ML20117M986

From kanterella
Jump to navigation Jump to search
Draft Probabilistic Seismic Safety Assessment of Diablo Canyon Nuclear Power Plant
ML20117M986
Person / Time
Site: Diablo Canyon  Pacific Gas & Electric icon.png
Issue date: 06/30/1977
From: Ang A, Newmark N
NATHAN M. NEWMARK CONSULTING ENGINEERING SERVICES
To:
NRC
Shared Package
ML20117M971 List:
References
FOIA-84-644 NUDOCS 8505170339
Download: ML20117M986 (50)
v • d • e


Text

{{#Wiki_filter:. o o !o e e I f 7 Meg o DRAFT A PR08A81LISTIC SEIS!!!C SAFETY ASSESSMENT OF THE DIABLO CANYON NUCLEAR POWER PLANT N f Yf - A. N-S. Ang and 5.M. Newmark Report to the U. S. Nuclear Regulatory Comission Washington, D.C. N. M. Newmark Consulting Engineering Services Urbana. Illinois June 1977 P g51g9e40914 3 gMOLLY34-644 PDR

\\ ~ ~..... - _ _ e g A, 3 s-yl' s s -( $q s> ,~ s ( } Contents I. 3 Introduction 1.1 Objectives and Problem Description 3 1.2 Premise and Special Consideratior.s s., I II. The Probabilistic Approach to Safety Assessment l <. y s 2.1 Bases of Approach s I 2.2 Determination of Seismic Hazard 2.3 Determination of Seismic Resistances P 2.4 Calculation of Damage Probability ( III. Pertinent Data and Information Base 8*. DataforSeismicHazardAnalysib-3.1 3.2 The Seismic Hazard Curve F 3.3 Response and Capacity Determinations l IV. Calculated Probabilities 1 l' 4.1 Damage of Structural Components \\ 4.2 Damage of Piping System 4.3 Piping Damage and Malfunction of Light Equipment g j 4.4 Piping Damage and Failure of Heavy I Equipment ') V. Sutunary i l-i ~ i h ( 1 .) k q 4- - 1

  • I*

-~- +-.. -...... ~..,.... _ _ 1 I J Foreword ~ This nport repmsents the result of a study on the evaluation of the seismic safety of the Diablo Canyon nuclear power plant, perfonned by N.- H. Newmark Consulting Engineering S' rvices for the U. S. Nuclear e Regulatory Comission. Refemnces to some of the material and publications were provided by Dr. W. J. Hall during the course of the study. The analysis of the seismic hazard was performed with the assistance of Dr. A. Der-Kiureghian. l h e + s 0 M

1 ^ A PROBABILISTIC SEISMIC SAFETY ASSESSMENT.0F THE DIABLO CANYON NUCLEAR POWER PLANT by A. H-S. Ang and H. M. Newmark I. Intmduction 1.1 Objectives and Problem Description The levels of safety against seismic haza'rds of certain critical 4 components or subsystems of the Diablo Ca'nyon nuclear power plant facility are evaluated quantitatively in terms of the respective annual probability of damage. The evaluations were perfonned particularly in the light of the recently discovered Hosgri fault, as well as other known fault zones in the region of the plant. At its closest point, the Hosgri fault is approximately six kilometers from the plant. ~ The power plant is located in an active seismic region of Southern California. A number of fault zones have been identified in the general region within 100 kilometers of the plant, including the San Andreas, the San Simeon, the Rinconada and other faults. The locations of these faults relative to the plant site may be seen in the geographical map of Fig.1; the shortest surface distances of these various faults to the plant 4 vary from 6 km for the Hosgri to 88 km for the Santa Ynez fault. -The probabilities are detennined in terms of the annual probabilities of specific adverse events; these may be the damage of major structural components or pipe joints. However, the event may also be the combined . failures of two or more safety related systems whose failure could lead to accidental release of radioactive material; examples of the latter may be the occurrence of a crack in a pipe joint combined with the malfunction of i . man

A* 2 an electrical shut-off control equipment, or the combination of a pipe break with the malfunctinn of the emergency core cooling system (ECCS). 1.2-Premise and Special Considerations The probabilities calculated herein pertain only to the safety of the plant against seismic hazard; no other environmental or man-made hazards are considered aside from that of earthquakes. In this regard, irrespec- ~ tive of the correlations (or lack thereof) between the various conponents in the piant, they are, of course, all subject to a common environment, namely the earthquake ground motion of the site; consequently the failunts of the' various components or subsystems in the plant will be (at least) partially-correlated, even though the resistances or limiting capacities - (fragilities) of the various elements to withstand ground motions may be uncorrelated or statistically independent. The problem is obviously one of low probability and high consequence. Because such extremely small probabilities are involved, the actual numerical values of the' calculated probabilities should not be taken too seriously; only its order of magnitude may be significant. The results of the calculations are intended to provide a quantitative measure of the level of safety underlying the specific components of the plant. The results, however, do not and can not in themselves answer the question of whether or not the plant is safe enough. Nevertheless, the results presented herein should provide a quantitative basis for the resolution of this question. 9

  • re

,w --w,,-,...-.,,. ---,,,p.-,,w,. ..-,e...-w,-,,ey,,. -.,,3,--.3,-,e-,e,-.,..v~..-%, 9.-m.-,,w r, .-n-,-ww r., - - - - -w-,

._4 ~ 3 II. The Probabilistic Approach to Safety Assessment 2.1 Bases of Approach _ Failure or damage of a structural component, or safety equipment, would occur when the seismic ground motion exceeds the capability of the component to resist such motions. The maximum ground acceleration that could occur at the site of the power plant over a finite time period is clearly not predictable; similarly, the seismic capability of a major structural. component or the fragility limit of a piece of equipment may only be described with a random variable. Consequently, the probability of the resistance being less than the conceivable levels of seismic motions at the site is the relevant measure of safety. The calculated probabilities must necessarily depend on the bases under which the structural components wore designed, and the various safety-related equipments were qualified, for seismic resistances. In this regard, it is stipulated that the major structural components of the Diablo Canyon power plant were designed according to standard practice for the design of nuclear power plant structures, and the ss 1 safety equipments were qualified in accordance with recomended practices (e.g. IEEE 1974,1975). Because extremely small probabilities are involved, they cannot be estimated directly on the basis of statistical observations, nor verified on the same basis. However, the required probabilities may be deduced through a proper synthesis of available information and statistical data, using established physical relationships of the underlying problem. In j this regard, it is important to recognize that the derived probabilities

