ML20212L249
| ML20212L249 | |
| Person / Time | |
|---|---|
| Site: | 05000000, Diablo Canyon |
| Issue date: | 11/24/1976 |
| From: | Allison D Office of Nuclear Reactor Regulation |
| To: | Stolz J Office of Nuclear Reactor Regulation |
| Shared Package | |
| ML20150F500 | List:
|
| References | |
| FOIA-86-391 NUDOCS 8608250262 | |
| Download: ML20212L249 (29) | |
Text
{{#Wiki_filter:i o.' <. ). ) [O.NI Di s tributio_n, A k Fi1V ' O t t 1976 hC Local POR LWR-1 File I. Sihweil MESt0RANDUM FOR: John F. Stolz, Chief J. C. Stepp Light Water Reactors Branch No.1 K. Kapur Division of Project Management J o tellotte FR0ft: Dennis P. Allison, Project Manager ACRS (1) Light Water Reactors Branch No.1 Division of Project Management
SUBJECT:
SEISMIC REEVALUATION OF THE DIABLO CANYON NUCLEAR POWER PLANT The enclosed report prepared by the NRC Staff was provided to the Advisory Committee on Reactor Safeguards for consideration at the Comittee's meeting on November 13, 1976 concerning the Diablo Canyon Plant. gyrr.1017.ed B7 M s y ent.h P I M Dennis P. Allison, Project Manager Light Water Reactors Branch No.1 Division of Project Management e cc: Pacific Cas and Electric Company Ms. Elizabeth E. Apfelbert Attn: Mr. John C. Morrissey 1415 Cazadero Vice President & General Counsel San Luis Obispo, California 77 Deale Street 93401 San Francisco, California 94106 Ms. Sandra A. Silver Philip A. Crane, Jr., Esq. 5055 Radford Avenue Pacific Gas and Electric Company North ifollywood. California 91607 77 Beale Street San Francisco, California 94106 Mr. Gordon A. Silver 5055 Padford Avenue Andrew J. Skaff, Esq. North l'ollywood, California OlC07: California Public Utilities Comission 350 ficAllister Street Paul C. Valentine San Francisco, California 94102 400 Channing Avenue Palto Alto, California 94301 Mr. Frederick Eissler, President Scenic Shoreline-Preservation Yale I. Jones, Esq. Conference, Inc. 100 Van Ness Avenue 4623 More Mesa Drive 19th Floor Santa Barbara, California 93105 San Francisco, California 94102 060m90;?6;.' 06cno 1 o*
- ec e
- Pl)H FOIn
-.p p - DATrison:k1j H m H36-y,1 m f'y Cy d,p- ..,..me Fcrr# FG 99, f aa,e > -.--- h h I Feven Arc.lle iRev. 9 99) A304 0240 W w. e. eovammment paeuttue opreens sere.ese.see
'+ ~ ~ ): .} -NOV 2 4 B7s slohn F. Stolz, Chief cc: James A. Georcaris Center for Law in the Public Interest 10203 Santa Monica Boulevard Los Angeles. California 90067 its. Raye Fleming 1746 Charro Street San Luis Obispo. California 93401 fir. John Forster 985 Palm Street San Luis Obispo, California 93401 lir. ililliam P. Cormiell P.O. Cox 453 florro Bay, Calf fornia 93442 !!r. II. J. Lindblad, Project Engineer Pacific Gas and Electric Company 77 Beale Street San Franc *sco, California 93442 trs. Thelma Hirdler 811 Fair Oaks Avenue Arrowo Grande, California 93420 !!r. Westing,C. Gan91off house Electric Corporation P.O. Box 355 Pittsburgh, Pennsylvania 15230 Thomas J. Hirons Los Alamos Scientific Laboratory Group TD-6, MS 226 P.O. Cox IF63 Los Alamos New f!cico 07FA5 (* e 9rPOS E > eW ANe 888 > -"-'aa Se,e > Paves ABC.)14 (Rev. 9 53) ASCM 0240 W v. o. oovannesent eneastisee opriers este.ese see
.) f. l.
