ML19352A789
| ML19352A789 | |
| Person / Time | |
|---|---|
| Site: | Summer  |
| Issue date: | 05/27/1981 |
| From: | Nichols T SOUTH CAROLINA ELECTRIC & GAS CO. |
| To: | Harold Denton Office of Nuclear Reactor Regulation |
| References | |
| NUDOCS 8106020117 | |
| Download: ML19352A789 (16) | |
Text
-
SOUTH CAROLINA ELECTRIC a gas COMPANY
^
I f 19, est ornes som re.
CoLumstA, Soutw CARouMA 29218 N
' T. C.NICHOts, Jn.
w o e......... c... c..c~
May 27, 1981 4
p p9 f bp 1
o.5-Mr. Hatold R. Denton, Director f,
Office of Nuclear Reactor Reguistion
(\\ '
U. S. Nucicar Regulatory Commission Washington,LD. C.<20555
Subject:
Virgil C. Summer Nuclear Station Docket No. 50/395 Supplemental Seismologic investigation
Dear Mr. Denton:
South Carolina Electric and Gas Company, acting for itself and as agent
. for South Carolina Public Service Authority, herewith forwards forty-five (45) copics of Appendix XI of the report entitled " Supplemental Seismologic Investi-gation" as previously submitted on December 12, 1980. This material is submitted at the Staff'c re. quest to provida additional information for the estimates of reservoir induced scismic ground accelerations postulated to occur from a range of hypothetical seismic events.
Very truly yours,
,/
T. C. Nichols, Jr.
RBW:rh cc:
V. C. Summer G. H. Fischer T. C. Nichols, Jr.
Dr. John Ruoff D. A. Nauman W. A. Williams, Jr.
R. B. Clary A. R. Koon A. A, Smith H. N. Cyrus J. B. Knotts, Jr.
j J. L. Skolds ggdA B. A. Bursey L>
0.'S. Bradham g
H. E. Yocom Dan Cash Andy Murphy
/
/
Phyllis Sobel
/
ISEG NPCF File B106020 l
l' APPENDIX XI ESTIMATES OF RESERVOIR-INDUCED SEISMIC CROUND ACCELERATIONS J
I e
=
reserh e %,
g y 'ev
, 7
f
_1_
GROUND MOTION MODEL The mathematical model used in this. study to estimate peak ground accelerations during reservoir-induced earthquakes is based on a simple physical representation of the earthquake source and of the crustal properties between the source and the site. The validity of the model to accurately predict characteristiles of seismic shear waves has been demonstrated with a wide range of earthquake data in California (Hanks and McGuire, 1981).
For applications at Monticello Reservoir, all earthquake and crustal parameters used in the model have been verified with observations from the area. The most critical of these, for estimates of peak acceler-ation, is the static stress drop, &c.
For high frequency ground motion at Monticello Reservoir, a value of Ao of 25 bars has been shown to be appropriate and conserystive by comparison of theory with observations I
from three small earthquakes in 1978 which triggered a strong-motion accelerograph. This value of 25 bars, calculated from strong motion records and appropriate for estimating strong ground motion, has been confirmed independently by J.B. Fletcher of the U.S. Geological Survey.
Professor L.T. Long of Georgia Institute of Technology also concludes th.t the use of 25 bars for the stress drop during large events is o
conservat ive.
In California, a stress drop of 100 bars is more appropriate, because these events originate and relieve stress from depths of 10 to 15 km, where crustal stresses are likely to be large. The smaller value of Ao at Monticello Reservoir is consistent with the observation that seismicity there is shallow (98 percent of recorded events occur at deptha less than 2 km) and with the belief that large stresses cannot be generated at such shallow depths. Thus the strong motion records at i_
Monticello Reservoir are themselves the best data for estimating earth-t quake stress drops, for the purposes of predicting strong ground motion in the region.
I l
i 4.,,,-
s
.~
i y
,-n-- -..,, -
_v,
-.-4 4
JP r
. LARGER MAGNITUDE EARTHQUAKES The conclusion of the. Applicant isthatamagnitude(g)4 earthquake is an appropriate, realistic upper-bound event which can be induced by the Monticello Reservoir; the U.S.N.R.C. staf f concludes that l
an M = 4.5 earthquake is a conservative estimate of the upper-bound g
event. Because any such event is rare, it is highly unlikely that empirical strong motion data will be soon available to determine source
[
parameters directly. Hence' the estimation of source parameters, in j
particular the static stress drop,' for these larger events must be made using observations during smaller events, physics 1 considerations re-garding earthquake locations, and analogies with other reservoirs.
t As discussed above, recorded events of h = 2.8 indicate that a stress drop of 25 bars is appropriate for characterizinn strong groun,1 motion during these earthquakes.