4 should be technologically credible, in the sense' that it.is based on the logical synthesis of various pieces of infomation, each of which is based on objective data or can be individually judged to be technically sound or reasonable. In perfoming the probability calculations, methodologies for detemining the seismic hazards (e.g.' the maximum ground motions) at the site as well as for detemining the description of the resistances are required. The methods used for each of these purposes are described below. 2.2 Detemination of Seismic Hazards Realistically, over a finite time period, it is not possible to state with any percision the maximum ground acceleration that can be expected at the site of the Diablo Canyon power plant. The description of the ground motion intensity may be developed only in terms of probability -- specifically, for exanple, the probability of exceeding specified acceleration levels in a year, or the annual probability of exceedance. The site of the power plant is subject to potential earthquake hazards from several well-defined fault systems, including the recently discovered Hosgri fault which is six kilometers offshore from the plant at its closest point. Although all the existing faults in the region are potential sources of earthquakes, the high intensity motions that may be expected at the site would largely be the result of events on the nearby faults, especially the Hosgri. In any case, the total seismic hazard at the site would be the cumulative contributions from all the sources in the region. The deter-mination of these hazards can be performed systematically. In the present case, this determination is based on the recently developed fault-rupture

i 5 j model of DerKiureghian and Anq J 977). The details of this model will not. be d'escribed here; it will suffice only to point out the main features of this model. The DerKiureghian-Ang model is physically consistent with the important characteristics of an earthquake event. It is based on the assunption that an earthquake originates as a fault break at its focus and propagates as an intermittent series of ruptures or slips in the earth's crust, and that the maximum intensity (e.g. acceleration) of ' round shaking at the site is g determined by the rupture that is closest to the site. Accordingly, the maximum ground acceleration at any point in the neighborhood of an earth-quake attenuates with the shortest distance from the rupture. Finally, f the occurrence of earthquakes with magnitudes M > 4 constitutes a _ Poisson process and, consistent with Richter's magnitude law, the magnitude of an earthquake is described with a truncated exponential distribution (Rosenblueth,i 1973). j l Schematically, the major fault systems in the region of the Diablo ,i

/

-Canyon power plant is depicted in Fig.1. For each of the faults indi- 'U i cated in Fig.1, the frequencies of earthquakes of magnitude M > 4 are j / assumed to be the same as those observed historically in this region; 7 < / furthermore, within a given fault, earthquakes are equally likely to - occur anywhere along the fault. Also, during a given earthquake, the length of the fault rupture will be a function of the magnitude and appropriate relationship for this purpose must be specified. The maximum intensity at the plant site, of course, will also depend on the distance to the closest rupture for an earthquake of given magnitude; as illustrated in Fig. 2a, if an earthquake occurs on cne of the faults of "r + - - - _ _.,

~ o e 6 such a magnitude as to produce a rupture length 1, then the maximum acceler-l ation at the site will be detemined by the distance d as shown in Fig. 2a. Besides the hazards of earthquakes from the known faults, earthquakes may occur also anywhere spatially within a nundred-kflometer radius; that l 1 is, the epicenter may not necessarily be on one of the known fault systems. Earthquakes from these latter sources are assumed to have fault ruptures that may propogate in any direction from its focus as illustrated in Fig. 2b; the maximum acceleration, however, is still detemined by the shortest distance d to the causative fault. Appropriate relationship for the attenuation of the maximum acceleration with d, therefore, will be required. In using the fault rupture model of DerKiureghian and Ang, a number of physical relationships as well as parameter values must be specified for the site in question. In this regard, the specific relationships and parameter values used for the Diablo Canyon plant site are based on infor-mation and data directly pertinent to the site. With this model, the known physical relationships as well as all historical data of earthquakes and pertinent seismological information are synthesized systematically in a manner consistent with the major characteristics of earthquakes. Of course, the validity of the results will still depend on the reliability of the data and information used with the model to obtain the calculated results. In the present case, the bases for such infomation and data are described in Sect. 3.1. On these bases, the total hazards of the Diablo Canyon power plant site _ is then obtained as the sum of all potential sources within a region of 100 kilometers.,The result is the seismic hazard curve representing the annual probability of exceedance associated with specified maximum ground accelerations. ~ - ' ~

O O 7 2.3 Determination of Seismic Resistances The capabilities of the structure and equipments to resist specified levels of ground motions will obviously influence the probability of damage. However, for the purpose of calculating the underlying failure probabilities. - it is not necessary to know the actual resistance capability or fragility limit of a specific structural couponent or piece of equipment; it will be sufficient to know its resistance capability relative to the maximum motion to which it is subjected. This information would be contained in the factors of safety used in the design of the structural cocponents; similarly, the margins or degrees of conservatism underlying the design of the struc-ture or equipment would also provide equivalent information. In other words, one of the parameters required in the determination of the damage probability _ is the relative value between the median (or mean) value of the resistance 4 to the specified design earthquake (i.e. the safe shutdown earthquake originally prescribed for the design of the plant).

  • \\

Another resistance parameter that has a bearing on the damage prob-ability is the degree of uncertainty underlying the prediction of resistance, which may be expressed conveniently in terms of the coefficient-of-variation. One of the main tasks, therefore, in the evaluation of the damage prob-abilities is the assessment of the various sources of uncertainty associated with the prediction of seismic resistance or fragility limits. Available observed data, including information on ranges of observed measurements, umy be used in this determination. Invariably, however, the available data and information will be insufficient to provide completely objective bases for assessing the underlying degree of uncertainty. In this light, it -' will be necessary to augment the available information with engineering e---- _._y- . -,, - -, - =

o .e n, 8 judgment; the necessary judgments, however, may have'to be expressed in or translated into probability terms in order to derive the appropriate coefficients-of-variation. 2.4 Calculation of Damage Probability At a given level of gmund acceleration, say a,, the probability of damage of a component with resistance distribution function F (a) would R simply be s = F (a,) (2.1) pp R I which is the area in the probability density function of R below a, as shown in Fig. 3. However, since there could conceivably be a wide range of possible ground accelerations that may be described as a random variable with probability density function f (a), the probability of damage must.be A integrated over all possible values of a.

Hence, i

F (*) # (a) da (2.2) pp R A = o It may be observed that the density function f (a) is the negative derivative A of the seismic hazard curve shown in Fig. 4a. Invariably, numerical integration of Eq. 2.2 is necessary; this may be performed as, [ F (a) aF (a) (2.3) pp = R A all a's wMm. AF (a) = F (a + y) - F (a - f), A A A 1

I 9 j 1 is the difference in the ordinates of the hazard curve over a small acceleration interval aa as shown in Fig. 4a. AF (a) involves the difference between two small probabilities In Eq. 2.3, g in the hazard curve; because of the lack of high degree of precision under-lying the hazard curve, Eq. 2.3 may accumulate significant error. An alternative but equivalent formulation for p is, p pp= [1 - F (a)] f (a) da (2.4) A R o in which 1-F (a) is the exceedance probability of acceleration a, as given A by the hazard curve of Fig. 4a. Numerically, Eq. 2.4 can be evaluated as, pF [ [1 - F (a)] aF (a) (2.5) = A R all a's where, AF (a) = F (a + f) - F (a - ), is the probability that the p, p, R resistance will be in a cmall interval aa, as represented by the incremental area shown in Fig. 4b. Prescribing the lognonnal distribution for the resistance R. we__ obtain FI"}" } ('} R cR in which 6(x) is the standard normal probability, and, c = dn(1+n ) (2.7) 2 a where r is the median of R; and O is the coefficient-of-variation repre-R senting the uncertainty in R.

