ML16340A944
| ML16340A944 | |
| Person / Time | |
|---|---|
| Site: | Diablo Canyon  |
| Issue date: | 05/05/1980 |
| From: | Kuo P, Rothman P Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML16340A945 | List: |
| References | |
| NUDOCS 8005070376 | |
| Download: ML16340A944 (56) | |
Text
UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION BEFORE THE ATOMIC SAFETY AND LICENSING BOARD In the Matter Of:
PACIFIC GAS AND ELECTRIC COMPANY (Diablo Canyon Nuclear Power Plant Units 1 and 2)
Docket Nos.
50-275 O.L.
50-323 O.L.
JOINT AFFIDAVIT OF ROBERT L.
ROTHMAN AND PAO-TSIN KUO Robert L. Rothman and Pao-Tsin Kuo, being of legal age and duly sworn, depose and say as follows:
I'.
Pao-Tsin
- Kuo, a structural
- engineer, is employed by the Office of Nuclear Reactor Regulation, U. S. Nuclear Regulatory Commission.
Dr. Pao-Tsin Kuo testified at the seismic hearings in the above captioned case.
A copy of Dr. Kuo's professional qualifications is bound into the transcript following TR. 8697.
Robert L. Rothman, a seismologist, is employed by the Office of.Nuclear Reactor Regulation, U. S.
Nuclear Regulatory Commission.
A copy of Dr. Rothman's professional qualifications is attached to this affidavit.
Dr. Kuo's contribution to this affidavit is with respect to structural engineering matters and the DCNPP design response spectra and Dr. Rothman's contribution is with respect to seismological matters raised by Joint Intervenors.
2.
The purpose of this affidavit is to provide our present'valuation of the preliminary seismological data from the Imperial Valley Earthquake of October 15, 1978 as it may affect the design response spectra for the Diablo Canyon Nuclear Power Plant (DCNPP).
We wish to emphasize that our analysis 'is based on preliminary data available from the Imperial Valley Earthquake; however, this data does not affect the design response spectra for DCNPP.
Back round:
The new data from the October 15, 1979 Imperial Valley Earthquake provides a
new data set for a single earthquake of magnitude M
= 6.9 or greater recorded s
at distanc'es comparable to the distance from the Diablo Canyon site to the Hosgri Fault.
The Pacoima Dam record of the San Fernando Valley Earthquake (M
= 6.6) was used to develop the design response spectra for DCNPP.
3.
The United States Geologic Survey (USGS) has not officially released their strong-motion records and response spectra for the Imperial Valley Earth-
'quake of 15 October 1979.
The USGS is still working on the preparation of the data.
The preliminary version of the data which is unofficial may contain errors and is subject to amendment.
The strong motion data has not been thoroughly evaluated.
It is currently under study by the seismological community.
No final conclusions have been reached as to the meaning of the data.
Such studies will very likely continue for several years.
Geolo ic Conditions:
4.
Although the distance from the Diablo Canyon site to the Hosgri fault
is comparable to that from the Imperial Valley fault to some of the strong motion recording stations, the geologic conditions at the recording stations are significantly different from those at the Diablo Canyon site.
The strong motion recording stations in the Imperial Valley Earthquake area are at sites located on over 300 meters of alluvium soil which is underlain by approximately 5.5 kilometers of sedimentary rock.
The Diablo Canyon site is underlain by a relatively thin section of sedi-mentary rock overlying the basement which consists of the Franciscan Formation.
5.
The geologic condition at a site can have a significant effect on the nature of the strong motion at the site; for example, through phenomena such as amplification effects produced by soft unconsolidated materials.
Thus, the geologic conditions in the Imperial Valley Earthquake area would be expected to affect strong motion data in a manner different from that expected at the Diablo Canyon site.
~Focusin 6.
Focusing is a signal amplitude directivity effect resulting from rupture propagation.
Dr. Brune's affidavit of February 29, 1980 (para.
5, page 5) states that three recent earthquakes have shown an asymmetric pattern of strong motion and damage, suggesting that the phenomenon of focusing by rupture propagation was operating in these earthquakes and further that this phenomenon is more common than previously assumed by many seismologists.
I t
A
7.
The focusing effect occurs to some degree in all earthquakes.
For those cases where there is fault rupture toward the location of an instrument, the resulting strong motion recording may reflect the focusing.
The response, spectrum for the Diablo Canyon Nuclear Power Plant was developed by Dr. Newmark by enveloping the response spectrum of the Pacoima Dam record of the San Fernando Earthquake.
The Pacoima Dam record had high peak accelerations.
These peak acceleration values were, in part,,caused by the focusing phenomena and therefor e the response spectra for DCNPP also reflect this focusing effect.
8.
For the Imperial Valley Earthquake the accelerations are generally higher along the fault to the northwest of the epicenter in the Imperial Valley.
The pattern of strong motion shows, however, considerable complexity, which is not explained by a simple propagating point source (Brune, et al, 1979).+
As an example of this complexity, the California Division of Hines and Geology Heloland Overpass Station which is located only 1 kilometer from the fault rupture and is northwest of the epicenter experienced relatively low (0.23g vertical, and 0.25g and 0.32g horizontal) peak accelerations in the free field.
Vertical Acceleration:
9.
At distances of less than 10 kilometers from the fault both the Imperial Valley Earthquake of October 15, 1979 and the Gazli Earthquake of May 17, 1976 had peak vertical accelerations which were higher than the horizontal
accelerations.