Seismic events at Monticello Reservoir have all been shallow (98 percent less than 2 km) since initial reservoir I
filling, and have not deepened with time. This is consistent with
[
cructal stress measurements in two deep bore holes in the area, which indicate a laterally variable but shallow stress barrier at a depth of less than 5 km.
Thus it is likely that hypothetical,' larger magnitude h
earthquakes, if they oscur, will be confined to the shallow crust or to depths below the stress barrier, i.e., greater than nbout five km.
If l
they are near the surface, they can relieve only the relatively low j
crustal stresses in the shallow active seismic zone. It should be pointed out that both stress observations in the two bore holes and tectonic models of the observed seismicity indicate that the largest j
~
deviatoric stresses a-e at very shallow depths (less than 0.5 km).
Events which have generated strong motion records have initiated at these f
very shallow depths; it is conservative to assume that stress drops j
observed for these events will apply to larger, and necessarily deeper, l
f fault surfaces. Therefore, on the basis of physical coasiderations it is felt that observations during g = 2.8 earthquakes can be used to j
estimate earthquake stress drops conservatively for larger, hypothetical i
events.
i l
i i
t
?
, Some strong motion data are available for reservoir-induced events at Heinfengkiang Dam in China. 'Tais area is not analogous to Monticello Reservoir, in that Hsinfengkieng Reservoir is much larger (11.5 x 10' 3
9 3 m and 49 m); seis-a ) and deeper (105 m) than Monticello (0.5 x 10 micity at the Chinese site is also deeper (up to 10 km) than observed at Montieello. However, the Chinese data do allow investigation of earth-quake source parameters over a range of magnitudes.
The strong motion data are reported by Hsu et. al. (1975). They consist of 28 evencs recorded over a nine-year period on the dam and in the adjacent free-field on a rock site. The data are reproduced in Table l
1.
No uncertainty is reported for magnitude, depth, or distance deter-minations, and the peak acceleration values listed are apparently the larger of the two horizontal componente. The magnitude values are reported as M, but it is unclear how these were determined. Values of 3
physical characteristics of the crust (density, shear wave velocity, Q) l are not readily available for this area. The strong-mot ion instrumen-tation is well-described; the instruments provide linear response up to 35 hz frequency.
To estimate strong ground motion stress drops for these earthquakes we use a procedure similar to that for the Monticello events. The theory used for predicting peak accelerations was invoked to determine, for observed peak accelerations, what stress drop for each event (with an independently determined magnitude and location) was consistent with the observations. For the Chinese data, crustal characteristics (density, shear wave velocity, and Q) were assumed to be similar to those in South Carolina: The relationship between ( and Mg derived by Nuttli i
(1979) was assumed to apply:
g = 0.625 M3 + 2.28 l
although obviously this a tenuous assumption: the M reported by Chinese g
authors may be more equivalent to a local magnitude than a surface-wave l
magnitude. None of these assumptions affects the conclusions reached here.
f
. Shown in Figure 1 is a plot of aa derived from the peak acceleration versus magnitude (both the M and M scales are shown).
Some of the g
g scatter in Figure 1 results from the use of only the larger of the-two horizontal peak accelerations, rather than the average or a measure of the root-mean-equared acceleration. Any uncertainty in earthquake focal depth oc epicentral location also contributes to scatter in derived values of stress drop.
In spite of these sources of uncertainty, the data show no general increase of stress drop with magnitude, at least over a range of two or three magnitude units. This gives confidence to the assumption made at Monticello Reservoir, that strees ecops observed for M. 2.8 earthquakes are representative of those to be expected during larger shocks.
(It is worth pointing out that the 26 bar stress drop shown for the Mg = 5.1 event is heavily dependent on the 9.9 km depth reported by Hsu et. al.. 1975. If the source-to-site distance is less, as estimated by Bolt and Cloud,1974, the estimated stress drop is sub9tantially reduced.)
The Chinere data were obtained in a different tectonic regime from that of Monticello Reservoir; the seismic events ' were induced by a reservoir much larger and deeper than Monticello. However, these data do indicate that stress drops appropriate for estimating strong ground motion do not vary over a range of magnitudes.