10 The damage probability described above applies to individual compon-ents, where AF (a) is evaluated for suitable increments of a. However, in R the case of a piping system, the failure of any critical section along the pipe would constitute failure of the entire system. In such a case, if there are n independent critical sections in a pipe, the resistance distri-bution of the piping system would be F (a) = 1 - [1-F (a)]" (2.8) R j in which F (a) is the resistance of one critical section. j The incremental probability AF (a) een comes R AF (a) = n[1-F (a)]N F (a) (2.9) A R j j Furthermore, it is generally recognized (e.g. WASH 1400) that failure of two or more systems would be necessary to cause release of radioactive material in the plant; for example, the joint failure of the piping system and the electrical safety control equipment. In such a case, the cumulative j distribution function of the resistance becomes F (a) = {l - F (a)]") F (a) (2.10) R j E where F (a) is the distribution function of the equipment fragility. From E which the incremental resistance probability, therefore, is AF (a) = {l - [1-F)(a)]") AF (a) + (2.11 ) R E nF (a) [1-F (a)]"~I aF(a) E j j In E,qs. 2.8 and 2.10, the distribution functions of the respective f seismic resistances are also assumed to be lognormal. L

s 9-11 III. Pertinent Data and Information Base As indicated earlier, specific physical relationships and parameter values are needed in the calculation of the required probabilities. The information base used for developing such relationships and for evaluating the specific parameter values are stamarized below.~ In every case, infor-mation and data pertinent to the Diablo Canyon site are used for these purposes. However, where objective infomation is not available or sufficient, subjective judgments are introduced to augment the available information base; when necessary, such judgments would generally tend to be on the conservative side. 3.1 Data for Seismic Hazard Analysis Obviously, the level of seismic hazard at the Diablo Canyon power plant site would depend on the seismicity of the major faults in the region 1 surrounding the site. The locations of the major faults have been identi-fied geologically; these are sumarized in Fig.1. At the closest point, the surface distances of these faults from the Diablo Canyon plant as well as the lengths of these faults are shown in the. table below. Besides the ^ known faults, it is assumed that earthquakes could originate also at any point within a 100-km radius from the plant; fault ruptures from such earthquakes could propagate in any direction. Potential sources within this 100-km radius am represented by the four annular areas described below, s ~ l

12 Seismic Surface Distance Fault Length, Source from site, km km San Andreas 75 400 Santa Lucia Bank 48 55 Rinconada-0zena 30 195 Santa Ynez 88 130 San Simeon 25 1 32 Hosgri 6 107 Combined Hosgri and San Simeon 6 220 g 1 12 2 37 3 62 4 90 Seismicity of Each Fault -- The frequency of significant earthquakes i.e. earthquakes with M > 4, are different among the various faults in this region. As expected, the average occurrence rate of such earthquakes is highest for the San Andreas fault. In any case, the occurrence rate of significant earthquakes for each of the faults are detennined or evaluated on the basis of the epicenter map for this region, as shown in Fig. 5 which is reproduced from Volume II of the Final Safety Analysis Report (1974) for the Diablo Canyon power plant. Fig. 5 represents the epicenter d map for this region for the period 1934 through 1971. In the case of the i Hosgri, the relocated epicenters reported by Gawthorp (1975) are also used. In evaluating the average occurrence rates for the respective faults, each I

O e j 13 i epicenter shown in Fig. 5 is assigned to one of the faults within about In this way, an actual count of the epicenters for the 5 kilometers. respective faults are made, from which the average occurrence rate fcr each fault is then determined. On this basis the annual occurrence rate 4 for the various faults were obtained as shown in the table below. of M 2 For the areal sources, the occurrence rates indicated below represent those for each of the annular areas. Seismic Annual Occurrence Upper-Bound Source Rate of M 2 4 Magnitude, mu San Andreas 6.0 8.0 Santa Lucia Bank 0.92 7.5 Rinconada Ogene. 0.92 7.5 Santa Ynez 0.41 7.5 San Simeon 0.46 7.5 Hosgri 0.22 7.5 Combined Hosgri 0.68 7.5 & San Simeon Annular areas: 1 0.09 6.5 2 0.27 6.5 i 3 9.45 0.5 4 5.40 4.5 Frequency Distribution of Magnitude -- The magnitude of significant earthquakes (i.e. with M > 4) can be describt.d with the shifted exponential distribution (Rosenblueth,1973), as fol%.5: F (m) = k[1 - e"8(*~4)]; 4 5m5m (3.1 ) u g

14 where k = l-e~0("u" ), in which m is the upper-bound magnitude. Because y the results could be sensitive to mu (DerKiureghian and Ang,1977), reasonable values of m must be specified for each of the potential earth-u quake sources; in the present case, the m, assigned for the various sources in the region are as shown in the table above. The parameter s in Eq. 3.1 influences the relative frequencies of different magnitudes, and therefore must also be carefully evaluated. This may be obtained as the slope to the magnitude recurrence curve for the region. The magnitude recurrence curve is.a plot of the earthquake magni-tude M versus the annual number of events with magnitude greater than or equal to M. The frequency of different earthquake magnitudes in this region has been summarized by Anderson and Trifunac (1976); on the basis _ of these data, the recurrence line shown in Fig. 6 is developed. The recurrence line shown in Fig. 6 can be observed to be slightly conservative, as it is above almost all of the data points plotted on this figure. The slope of this recurrence line yields a value of 8 = 2.01. The Attenuation Equation -- The attenuation of the maximum accelera-tion as a function of the shortest surface distance to the fault rupture (or causative fault) is needed with the seismic risk model of DerKiureghian and Ang. For the Southern California region, data of various earthquakes representing the maximum acceleration as a function of the distance to the causative fault for various large-nsgnitude earthquakes have been reported by Page, et al (1972). These data are summarized in Fig. 7. Page, et al (1972) also observed that the maximum ground acceleration attenuates with the shortest distance d to the slipped fault at a rate of d'I* to d .0 On the basis of these data, the following attenuation for maximum accelera-tion is developed:

o ~ l 15 a = 1.35 e.67M(d + 15)-1.75 (3.2). 0 The maximum accelerations given by Eq. 3.2 may be compared with the observed data for this region in Fig. 7. It may be eghasized that in Eq. 3.2, d is the horizontal or surface distance to the slipped fault. fI Rupture Length Equation -- The rupture length is obviously a function of the earthquake mhgnitude. For this purpose, the rupture length equation This m1ation-proposed by Patwardhan, Tocher, and Savage (1975) is used. ship can be repmsented by I=10(0.91M-4.62) (3.3) which gives rupture lengths in kilometers. Specific rupture lengths given by Eq. 3.3 for various magnitudes can be seen in the following table. Magnitude Rupture Length, km l 4 0.1 5 1 6 7 6.5 20 7 56 7.5 160 8 457 3.2 The Seismic Hazard Curve Using the data described above. and with the DerKiureghian-Ano seisInic risk model, the annual probability of exceedanca for all levels of ground acceleration (up to 1.59) are calculated. The resul t is thg_______ seismic hazard curve for the Diablo Canyon power nlant :ite. e i

a I 16 L Three specific cases were considered as follows: (1) Considering all seismic sources in the region. (2) Considering all sources in the region, and assuming that the Hosgri and the San Simeon faults constitute a _sjngle_. fault..This case assumes that the Hosgri is an extention of the San Shnn fault. (3) Considering only the Hosgri fault. This case calculates the I risk associated with earthquakes _or_iginating o'n the.Hosgri _ fault _only The resulting seismic hazard curves for all the three cases described above are" summarized in Fig. 8. The annual exceedance probabilities for specific values of the maximum ground acceleration, ranging from 0 to 1.5g, are also presented in Table 3.1. The results presented in Fig. 8 also included the effects of the uncertainties underlying the seismic risk model. The most significant of these uncertainties is that associated with the attenuation equation, Eq. 3.2. From the scatter of the observed accelerations shown in Fig. 7, the coefficient-of-variation representing the uncertainty underlying the above attenuation equation is estimated to be around 30%. In short, the hazard curves presented in Fig. 8 has been corrected for the uncertainties associated with the seismic risk model including in,particular, the major uncertainty underlying the attenuation equation. According to the results of the three cases considered herein, the following information (see Fig. 8) may be inferred: (1) At low levels of acceleration, there is little difference in the seismic risk whether or not the Hosgri is an extention of the San Simeon fault. However, at the higher acceleration levels (> 0.5g), there is noticeable difference in the risk between whether the Hosgri and the San Simeon faults are two separate faults or constitute a single longer fault. The difference is about 0.1g for return periods > 2500 years. 7