1.0 INTRODUCTION
We have obtained the advice of the U. S. Geological Survey for geology and seismology considerations and Dr. Nathan M. Newmark for structural engineering considerations. We consider both sources to be the best available to us, and we have accepted their recommendations. These were presented in Supplements 4 and 5 to the Safety Evaluation Report. Some additional considerations with respect to the applicant's proposals and our consultants recommendations are discussed below. 2.0 EFFECTIVE ACCELERATION ~ The Postulated Earthquake The Hosgri fault is relatively long and shallow. The tectonic =2 7 ' characteristics of the region indicates that, if a' magnitude 7.5 ~ earthquake should occur, on the Hosgri fault, it would involve pre, dominately strike slip motion. For this magnitude,and a shallow strike slip mechanism, the length of fault rupture would be at least several tens of miles and possibly many tens of miles similar to the long strike slip breaks that have occurred in earthquakes on the San Andreas fault. Strike Slip Earthquakes Generally the stresses which lock faults are believed to be lower for strike slip faults than for reverse faults. In the case of a reverse fanit the two sides of the fault are being forced together by tectonic stresses, j This increases the effective stress on the fault as regional stresses increases. Under these conditions the stress can reach extremely high 1evels before the frictional force is overcome and the fault s12.ps. In j 1 the case of strike-slip faults, because the forces are generally parallel Eh l
..__...._.-_u....._.. ) ) -2 with the plane of the fault, the levels of effective stress are not as high,as those involved in reverse or thrust faults (Thatcher and Hanks, 1973). Evidence of lower effective stress on faults of the San Andreas system may be seen in the determination of length of rupture versus magnitude. Two such curves were summarized by Hofmann, (1974). Figure 2-1 illustrates the better data of Ambraseys and Tchelenko (1968), which indicates a very wide range of rupture lengths versus magnitude. Figure 2-2 is from Algermissen et. al., (1969), whose data are restricted to the strike-slip San Andreas fault. The latter curve lies approximately along the upper bound of the Ambraseys and Tchelenko data. This suggests that for strike slip faults, much greater lengths of rupture are required on the San Andreas fault than for the entire available data set to generate the same magnitude earthquake. Hence, the higher effective stress across other kinds of faults may be a contributing factor to the generation of large magnitudes from short, rupture lengths. This effect may also be observed in the data of Bonf11a (1970). Based on the above it appears that strike slip earthquakes of the San Andreas system have lar8e source dimensions and may have correspondingly lower effective stress. Near Field Earthquakes The Diablo Canyon site would be in the near field of the postulated event, the distance to the' source would be small compared to the size of the source. In this situation the energy available to contribute to peak acceleration is limited to the energy released in a short segment ~' of fault rupture, the length of which equals the distance to the source
- - :. z : :, ~.-.:- ,..--.. =.:: - ^. ---. - =---.. u ) 3 (Brune,1979). Thus, a large near field earthquake can % expected ] to produce smaller peak accelerations than would be indicated by: (1) extrapolating from distant events where source size is not large' compared to distance, or i (2) extrapolating from closer events of small magnitude with small source dimensions. Further, the design significance of peak acceleration is different for I near field events. Instrumental records close to the source ir41cate relatively high values for the highest acceleration peak, with rapidly declining values for subsequent peaks. Further, the higher peaks of ten do not occur in sequence. This contrasts with recordings l from distant events where subsequent peaks may be nearly as high as g j the highest peak. This would suggest that in the.near field the l effective acceleration can be lower relative to the maximum peak l expected and yet provide an adequate representation of structural response. I Intensity Data i There are no instrumental records of ground motion close to the source of j earthquakes as large as magnitude 7.5. However, intensity data, based on observed effects and damage, are available for such events as well as j smaller quakes. Correlations between acceleration ad intensity have been made based on available data. Although there is a great ' deal of scatter in the correlations, they are useful in bounding the level of I effective acceleration. We normally use the correlations of Trifunac and Brady, (1975). i l L . -...~..- .:.-.=. - --.
m
- ,6 df
') 4 The 1906 San Francisco earthquake of magnitude 8.3 provides an example of a large strika-slip earthquake. In this case,'Rossi-Forel intensity I or greater occurred only within about a mila and a half of the main rupture of the San Andreas fault. At 31/2 miles from the San Andreas fault Rossi-Forel intensities of II and less were observed along the main break. This corresponds to the Modified Mercalli Intensity of VIII ~ (USGS Circular 1279). The nean acceleration from the Trifunac and Brady curves for Modified Mercalli VIII is approximately.25g. It is difficult to determine the first and second senadard deviations because of a lack of data. The data for MK VIII alone is considerably less than that which is derived by the 1975 Trifunac and Brady straight line extrapolation ~ from smaller intensities. However, considering the - E straight line extrapolation, it appears that the second standard devia' tion of acceleration associated with MN VIII is about.54g which is very close to the original effective acceleration used for the Diablo Canyon Plant. The second standard deviation of acceleration would include virtually all the scattered accelerations observed for a given Modified Mercalli intensity. The Trifunac and Brady 1975 second standard deviation value exceeds the largest acceleration which has been associated with IN VIII. Thus the effective acceleration from the magnitude 8.3 on the San Andreas fault may have been as low as about.54g at a distance of'3 1/2 miles from the fault. e G .. m. .?-
- 4
- e***.
