,.wnia_as rennam-nv s e m i a n .32 o na ca -31

'950 1'0003-70 T'0003-70 f*0003-20 T*0003-70 t'0003-20 t*0003-70 senac3.ne ,*aer3.nn e? nema.n2 revr&3.nv ressra av

'40C 1*00C3-ic t'00C3-20 T*0003-ZO t'0003-70 1*0003-20 r*1613-05

,ennaz.ns rennsa as sem ara ms ,eerca-av renira-av e

7 e 3[SASU soup [4)M[ JJSbRSusgas RJS $)ASu M)%q J8Sdts))AS dJogegsttttes

) )na56ygs( O*OLZZ' 0'0229' 0'0999' 0*lZl0* 0'lltI' O'L6lt* 0*lltI' 0*LZl0' 0'0999' 0*0249' 0*0lZZ i

J 9-Z 1

1V813 9-L i

N(([s)OUS nu$$ g geWt64 $3W%8 AE Jtsp[K 0} d[tuS 14Ae[ JJt65[5) gas W33*LeJ3%50u 3oup$) gout [ JJebneu3# oJ 033nasuos.

6 -

uO i 00ci-F v 0003-10 1 0003-rO t'0 03- 9 .i'09eB -nO t'stM-et

' t'9001-00_ 2 t*9653-AZ__.1*a111-21. _Z'0163-81 t'0953-DI

• 80* t'ecel-70 t 0003-20 t'0003-20 T*0003-20 1*0001-70 5.' t t 43 -0 (

2*t913-02 9'*2913-0Z 1* tot 3-01 2*5213-01 t *12nE-01

• OCC t*00f3-F0 1*0003-20 1'0003-FC f*0003-FO C*lF63-0( t'titI-02 t'90t3-02 8'9543-02 1*L003-D1 1*0513-01 5*t6L3-01 -

• 600 f

• 00 0 3-7 0 t*0003-20 t'0003-20 1*0003-20 S '6 273-o t T*6959-OF t E&03-A2

' T '0413-A1 _ 2'1023-01 . 1*5 6F 3-81 5*6263-91

• 65C T'00C3-20 t'0003-20 T*0003-F0 1*t053-0E 6*0913-05 t'1853-02

. t

• t5 6 3-22_ _1'11h3=DI_ 2't103-01 A*1Z63-01 9't923-41 I

• o_C. .C t*00c3-70 t'0001-20 1*0003-FO **0tt3-0( l'tttS-OF t* tim-0 2

. s 'rra a -aT 2?tt&3-A1 . 1'0503-01 t?9913-61 9'4653-91 T*050 t*00C3-F0 t'0003-70 T*0003-F0 5*9503-0E T'6663-02 t'0493-07

.L *ot22-AI 7 *nira-31 1*5113-01 1._1113-81

• L'11hE-A1 1*t00 T'00r9-70 T*0003=fO 1'0003-20 d'9&t3-0( 2'5943-02 1*2113-01 2'1191-01 E'6gt3-01 5* tit 3.eA2 9 ' t.f03..-92 1'issa-A1 t*15n C0lf3-70 1'0003-20 F*0103-0E T*t093-OF t'tttE-02 1 t'5003-01 t' 2'tt51-03 t.'t103-D3.. S*11Z3-A1 S '0063 .A1

.'9.EM -97 -

1*F00 1*00C3-20 1*0001-70 t'0503-0( 1'5013-02 t' rots-OF tt 1'0891.21_ .1?!2E3-31.. n*assa.ae 4. '___53...0. 2 e* e rsa.01.__ E 'E113-A3.

T *0003-70 t' 00 03 -2 0 9'0523-ot F tt03-92 s'le M-ti 1 ' F_5 0._

_.L ' tn c a -31... r'tfri.av c'snaa.av e'*aret-al s'92Z3 31 t'12M .9_1 ..

r T' toc t*55C3-06 t'E013-0( 6 1123-OE 9613 92 9*65t3-07 2?6113-01_ .E*&002_-as_ g.'.egt3

. 31_ F'rava,32 s' 0 ' 0103-42_ 1..*_t t.03 --. -_t.5 f*

0 **6603-04 t*2003-0( 1'1193-02 t' t'5513 =91

,2 'tf23.gg , g' zag 3.ag_,_eeaam3.ne ts*' t6 5 3-92_. . .g,.

ansa.a3 25 lB-92

'aeam-av __

. t.'_6 0.__

0

_e'0473-09 t' tic 3-ot 1'6063-02 t'0653-02 6'1003-07 t'1043-01 r'a. ras-at a*0t&3-01.... 0146am_av s' ann.av ,.s'tica-av

,_.1*650 6'F043-04 5 0413-0( 1* **6653-92 , 1'0193-01 r'ats.av &'* a rca .a v .g.0503-02

• 1nat.at n' vit3.a J s'rasa-av E' 9 t M_-01_

e r ,

][SASU 30up)))0uf[ JJ8bntup8s WJS $)ASU M)%(1 atsd93))Ae dJogeqSt)) gas

) )na56 4igs( 0*OLZZ' 0*0348' 0'0999' 0'lZl0* 0'lltI' 0'L6lt' 0'litl' 0'lZlO' 0'0999' 0*02lS* 0'0lZZ i

8-E

n913 9-L Wtitstout nu64 E O'W963 $)W%5 g 3335[K OJ d[tuS 14A4[ JJe6tt63)as .

yssatsavggou 3oup53goust JJebneub 04 0ssaJauos.

6 -

t ece t *6s u-ot . 74n-ot r 8903-01 s ttu -Or i*:503-ei r ' anI.-u_..