17 (ii) For the higher acceleration levels, say.> 0.59, the seismic e \\f hazard at the Diablo Canyon power olant ef ta ic vi t'nlly -tinly4%, the Hosgri fault. At the lower levels of acceleration, the contributions from the other sources are more significant. ~~ 3.3 Response and Capacity Determinations With regard to the evaluation of the structural conponents, it is the seismic capacity of a component relative to the maximum applied earthquake motion that is relevant. In describing the capacity of a structural component,-it will suffice to know the factor of safety used in its design, as this then determines the design capacity relative to the specified design earthquake, normally prescribed as the safe shutdown earthquake (SSE). Structural components of nuclear power plant structures are be5ieved to have an underlying factor of safety that is about three times that for ordinary structures (Newmark,1975). Thus, the actual factor of safety for_ nuclear plant structures would be between 10 and 20; i.e., the median ymie=4c Nsis+= as would be 10 to 20 times the design _ earthquake disturbance. Alternatively, as in the case of equipment,s, some reserve capacity is nonnally assured on the basis of qualification tests. According to the IEEE Standards (1974,1975), light equipments are normally proof-tested for vibrations with maximum acceleration at the equipment mounting that is 10% above that expected during an SSE. The capacities of heavy equip-ments, however, are determined through calculations. The mean capacities any be determined by assuming that there is a relatively rare chance that an equipment that has been qualified or calculated according to IEEE en m 7-

o e 18 standards may fail. In this regard, it will be assumed that light equipments such as electrical or electronic safety control devices wil have a failure probab_ility of__0.01, whereas heavy mechanical equipment will be assumed to have an underlying failure probability of 0.001. On this basis and with a specified SSE of 0.4g, the mean seismic capacity of equipments are then determined. With regard to the seismic capabilities of the piping system, it will be assumed that each critical section of a pipe is designed to i withstand the design SSE with an underlying failure probability of 0.001._ Pipings and heavy mechanical equipments can undergo some inelastic deforma-tions, and for this reason are ssumed to have a failure probability of 0.001; whereas light electrical equipments probably do not have any reserve ductility and therefom would have a higher failure probability, namely 0.01. Homover, very high acceleration levels can be expected in the case of light equipments when subjected to the floor spectrum within the plant. The degree of uncertainty in both the response prediction.,and capacity determination of the structure as well as the various equipments are also important in evaluating the damage probabilities. The various sources of uncertainty and their combined effects are assessed systematically below. Uncertainty in Structural Response Prediction and Capacity -- In the case of the structural response prediction, the several sources of uncer-tainty are as follows: (1) Effects of Site Dependent Factors on the Definition of Ground Motions -- This would include the effect of local soil conditions and geology on the predicted ground motion at the site as detennined by the seismic hazard curve described earlier. Information that can be used to assess the level of uncertainty associated with these factors is scarce. c --e- ~a

e e .e 19 Based largely on subjective judgment, this uncertainty is estimated to be on the order of 20%, expressed in terms of coefficient-of-variation (c.'o.v.). (ii) Variability in Amplification Factors -- Mohraz, Hall, and Newmark (1972) in their study of a large number of earthquakes have reported the 50 and 90-percentile values of the aglification factors for displacement, velocity, and acceleration; from this study, the 50% and 90% values of the respective amplifications are as follows: Disp. Vel. Accel. Amp 1. Aml. Amp 1. 50% value 1.40 1.66 2.11 90% value 2.21 2.51 2.82 Assuming that the amplification factors can be described as Gaussian random variables, the variabilities for the respective amplifications (expressed in terms of c.o.v.) are: a = 45% d 0,= 40% n,= 26% Therefore, a coefficient-of-variation of 30% to represent the uncertainty associated with the variability in the amplification factor for acceleration appears reasonable and will be used in this study. (iii) Uncertainty in the Specification of the Structural Damping -- The uncertainty associated with the specification of damping coefficient used in the response calculation of a structure may be of the order of 30%, as has been suggested by Newmark (1974). (iv) Modeling of Structure for Dynamic Analysis -- Because of the complexity of a nuclear power plant structure, there is bound to be signifi-cant uncertainty in modeling the structure for dynamic response analysis.

e 20 This should include.the effect of structure-soil interaction when pertinent. It is conceivable that this uncertainty could be of the order of 30%, which Since the Diablo would include the effects of soil-structure interaction. Canyon power plant is built on rock, where the effects of soil-structure ~ interaction would not be significant, the 30% assigned herein could be on the conservative side. On the basis of the indivfdual uncertainty levels assessed above, the total uncertainty associated with the prediction of structural response, _ therefore, is .202 + 0.302 + 0.302 + 0.302 n t-as C 56 = p With regarc to the ancertainty underlying the estimation of the capacity ter structural compos.ent, it is assumed that this uncertainty of the structura will be comparable to that associated with. the prediction of the resistance I to blast forces (Ang, iia 17, and Hendron,1974), which has been estimated to

  • 1 he 30%.

fTherefore, the total uncertainty underlying both the response prediction and structural capacity tecomes e s 0.562 + 0.32 0.64 = Estructure From which, the parameter c of Eq. 2.7 is, 2 n(1 + 0.64 ) = 0.59 = 5tructure Uncertainty in the Response Prediction and Resistance of Piping Systems -- The forcing function on a piping system would involve the floor response spectra of the structure. Accordingly, any uncertainty in the ap _ 41

e 21 prediction of the response of the piping system wo'uld be above and beyond those estimated earlier for the structure. This additional uncertainty should include those underlying the in-s'tructure motions such as the floor response spectra that constitute the input to the response analysis of 'the piping system, as well as any imperfection associated with the modeling and response analysis of the piping system itself. Considering the complexity of pipings in a nuclear power plant, the uncer'rinty associated with these factors can be expected to be quite large; conceivally, this could range from 30% to 50%. For the purpose of this study, a coefficient-of-variation of 50% will be assumed to represent the level of uncertainty in the pre-diction of piping system response. Therefore, the total uncertainty associated with the piping system response would be Ses= 0.562 + 0.502 = 0.75 With regard to the resistance of a pipe, it is reasonable to assume that pipes can take some inelastic action; for example, it is not unreason-able to expect a ductility factor of 1.5 for pipes. Accordingly, the uncertainty in the resistance of a piping system would be comparable to that of a structural component, and thus a coefficient-of-variation for the piping resistance of 30% may be assigned. The total uncertainty associated with the response calculation and resistance of a piping system, therefore, is pp,=/0.752 + 0.302 = 0.81 o j Thus. 2 r.p9p, = [in(1 + 0.81 ) = 0.71 e-oy w -,w-,-~ ,,,,ev- ,-n e ev4 -,one-4,m,- 4m.r >-~9,--mmr-+-ser--- ---n-,--