- 7 a.* * * * * * % _ _,,.
......,m 7., i .) 'y 5 Another example is the 1927 Point Arguello earthquake of magnitude '7 1/4. There is disagreement about the location of this earthquaka and its mechanism. However, the possibility that it occurred on the Hosgri fault was one of the reasons for setting the magnitude j for the postulated event. If the isoseismal map of the 1927 earthquake were moved northward along tle Hosgri fault to the plant site, the highest observed intensity, at a distance of 31/2 miles from the Ault, would be the same value as discussed above, Modified Mercalli VIII. Pacoima Dam Record The Pacoima Dam accelerogram represents the largest peak acceleration i yet recorded, and was recorded in the near field. Accelerations from the magnitude 6.6 San Fernando earthquake at approximately 3 km from the fault reached a maximum peak value of 1.2. Other peaks occurred 3 at lower values (Table 3 Geological Survey circular 672). The source was a thrust fault where effective stresses are expected to be at a high level. The records, when filtered to eliminate frequencies of 8 hertz and above,which approximates the response of older strong motion instrumentation, yields a peak acceleration of 0.9g. It has been proposed that the dam abutment, where the instrument was located, amplified the accelerations because of geometry and further amplified them because of its vibrational response to the shaking. ' The abutment was also damaged by the shaking, resulting in fractured rock including a fracture of the pier beneath the accelerograph. ~ t i i . _ -, _ _ _ _ _,.,. _ _ _. _ _ _ _ _ _ _ -. - - _.. ~.
4 .l 3 3 6 Extremely high accelerations are commonly observed in the laboratory on rock specimens undergoing failure in triaxial testing machines. Papers by Trifunac and Hudson (1971) Boore (1972) and Bouchon (1976) all agree, using various theoretical approaches, that the Pacoima Dam record was amplified because of topography. Even so the spectrum for the Pacoima Dam earthquake as it was recorded can be enveloped, 8 over all frequencies but the very highest ones, with a Regulatory Guide 1.60 spectral envelope achored at 0.75g. Other Instrumental Data Instrumental records have not been made for very large earthqukkes at ~ very close distances to the source. Records are available only for smaller earthquakes and/or at greater distances than we are discussing here (USGS Circular 672). Ranks and Johnson (1976) indicate that high frequency accelerations are independent of magnitude for near field earthquakes (distances less than 10 km). Numerous correlations have been developed relating magnitude, source distance and acceleration based on the existing records. A number of problems arise in attempting to use any such correlation for the small source distance, large source size and large magnitude appropriate for Diablo Canyon. l Using any of the correlations to estimate the accelerations at the 1 Diablo Canyon site would involve extrapolation beyond the existing data I
} ') ~ ~ 7 set. Thus, the correlations cannot be tested against data for ' these ' conditions. Further, for many of the earthquakes with smaller source distances, there is uncertainty concerning the horizontal source' distance which could affect the extrapolation. One curve, that was produced by Donovan (1973), attempts to establish standard deviations for accelerations as a function of magnitude and distance. The one standard deviation value extrapolated to. the near field for a magnitude 7.5 earthquake results in an acceleration of about.7g. We normally consider the acceleration level at one standard deviation above the mean to be an. acceptable wn. w ~~ anchor point for the ground response spectrum.