8*1261-01 $'t693-01 t*109=-01 t't061-01 6*t163-01.-

• f s' Cts;-6f 1'C*04-CF i*5405-02 t*C0t3-02 1*ttt3-01

• 'C573-01 s'1+93-01 t*t693-01 t ' 562 3-01 t'1113-03. .t'5991-61--

t*eP' r*t091-0( 1*2613-07 t*tGOi-02 9*t4(3-02 t 9213-01 .r. '0 99._1-01

• tit 3-CT **0tC3-01 t'**t3-01 9*&t53-01 A5 613-11.

1'

• Ptt?-rf t'6813-02 **t013-02 t 6613-02 t'ttF3-01 C

• 1Z 63 -01

• • 11E3-01 9*16A3-01 1'40 83 -33,, .. g'ot gA3-g 3_ a

• ec sa-a e 1*t00

• a9f4- t*6193-OF **SF&3-02 1'08F3-01 2*0693-01 8*t163-91 5 '* OIC013 .01 S'9633-01 $

• 1113-g3 geeLa3.gg_,, geetsa-av 1.* t(0 s'ttra-C( 7*2649-02 69 t'FF63-01 l'F913-01 t '4463 -01 s'C901-01 9*6201-01 A1753-02 113-gl s'2323-31,,, s'seta ,at _

t'000 s ' te e 3 -0 ( l*4973-02 M69G3-02_ l ttel-ot 2s0603-9 1 L *tul-01 t'2 tL J-01 t*4193-&1 6'*1181-31... na n -a t . t'u63=D1 t*46 A'0883-rf t*It53-0T J*6013-02 t'6563-01 2*ttE5-01 b

• 6263-01 J*6613-01 8'9003-03 6* tats-a! &' A113-al. t're13-01 t*600 t'ttes-eg geert3-OT 9*Et43-02 t'LFSI-01 2'62(3-01 t'.5_993.

• t*2ctI-01 t'1203-01 6*SF03-01 e't843=DI A*4912-D1 t * 'CC 1*t641-07 **tl&3-02 4'9953-07 t 4063-01 t'F193-01 t ' 9693 .01. .

. 9 6121-01_

• t*A133-AI 1*A903-01 A5 583-01 &'8861-6t F*000 t'9tt3-07 ._s'0203-02 t'0993-01 t*0649-01 t*t913-et e

• st e n h t J' 17 ot -n 1 _A' o J &1 . &' 9101=0 3.. .&*&a13-02 . 6

• 1213 =91 s'05c **6 tit-OF 5*tp03-07 t F123-01 F't 6F 3-01 t'4043-01 5*t065-01

. 216513.ni g.L012sA3

• &119J=&1 t '916 3=42 . . A' AB -81 -

4 i'100 #*rtt9-et 9*5961-02 l'tt&3-01 r 06F3-01 t'6669-01 S'9603-91

.u tiu-ea. .aesaa-n2 6 2:62-al . az'n03-a2 * * = = -a v j

3 --i tsO r'560=-rF 4 t6tS-ot 1 te n-n _ A's r 9663,ot_-

t re13-ot s utS-et J*t0E3-01 2*elg3,32 , ait s aa -at eam 33 a*Assa-as F*F00 r*69*3-07 9'7903-02 1*9863-01 2*6903-01 9'1693-01 r ..

4'9041-03 8'1622-g3 ,, a* ,m sa -n i g_etag3-31,__at'0643=41 s ac sa -a v e

e J e 3[4Aau soupt2) oust ;aabneuoges eJa 6gAou spag aasdosnAs dugeqgttnes 3 )nas6yts( O'OL3Z' 0'0329' 0'0999' 0*ltlO' 0'litl* 0*t6lt* 0*Litl' O'LZlO' 0'0999' 0'0349' 0'0133

)

t 8 71

)

TABLE B-2 Millstone Unit 3 Damage State AE Family of Plant Level Fragilities Acceleration Conditional Frequency of Occurence*

9 1.00n-n 1.00n-2 1.ent-20

.030 T.wnw-1.00n-n 1.000E-20 1.eoor-20 ,1.un-n.ww-rT taew-tr rmet 2,0 - - - - - - -

1.00u-20 1.00n-n 1.tou-n n.ette-n

.nu 1.00n -n

1. con-20 1 00n -20 1.s00E-20 1.000E-n 1.ovet-n 1.oen-a 4.WM-23 1 000E-20 1 000E-20 1. 000E -2 0 1 000E-20 1 000E-20

.ht3 1 000E*20 6.30ZE-06 1 0031-23 1 003E-23 1.000E-20 1.003ia20 1.303E-23 1.000E-20 1.000E-20 1.000E-20 1.000E-20

.20J 1.J00E-20 1.000E-20 4.S 62 E-04 1.052t-04 1 00 3E-20 1.bowi=20 1 000t=20 1 000E-20 1 000E-20 1 000E-20 1.000E-20

.2M 1.000E-29 1.0251-01 1 5964-04 3.290E-03 1.000 E-20 1.00dE=20 1.000E-20 1.000E-20 1.000E-20 1.000E-20 1.000E=20

.400 434021*09 2.730E 1 099E*02

~~ -

2.684t=09 4.994t45 1 00Jt*20 1.00JE=20 1.000E-20 1.000E-20 1.000E=20 3.022E-06~ ~ - - - - * -

.350 73799E-02 S.96bt-UT 6.002t-99 ~3.2991-5T' ~1&#011602 1.000&-20 1.000E-20 1.0005-20 1.000E-20 1.73SE-49 1.050E-04 400 3.973E*ST'

' - ~ ~

• 1 14TE-03 4.4231-07 T.189E-0T T.789t=TI

1. 00 0E-20 1 000E-20 1.000EJ20 1.270E -05 1.013E-04 1.163E-03 433 8.759E-62 2 2 Tit-ST*

4 695F-33 1 430E 7.657E-07 1.0044-20 1.000E-20 1.000E-20 1.092E-04 1.1SFE-03 4.729E-03 500 1.191E=01 3.409E-St 1 343E-02 3.476E-02 F.91EE-02 1.00 J E-20 1.000E-20 0.442E=0S 9.219E-04 3.740E-03 1 199E-0 2 -

