ML20054K488
| ML20054K488 | |
| Person / Time | |
|---|---|
| Site: | Big Rock Point File:Consumers Energy icon.png |
| Issue date: | 06/30/1982 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20054K485 | List: |
| References | |
| NUDOCS 8207020249 | |
| Download: ML20054K488 (13) | |
Text
'
Assessment of Possible Amplification at Big Rock Point Introduction In the initial (Memorandum from R. Jackson to D. Crutchfield, June 23, 1980) and final (Memorandum from R. Jackson to W. Russell, May 20,1981) recommendations for site specific spectra at SEP sites it was noted that possible anomalous site conditions at several sites (Palisades, Lacrosse and Yankee Rowe) were not taken into account.
Since then it has also been noted ( Affidavit of Richard L. Emch, Jr., NRC, in support of Motion for Extension of Time, May 3,1982) that Big Rock Point may also have possible anomalous site conditions. The purpose of this review is to determine whether, and if necessary to what extent, the spectra t
recommended in the original review for Big Rock Point need to be modified to take into account the site conditions.
Our recommendations are based upon the material reviewed for the June 23, 1980 and May 20, 1981 memoranda and the following additional items:
l 1.
Assessment of the Need to Correct the Probabilistic Spectra Developed for the Big Rock Point Site to Account for the Sites Soil Column - J. C. Chen and D. L. Bernreuter (LLNL), June, 1982.
2.
Letter from D. L. Bernreuter (LLNL) to L. Reiter, June 17, 1982.
l 3.
Menorandum from G. Lear to R. Jackson, Evaluation of Geotechnical Data for Developing Seismic Design Basis for the Site, June 17, 1982.
4.
Letter from_R. A. Vincent (Consumers Power Company) to D. M.
Crutchfield (NRC), May 5, 1982 with attached report:
Site Specific Response Spectrum Big Rock Point Nuclear Power Plant, Weston Geophysical Corporation, May 1982.
t 8207020249 820630 PDR ADOCK 05000155 l
P PDR
2_
Recommendation It is recommended that the original spectra put forth in the 1980 and 1981 memoranda are still appropriate for use at Big Rock Point. This conclusion takes into account that in deriving the spectra for the SEP, significant uncertainty was allowed such that the amplification present at the Big Rock Point site has already been generally accounted for.
The very low seismic hazard in Northern Michigan indicates that the chance that there will be earthquake ground motion of any significance at Big Rcck Point is extremely small.
Statement of Problem I
Generalized studies of, ground motion have shown that shallow soil sites record higher peak accelerations or high frequency motion than rock or deeper soil sites.
Campbell (1981) for instance, found that estimates based on the strong motion data set recorded on sites with 10 meters or less of soil over rock showed an average 84% increase in peak ground acceleration when compared to estimates based on data recorded on other geologic classifications. The Big Rock Point site is a shallow soil site.
It is underlain by approximately 13 meters of sand, gravel and glacial till overlying limestone.
In the approach utilized in developing spectra for the SEP, sites were classified either as " soil" or " rock".
Big Rock Point was classified as a soil site.
In general the western U. S. data set, upon wh:ch the recommended soil site spectra are mostly based, is made up of data recorded at deeper soil sites. The problem that has to be resolved is whether, and if necessary to what extent do the SEP Big Rock Point spectra have to be modified to reflect these shallow soil conditions.
In the following review we will assess
. not only the extent to which amplification occurs at Big Rock Point, but also the extent to which deviation from " typical" soil c'onditions were allcwed for in the original recommended spectra.
Both these factors must be considered in making a reasonable estimate as to the apprcpriateness of the nriginal spectra.
Site Amplificatinn at Shallow Sites:
Real Records Chen and Bernreuter (1982) address the amplification problem by examining amplification of earthquake ground motion at shallow soil sites as compared to nearby rock sites.
The comparison with rock is made so as to have a common and convenient reference point for all the spectra, including the original recommended spectra.
They first examined station pairs / nearby shallow soil and rock stations) which recorded ground motion from the 1975 Oroville, California earthquake and the 1976 Friuli, Italy earthquakes.
In the Oroville, California sequence the shallow site, DJR, showed an average amplification factor of 2.8 for the peak acceleration and maximum spectral amplification at 4.8 at 6.5 Hz when ccmpared to two nearby rock stations.