3.0 CONCLUSION
Based on the foregoing considerations, we consider 0.75g to be an acceptably conservative effective acceleration for reevaluating the Diablo Canyon units in consideration of a postulated earthquake of magnitude 7.5 centered on the sector i of the Hosgri fault nearest the plant site. '.i o O e D
6 ) References Algermissen, S. T., and staff (1969a) " Studies in Seismicity and Earthquake Damage Statistics," Three parts, Summary and Recomandations, 23 pages; Appendix A,142 pages; and Appendix B, 68 pages, Prepared for the Department of Housing and Urban Development. Office of Economic Analysis by the Staff and Consultants of the Department of Commerce, ESSA, Coast and Geodetic Survey. Ambraseys, N. N. and Tchalenko, J., " Documentation of Faulting Associated with Earthquakes" (unpublished), 1968, Department of Civil Engineering, Imperial College of Science, London, England. Barosh, P. J., "Use of Seismic Intensity Data to Predict the F.ffects of Earthquakes and Underground Nuclear Explosions in Various Geologic Settings," Bulletin 1279, 1969, U. S. Geological Survey, Washin8 ton, D. C. Bonilla, M. G. and Buchanan, J. M., " Interim Report on Worldwide Historic Surface Faulting," open file, Series No. 16113, 1970 U. S. Geological Survey, Washington, D. C. Boore, D. M., (1972) "A Note on the Effect of Simple Topography on Seis-mic SH Waves," Bulletin of Seismological Society of America, Vol 62, No. pgg 1, pp. 275-284 Bouchon, Michel,1976, " Discrete Wave number Representation of Seismic Wave Fields with application to various scattering Problems" Ph.D. Thesis, Massachusetts Institute of Technology. Brune, J. N., (1970) " Tectonic Stress and the Spectra of Seismic Shear Waves from Earthquakes," Journal of Geophysical Research, No. 75, pp. 4997-5009. Donovan, N. C., "A Statistical Evaluation of Strong Ground Motion Data Including the February 9,1971 San Fernando Earthquake," Fif th World Conference on Earthquake Engineering, Rome, Italy,1973. Hanks, T. C., and Johnson, D. A., 1976, " Geophysical Assessment of Peak Acceleration" Bulletin of the Seismological Society of America, Vol. 66, No. 3, pages 959-968. Hofmann, R. B., " State-of-the-Art for Assessing Earthquake Hazards in the United States; Factors in the Specification of Ground Motions for Design Earthquakes in California," U. S. Army Engineers Waterways Experiment Station, Miccellaneous Paper S-63-1, June,1974.
....2_...--.- ... : ::- :. = - ) 1 2 Page, R. A., Boore, D. M., Joyner, W. B., and Coulter, H. W., " Ground Motion Values for Use in the Seismic Design of the Trans-Alaska Pipeline System," U. S. Geological Survey Circular 672, 1972. Thatcher, Wayne, Hanks. T. C. (1973), Source Parametera of Southern California Earthquakes", Journal of Geophysical Research Vol. 78,- No. 35, pages 8547-8576. Trifunac, M. D. and Brady, A. G., "On the Correlation of' Seismic Intensity Scales with the Peaks of Recorded Strong Ground Motion," Bulletin, Seismological Society of America, Vol. 65, 1975, pp. 139-162. Trifunac, M. D., and D. E. Hudson,1971. Analysis of the Pacoima Dam accelerogram, San Fernando, California, carthquake of 1971, Bull. Seism. Soc. Am., g, 1393-1411. I s t l o
.~,,-., .__:.--.:.=1.=. 1 I I I l 1 l l NOTE: DATA POINTS ARE FROM ALGERWISSEN (ISet) [e~ ~ ~ 9 / / / j- / / E Q/ l l.~ -. 8/ l v ~ / / = L / I / e 0 f4 E /v e v e i / 3 l I j h"'* / I / 7 p% l / nduas 2-2 od I / s.W l / FIGURE 2-1 1 gM __/ e $@W I I I i 1 1 I I o 1 z s 4 5 EL 7 3 e MAGNITUDE i Fis, 1, length of surface rupture versus magnitude l l 1
- Uh 4
.3 \\ ~ ) 1000 7_ 500 lo / / o c O f /
- /f I
J ,/ /,y - s f // 2 / // o / x 3, / // j a. l
- / /'
..-,,c,m ~ ~ 7/ g I / y t I / ~ LEGEND' l ~ ~ y woRLD /* ~~ g ,/. p/, --~~ STRIKE SLIP se g
NORTH AMERICA 8
f I i i mo /' l 1 ,/ /y' c 'g to a f f l, e /. e // I l // e. i/ // l /. l l. s / e' I j l I ~ l 1 2 3 4 5 6 7 - 8 9 EARTHQUAKE MAGNITUDE ' IGURE 2-3 Length of surface rupture on main fault as related to earthquake magnitude. From Bonilla and Buchanan;21 boundary of applicability added n ? e e
~ a ) T 1
- 3. 0 SEISMIC DESIGN BASIC RESPONSE SPECTRA'(IN THE' FREE FIELD)