.gs0 4.663thet-

~2.963EsOT' 6.701E-t?" 1.377E-91* Fi? 17t=T1-1.003 E-20 1.000E-20 7.200E-04 3.414E-03 1 014E-02 2 972E-02

.400~ -

I S . 0t1E *WF-- litist=et- 7.t*TE41- -37721E=9t- f.797Eatt --

1 000E-20 1.000E-20 1.9048-03 F.111E-0 3 1.90 M-4 2 4.494E-02

.450- 9.169t=TF - -1.ii.e ,i - F.096t=91 9.**9E-91 F.979t*02 3.000E-20 0.302E-04 4.467E-03 1.1946-02 3.420E-02 7.20M-02

.700 7.75 7t41- ts 3 set =0T--T. 939t=91----

-- - -1 tTF9E*tt--P.779t=et 1

L ...

=

5 w

• Eleven conditional frequencies are given with respective probabilities (weights) 0.0122. 0.0278. 0.0656. 0.1210. 0.1747. 0.1974. 0.1747. 0.1210

( 0.0656. 0.0278. 0.0122 f

B-5

TABLE B-2 Millstone Unit 3 Damage State AE Family of Plant Level Fragilities Acceleration Conditional Frequency of Occurence*

9

.FSO 2.S t u-0 3 9.3?M-03 1.e426-41 1.

1. 0.a . , 0 E 1- -2 3.erte-e 0 t - 9.99ee 1 2.4SS,E-el

.22 e-et- .S.32M-e2

.ua-G

. 00 S.01tE-04 4 4SSE-03 1.467E-02 3.620E-42 F.923E-02 1.41M-41

2. 341L-01 3.747E-01 S.334E-91 6.929E-41 9.SP2t=91 - ---

'.S$c 1.0e.t-JA F.4 tit =03 2.1746-02 S.052E-02 1.012E-01 1.0166-01 t .991k- 01 e.434E-01 e.027E-01 7.53FE-01 5.teft-41 -- --

.400 4.04SE-ds 1.32SE-02 3.207E-02 6.St6E-02 1.30SE-01 2.2446-41 J.SOPL-J1 S.074E-01 6.44SE-01 0.032E-91 9.10M-91 *- -

.gsJ S.444i-JJ 1 014E-02 4. 26S E-0 2 8.704E=42 1.616E-01 2.682E-41 4.0961-01 Sets 2E-St F.10 M-41 8.4209-91 9. 39 90 - - ----

1 000 F. 72 S E-03 2.406E-42 S.493E-02 1.091E-01 1.949E=41 3.12 M-41 4.SS!E*01 S.1849-91 7.6 30t-91 S.FSet=9t- 9.9129-9t 1.0$0 1.237E-02 8.293E-02 7.044E-02 1.33SE-41 2.306E-01 3.5 7M-C1 S.0079-91 b e t* M-91 9.927t =41 9.90 M-Ot- -9 47G G 1 100 1. 63 F E -02 4.166E-02 8.649E-02 1.S90E-41 2.67M -01 4.4215-41 S.S62E-01 F .991 E-S t -- 8.319t-01 9.299E-9t- t.GG G 1.150 2.0854-02 S.172E-02 1.0 46E =01 1.8636-41 3.041E-01 4.494E-91 6.007E-01 7.474E-91 8.e221-91 9.3799-91 9.ek9E=9t- -- --

3.200 2.615E-02 S.311E-02 1.234E-01 2 191E-01 3 419E-01 4.4??E-41 l 0.421E-01 F.813E-91 e.95tt=01 9.499E-01 9. peste- - - --- - -

1.250 3. 7.693k-02 1.4S4E-01 2.4 S M-41 3.794E-41 S.2918-41 L 6.3e>E-02 01t-*1 set 1*E-er 9.9999-et 9. 9ee-et- - 9.;--- -;;

1.300 4.114E-02 9.131E-42 1.674E-41 2.77eE-41 4.1796-41 S.649E=41 F.W21-9f--T.38ft 9 299t*@t-

• 95992t=91--T.%;-;;

3 390 4.94?f-02 1.0F0E-01 1.914E=e1 3.890E-01 4.$ 3M-41 6.0696-41 7.4721*tt - ts61 M -91 Tit *00-ft 93 799t=91--- 94 7,;^ 6.

3 400 S.871E-02 1.241E-01 2.162E-41 3.41SE-41 4.900E-41 6 420E-41 7 76 31-91 -9 s titt*St - 9.999t=91-* 9n F99t=9t-- - ts G G -0.

1.490 6.91SE-02 1.42SE-01 2 422E-01 3.743E-41 S .ZS M-41 6.769E-41 9 029t *91- 84999t=9t- 9.949t=91- 9.9 99t=9t*

• 7. ^.G -;;

n=u=

A

• Eleven conditional frequencies are given with respective probabilities (weights) 0.0122, 0.0278, 0.0656, 0.1210, 0.1747, 0.1974, 0.1747, 0.1210 C- 0.0656, 0.0278, 0.0122

(

B-6

TABLE B-2 Millstone Unit 3 Damage State AE Family of Plant Level Fragilities Acceleration Conditional Frequency of Occurence*

9 8.032E-02 1.620E-01 2 689E-01 4.072E-01 S.S99E-41 7.0 F N-41 1.940 9.118 E-01 9 630E-01 9.571E-01 9.973f=GE B.265f-01 4 236E=02 1.826L-01 2.96H -01 4.394E-01 S.924E-01 7.3696-01 1.SSO 9 699E-01 9 59FE-01 9.979E-01 8.443E.J1 9.266E-01 1.0S35-01 2.0*ti-01 J.242d 01 4.F21E-01 6.234E-01 7.6 37E-01 g.e00 9.984E-01