This conclusion was also reached by Hanks and McGuire (1981), using California earthquake data.
a h
e l
t r
l i
-*.-..A
.m W
e
-T-
GROUND MOTION EFTINATES The Chinese strong-motion data chown in Table 1 provide one refer-
'ence point for' directly estimating peak acceictations during small magnitude earthquakes. To accomplish this, the free-fleid peak acceler-
=
ations in Table 1 were regressed on earthquake magnitude (M computed j
from M using Nuttli, 1979) assuming an R distance dependence, which
~
3 is appropriate for seismic shear waves (McGuire and Hanks, 1980).
The resulting equation is:
i in a = 1.21 + 1.02 M - 1.5 in R where R is focal distance and a is the larger of the two horizontal 1
peak accelerations (which is about 15 percent greater than the average of the two horizontal peak accelerations). A comparison of estimates using this equation with peak accelerations estimated previously is shown in Table 2.
For the same magnitude and distance, the Chinese data indicate f
peak accelerations approximately one-half tho'se. estimated for Monticello Keservo ir. This 16 not surprising in that tha average stress drop for j
the Chinese ea-thquakes was approximately 10 bars (Figure 1) whereas a 25 bar stress drop was used for the Monticello Reservoir estimates.
f 5
E I
i i-
~
I l
I i
i f
l r
isn.
I
- - 1 SATURATION OF PEAK ACCELERATIONS When peak accelerations are estimated at locations close to the source of energy release of earthquakes, the saturation of peak acceler-ations with distance must be considered. That is, as the source-to-site distance decreases close to the source, peak accelerations do not in-crease at the same rate as at farther ranges. The theory used here is strictly applicable only in the far-field, defined as distances greater than twice the diameter of an equivalent circular source. At closer distances, peak accelerations increase, but on.ly slowly, with decreasing i
distance.
In estimating peak accelerations for different magnitude earth-quakes, it is logical that larger magnitude events saturate at larger distances than smaller magnitude events. The near-field, in terms of geometrical ef fects, extends over a larger volume for larger events, as a result of their larger fault surface. Also, if dif ferent magnitude e2rthquakes all are accompanied by surface fSulting, the effective point source of energy release for larger magnitude events is deeper (assuming steeply dipping faults and approximately circular fault surfaces). The estimation of peak acceleration saturation, then, requires determination of distances for different sagnitudes within which peak accelerations are invariant in the mean.
i -
An examination of available earthquake data leads to the conclusion that extrapolation of the far-field model to a ecurce-to-site distance of one source diameter gives a reasonable approxication to the acceleration saturat ion level.
Some of these data are shown ir. Figure 2, for Cali-fornia earthquakes with M = 6.5.
Also shown are the peak acceler-ations predicted by the theory applied with parameter values appropriate for California. Within about ten km (approximately the sourcs diameter),
peak accelerations are apparently saturated, and the theory predicts the observed mean peak acceleratio:as of about 0.4 g.
S b
i i
i O.'
M 4
v e+
.e4 4,
e mw, 9
,n,
.,--w.
,-, - - ~,
jt
-7_
For earthquakes at MontIcello Reservoir, an g = 4.5 event with Ao a 25 bare has a source diameter of approximately 2 lus. Using this distance, the predicted saturation level for peak accelerations is 0.21 g, as reported earlier.
Earthquakes larger than 4.5 are not considered poesible by the Applicant or by the U.S.N.R.C. staf f, as discussed above. The observed depth of the active seismic zone and the inferred shallow stress barrier virtually prohibit events larger than ( = 7.5 from occurring. For example, earthquakes of ( = 5.0 and 5.5 would have fault ing diameters of 3.6 and 6.3 km, respect ively. A blind application of the distance limits discussed above yield peak acceleration levels of 0.17 g and 0.13 g, respect ively. This does not imply that saturated peak accelerations decrease with magnitude; rather, other factors are important. For example, Mg, = 4.5 earthquakes rarely rupture the ground surface, uhorcas, g = 5.5 events often do.
Thus, it may never be possible to record "near-field" surface accelerations of an g = 4.5 earthquake. For this reason the peak acceleration estimate given ~above for g = 4.5 is considered quite conservative. - Selection of even closer distances than one source diameter for the Q = 5.0 and 5.5 events yields mathematically higher accelerations, as shown in Table 2.
These estimates are con-sidered conservative
'.n light of the discussion presented above.
A further reason for the use of dist.ances of 3 km or greater for Q > 5 is that the probability of occurrence of such a large evr.nt at j
such a clese distance is extremely low. An approximate calculation of the probability of such an occurrence can be made by coesidering the 4
history of accepted cases of reservoir-induced-seismicity in the Piedmont province of the eastern United States. The resarvoirs involved are as follows:
~,, -
---s
Water ~
Largest Reservoir Depth Volume Age Lgnitude Clark Hill 54 m 3.5 x 10' m3 29 yrs.