Under the applicant's (Dr. Blume's) approach an acceleration of 0.75g was used as a nonnalizing value for time histories of the strong j motion of eight selected earthquakes recorded on rock close to the epicenters, thus providing what the applicant considered to be the best available models for the Diablo Canyon conditions relative to the Hosgri fault zone. The magnitudes of the eight earthquake records used are the greatest recorded thus far on the rock close to the earthquake source. They range from 5.3 to 6.6 in magnitude. The procedure followed was to statistically develop the spectral response j based upon these eight earthquake records. Dr. Blume used a mean plus one half standard deviation (sigma) values based on these C% i eight recortis. At our request, the applicant made adjustments to the final response spectrin for periods above 0.4 sec. to account for the greater long period energy expected in a 7.5 magnitude shock as compared to the lower magnitudes in the available earthquake records. j The approach of our consultant, Dr. Newmark, was discussed in Supple-ment No. 5 to the Safety Evaluation Report. Without adjustment to account for building size effects (i.e. for Tau = 0), his recommended spectra are generally consistent with Regulatory Guide 1.60 where 33 diverse earthquake records were used to develop design response l spectra and with NUREG 0003 (Ref.1) where 56 diverse earthquake l j mcords were used. These spectra employ mean plus one sigma values based on a large number of diverse records. Figure 3-1 compares Dr. Blume's basic results with Dr. Newmark's l basic results (no foundation size adjustment and no ductility adjustment). l As is often the case where one cr a few earthquakes are used instead of a large number of earthquakes, Dr. Blume's result has a more sharply peaked shape than Dr. Newmark's result. On this figure, M # W -35Vl ..=.....x~. F. -) ) _2 Dr. Newnark's result is the more limiting. The maximum values / at about 4 hertz are virtually identical. EFFECT OF FOUNDATION SIZE The spectra shown in Fig. 3-1 is considered app 1tca51e. to tre~ free. field conditions (i.e. effective design spectra for structures with small foundations). However, it has been observed that structures on large foundations experience reduced motion from high frequency waves as compared to free field motion and motion of structures on small foundations, particularly those associated with the support of seismic instrunents. A rational explanation for this phenomenon was presented by Yamahara in 1970 (Ref. 2). Similar procedures were developed independently by Ambraseys (Ref. 3) and by Scanlan (Ref. 4). E21lIlM Verification of this phenomenon is indicated by the response measured ~ in the Hollywood Storage Building compared with the response computed from records in the free field about 112 ft. away from the nearest corner of the building. Response spectra for the storage building basement and for the parking lot in the east direction are shown in Fig. 3-2 as reported during the 1971 San Fernando Earthquake. It can be seen that there is a significant reduction in the res-I ponse spectrum for the building as compared to that for the parking i lot for periods less than about 0.4 sec., whereas, for longer poriods, the response spectra are practically identical. Similar effects are observed for the response spectra in the south direction. It can be seen that the high frequency components of the response spectrum are attenuated by a factor of 2 to 2.5 in the range of frequencies higher than 2.5 hertz. Yamahara in Japan made similar observations during the Tokachioki ~. earthquake of 1968 (Ref. 4). The following observations were made by Yamahara: .. I
.. ~ ) .) (1) The maximum acceleration amplitudes of the building foundation were always less than the maximum free field acceleration ~ of the surrounding ground. If the records of the building foundation had indicated the response characteristics of the building, the amplitude of the building foundation would have been larger than those of the ground, due to elastic defomations of foundation soil. (2) The natural period of the building rarely appeared in the records of the building foundations. The period that appeared most frequently in the records was not the natural period of the building but the predominant period of the adjacent ground. (3) If the input vibration frequency of ground motion is relatively high, the affective input power to a building with a large foundation is greatly decreased because there is a large phase gggan; difference among the movements of different points of the building foundation. This is why ground motion having high frequency content does not usually cause severe response of a building, as it is shown by the current