3. 671E -01 9. 37 FE -01 9.754E-01 9.918E-01 1.19 2a-01 2.tt6E-01 3.92SE-01 S.038E-01 6 53M-01 7. 08 N =41 1.650 9.472E-01 9.799E-01 9.9 3S E-01 9.957E-01 8.846E-01 1.340E-01 2.494E-01 3.8092-01 S.34tE-01 6.82M -01 8.t eH-41 1.703 9.5145-01 9 837E-01 9.948E-01 9.990E-51 9.000 E-01 1.496u-01 2. 7 3 FE-01 4.09 M-01 S.644E-01 F.091E-01 0.309E 41 1.ygg 9.997E41 ~ ~

9.13SE-01 4.623E-01 9.548E-01 9.939E-01 1.694E-01 2 941E-01 4.37FE-01 S.939E-01 F. M M-01 s.49M-41 1.s00 9.68 3E-41 9.893E-01 9.968t-01 9.9990-07 ~

~~

9.2537-01 1.029E-01 3.228E-01 4.698t=01 6.218E-01 7.SF8E-01 8.661E-41 1 850 9.913E-01 9.9 71E.01 9.996E-S E 9.317t=01 9.73M-01 2 00m&=01 3.479:=01 4.9 39t -01 6.4 84E -01 7.79 FE-01 8.012E-01 1.990 9.FF6E-01 9.929E-01 9.990E-01 9.99FE-01 9.44FE-01 2 .14 4 E -01 3.731E-01 S.20FE-01 6.F*2E-01 8.00 0E =41 8.94 H-41 l 1.950

- 9.StSt-v1 9.812f=01~ 9.9918-01 9.981E-91 9.9978-91 -

2.376 E-01 3.9tAE-01 S.472E-01 6.982E-41 S .190E-41 9.070E-91 L 2.000 F.*T3t-tr T.TS9E-41 9.9tSt*tt - TTT99tC

-- T.997tatt 2*M'E*01 4.23M-41 S.F31E-41 F.200E=41 8.hM-41 9.179E-41--

2---0$0 T .9ttraft- T.te9t47--14991t=01- TETviteet- T.999t=91 -- - -

2 .76 t E-01 4.40$t-et 5.991E-01 7.422E-41 8.S26E-01 9.2766-01 2.100

-- -T .?Utt arW1 - 9 499tt=97-- -T;T79E -e r

• TIT 9M =9t- - T'.999t =97---" -

~

2. 6E-01 4. n M-4 6. n u-01 7.62M -.1 . 6 F M =.1 9.MM 1

,1,0 g - *T;797t=WT 9'.TO9t=St v.vais-va TTT99f*TT--T.... -us 3*I68E*01 4.9FFE-01 6.497E-01 7. 81 M-41 8.844E-41 9.4 44E -41***

2.200 9.9tyt-wa v.v895-91'-* 9.TTSt=9t- 7.999tW ----

"- - 977921 91

~

W

• Eleven conditional frequencies are given with respective probabilities (weights) 0.0122. 0.0278. 0.0656. 0.1210. 0.1747. 0.1974. 0.1747. 0.1210.

( 0.0656. 0.0278. 0.0122

[

B-7

1V81'3 9-E i

Wtitstoue nu)g g 0 we6e S**** S3 Jsut[g o} d[ tug 1eAot Jae6ttgggos l

ysseteaeggou 3oupsggoueL JJebnnust og ossnasusa, 5 -

'050 t'0001-20 1*0003-FC f*0003-20 1*0003-70 1'9003-20 T

• na M-Jn 1*0203-20 t*0003-24 1'0003-20 .t'0003 ,20 _._1*0001-FO 1*00C3-70 1*n003-20 t'0003-20 1'0 003-2J

-- *I00 1 ' 00 02 -2A. 1'GbL3-20 1'0003-20 t'0003-20 1*0003-20 1*00C3-20 T '0 003-F0

• IM 1*oCns-FG t * #t 00 9 -7 0 1*0008-70 f*C0CI-?0 1'0003-70  !*0003-20 t*0Cc3-20 t*uc03-20 1*0003=l0 1'00C3-20 9*t023-09

• 1. 'C 1*00C3-F0 1*3001-70 t'0003-20 1*0004-20 1*0003=lO 1'0003-20 I*00CA-20 T'CCC3-20 6*2t63-04 s'0913-09 2*1E03-06 --

• I'" t'0004-10 1*00c4-20 t'r003-70 t'0003-70 1'0003-20 1*0003-20 G't981-09 6*6103-05 t*9613-06 6'0103 01 1*0003 .20-

• t00 t 00 c 3 -20 t'c001-FC t 0003-20 1*0003-F0 1'0009-20 t '96 93 -0 4 2'*0t%3-05 2*t243-04 t5013-01 4*1263-0( t'1013-02 --

• "0 f*0org-te t'r003-10 1*0003-20 t'0003-70 6*tttI-01 l --

9*t923-AA t*0Eal-ot t*0493-02 t*1513-02 t*0563-41 d' LFtI C 6. '

• 400 1*0003-10  !*0003-70 1'0003-70 t'6363-05 t*t593-06 E '53 23-ot . t*St23-22 t*2503-C2 t'8003-02 2*E163-01 t'.6663 - -

I '

• 1 00c3-20 1*c0rI-70 1*tttI-06 2'CS*3-06 2*9153-0( t'L663-0(

• Zt613-02 G*2623-02 1'1291-01 2'1693-01 4'0023-01 --

1

" t*00C4-70 t'vtfi-06 t'C111-0( t'CttI=0f 1'0643-02 F*6493-02 l S'Ct11-C2 1'2443-et 2*2182-CI t'9673-01 6*&t93-01