e .. ~ _.. 22 Uncertainty in the Response Prediction and Fragility of Et dipment -- In the case of equipments, the additional uncertainty in the in structure - motions as well as in the modeling and response analysis' of the equipment would be the same as those of the piping system; these uncertainties, of course, would again be above and beyond those underlying the response I 4 prediction of the structure. In other words, a coefficient-of-variation of 50% may also be used for the response prediction of equipment. Accord-L l ingly, the uncertainty associated with the response prc<'.iction of equip-ment would also be, 4 e /0.562 + 0.502 0.75 ay,= = Information on the seismic capacity of equipments are generally t I proprietary (IEEE Standards,1974). For this reason, specific information t i on the fragility limits of equipments used in nuclear power plants are quite rare. However, infomation on fragility limits of equipments sub-- j jected to blast-induced nuclear ground shocks is available Newmark, et al (1963); according to Ang, Hall, and Hendron (1974), the coefficients-of-4 i variation associated with these data are 60% for heavy mechanical equipments, [ and 80% for light electrical and electronic equipments. It follows then that the total uncertainty underlying the response prediction and fragility i of a heavy mechanical equipment is HE = [0.752 + 0.602 = 0.96 D From which, 2 cHE = n(1 + 0.96 ) = 0.81 I e e >+ i---*,y .y.- - -, -., ,.yy._, -y y-,,,,,,i,

.rw-.---ps.,,,,y.w,y pe,s r-g.--.,.,--%,--e,-,--,

y3-., ,e-yq wwgww-,-w-=

r

'.,n 23 Similarly, for light electronic and electrical-equipments, the total ~ up.:ertainty would be SE 0.752 + 0.802 = 1.10 =

and, 2

LE = /An(1 + 1.10 ) = 0.89 C The levels of uncertainties assessed and determined above are summarized in the following; all art expressed in terms of coefficients-of-variation, i Sunnary of Uncertainties f

Response

Resistance ' Total c. l Structure 0.56 0.30 0.64 0.59 Piping 0.75 0.30 0.81 0.71 Equipments: Heavy Mechanical 0.75 0.60 0.96 0.81 Light Electrical 0.75 0.80 1.10 0.89 IV. Calculated Probabilities With the formulations described in Chapter 2 and the specific relation-ships and parameter values described in Chapter 3, the probabilities of specific adverse events involving the damage of major structural conponents, piping system, or safety equipment, or combinations thereof are evaluated. In all cases, damage is defined to mean that the seismic capacity of the couponent or equipment is exceeded by the maximum seismic acceleration occurring at the site. Also, in all cases, these probabilities are calculated and presented in terms of the annual probability of damage.

l 24 p It may be emphasized that the calculated probabilities are based on. I certain tacit assumptions regarding the manner in which the various components are designed, as described in Chapter III. Such probabilities are necessarily dependent on the criteria under which a conponent is ~ designed; for exanple, on the factor of safety and on the safe shut-down earthquake specified in the design of such conponents.. In this regard, l l a range of design conditions were assumed; specifically, factors of safety b ranging from 10 to 20 am assumed to be the conditions under which the major structural components of a nuclear power plant are designed. In the case of equipments, and the piping system, it is more difficult to deter-I mine the factors of safety underlying their design; equ'ivalently, however, and for purposes of this calculation, it is assumed that the piping system is designed to withstand the specified SSE of 0.4a with an implied damage probability of 0.001. Similarly, the equipments are assumed to be de_s_igned to withstand the specified SSE with an underlying failure probability of 0.01 for light equipments and a failure probability of 0.001 for heavy mechanical equipments. 4.1 Damage of Structural Cogonents The probability of damage of specific structural components is evaluated using Eq. 2.5, in which 1-F (a) is the annual seismic hazard A curve of Fig. 8 and F (a) is assumed to be lognormal. R l The results of the calculations are sunnarized in Table 4.1, which shows l the incremental probabilities for accelerations between 0 and 1.5g in incre-ments of 0.12g. The relevant damage probability is then the sum of all these incremental probabilities. In this case, the probability of damage of a major structural component is evaluated for factors of safety of 10 and 20; c 1

= ~ 25 also depending on whether the Hosgri and San Simeon faults are separate faults or constitute a single fault, the results are slightly different. The final results can be summarized as follows: 4 ^ Annual'Damane Probability of' Structural Component Separate Hosgri Combined Hosgri & San Simeon & San Simeon Safety Factor = 10 1.29 x 10-6 1.88 x 10-6 Safety Factor = 20 3.31 x 10-8 5.12 x 10-8 j_ The nespective contribution to the total damage probability from the different ranges of ground accelerations can be seen in Figs. 9 and 10; the step size shown in these figures represent the incremental prob-abilities associated with the respective acceleration increments. On this basis, the range of ground accelerations that has major contributions to the damage probability may be observed. It may be emphasized that the probabilities calculated above pertair to the damage of individual structural conponents. In general, it does not necessarily mean the collapse of the cogonent. Moreover, in the case f, of redundant structures, the damage of a single structural component map ~ / not be of serious consequence to the structural system. However, common design and construction procedures, plus stringent inspection requirements to achieve unifonn resistance quality, would tend to increase the correla-tion between the resistances of various conponents. For this reason, damage to one structural component may likely cause similar damage to other components in the same structure. 4.2 Danage of Piping System In the case of a piping system, the resistance along the length of a pipe will generally be correlated; however, because there are numerous -p.. .-~,,,..-, -, - -

3 26 velded connections of a piping system in a nuclear power plant, many of which could be potential locations of weakness, it is likely that some of 1 the pipe sections may be uncorrelated. In a given piping system, it is l ' probably reasonable to stipulate that there could be several, say n, independent sections of potential failure locations in the system. Clearly, the occurrence of damage (for example, the occurrence of a macro crack at a welded joint) in one section is ta'ntamount to failure of the piping system; hence failure in this case is determined by the weakest among the n independent sections. Accordingly, therefore, for the piping system, the resistance distribution function would be defined by Eq. 2.8, in'.which F (a) is the distribution function of the resistance of one pipe j section. I The value of n would probably range between 1 and 10. Calculations for n = 1, 5, and 10 are sumarized in Table 4.2. Again, the incremental probabilities in each case are shown in Table 4.2 for accelerations ranging from 0 to 1.5g in increments of 0.12g. The respective damage probabilities is the sum of all the incremental probabilities. It can be observed on the basis of the results sumarized in Table 4.2 that the damage probability of the piping system, depending on the value of n, is as follows: .n_ Separate Hosgri Combined Hosgri & San Simeon & San Simeon 1 4.08 x 10-6 5.96 x 10-6 d 5 1.96 x 10-5 2.86 x 10-5 10 3.76 x 10-5 5.46 x 10-5 In the r.asa of the piping system in a nuclear power plant, it is probably realistic and reasonable to assume a value of n = 5. On this basis, then the damage probability of the piping system would be of the order of 10-5 annually.