methods of calculation, even if the acceler-ation of the ground motion is fairly large. Yamahara developed an analytical method for numerically estimating the input loss because of the size of the foundation. He applied this method to Tokachioki earthquake record in the free field and obtained a reduced effective input for the building. He compared this effective input with the actual observed record at the ground floor of the building and showed that the two motions were similar to each other. The applicant's consultant used Yamahara's technique on the eight earthquake records which he considered the most suitable for the Diablo Canyon site. Using the time histories of the eight records nomalized to 0.75g, the acceleration was averaged over the time required for the waves to pass through the foundations. New effective response spectra were then developed from the modified time history P k h
w.- ".. : ' s:_ z -.-.. _ ') Y 4 motions and asponse spectra. These spectra wem then smoothed and adjusted as before for nagnitude effects. The reduction in response am a function of. frequency and are consistent with the mathenatical models of Yamahara, Scanlan and Ambraseys. They are also less than those observed in the Hollywood Storage Building. As discussed below, a comparison of the results obtained by the appli-cant and our consultant indicates that our consultant's results are more limiting and we have adopted them. Our consultant, Dr. Newmark, performed a similar calculation to obtain a reduced response spectrun for the Pacoima Dam mcord. He found that the response spectrum was reduced by a factor of 1.2 to 2.5 above a frequency of 2 cps. In his recommendation to us he utilized a mduction factor applied unifomly in the acceleration limited portion of the spectrum. The mductions am. pugosely' Rept ' 8 915'!* lower than the average value calculated for the Pacoima Dam record and those observed at the Hollywood Storage Building. l l Some additional tilting and, torsion may result as a consequence of the nonsynchronized earthquake motions. At our request the applicant has agreed to consider the additional tilting and torsion when using the Yamahara procedure.. Where the stress increase due to torsion is significant, torsional analysis shall be conducted. The analysis shall take into consideration the inertial effects and the natural modes of torsional vibration. DUCTILITY CONSIDERATIONS We have also allowed the use of ductility in developing the final seismic input. Slight excursions beyond the yield point are allowed under certain conditions when checking the plant for the short duration effects of rare events such as the hypothetical 7.5 magnitude earth-quake nearby on the Hosgri fault zone. The ductility factor is the
- a..a
... = - - -... -..... . ::= ,i -) 5-maximum useful (or design) displacement of a structure to the effective elastic limit displacement, the later being detennined not from the actual resistance-displacement curve but from an equivalent elasto-plastic function. This equivalence requires that the energy absorbed in the structum (or area under the resistance-displacement curve) at the effective elastic limit and at the maximum useful displacement must be the same for the effective curve as for the actual relationship at these two displacements. For conventional buildings in California, input motion is predicted on the assumption that the buildings will develop a ductility of 2 to 5. Accordingly, we consider the choice for the ductilit;y factor of 1.3 for all Category I structures to be conservative. To illustrate the effect of ductility mfer to Figure 3-3 where D, V and A refer to the bounds of a typical elastic spectrum while the symbols D', V' and A' to the bounds of a reduced elastic-plastic spectrum for acceleration. For a ductility factor 0f u, the' elastic response spectral acceleration is decreased by a factor of y up to a fre-quency of 2 hertz and by the factor of J2>( -1 between 2 and 8 hertz. There is no reduction above 33 hertz. Between 8 and 33 hertz the reduction is linear. Some judgment was used in selecting a ductility factor of 1.3 for use in the Diablo Canyon reevaluation. Observation of the perfonnance of structures in earthquakes, interpretation of Laboratory tests, including those on earthquake simulations and shake tables, observa-l tions of damage to structures an'd structural models in nuclear tests, including damage from both air blast and ground shock, are all pertinent factors in arriving at a judgment as to the appropriate ductility factor to be used in design. Based on these judgments and in accordance with the advice of our consultant we have concluded, as was stated in Supplement No. S to the Safety Evaluation Report that: l (1) The applicant may use a ductility ratio of up to 1.3 to reduce j the response spectra developed by the applicant. The end result, however, must not be less than the spectra recomended by our consul tant.