• G" y'6213-06 teggea-et *'F1h3-0( I'1683-02 E'0163-OF 9

. lt 2453 2'1263-01 t*9983 01 E'2823-01 t'2 2E 3 -01 9*6 _03-07. -

'900 1'td53-ot ***t04-01 t ret 3-OF t*0603-02 9'd%13-07

.Z'2i&3-01 t ' %t t 3-01 &'*2123-01  %*lC23-01 8*1223 41- 1'_2693-01

" **t6as-ot t'rtt1-ot s'05*3-02 9'ser3-ot t're09-01 2'1693-01 i

t'65a!-01 6'0191-01 9*6693-01 4*6tta-01 6'0903-01 l

• lCf 1*0ffa-ni 7*tteS-PT **rit 1*te*4-01 l'0613-01 6 4:s -ti 5 2 5;-0t 4 40.3-07

-01 e rets-Ci 6'sisE-03 t.'iL 63 -

, 3teAes souptggouni ;aebnousses em 65Aeu m54g esdooglAo deleq)t646es

) ) met @gs( 0'0lZZ' 0'0349' 0'0999' 0*lZLO' 0*titl* 0*t6lt* 0*titl' 0*ltl0*

0'0999' 0'0318' 0'0133 '

)

i . .

i 8-8 f _ _ _ _ _ _

TABLE B-3 Millstone Unit 3 Damage State SE Family of Plant Level Fragilities Acceleration Conditional Frequency of Occurence*

9 1.wS*s-4t 4.23$t=32 1.044f-01 1.8406-01 3.011E -01 9.410E-01

. F30 9.2071-01 9.760t*02 -- -

S.943t-Ut F.3472-01 9.91St=01 e

J . 4154 - J e 4.37Fa-02 1 62et-01 2 680t -01 4.045t-01 S.5146-01

.309 9.167f-01 f.SOFE-01 9.etet=01 4 9'It -11 9.195t-01

. 2e,. ,4 1./s/r-61 c.F77L-ut 3.57$4-01 S.0$st-G1 4.5396-01

.-3 f . 91): - ? ! 1.9sSt-31 v.esu.-01 9.F795-01 9.9446-01 4 . 011 --si 1.782s-J1 3 0026-01 4.4316-01 S.993t-01 7 3746-31

.vas 4.!3SL-31 9. tele-01 9.9F3(-01 1 .431i '11 9.ht3L-01 1.4/ar-J1 f.64)t-01 A.FSef-01 3. 302 E -01 6.02SE-01 8.070k-01

.ws; 9.9eff=01 6.41tt-J1 9.5191-01 v.010E-01 9.9)st-01 1 50 st- J1 2.9s*E-01 4.510t-01 e.099E-01 7.52F4 01 4 414t=0!

g e og a 9. 99 4 t -01

9. 3/ f t - 11 9.71tt-01 9.095t=01 9.iett-01 a.nta si I.stJt-01 S.2*bt-01 e.cott-01 B.1136-01 9.021E-01 s.J W 9.Sett-J1 9.190t=01 9.942t=0! 9.9eJE-01 9.99 Ft-01 t .*t s t -s 1 e.29et-01 S.9436-01 F.42*E-01 e .S S SE-01 9.320E-01 1.10J 4.907F=01 9.90e6-01 9.992E-01 9.999t-01 v.7/1*=41 4 94 s - J s *.3as.-01 e.480r-01 F.9s*E-01 s.9546-01 9.5366-01 r 1.g3, 9.99et=01 9.999f=01 9.9/et->1 9.4919-01 9.iest-01

. .

• d s t - ~, 6 s."J16-31 F.1546-31 8.34FE-01 9.2146-01 9.6871-01

..fu, 1.000t000 s .-M t 'l D.89Ft=01 9.#916-01 9.99st=01 I 1.g3, * .J son s 1  %.12st-J1 F.6616-01 8.FS4t-01 9.452E-01 1.00 0t +00 9.790t=01 e .9 s tz -01 9.9ste-01 9.99et -01 9.999t-01

• .9 set-Ja s.s77t-01 d.101E-01 4.0496-01 9. 611E -01 9.4616-01 3.30s #.990t=01 9.998t=01 1 000t*00 1.000t*00 1 a . s ,.

9.959t-11 L e ttak->1 7.194c-91 9 4FSi-01 9.2744-01 d.72e6-01 9.906E-01

9. 7FSt= s t 8. *v e t ='J 1 9.vv9t-01 1 000E *00 1 00Jt*00 F.*ssi-31 s.Fevi=01 4.459e=01 1.e041-01 9.9406-01 l 1...., * . ss Jt - il a . en t = ' t 8. P1 h =0 ! w.vww8-01 1.000t*00 1.000t+00 h .1. *: . a tesf%s-)1 v.J47t-31 4.S946-01 4.6641-01 9.961t*01 L ...,,

e .+e t -

• 1 *.ew r.01 1.psode20 1.J69t+00 1.000t*00 M

L L

• Eleven conditional frequencies are given with respective probabilities r (weights) 0.0122. 0.0278, 0.0656. 0.1210, 0.1747. 0.1974. 0.1747, 0.1210, l, 0.0656. 0.0278. 0.0122

{

B-9 1

l

0I-8 ,

ZZt0'0'9220'0'9990'0 )

'0LZl*0 *ltit*0 *tl6t*0 *ltit*0 'Oltl*0 '9990'0 '9420*0 *23l0'0 (styngen) segutigeqmd eAnsedsea up uoA66 eJe sepuenbeJJ Leuontpuos ueAet3 ,

)

J 9

e e(*.tt;*! -

CC+3C00't 00+7C00'1 CC+20sC*1 CC+3CCC'1 ' ' ' '

00*3000'T 00*3000'1 10-35*6'* 10 F.3t** TC-ia* *t it tre / *

~

00*3C00'1 00*3C00*1 C0*!J00*1 00* 3 Cec

• I 00*3000'T 00+1000*1 10-3466*4 to-sta#** 1 r-in te * #. (In-C

• s(6*

sics' .t 't1

00*3000't 00*3000't 00+1000't PC*3C0C*1 PC + 'f CC ' !