.~ 27 .The' inemmental probabilities representing the~ contributions to the ~ - total damage probability from the various acceleration increments are portrayed in Fig.11 (for n = 1). 4.3 = Piping Damage and Malfunction of Light Equipment An adverse event that may likely lead to release of radioactive material is the joint occurrence of damage somewhere along the piping system and the failure or malfunction of the safety control equipment. The pmbability of this joint event can be calculated assuming that the-resistances of the piping system and the control equipment are statistically ' ndependent, but are subject to the comon seismic environment of ground i acceleration. In this case, therefore, the joint resistance distribution s function is given by Eq. 2.10. Again, three values of n for the piping system are used in the calculations. The results are sumarized in Table 4.3, which shows the incremental probabilities for all levels of: ground acceleration, yielding the annual probability of this joint event for three values of n as follows: = 3.01 x 10-7 n = 1; pp p = 1.40 x 10~0 n = 5; p = 2.57 x 10-6 n = 10; pp x .These probabilities were obtained assuming that the Hosgri and San Simeon l faults are separate faults. The relative contributions to the total damage probability from the 1 Various increments of ground acceleration can be seen from the different t step size of the curve in-Fig. 12. 4 m

r I e, ,a

  • I I

l l 28 4.4 Piping Damage and Failure of Heavy Equipment Another adverse event of concem is the joint occurrence of a major ~ pipe damage somewhere in the piping system of the plant and the simul-taneous malfunction of certain heavy safety equipment such as the emerge'ncy core cooling system (ECCS). The calculations involved here are similar to those described in Sect. 4.3, with the exception that for a heavy mechanical equipment such as the ECCS, it is assumed that the equipment is designed to withstand the SSE of 0.4g with an underlying failure probability of 0.001, in contrast to 0.01 for a light equipment. The results of the calculations are sumarized for this case in Table 4.4, showing the incremental probabilities for the entire range of In this case, the probability of this adverse accelerations of interest. event, for the three values of n, are as follows: = 8.39 x 10-8 n = 1; pp = 3.85 x 10-7 n = 5; pp = 6.97 x 10~7 n = 10; pF The results for n = 1 are also portrayed in Fig.13, showing the t relative contributions to the total damage probability from the different acceleration increments. V. Summary Seismic safety of the Diablo Canyon nuclear power plant has been systematically evaluated, in tems of the annual probabilities of adverse events involving the damage or malfunction of one or more components in the plant; such events are those that could potentially lead to release f ~ f

.o 29 of radioactive material. The specific events and associated probabilities can be summarized as follows: Damage to Structural Components -- 3.3 x 10-8 for safety ' factor = 20; I to 1.3 x 10-6 for safety factor = 10. Damage to Piping System -- 4.0 x 10-6 for.n = 1; to 3.8 x 10-5 for n = 10. Damage to Piping and Malfunction of Light Electrical Equipment -- 3.0 x 10-7 for n = 1; to 2.6 x 10-6 for n = 10. Damage to Piping and Malfunction of ECCS -- 8.4 x 10-8 for n = 1; [ to 7.0 x 10-7 for n = 10. The probabilities calculated herein pertain to all levels of potential earthquake hazards at the site. In obtaining these probabilities, a careful analysis of the seismic hazard for the site was performed using a i fault-rupture model for this purpose, giving the probability of exceedance for all significant ground accelerations expected at the site. Also, a careful analyes of the seismic resistances of major structural components and equipments were perfonned, including an-tassessment of the uncertainties underlying the respective resistance functions. The relevant probabilities are then evaluated by integrating the damage probabilities associated with all accelerations from 0 to 1.5g. References 1. Anderson, J. G., and Trifunac, M. C., " Uniform Risk Absolute Accelera-tion Spectra for the Diablo Canyon Site, California," Rept. to the ACRS, U.S. Nuclear Regulatory Comission,1976. 2. Ang, A.- H-S., Hall, W. J., and Hendron, A. J., " Uncertainty and Survivability Evaluations of Design to Air-Blast and Ground Shock," Final Report Contract DACA 88-73-C-0040 to U.S. Army Construction Engineering Research Laboratory, June 1974.

r 2. 30 Der-Kiureghian, A., and Ang, A. H-S., "A Fault-Rupture Model for s 3. Seismic Risk Analysis," Bulletin, Seismological Society of America, -August 1977.- Inst. of Electrical and Electronics Engineers, Inc. "IEEE Standard for 4. Qualifying Class IE Equipment for Nuclear Power Generating Stations," IEEE Std. 323-1974, 1974.

5. -Inst. of Electrical and Electronics Engineers, "IEEE Recomended Practices 4

for Seismic Qualification of Class 1E Equipment for Nuclear Power Generating Stations," IEEE Std. 344-1975, 1975. Mohraz, B., Hall, W. J., and Newmark, N. M., "A Study of Vertical and 6. Horizontal Earthquake Spectra," Contract AT(49-5)-2667, Atomic Energy Comission, N. M. Newmark Consulting Engineering Services, Dec.1972. Newmark,~ N. M., " Comments on Conservatism in Earthquake Resistant . 7. Design," Report to U.S. Nuclear Regulatory Comission, ' Sept.1974. 8. Newmark,- N. M., " Probability of Predicted Seismic Damage in Relation to Nuclear Reactor Facility Design," Draft Rept. to Nuclear Regulatory Comission, Contract AT(49-24)-0116, N. M. Newmark Consulting Engineering Services Sept.1975. Newmark, liansen, and Associates, " Vulnerability Handbook for Hardened 9. ~ Installations," Vol. I, Final Rept. Contract AF 33(657)-10287.Dec. .1 %3. Pacific Gas and Electric, Final Safety Analysis Report, Units 1 and 10. -2 of Diablo Canyon Site, Vol. II, USAEC Docket Nos. 50-275 and 50-323, 1974. 4, Page, R. A., Boore, D. M., Joyner, W. B., and Coulter, H. W., " Ground - 11, Motion Values for Use in the Seismic Design 'of the Trans-Alaska Pipeline [ System," Geological Survey Circular 672, 1972. 4 Patwardhan, A. S., Tocher, D., and Savage, E. D., " Relationship Between 12. Earthquake Magnitude and Length of Rupture Surface Based on Aftershock - Zones," 70th Annual Mtg., Seismological Soc. of Am., Earthquake Notes, Eastem Section, Vol. 46, No. 3. July, Sept.1975. 2 Rosenblueth, E., " Analysis of Risk," Proc. 5th World Conf. on Earthquake 13. Engineering, Rome, Italy,1973. J l a ~

y 31 ~ l] Table 3.1 ~ Annual Probability of Exeedance. 1 i .All Sources Annual Separate Combined Hosgri Max. Ground Hosgri & Hosgri & Only Accel., o San Simeon San Simeon -2 5.1 x 10-2 6.0 x 10-2 1.6 x 10 O.10 7.5 x 10-3 9.2 x 10-3.0 x 10-2 l; 0.20 1.9 x 10-3 2.'4 x 10-3 1.0 x 10-3 0.30 ~ 7.5 x 10-4 1.0 x 10-3 5'.0 x 10-4 0.40 -4 3.6 x 10-4 d3 x 10) 2.9 x 10 0.50 t -4 3.1 x 10-4 1.8 x 10~4 0.60 2.0 x 10 1.2 x 10'4 1.9 x 10'4 1.2 x 10-4 0.70 -5 -5 -),5 x 10'4 9.5 x 10 0.75 9.8 x 10 ~ 7.8 x 10-5 '1.2 x 10~4 7.7 x 10-5 0.80 5.1 x 10-5 7.9 x 10-5 5.1 x 10-5 0.90 -5 5.0 x 10-5 3.2 x 10-5 1.00 3.2 x 10 -5 3.2 x 10-5 2.1 x 10 1.10 2.1 x 10 1.2 x 10-5 1.9 x 10-5 1.2 x 10-5 1.20 7.4 x 10-6 1.1 x 10-5 7.4 x 10-5 1.30 5.0 x 10-6 7.2 x 10-6 5.0 x 10-6 1.40 -6 2.7 x 10-6 4.2 x 10-6 2.7 x 10 1.50 i l l 4 l t b I

^ *

  • ~ -.