) ) (2) On the other hand, no ductility reductions should be made to the spectra developed by our consultant. If ductility is to be ~ utilized with these spectra it should be justified on a case by case basis for each portion of the plant where it is to be utilized. A cogarison of the applicant's msults and our consultant's results is shown in figures 3-4 through 3-7. For each structure the appli-cant's results are shown before and after reduction to account for ductility. However, only one curve is shown for each structure for our consultant's msults, since ductility is not to be used to reduce these results. It can be seen, for small structures (Tau = 0),that our consultant's recommendation is the more limiting. In addition, for other struc- ,,,s tures, after reducing the applicant's spectra to account for ductility, our consultant's recommendation is the more limiting, except at very high frequencies (above 25 to 29 hertz). The most significant struc-tural msponse modes are all below 25 hertz. Accordingly, for the purpose of structural design, our consultant's results are clearly the more limiting and we have adopted them. This cogarison is also valid for the purpose of designing ductile equipment such as pipe supports and piping systems. The structures may not yield and thus the structure's motions may not be reduced to the extent that would be indicated by the use of a reduced ground msponse spectrum in the analysis. However, in this event, the ductile equipment would exhibit the additional ductility needed (as compared to the analysis case). Accordingly, as was the case for structural design purposes, our consultant's results are more limiting and we have adopted them.. I The comparison is somewhat different where equipment behavior under seismic loading cannot be considered ductile. One example would be i 4 l l
)' 7_ ) electrical miays where the seismic design was verified by shake testing. In these cases it cannot be said that tiie equipment would mobilize the needed additional ductility in the event that the struc-ture did not. Accordingly, if one were using the applicant's results for these cases he would use the " elastic" spectrum without reduction for ductility. This would then be more limiting than our consultant's results in a frequency range of about 2 hertz to 7 hertz as well as for fmquencies greater than about 25 hertz. The end result would not be adversely affected so long as these frequencies were not the important ones for the equipment involved and/or substantial margins were included in the original equipment design basis. We believe that this will usually be the case. However, since we have not yet reviewed the details of the applicant's work concerning equipment qualification we cannot say, with mgard to non-ductile equipment, whether or not use of the applicant's result might be more limiting for any particular es;;; i tems. We will consider this in our review of the results of the reevaluation. CONSERVATISM There are several conservatisms used in the seismic design of nuclear power plants. These have previously been discussed in various forums. We have discussed here two relaxations relative to the usual case: (1) reducing response spectra to account for foundation size and (2) allowing some credit for ductility effects. We believe that these two items are technically justified. The other " usual" con-servative aspects remain in effect, however, providing what we believe to be substantial margins for the seismic reevaluation of the Diablo Canyon plant. AUDIT We conducted an audit of the plant's seismic design in 1975. At that time the audit considered:
4
- y