' " ' ' ' l 03*3000'T 10-1666*e 10-346e't to 84t*** 10-isne'e 1e atst**

00*3000'1 00*3C00'1 00*1000't 00*3'0C't 0 0

• J vC ' I 10-5f66*6 * C ' ?

0043000'T 10*1466'6 10-lee 6*e in=3*l6'A 10-3550**

00*3000'1 00*I000't 00*iOOC*1 00*iC0C*1 C C + 3 C CC ' l I'F**

00*1000'T 10-1966'6 10-30th't 10-stGe't 1p=aitE*t f r -ist(

• e 00*3C00'1 00*3000't 00*3*0C'1 CO*3;CC'1 CC el: o't

'I l

00'*1600't 10-3466*6 10=a9ph'e in-3pte** 10- hed** 10 srt;*.

C005000'1 00*1000't 00*3C00*l CO*1CCC'1 GC + 1C CC ' 1 .

'*I a

10-3666'6 10-3966't 10- 4 C e 6 *

• 10-itit** 10-44tt't 10 .* f % '

• 00*30C0'1 00*3000't 00*i000'1 00*3*:0't C o

• 2C 0C ' t

"*I i

10-3666'6 10-3666'6 10-3146*6 10-3t6e*6 10-if99't TC- toge* p

.00*3000't 00*3000'1 00*!000't CO*3C00't Oc'Itc0*1 '# *3 10-3666*6 10-3166'6 to-lete't 10-1968'* IC-4ett'* to ******

00*3000't 00*4C00't CC+1:00'1 CC*3CCC'1 CC+ CCs't 053*l 10-3866*6 10-3486*6 10- 3 F 6 6' 6 1C-ions't 10-**66'6 10-0tC***

00*3000'I 00*I000't Oce3000'1 CC+30Co*1 1C ' eta *t.

• 10-1966'4 10-4194't 10-361t't 10-3%+d't 10-*itt'e In anp1* e *0t*I 00*3000'1 00*1000'1 CO*s*0C'l CC*iCCC't IC 2Ce6*t.

"%'*3 10-3*66'6 10-3116*6 10-Ide9'* TC-fut** 10- M e t 1

• f. 10 Me&*/

00*30C0'1 00*1000*1 CC*it0C*l 00+3000*1 1C-lit 6'e 10-7066*6 10-3946*6 10- 1 t * * *

• 10--454*e 10 eete*6 to air ***t W **

00*3000't 00+1000*1 0C+1000'1 1C-16th't. 10-2444*6 C56*I 10-3486*6 10-1956't 10-3*ed*6 10-3&F6'e 10-36t**e 10-4160'd 04*1C00'1 00*3000'1 00

• 3 30 C* t . ta-aass*a 19-3 iO-i 5im.'.a . **5

30-3s46 6 10-3606 4 10-3604'6 10-3 51 6 iO-liet's 6

esausan330 Jo A3uenbeJ3 teuontpuo3 uoneastessy seittt66 sad toAs1 gueld 40 KLtwed 3S **e%S e6ew 0 g 2 gun suogstt&W E-8 318V1

TABLE B-4 Millstone Unit 3 Damage State TE Family nf Plant Level Fragilities l

Acceleration Conditional Frequency of Occurence* I 9

.050 1. 0JE-2e 1.000E-20 1.Geot-2e 1.00eE-2e 1.ooeE-2e a.ee E-te

- -t.=9e=et--n 999e-et- - 1 990e-t9- a,999e-ee- -t.99w-et--

.100 1.000E-20 1.000E=20 1 000E-20 1 000E-20 1.94W -20 1.000E-20 t.0021-to 3.06SE*0s 1.4101-09 9.732E=99 f.5459-99 ----

. 34 3.74dt-19 1.262t-JS 1.249E-07 8.8$4E-07 4.479E-06 1.7098-05 3.019t-Os 1.22et-04 2.72st=04 9.F09E=44 1.4 9et =99 -

.400 2.012 E =J 7 1.944t=0e 6.820E-06 2.104t=0S S .3 9 FE =89 1 1576-04 t.363t=04 4.F99t-04 1.181E -0 3 9.19et-93 3.09tt-92 -

.2SJ S.488t*ut 1.7466-03 4.040E-05 0.S71E-OS 1.6046-04 3.694E-44 9.t7staev- 1 931E-03 f.9ttt att - 8.9 t et *7t 2. 499E =9t---- - -

l .AJO 2.S13t-95 t .Se it **T--

9.496E-05 tT909t =92 9.912E-09 1.9tFt-9t 1 932E-64 S.2246-44 t .99tt =9t-- -t a999t-;;

1.23M-63

.33J 4. welt-06 1.2216-04 2.e 12 E-04 7.301E 44 2.0 32E-03 8.7596-43

--T(131te hetteatt - 2 4999t*91 - 9 sit 99=9t --ti FtSt 4i o

400 1.12 5 E -04 3.e79t-04 9.294E-94 2.972E-43 9.7 3?E-4 3 ' 3. 31eE-42

-t17 tit **f- tsis9t-9t- *1~.T79t*t t-- 7.93tt*9t- 9s M ;-e; -

430 3.941 E *4

• 1.181k-03 5.3S9E-03 1.1275-42 3.890E-42 4.820E-02 I

FTittt **r* 't.11st*e 1 F.t99t*91 - 94telt=98- 9.9tet-et - -- - --

,,0g 1.100 t-0 3 4.381t=03 1.183E-02 3.2 70E-42 7.740E=82 1.8496-41

, 3 5 797t-97

• 8.413t =01 9.814E -91 9. 7 28 E - 9.99 fE =01-- -

.$3J A . 2346 -O s 1 164k-02 3.94SE-02 7.493E=42 1.97 M -91 3.17 tE -41

~

- 1. m u v . T.973t*98- STT99t=91 T.929t*tt- M499et*9t

{ -

.e 9. ,,E-.3

--TTTitf=9t-2.. 1E=.2 T.97?t w.