4.3 Jocassee 107 m 1.4'x 10' a 8 yrs.
3.2
~
Keowee 50 m 1.2 x 10' m3 12 yrs.
3.7 Menticello 49 m 0.5 x 10' m3 4 yre.
2.8 Total 53 yrs.
This history indicates that the return period for a magnitude 4.3 earthquake at a Piedmont reservoir with seismicity is on the order of 50 years.
In fact, the return period very likely is much longer because many reservoirs with suspected esismicity have not been included here.
Using a Richter b-value of 1.0 (which is conservative; reservoir sels-micity generally exhi'oita higher b-values than 1.0) laplies that a magnitude 5.3 event would have a return period of
- 500 years, assuming such large events could be induced in a tectonic regime with no known
.act ive faulting. Assuming a radius of influence of Monticello reservoir u 9 km (the largest dimension, which is a common assumption), the probability that the energy center of a magnitude S 5.3 earthquake occurs within 3 km of the nuclear station site is about 0.1.
Thus, the combined probability for the occurrence of a large (g = 5.3) event at a close d8. stance (< 3 km) is conservatively estimated to be about t
1/5,000. This is of the same order as probabilities of occurrence of tectonic events and, considering that conservatism involved in this calculation, is further justification for the use of 3 km discance for estimating ground accelerations for hypothet1:al, g S 5 induced j
~
l l
i
[
~...
l c_
g---.m.__..____
~
-6 tw
n 1'
-9
SUMMARY
- Consideration of1 strong motion data from Monticello' Reservoir,
- the. tectonic regime associated with the reservoir-induced earthquakes, f and data available from Heinfengkiang Reservoir in' China indicate that a stress drop of 25 bars'i: appropriate to characterize strong ground 5;
motion -for a wide range of magnitudes at Monticello Reservior. The peak accelerations predicted with the model developed here are. consistent with
. those recorded at' Hsinfengkiang Reservoir, given the dif ference in stress drops for the two areas. The use of one source diameter as the distance within shich peak accelerations saturate' is considered appropriate and -
conse rvat ive, in particular for the M 4.5 events Which are con-
=
g sidered the largest earthquakes possible at Monticello, as these shocks quite likely will not rupture to the ground surface.
Probability con-
'siderations indicate that larger. events (M -> 5) at close distances -
(< 3 km) are extremely rare, if they are considered possible at all.
e J
i l
I dwemh a ww-
- e. + -
e w
.-,e Tv-e-e--
- eed--
-a
==--***
REFERENCES
- Bolt, B.A., and W.K. Cloud (1974), " Recorded Strong Motion on the Hsinfengkiang Dam,. China," Bull. Seis. Soc. Am., 64, 4, op.1337-1342, Aug.
Hanks, T.C. and R.K. McGuire' (1981), "The Character of High Frequency Strong Ground Motion," submitted to Bull. Seis. Soc. Am., May.
Hsu, Tsung-Ho et. - al., -(1?75), " Strong Motion observation of Water-Induced Earthquakes at Hainfenghiang Reservoir in China," Inst. of Eng. Mech., Academia Sinica, May, 16 pp.
McGuire, R.K., and T.C. Hanks (1980), "RMS Accelerations and Spectral Amplitudes of Strong Ground Motion During the San Fernando, Cali-fornia, Earthquake," Bull. Seis. Soc. Am., 70, 5, pp. 1907-1920, Oct.
Muttli, 0.W. (1979), "The Relation of Sustained Maximum Ground Acceler-ation and Velocity to Earthquake Intensity and Magnituds," U.S. Army Waterways Exp. Sta., Vicksburg, Mir c. Paper 2-73-1, Ne 16, 74 PP-f
~
I 9
i a,,
t A
g-
,a
--,y y, - -
---=e,,----*--w---ew--<w
-e-o,-, --+ww-~w-m-ve*
:---e-*-
i--
2.
I l
TABLE 1 Strong Motion Data from Hsinfengkiang Reservoir (after Hsu et al, 1975) 1 D.rt h.
Tiine*
Fomt Epirentral g,;g
.puhe Ikic'
.leptli di=tance
.%rt y roun.lation Tv' of y'
'h'f' M.
br. min. um.