~. (1) adequacy of the original design (0.4g). 1 (2) adequacy of the upgrading to 0.5g, as compared to the original design. We plan to conduct a similar audit in the near future to assess the adequacy of the current reevaluation program. The nature of scope of this audit program will be described in more detail separately. RESULTS The applicant and our consultant have used two diverse approaches to derive ground response spectra, based on an effective acceleration of 0.75. Based on the factors discussed above we have evaluated the 9 procedures, we consider each to be appropriate and we have accep,ted our consultant's recommendation which for most cases is more limiting pg than the applicant's proposals. Accordingly, we consider the results to be conservative and to be supported by both diverse approaches.
. w .u... -..: _. = , :.. _..._ a . u.x:. .. _ :...; = : = -...: - _ _ _ _. ) i. ' REFERENCES 1. W. T. Hall, B. Mohraz and N. M. Newnark, " Statistical Studies of Vertical and Horizontal Earthquake Spectra," U.S. Nuclear Regulatory Commission Contract AT(49-5)-2667, Report NUREG-0003, January 1976. 2. H. Yamahara, " Ground Motions during Earthquakes and the Input Loss of Earthquake Power to an Excitation of Buildings," Soils & i Foundations, Vol.10, No. 2,1970, pp.145-161. Tokyo. 3. N. Ambraseys, " Characteristics of Strong Ground Motion in the Near Field of Small Magnitude Earthquakes," Invited Lecture, Fifth Conference European Committee for Earthquake Engineering, Istanbul, Sept.1975. 4. R. H. Scanlan, " Seismic Wave Effects on Soil-Structure Interaction," Earthquake Engineering and Structural Dynamics, Vol. 4,1976 pp. 379-388. k
i DIABLO CANYON UNITS 1 & 2 FIG. 3-1 COMPARIS0NOFHOSGRI7.5M(BLUME&NEWMARK)BASICSPECTRAINTHEFREEFIELD j 2.0 -- q r = 0.0. 7% DAMPING i 6 1 HOSGRI 7.5M/BLUME [ 's 'f g HOSGRI 7.5M/NEWMARK
- 1. 6 -- I N - - - - - - --
m o y . ~p,, a t y j \\ .r M MISC. SMAtL STRUCTURES l ~ m I N N 'f \\
- e N
/ r u-s - \\ I \\ g
- 1. 2 -
g s = { s N 5 t-5 i 's N i s 's N, w t I '~ 0.8 -- .5-w M 0.75g ao 1 g 4 1
- 0. 4 -
l e-9 l i i 0.0 j i 0 4 8 12 16 20 24 28 32 36 40 FREQUENCY, HERTZ i,
- s
4.4,e-E-ewe ea-4 hwe we e. es --mee< -w 9m ,+em--weeme=e=w.- .,,,me-o-w-m ,a g O
- P *9 we
~ ) . pyg, 3-2 SAN FERN ANDO EARTHQU AKE, FEB. 9,1971 - 0600 PST HOLLYWOOD STORAGE BASEMENT AND P.E. LOT, COMPONENT EAST, DAMPING VALUE 2% OF CRITICAL 400 200 100 80 60 40 p 3 20 ['. i 5 IA 'Ml 10 Y j{f '; 8 6 / ,a g z s ,l 4 ll s I / o !l / O. ( / / 0.s -l 0.4 gN ~ Basement - - - - - P. E. Lot 0.2 I I I I I I I I I I I I 0.1 0.4 0.060.081.0 0.2 0.4 0.6 0.8 1.0 2 4 G 8 10 20 Period, sec ) e
FIG. 3-3 IN ELASTIC D ESIG N SPECTR A G EN ER ATIO N t y N SITE ELASTIC DESIGN g SPECTRUM [ 's l yi l f- --__m / s / g D'/ Q' i / g k / INELASTIC ACCELERATION g i / DESIGN SPECTRUM N /\\ \\ i / / / \\ \\ 6 L 9
- 4 y
+ !i f FREQUENCY
i DIABLO CANYON UNITS 1 & 2 ~ FIG. 3-4 COMPARIS0N OF HOSGRI 7.5M (BLUME & NEWMARK) SPECTRA ~. 2.0 T = 0.0. 7% DAMPING HOSGRI 7.5M/BLUME with u=1.3 's ] - - - - HOSGRI 7.5M/,8LUME 'f g \\ -- HOSGRI 7.5M/NEWMARK 1.6 7 -- v \\, W \\ g / \\ MISC. SMALL STRUCTURES D I N N 'f l \\. L N i v x \\ i 1.2 - \\ \\ 's N s \\ I N 'N s a wr~_ -N
- 0. 8 -
w ~ - 0.759 a Q 0.4-e 0.02 i 0 4 8 12 16 20 24 28 32 36 40 FREQUENCY, HERTZ y i
I t h .. I DIABLO CANYON UNITS 1 & 2 FIG. 3-5 COMPARISONOFHOSGRI7.5M(BLUME&NEbMARK) SPECTRA ~
- 2. 0 _
r =.04. 7% DAMPING HOSGRI 7.5M/BLUME with p=1.3 ,\\ t - - - - - HOSGRI 7.5M/BLUME HOSGRI 7.5M/NEWMARK 1.6 - ~s / N CONTAINMENT STRUCTURE ,e / \\ I h [ ~ 'l V 'o \\\\ 1.2 I,l ss = g p
- a?
f Y, I b b a r ss 08 'N a x-w 0.67g .o s 3 N.' } l 8 8 N a 0.60g O 0.4 i l l
- 0. 0 1
0 4 8 12 16 20 24 28 32 36 40 i FREQUENCY, HERTZ t i g .g
i j.Il 4 i i > !3 othI I! [ !, ~ v J . ^ 0 4 g g 3 6 ,6 3 6 5 0 0 23 AR 3 TC 1 8 ^ 2 E = P p S G N h ) I t K P i ~ R M w K G A A R N M D E E A I W M M M D 4 2 2E U U W L N 7 L L E I B B N U / / / B Z 1 5 M M M T 5 5 5 R Y N-E 5 R S M TU 0 E 7 7 7 A H I I L 0 N B = I I I L 2 U( R R R I Y t G G G X C , N M S S S U N O O O A E O 5 H H H Y U N 7 Q A E C I R 6 F R O G 1 Ns. L S B O A H I D F O. \\N NO 2 S N\\ 1 I R A f\\ PM OC 8 x 6 \\ 3 N G s I F ,4 -/ -(' / iil j i l 0 6 2 4 0 0 0 0 2 1 1 UEm=EJ l 5UWw E !i
il1 ' i 1 h l-{/ILpi t
- e. f[
i 4 v d ',d 0 3 ' 4 9 g 6 4 0 3 5 5 0 0 ~ 23 3 1 8 A 2 R = T u C E h P t i S G w K N R G ) I E EA N 4 K P M MM R M U UW D I 2 N-A A L LE L M D B 8N I l / // U E 1 M MM B 2 N 7 5 5 5 E 7 77 N 8 0 Z 1 E I M
- 0. k I I B
2 T S U R RR R ~ R G GG U T L = E I B S S S T H O O O N ( t H H H ~ U M Y N 5 NC O 6 E Y 7 1 U N s' Q I A C R E G R O S F L O B H \\ w A F I 2 D O 1 \\s' N O S I R A s P M N O 8 \\ C 3 \\ 7 s y 3 \\ \\ \\ G g 4 I F --f / / g, I a ' 0 g 8 4 0 6 0 0 1 2 Cg" "u* g.3$!< gd*Q E 3 S}}