In the Friuli, Italy sequence the shallow site, Forgaria, showed an average amplification factor of 2.2 for the peak acceleration and a maximum spectral amplification of 4.2 at 2.8 Hz, when compared to a nearby rock station (San Rocco).
These amplifications should be treated with caution since:
1.
Another shallow soil station at Oroville, OMC, which was closer to the earthquake epicenters than DJR recorded significantly less ground motion than that of DJR.
At OMC peak ground motion was at about the sare level as that shown at rock and deeper soil sites (Table 111.2, Chen and Bernreuter, 1982). There are obviously other factors such as
. radiaticn patterns or focusing which may'be involved. Comparing station DJR to either a rock or deeper soil station while not utilizing station OMC nay yield inappropriate amplification factors.
2.
According to Figure V. 9 of Chen and Bernreuter (1982), the impedence cortrast or difference in raterial properties between the shallow soil and underlying rocks is much less at Forgaria than at DJR.
The amplification at Forgaria with respect to rock however is similar to that at DJR.
Simple theory indicates that for a given depth, the sharper the impedance contrast, the greater the amplification.
Other factors such as the geometry of the rock strata (Figure III.6, Chen and Bernreuter,1982) ray be affecting the relative amplification observed at Forgaria.
Amplification at Big Rock Point Chen and Bernreuter (1982) computed the amplification at Big Rock Point with respect to rock using the SHAKE Computer Program which assumes that all incoming waves are vertically propagating horizontally polarized They assumed four different models of the subsurface properties waves.
i at Big Rock Point covering the range of actual measurements. These models encompass the site conditions interpreted by the NRC staff (Memorandum from G. Lear to R. Jackson, June 17,1982). The amplification factors for peak accelerations in these models ranged from a factor of 1.9 to 2.2.
The maximum spectral amplification varied from 2.2 to 2.8 at frequencies from 10.5 to 14.0 Hz (Table V.2).
The spectral amplification as a function of frequency (Figure V.7, Chen and Bernreuter, 1982) may be approximated by a simple curve showing no amplification for frequencies less than 3 or 4 Hz, gradually increasing
l
. i amplification starting from 3 or 4 Hz reaching a factor of 2 at 10 Hz, i
and a factor of 2 amplification for frequencies greater than 10 Hz. As indicated above the actual peak spectra amplifications were achieved in l
l the 10 to 15 Hz range.
Variations in assumed properties for each model i
would result in higher or lower amplifications at different frequencies.
i i
I Uncertainties associated with SHAKE computations are well known.
l 1.
SHAKE disregards other wave forms such as surface waves which make up a significant element in strong ground motion.
2.
SHAKE uses only a linear approximation of the soil's non-linear
)
properties.
3.
Chen and Bernreuter (1982), attempted to validate the SHAKE program.
They compared the actypl records recorded at Forgaria with those recorded at the nearby rock station San Rocco after being passed thru a l
SHAKE model of the subsurface conditions at Forgaria. The results were mixed, some elements of the comparison were reasonable while others were l
1 l
t poor. The real data showed different amplifications for different i
i horizontal components, and peak amplifications at different frequencies then those predicted by SHAKE.
(
(
Uncertaintv Associated with Site Effects in the Recommended Spectra t
l In order to utilize the information on amplification discussed above and
[
i apply them to the original calculations for the SEP carried for Big Rock l
Point we must take into account two factors: 1.
Since the LLNL
[
amplification studies discussed above all related to amplification at r
....e..
)
..-. -..,. -, - _,.. _. - - _.. _ _,.. -. -.. _ _,. - _,,...... _, ~.. _.
. shallow soil sites as compared to rock, sites, we must first recalculate the SEP Big Rock Point spectra assuming it is a rock site.
2.
The appropriateness of the assumed uncertainty in the original calculations when applying a site specific amplification factor.
It is a fairly routine matter to recalculate the spectra for a rock site at Big Rock Point since rock site conditions exist as an option in the original calculations for all sites. The treatment of uncertainty, however, is more complex.
The original uncertainty or dispersion associated with the ground motion was II=0.9 + 2r, whefe r represents the standard deviation associated with the natural log of the spectral acceleration.
This was based upon uncertainty in the source, travel path, site, building effects and in the use of intensity as an intermediate parameter.