... 4E-.2 v.iwii wi -Ti979t W *T.. .. w a

1. 4,, E .1 2.7. -.1 . 7,S.=.1

.eSe 2.092E-02 S.730E-02 1.204E=41 2.3095-91 4 0754-41 4.264E-41 p ---

TT117t=9t- Tatttt*TS---93929t =tt- 9.999te tv99eE *99----

.700 4.78SE*02 1.e44t=41 2.9444-41 S.91SE=41 7.5995-41

-- 3.ie.e .i t.;; .-;;

,  ; . ;;;. - ;; -,3. .;...-;.

9 64E -41  ; .; _ x a;

~

'nu F

L

• Eleven conditional frequencies are given with respective probabilities

{- (weights) 0.0122. 0.0278. 0.0656. 0.1210. 0.1747, 0.1974. 0.1747. 0.1210 0.0656, 0.0278. 0.0122

{

B-11

1V813 8-#

.e e

W)ttsgous nuig E OWW96a $tt4313 jemj[K uJ d[tU4 laAat JJe6gtgnas 1 i I

yssalaavpou 3oupg n ouvt jaabnausA oJ ossnaausa, 6

,. =-er i. on-n 45a 64*5581-03 t..n-u.

iA3-03 .r m*

nu-0* -Oi3 .i e0003 s u0 -n. . n, e 003 +&A . -e s s a-u. . .

• tot i 7543-03 7. *. 9 0 3-ni 1*.FF3= . i%3=et L 80(3-03 t .%*.3-03 4:3-03 ...i 43=ti 03-03 e003 00 e003.ei 00

C efi%-03 f 44f4-Oi 6-(.ia 03 i*2743-01 . it03-03

. 5 03 =e v .

s' tit $-03 t etti-03 i 0003 00 i 00e3 00 i 0003 00 e*  ? grp-r t .*gs.3-ot

- sanat-03 s i.n ot e rd93-Oi . 0093-ei

• .'ti-ei t etd;-05 5 0003*00 i 60(3*0c i 0003 00

* e 5-ri 1 #e -D

• t .999-03 9 8fli-03 . 9in-05 . .i*3-ei t*titi-bi tt-P.ts!=Di  :*000!*00 i 0003.eD i e003 00

, ggga-ov . irca-ni f F9n- 6 0093-03 i nne

, m i-es i ecn 00 i 0002.Oi

. tiFI-ot 00 i 00e3 00 3 0003 00

. 9.93-ei

, 080

.c-%ru-pi l- 544-05 8 8404- t 9:98- . .0 3- e 093-es 43-03 i 003 00 0003.03 00 i 00C3.01 00 i tee 3.03 00

, ioc . . .. u.c i .. n u- .cin- .c 3-03 e.sn-u . o.n .u--

i 000a.0c ieDn.05 os .een.u cO .ee 03.n eou.u

, ua , . r 5,. ti , . m -05 . m. -w . a n-

....n.n u00u.0c i0en.n m.n i m 3 se nne3.u ee . aw .u -

i irc ,...n- . v m.w .esu-w ..su=w , un- . ...a -o i i ecn.03 ee i eou.0e o m i.0e 0003 00 i nu.03 u l ir- ...eri  : . 5.ei.ci ... .i-03 ...,,3-03 ... 43-03 e e3.ee etu.00 i 00c3 00 1 0en .00 i ne3.eO neen.u I

i roa ..n. . 4.ci. i ...ni w ...u B -w . un-i0cu.n 00 ax.00 um3 00 um3.n i neb.au-.. - --.m .ee .

ot . r4 3-w . u n-03 . .w u -e s ....n-n oen.n u

. i onacs .i ucam . i ua3.u :.ee n .u . i m na u - . m 3. n._ ..

i. .ac . siw -ni . n u-w . un-w ...n S - i . m ...

- ucotan uecn.na , esa3.u tieSu.u00. i o m .u -. - .e m .ee.-.

.e .. 3-03 ... m -Oi. ..un-w ..un -n i em. e .

.eomu.n.. . n.-.m_ . n. - --

i

- uun.n iun.u , w nn u  :.eun.n 1

)

. 3LaAau 30up[pou9[ JJabnauslas WJS 6)Aau M)gy Ja5da3p Aa daoqeglitnas

)Ma)6yts(00lZZ'00249'00959'0.l2l0*0LLtl'0l6lt*0lltI'0L2lO' 0 0999' 0.0229' 0.0lZZ

]

8-12 f I

TABLE B-4 Millstone Unit 3 Damage StateTE Family of Plant Level Fragilities Acceleration Conditional Frequency of Occurence*

g .-

i .wo ..ern-n a.un. t. 4. a. 0 e.ee

. ..n i . et-e.i em .. i .ene= ..me

s. m .ne - tmee n.--as: :: 2

. . sw .. r.e-at ..ww.es . .me -n a.me . a . ..e i.e es.t.

a.ms.n i .em . i .em .w .me i .e a - --- -

i.... . . . . . i -u a. m e i.een. a.e m ...