(kan)
(km) 8 K
tjur.tson 1
!!Nki.5.19 10-44-24.5 2.4 4.5 1.8 N.o G2.1 2
11%)G.5.3 10-UN-18 3.2 7.2 2.A 34.7 9.8 81.8 3
I!hkl.9.19 00-52-3.1 3.3 4.5 2.6 41.0 lu.4 105.7 4
1147.7. 2 194)7-31.2 3.6 5.3 3.3 64 *
- u 248.0
!!G.3.7 1G-54-1.~.G 3.5 4.=
2.9 32.5 22.0 124.0 a
IthW.3.19 02-::8-29.9 3.5 6.5 1.6 7.7 4.7 54.9 7
IW.A.M.2 21-27-30.9 2.8 3.2 1.2 -
S.3 27.G 199.9 4
1:w.st.4 04-5G-40 2.6 3.4 0.8 31.4 24.0 124.0 to 19 tis.b.23 12-45-13.9 3.7 6.4 2.0 70.2 41.5 33.3 l's 1970.2.19 04-47-7.3 1.7 9.o 4.s 20.2 Ild.A 11 197U.4.19 21-23-50.5 3.5 4.0 1.0 56.5 55.7 6pG.5 12 1970.4.19 21-23-59.8 1.1 4.0 1.0 23.7 23.4 60.4 13 1970.4.19 21-24-05 2.1 4.0 1.0 47.1 30.3 177.9 14 1970.5.9 00-01-32 2.8 3.7 2.5 47.5 40.5 516.0 15 1970.10.3 23-38-10.5 3.5 2.9 2.0 92.4 58.1 493.5 14 IS'lJ.10.3 "3-38-23.G 1.3 f.9 2.0 34.7 In.2 215.0 17 1971.1.2 07-41-54.5 3.0 3.4 1.8 62.8 39.4 255.0 IA 11871.1.2 08-23-59.6 3.1 3.3 2.'
58.1 33.6 159.5 19 Ifs 71.2.25 13-09-50 3.5 8.0 1.9 G8.6 480.0 W
1971.10.22 17-57-08 3.2 6.5 1.2 69.9 3G.9 491.0 el 1971.11.15 07-45-45.8 2.3 3.0 0.8 29.0 16.5 116.5
~ :::
1972.12.18 17-05-20 4.5 9.9 4.9 74.0 56.0 597.0
- 3 1973.3.11 20-47-31 3.0 7.2 1.8 20.1 11.9 152.0 24 1973.6.2 20-42-1.1 2.8 8.2 1.0 10.3 7.9 56.3 2*.
1974.1.24 05-43-16 3.0 4.8 4.0 30.9 19.7 229.0 DL 1974.3.1 06-14-46 3.1 6.9 2.5 15.4 6.6 32.0 i
27 1974.6.22 18-23-37.7 2.2 5.8 2.0 12.9 4.3 54.4 I
o 1974.8.16 07-33-58 3.0 6.8 2.5 4.3 50.9
- luing leant time.
s e..e a-
' -, +
-n y --
-ra wm-m-
- e e-s-
v n
.7, ;, ~ - -,
c ;,.
w-
- ;.7
+
~
ly
, ' J m, --
.a.'
i I
~
TABLE 2.
COMPARISON OF PEAK' ACCELERATION ESTIMATES EstimateLUsing Est imate Using JN R,. km.
Brune Model Chinese Data L
4.5
.2 0.21 g-
- o,-12 g
'4'
'5.0 3
0.20'g o,11 8
-55~
4 0.22 g
-0,11;g 4
s' k.
n h
r t
?
?
+
i E
e 1
+
t x-g
-, ~.
-i
+-
~
-..-.r.
13_
.t 6
-6 5-e
~
4-ee o
e e,*,
'e M
M e
3 3_
L e
e
-4
- e. e 2-
-3 e
l_
I i
a 0
10 20 30 40 ACI FIGURE 1
{
Stress Drop versus Magnitude for I~
Hsinfengkiang Reservoir, Data
_ ~,.
g w
4,.-e-
4
_gg_-
\\;,
- 2000 e
-1000-900-800-a 7004J U
600-s 50042 e
a m-(J e
U U
g S
O t
[J
+
O N
U 00 U
x
-300-O B
O e g g
La m
0 CD U
O 5
- p. 200-0 + 0* s e
S 0
o o,g 0 0 z
ID H
e H
O 5
c 100-E 90-e La 80-d
'70-
-u 60-C y
50-T LU Q (,.
C-KEY:
O IMPERIAL VALLEY (1979) 30-o SAN FERNANDO (1971)
+ CALIFORNIA hOTHERS l
20-l l
l 10,
. i 1
2 3
4 5
6 7
8 910 20 DISTANCE (KM) i FIGURE 2 Peak Acceleration vs Distance for Q -6.5 Earthquakes r
,---9.-
.m.
-.-9--
-