It was felt in our original review that 7= 0.7 + 30 would be more appropriate. The difference between the two estimates are small and counterbalanced by other assumptions. This is discussed in detail in the memorandum from Jackson to Crutchfield (1980). The dispersion is due in some part to variation in conditions from site to site within the generalized soil or rock data set and may be too large when applying an. amplification factor relating to a specific site condition (13 meters of shallow soil over rock) at a single site (Big Rock Point).
In order to determine the dispersion at a single site Chen and Bernreuter (1982) carried out several studies. They compared the ground motion recorded at two deep soil sites (Ferndale and El Centro) from earthquakes and found a systematic difference between them.
For an earthquake of a given size
at a given distance the data implies that Ferndale would record a peak acceleration approximately twice that at El Centro. Chen and Bernreuter (1982) then evaluated the dispersion associated at these and other q
individual sites from suites of earthquakes. They found that the dispersion at individual sites ranged from 810.26 to D'= 0.67. The stiffer sites showed less dispersion then the sof t sites. The average dispersion for a single site was about 010.45. Chen and Bernreuter (1982) suggested the use of a reduced standard deviation when a site specific correction (amplification) is applied to Big Rock Point. They then recalculated the appropriate probabilistic spectrum (see Memorandum from Jackson to Crutchfield, June 23, 1980 for details) at Big Rock Point assuming it wasta rock site with a standard deviation of I
l CT=0.45 + 3 cr (Figure VII.2).
They point out that the use of CT=0.45 reduces the estinated ground motion by aproximately 405 when cocpared to t
the original calculations.
T The staff evaluated other individual sites and also found that cr=0.45 is an average estimate of dispersion associated with ground motion at such sites.
The incorporation of uncertainty in ground motion with respect to site conditions is problematic. The difference between two " deep soil" sites, Ferndale and El Centro is of the same order as the difference between some of the rock and shallow soil sites discussed i
previously.
Ferndale and El Centro were both included in the original soil data set used to calculate the SEP probabilistic spectra, again indicating that a considerable amount of variation was allowed for originally.
tL______,____
. i Spectra from Real Records The original probabilistic spectra calculated for Big Rock Point were modified so as not to fall below the median level of a nearby magnitude i
5.3 earthquake (Memorandum from Jackson to Crutchfield,1980). The records used were generalized soil records which were somewhat different than the records used in the probabilistic calculations described above i
r l
in that foreign (Japanese and Italian) data was also included. This data set included records from Forgaria, the shallow soil site evaluated in Chen and Bernreuter (1982), at least one Japanese site (M-53), that i
has been characterized as shallow soil (NUREG/CR-1660) and several western U. S. sites that are shallow or near-shallow soil sites. Thus I
while the data set was not specifically chosen for shallow soil i
r l
conditions, shallow soil conditions are included in the range of data used in deternining the median level of a nearby magnitude 5.3 earthquake.
I j
Using real records, the licensee (Weston Geophysical, 1982) computed a site specific spectrum for Big Rock Point from a nearby magnitude 5.3 event and found that the median level was below the spectrum recommended by SEP. Although they included obvious shallow soil sites such as DJR i
and OMC near Oroville the staff found that several sites really did not l
fit the conditions at Big Rock Point.
i The difficulty in calculating a site specific spectrum for a shallow soil site such as Big Rock Point is that there are a very limited number t
of records with the appropriate site, earthquake source, and distance I
I conditions.
, The direct use of real records in the SEP was to set a generalized (not site by site) minimum for sites in the eastern U. 5.
Its use was not meant to be the definitive tool as has been done in recent OL reviews.
It is the staff's position that for this purpose the initial estimate i
from real records used in the Jackson to Crutchfield memorandum (1980) is appropriate since it did include shallow soil sites in the generalized data set.
The licensee's study, while not definitive in itself with respect to shallow sites, does support the adequacy of the initial estimate with respect to a wide range of conditions.
l Historical Hazard Analysis The seismic hazard at Big Rock Point is very low. According to a recent compilation of historical and instrumentally recorded earthquakes (NUREG/CR-1577) the cTosest earthquake occurred at a distance of more than 100 km from the site and this event was of Modified Mercalli Intensity V or less.
In addition, Chen and Bernreuter (1982) performed a historical hazard analysis i.e. using only actual events in the i
historic record (not moving them) and a ground motion model which estimates ground motion (peak acceleration) at Big Rock Point from these I
l events.
They estimated the return periods associated with peak accelerations at the site.
Depending on the ground motion model used the peak acceleration associated with 4000 year return period varied r
from 0.039 to 0.1.