...a.-n

.w.. . a s .nn . .. ..

i.eem.e-n..

. a. m e. s .een.u -

. . < , , ..c...wi. i.een.oo a.me. i.e m ... a.me a.m m 4.

i .am.n ..

-n. i.een ... a.een. a. a.= - - - - -

i.,.. .. 14 .. .. a. a. e.w

.. .. m.. . -n.

. .i..m.t

. een ... a. m.Ge t.. _ee a.se.0.

. u ... . - -

( i.is.

. tee .

. .u n -n .. me -n i. m ..e s. m e.4 t. u e a. es a .sm.n m. m e.ee i.em m = > me - a ,*ee .eer-

4. u. .. -u . m. . a .nw . a. m a.=n.a- i.em. .. - s.em n - ii.me

. me .n.- - a ...ee ...

..mem

. o= .. . -on t.e m . .

i .me .= a .nw .e.

a.me .u

- i.me m- r.eeee n.een a.e m . 4

--e,me - a.mem

( .ua

.. r. -n i .mi .a i .m. . .. a . .m .=

i .un .a i . 0= m

t. m e.e.

s.me.ee-i..

s. en m. ----

a. e.m i...,. ..wn-n a.sen.e i.eew. . a.nu..e a.. ee.u - a. - e p i . em .n i. m i.e.e i.new .eme s. .w - - - -

L ..,x .. m e-at a. a. w . a. i. a. m i.e.Se.ee i . 4.e.n ...e a .. Gee.et- tsnn .ees.ee- te.e ee m G ...

t .30 4..,9e- 1. a.. Gee.e 1..eSe.ee 00s.06

-1.1s;..:.;; a. 00s

-i .00 0t ..e.t t.3.e.en... - easeee ve --tsegge.et-i.to. a.oon..ea t1.co t.

.ucte...n.- -t, a. m m. m e.4. i. m e

1. tees.ee m..ee - Sseet.e et- e.::: ;;

I 2.iSJ i. 1. 1.. 1. L.eece.ee i.eeese G t eSee.e.e 09 .. 1. G.ee.. eve.-- t. Se.e.e.eeve --tsGee see.et -ta:-  ;;

d.2.e 1.eena.e4 1. 4. 1. 00s..e 1..ees.e4 n. 00se 0 "8 sseevet * . .. 9.e.e -- 13.e40s. 00e-;; e.:._ -;; e.____.00 L

b

• Eleven conditional frequencies are given with respective probabilities

( (weights) 0.0122. 0.0278. 0.0656. 0.12100 0.1747. 0.1974. 0.17470 0.1210 0.0656. 0.0278. 0.0122 ,

f

,B-13

Wr ~

NORTHEAST UTILITIES cene,.i ori,ce. . seiden street. Beriin. connecticut c=

HARTFORD. CONNECTICUT 06141-0270 k L 1J [" [,N,,$ ,$~.~ (203) 665-5000 December 6,1984 Docket No. 50-423 B11336 Director of Nuclear Reactor Regulation Mr. B. J. Youngblood, Chief Licensing Branch No.1 Division of Licensing U.S. Nuclear Regulatory Commission Washington, D.C. 20555

References:

(1) Safety Evaluation Report related to the Operation of Millstone Nuclear Power Station, Unit No. 3, Docket No. 50-423, NUREG-1031, U.S. Nuclear Regulatory Commission, July 1984.

(2) B. 3. Youngblood to W. G. Counsit, Request for Information to be included in Program to Determine the Capability of Millstone Unit 3 to Withstand Seismic Excitation Above the Design SSE, dated August 15,1984.

Dear Mr. Youngblood:

Millstone Nuclear Power Station, Unit No. 3 A Program to Determine the Capability of the Millstone 3 Nuclear Power Plant to Withstand Seismic Excitation Above the Design SSE Reference (1), Confirmatory item (1), requested Northeast Nuclear Energy Company to utilize the results of the Probabilistic Safety Study (PSS) to document the capability of Millstone Unit 3 to withstand seismic excitation above the design SSE. In Reference (2) the Staff requested specific information to be included in these confirmatory studies.

This letter transmits 5 copies of the following report:

A Program to Determine the Capability of the Millstone 3 Nuclear Power Plant to Withstand Seismic Excitation Above the Design SSE, prepared for Northeast Utilities by Structural Mechanics Associates, November 1984.

This confirmatory study concludes that Millstone Unit 3 has the capability to withstand seismic excitation significantly beyond the SSE acceleration level with a high degree of confidence.

3 W f' d

,,,sww 0%[ e,.

m en i

t I Northeast Nuclear Energy Company trusts that the enclosed study resolves the Staff's concerns in this matter, and remains available to discuss any questions or concerns related to this inf ormation.

Very truly yours, NORTHEAST NUCLEAR ENERGY COMPANY, et. al.

By Northeast Nuclear Energy Company, Their Agent

. LL tC W. G. Coun(d Senior Vice President i

( U By: W. F. Fee Executive Vice President cc: Mr. Nilesh Chockshi, NRC, Structural and Geotechnical Engineering Branch Mr. Leon Reiter, NRC, Geosciences Branch Mr. Ashok C. Thadani, Reliability and Risk Assessment STATE OF CONNECTICUT )

) ss. Berlin COUNTY OF HARTFORD )

Then personally appeared before me W. F. Fee, who being duly sworn, did state that he is Executive Vice President of Northeast Nuclear Energy Company, an Applicant herein, that he is authorized to execute and file the foregoing information in the name and on behalf of the Applicants herein and that the statements contained in said information are true and correct to the best of his knowledge and belief.

,. A _2 -

Notiiry Pub Ny Commis:fon Expires Lbrth 31,1938 A