32 Table 4.1 Damage Probability of Structure NF Separate Hosgri & Combined Hosgri &, San Simeon San Simeon Accel., o F.S. = 10 F.S. = 20 F.S. = 10 F.S. = 20 0.12 2.341 x 10-9 1.590 x 10-12 2.765 x 10-9 1.879 x 10-12 -II 2.743 x 10-8 6.400 x 10-II 0.24 2.166 x 10-8 5.052 x 10 0.36 6.380 x 10 2.937 x 10 8.612 x 10-8 3.964 x 10-10 -8 -0 9 0.48 1.136 x 10-7 8.818 x 10-10 1.638 x 10-1.272 x 10 0.60 1.595' x 10-7 1.888 x 10-9 2.404 x 10-7 2.845 x 10-9 0.72 1.899 x 10-7 3.193 x 10-9 2.924 x 10-7 4.917 x 10-9 0.84 1.960 x 10 4.451 x 10-9 3.043 x 10-7 6.910 x 10-9 -7 0.96 1.765 x 10-7 5.210 x 10 2.749 x 10-7 8.111 x 10-9 -9 1.08 1.419 x 10-7 5.317 x 10 2.212 x 10-7 8.285 x 10-9 -9 1.20 1.050 x 10-7 4.788 x 10-9 1.637 x 10-7 7.465 x 10-9 1.32 7.174 x 10-8 3.968 x 10-9 1.119 x 10-7 6.188 x 10-9 i.44 4.636 x 10 3.048 x 10-9 7.234 x 10-8 4.756 x 10-9 -8 1.288 x 10 3.309 x 10 1.961 x 10-9 5.121 x 10-8 -6 -8 pp = e e y y +w- -,w-w-_v..e m --yp y--

s Table 4.2 Damage Probability of Piping, app I u Combined Hosgri & San Simeon Separate Hosgri & San Simeon Accel., g n=1 n=5 n=10 n=1 n=5 n=10 ~0 0.12 1.591 x 10 8.030 x 10 1.605 x 10 1.897 x 10-9.487 x 10-7 1.897 x 10 -7 -7 -6 0.24 4.043 x 10-7 2.021 x 10 4.042 x 10 5.123 x 10 2.561 x 10-6 5.121 x 10-0 i -6 -6 -7 -6 -6 -7 -6 0.36 5.507 x 10-2.750 x 10 5.492 x 10 7.433 x 10 3.712 x 10-6 , 7.413 x 10 O.48 5.844 x 10-7 2.908 x 10-6 5.781 x 10-6 8.42'7 x 10-7 4.193 x 10-6 8.336 x 10-6 0.60 5.699 x 10-7 2.813 x 10-6 5.534 x 10-6 8.590 x 10-7 4.239 x 10-6 8.340 x 10'O i 0.72 5.153 x 10'7 2.506 x 10-6 4.842 x 10-6 7.934 x 10'I 3.859 x 10-6 7.455 x 10-6 O.84 4.312 x 10-7 2.052 x 10-6 3.856 x 10-6 6.693 x 10 3.184 x 10-6 5.985 x 10-6 ~7 0.96 3.282 x 10-7 1.515 x 10-6 2.741 x 10-6 5.111 x 10'7 2.358 x 10-6 4.267 x 10-6 1.08 2.323 x 10'7 1.032 x 10-6 1.778 x 10-6 3.620 x 10-7 1.607 x 10-6 2.771 x 10-6 1.20 1.529 x 10~7 6.475 x 10~7 1.053 x 10-6 2.385 x 10'7 1.010 x 10-6 1.641 x 10-6 1.32 9.554 x 10-8 3.829 x 10-7 5.808 x 10-7 1.490 x 10-7 5.972 x 10~7 9.058 x 10-7 -8 3.326 x 10~7 4.662 x 10-7 1.44 5.666 x 10 2.132 x 10-7 2.988 x 10-7 8.842 x 10 -0 4.081 x 10-6 1.964 x 10-5 3.760 x 10-5 5.960 x 10-6 2.860 x 10-5 5.460 x 10-5 = pF

34 u i Table 4.3 Piping and Light Equipment Damage Probability app L n = 10 Accel., o n=1 n=5 2.498 x 10-II 1.249 x 10-10 7.854 x 10~9 0.12 1.154 x 10-9 5.771 x 10-9 1.154 x 10-8 0.24 6.287 x 10~9 3.141 x 10-8 6.274 x 10-8 0.36 1.572 x 10-8 7.831 x 10-8 1.559 x 10-7 0.48 2.781 x 10-8 1.376 x 10 2.715 x 10-7 -7 0.60 0.72 3.931 x 10 1.922 x 10 3.738 x 10-7 -9 -7 0.84 4.650 x 10 2.233 x 10 4.248 x 10'7 -8 -7 4.705 x 10-8 2.205 x 10-7 4.070 x 10-7 0.96 4.184 x 10-8 1.901 x 10~7 3.378 x 10-7 1.08 3.345 x 10-8 1.463 x 10'7 2.485 x 10-7 1.20 2.471 x 10-8 1.034 x 10-7 1.669 x 10~7 1.32 1.44 1.699 x 10-8 6.772 x 10-8 1.032 x 10-7 3.009 x 10~7 1.397 x 10-6 2.572 x 10-6 = pp e m n

A c 35 Table 4.4 Piping and Heavy Equipment Damage Probability App Accel., a n=1 n=5 n=10 0.12 6.030 x 10-13 3.015 x 10-12 6.030 x 10-12 0.24 7.08.6 x 10-II 3.542 x 10-10 7.084 x 10-10 j 0.36 6.379 x 10-10 3.187 x 10~9 6.366 x 10-9 0.48 2.226 x 10-9 1.109 x 10-8 2.207 x 10-8 0.60 5.034 x 10~9 2.492 x 10 4.920 x 10-8 -8 0.72 8.580 x 10 4.200 x 10 8.180 x 10-8 ~9 -8 0.84 1.184 x 10-8 5.701 x 10-8 1.088 x 10-7 0.96 1.357 x 10-8 6.381 x 10-8 1.183 x 10-7 1.08 1.352 x 10-8 6.177 x 10-8 1.106 x 10-7 1.20 1.185 x 10-8 5.225 x 10-8 8.967 x 10-8 1.32 9.499 x 10-9 4.021 x '10-8 6.581 x 10-8 1.44 7.027 x 10-9 2.842 x 10-8 4,4j4 x 10-8 8.386 x 10-8 3.850 x 10-7 6.971 x 10-7 pp =