The high value was determined using a ground motion 9
model that according to Chen and Bernreuter (1982) may over emphasize the distant (over 1000 km) 1811,1812 New Madrid Earthquakes.
- Indeed, l
using the most recent ground motion model (Nuttli and Hermann, 1981),
i results in peak accelerations on the order of 0.001g at a distance of l
1000 km.
Excluding the New Madrid events (which according to Chen and l
l
Bernreuter,1982, have estimated return periods on the order of 500 to 1000 years) results in a peak acceleration at Big Rock Point of 0.03g associated with the 4000 year return period. While no attempt is made to correct for completeness of the data or delineate earthquake zones these studies indicate that based upon 200 years of earthquake history the ground motion occurring at Big Rock Point has been very low and that simple projections of this history using current ground motion models, to long return periods on the order of thousands of years yields peak accelerations well below that originally recommended (0.1g) for the site.
Based on the above, the chance that Big Rock Point will experience earthquake ground motion of any significance is extremely small.
t Conclusions It is our position that the original spectra recommended for Big Rock Point are still appropriate.
This is based upon the following:
1.
Shallow soil sites in general do seem to amplify earthquakes ground motion.
However, the observed variation in this amplification is large ranging from very little (OMC) to substantial (DJR, Forgaria). These vartiations cannot be predicted simply from the soil rock impedance contrast or by simplified computer models such as SHAKE. Many of these observed variations in amplification may also be due to differences in
/
topography, wave type, radiation pattern or focusing.
2.
Data sets based on deeper soil sites such as those utilized in the original spectra also exhibit large variations.
The difference between Ferndale and El Centro for example is a factor of 2.
Sites when considered individually have an average dispersion equivalent to a
standard deviation of 7= 0.45 associated with the natural log of peak acceleration.
3.
SHAKE studies indicate that the average amplification at Big Rock Point with respect to rock sites is low at low frequencies and about a factor of two at high frequencies (greater than 10 Hz). SHAKE calculations are based upon simplifying assumptions that may not always l
be appropriate. They should used in a judgmental manner.
[
4.
In applying a correction to the Big Rock Point site it is appropriate to reduce assumed dispersion to that observed at single sites $'=0.45).
This was done by Chen and Bernreuter 1982, Fig. VII-2.
Applying a factor of 2 amplification at all frequencies results in a spectrun approximately equivalent to the original recommended spectrum t
(see attached figure).
i 5.
The original recommended spectra for Big Rock Point were controlled i
at high frequencies by a minimum level based on direct use of a generalized soil data set. Given the available data, the fact that the direct data set incorporated shallow sites and the observation that the shallow soil spectra developed by the applicant are below this minimum level, the original minimum level is still considered appropriate for use in the SEP.
6.
Extensive studies of amplification at Big Rock point may only be of i
4 marginal safety significance. The seismic hazard at this site is so low such that the chance that there will be amplified ground motion in excess of the previously identified spectr'um (Memorandum from R. Jackson to W. Russell, dated May 20, 1981) is extremely small.
,,n
-. ~ - - - -,
REFERENCES Campbell, K. W. (1981) Near Source Attenuation of Peak Horizontal Acceleration, Bulletin of the Seismolonical Society of America v. 71, pp. 2039-2070.
f'uttli, 0. W. and D. B. Hermann (1981).
Consequences of Earthquakes in the Mississippi Valley.
Preprint 81-519 American Society of Civil Engineers.
NUREG/CR-1577. An Approach to Seismic Zonation for Siting Nuclear Electric Power Generating Facilities in the Eastern United States, tiay 1981.
NUREG/CR-1f,60.
Compilation, Assessment and Expansion of the Strong Earthquake Ground Motion Data Base, September 1980.
9 e
9 m
8IC4tcCK PoItVT
/o00.0
- e i esi j
e S
6
[
,sf
/00 0,'
/
E a
s m
O fl lf OL b
ff Comm c}riolpcl c3 %
C.kSoQ
~
@M70 g
/
ing O 4S
,A
.t.)e &rve i
u-c.u
$ fo o f
ef E
r
-7 0 46 (Chen +6eme 1er b
/
.m) 5
/
=
m l
T 0
/< 0.
/
f
,o a
m W
.l l
O'0I O.Io
(,o
,n, PFA IOD-S ec.
D " O k het ulce Resp,m s p h d-B,3 ge go,,,
i
\\
n.