.t. i 36 l t 6\\ 1..' N 37 g ]\\ \\ \\ N \\ N e., N s 4;, \\ N \\ 4, 1 y \\\\' g, \\ i \\ \\ Diablo Canyon Ax \\ \\ ower Plant Site P \\ \\ \\ sN \\ \\ \\ \\ N \\ \\ / \\p \\, y yq y / \\ k \\ / X,' \\g g N o .sg 2nef -'g ~\\ f-w- N\\ u 12 2 121 12 0 119 FIG. I MAJOR FAULTS IN REGION OF SITE e

e-----,-.-ee--,-,,-,-----.4,-.,,.,w w--,--

m- -y.-em-...,y-.-w w --=,-w--m,,n9,-

i.. I } N \\ g Epicenter Epicenter / Y= / Rondom / Directions g Diablo Canyon d of Rupture l Diablo Conyon dg Power Plant Site d Power Plant Site g i ~ ~ I FIG. 2a EARTHQUAKE ON A KNOWN FIG. 2b EARTHQUAKE NOT ON A FAULT FAULT 9

38 j + a . { fa(a) into) l + 1 = co a -Area = Fn(c ) e FIG. 3 PROBABILITY DENSITY FUNCTIONS O I l

39 J i Seismic Hazard Curve, I-F (o) 4 l. -o .L AF (c) 1 A I g l d iI o l 0 W I FIG. 4a SEISMIC HAZARD CURVE l i I I v ta) l o 1 I i I I I w 8 Area = AFn(o) 1 FIG. 4 b RESISTANCE DENSITY FUNCTION

O a 40 h, \\ a. r. z [a*.; s,\\,.- a' ~.~ - r,.,. x a i.p g f N s,.f s t a a e s 6 a u s,,,..'..g ./ 5

  • f
  • s

..+. e /% ~. ~ :- - $.*,, s,.' \\,,,,,,,,a 4 *. ,.Q,,&.". 4 e,e-t .%s La., wa.,,, ?,- \\,./ ' ( n. l ,4 *. .1- - . " %< y/.,,...,,-y. s .{s :. ,, ~., - 7s.' 'hW n g'...f. 'e....l' = 1..,.;T . h, 4- ~t. .s + / \\,' .., w. h =armi)'s.

s., c,

..i. p. ~ ,is --{ s.f.I..s .g .h. /= / s. . K '* 1 --- ' / s t, t C W..~ g 36!

  • I
  • i.
    • att#

J ,s ).- .r ,j t,. g t, .3 s p .s.ss +

  • ,,44

.. /. . t, s... /

g. A,s/

s ....._. _. _. j,Wg.:_...,_!. :. _ _ _. _ _ _.._. _.=_. q,. \\! h ./".'t * "* Ausssa tearss saa tuis eerspo L. g 4*

f....
  • 'g' so use osas r tarf t
  • A, 4

,,eg4_ 5 , *M. s

,.[,v

) 2 2s,(. wA ::;, J.pg,gv,/. / :... ' ' ' + s'. i =F,,__.n._.,_._....--_. Satra eansa. .p.' s s'

  • i j

t .r.x. s s

  • W.

=. .i y e s i e n a \\, a ~ g

  • ==

\\ ). +., a.y an .p. 3

t.

.A

.e*

/ -?. e e on e

    • un or
,:,,~

.s p. l l F IG. 5 EPICENTER MAP OF REGION (Reproduced from Final Safety Analysis Report for Diablo Canyon Plant) i -,we-,--,,-,-,3-m.a- --.-,--w,- - - - - - ---.w,----m ,.w-*-.,,-w-sem.--.--,--e.e -c-..------

a 4 0 i 41 Ij' 100 _ i i i i i 4 m 8 S u.s D 2 7 m e 1 = c F m 5._ D a S .3C D fr I I f ~ m o p = 2.01 W 5 E Z O 7 E o.i c <t MS 9 l l l 1 I I k0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 Magnitude FIG. 6 MAGNITUDE RECURRENCE CURVE tu

42 m i i e i i e w N M = 7.5 1.0 N N s g M=6.5 g N N Eq. 3.2 N N 0.52 g s g ~ b.M km ~ s N \\ oO ' N N M = 5.5 s N O ~ O \\" ~ s g O g\\ \\ NO D s \\o >\\ ON \\ D 0.1 g N \\ \\ \\ g \\O g Oe o s NO N Gy O \\ \\ \\ Yg \\00 N 2 s e g o \\ k0 \\ a o O o\\E 3 g \\ k O g ~ Earthquake Magnitude A 5.0-5.9 CD bg O 0 6.0 - 6.9 E 7.0 -7.9 a O We GOO e B-I 0.001 10 10 0 Distance, km FIG. 7 ATTENUATION OF MAXIMUM ACCELERATION WITH DISTANCE TO SLIPPED FAULT (DATA REPRODUCED FROM PAGE, et al,1972) .{ -->w a m

=*s 43 i ~ \\ ( 4 ~ 30 s\\j 10 7 j ~ d, Separate Ho,sgri O 8"" 8I**'" All Sources I I --- Combined Hosgri \\ ~ \\ S Son Simeon s ~ -- Hosgri Fault Only ~ g 8 0 $ E {i S iO' s 5 \\ E 10-* ._m! =)Lt I E u 1 x t W o o t 2 _4 x O I \\ = 4 J \\ / 3 ! 4 O' io' 3 2 16 ' -.*,"[g f h \\ E a:

n.. n

,y g.,-p _ _w. -., s .,eev g g

  • s 4

-.000 ~ . coo s / s N ,o. .,cci / ,g. N g g g s l g j N l N \\ e i i i N t ios io-o f i 0

0. 2 0.4 0.6 0.8 1.0 1.2 f.4 Maximum Accelerotion, g SEISMIC HAZARD CURVES FOR DIABLO CANYON FIG. 8 POWER Pl' ANT SITE

1 44 is I e i2 it m s T9 x e a. 7 .3 O ~a E 6 D. s E <t 4 ~ 3 I il I I I I I I I O O.2 Q4 G6 0.8 ID L2 1.4 1.6 f.8 Acceleration, g FIG. 9 INCREMENTAL CONTRIBUTIONS TO DAMAGE PROBABILITY OF STRUCTURAL COMPONENT fSF = 10)

k Nm s,. 45 4h M r l e nnOO = '9 x b 5 ~ - 2.0 2 E O a = 4 ~ lD i t E I l I I I i i 1 I 'I O Q2 0.4 0.6 G8 - 1.0 1.2 1.4 L6 Le Acceleration, g FIG.10 INCREMENTAL CONTRIBUTIONS TO DAMAGE PROBABILITY' 0F STRUCTURAL COMPONENT (SF=20) ~,,--n,.,

D . g_. 46 E0 i i i i a une i m 40 g I e B o M 3D .u. =a 3 E 3 O 2.0 a g a 4 Ime 1.0 j I i i i i i i i i O R2 0.4 0.6 QS LO l.2 1.4 1.6 1.8 Acceleration, g FIG. Il INCREMENTAL CONTRIBUTIONS TO DAMAGE PROBABILITY OF PIPING SYSTEM (n = 1) ,-,--------,.--.---~-n- -a- ..,.,---,-----,--.-n,--n-.

,,.,---,.,_-,----....n-.,.,_-,,~,,--,..r

,.--n-,.,,,..e,-

I. i 3-- 47 e i i i i 4.0 to I u m e 2.0 = E* o o C e 4 .o l ~ I i i I I I l I' I O O.2 Q4 0.6 0.8 1.0 1.2 1.4 L6 L8 Acceleration, g FIG 12 INCREMENTAL CONTRIBUTION TO PIPING AND LIGHT EQUIPMENT DAMAGE PROBABILITY (n =l) N}}

Retrieved from "https://kanterella.com/w/index.php?title=ML20117M986&oldid=4282663"