These observations do not affect the DCNPP design response spectra for the vertical component of acceleration which is set at 2/3 of the horizontal value.
The reason that neither the Gazli Earthquake nor the Imperial Valley Earthquake affect the use of a design value for the vertical component of vertical acceleration of 2/3 the horizontal value is based on site specific considerations.
These considerations are earthquake mechanism and geology.
Earth uake Nechanisms/Com arabilit 10.
In considering the comparison between these two earthquakes (the Imperial Valley and the Gazli) and a possible earthquake on the Hosgri Fault, it is necessary to consider the earthquake mechanisms involved and to compare the geology of the areas.
One explanation is that of Helmberger (1980)~2/
who has been studying the mechanism of the Imperial Valley Earthquake, particularly the cause of the high peak accelerations on the vertical recordings.
He suggests that high-frequency large vertical accelerations of this type occur on recordings at distances up to 10 km made in sedimentary basins containing strong velocity contrasts where the fault has broken into the sedimentary layers from the basement to heights of at. least 1 or 2 km.
Helmberger and Hadley (1980)+ state that many of the strong motion records produced by Imperial Valley earthquakes display high frequency vertical accelerations preceding the direct S arrivals.
These arrivals generally show some dispersion with the highest accelerations produced by the highest frequency signals.
The corresponding velocity traces produced by numerical integration drastically reduces their significance with respect to the
S-.waves.
They have modeled the Imperial Valley Earthquake and found h
that this overall behavior can be produced in synthetic motions by positioning shallow shear dislocation sources above a layered transi-tional zone that generally occurs at the base of sedimentary basins.
Post-critical angle P-wave reflections produce very high frequency pulses with low frequency head waves, or dispersive effects.
These P wave arrivals reach the surface traveling nearly vertically and thus produce high vertical motions with negligible horizontal motions.
Hartzell (1980)-
has studied the faulting of the Gazli Earthquake and the following is a synopsis of his study.
ll. The Gazli Earthquake of 17 May 1976 had a thrust mechanism.
The Gazli Earthquake rupture started at a depth of 15 km and propagated almost entirely unilaterally upward.
The thick sediments in the source region (1
km) had an important effect on the local strong ground motion.
The sediments probably contributed to the high frequency ringing in the acceleration records.
The large-amplitude, high-frequency accelerations on the vertical component may have been due in part to the focusing effect of the vertically propagating rupture.
High-frequency resonances, caused by the faulting of the sediments, may also have had an effect.
12.
The Hosgri Fault is a right lateral strike slip fault which is approximately 5 kilometers from the Diablo Canyon site at its nearest approach.
The Gazli Earthquake was on a thrust fault.
In general, the stress drop for earthquakes on thrust faults are greater than those for strike slip faults.
Since neither the geology in the area of the Imperial Valley Earthquake (unconsolidated alluvium and a distinctive sedimentary section when compared to Diablo Canyon) nor the fault mech'anism for the Gazli Earth-quake (Gazli Earthquake is on thrust fault) correspond to conditions at the Diablo Canyon site, these two events with their high peak vertical accelerations do not impute the adequacy of the design response spectra of the Diablo Canyon Nuclear Power Plant.
13.
The "Tau" reduction is an engineer'ing equivalent factor associated with the averaging of accelerations over the area of the structure.
This is taken account of by use of a value of transit time determined by the effective width (the square root of the area, in general) of the foundation divided by the wave velocity.
14.
The Imperial County Service Building is founded on a group of piles, which implies the lack of a good soil base for a large structure.
This is important to comparing building response because the type of foundation affects how a building would respond.
DCNPP is not founded on piles but is built on top of concrete over hard basement rock.
The motions recorded at the Imperial County Service Building foundation, it appears, are the amplified responses from the earthquake motions through the piles.
It is evident from the records of the horizontal motions at the Imperial Service Building foundation that these records contain low frequency components of motions that do not exist.in the free field motion~ and that the building-pile foundation system is a low frequency system in the neighborhood of 1.54 Hz to 1.56 Hz and 2.24 Hz to 2.85 Hz in the EW and
NS di recti ons, respectively.
- 15. Conversely, the vertical acceleration at the Imperial County Service Building foundation is lower than that in the free field.
This is the expected effect of a soil-structure interaction system since there is little amplification of ground motion through the piles in the. vertical direction.
It is apparent, therefore, that the records at the Imperial County Service Building from the 1979 Imperial Valley Earthquake can neither validate nor invalidate the use of the Tau effect as an engineering equivalent factor.
These records cannot serve as an adequate basis to alter the conclusion based on available data to date and experts'udge-ment that "tau" reduction is a realistic and reasonable engineering t
equivalent factor used to develop design response spectra.
Im erial Valley Res onse S ectra/Newmark Res onse S ectra:
- 16. Strong motion records for the Imperial Valley Earthquake of October 15, 1979, from a total of 10 stations within 10 km were available from the USGS.
The response spectra of these records at 5X of the critical damping were compared with the corresponding Newmark horizontal and vertical design spectra anchored at 0.65g and 0.50g, respectively.
Only two horizontal response spectra (stations 5 and Bonds Corner Station) out of twenty slightly exceed, the Newmark horizontal design spectrum in the frequency range of interest (2-33 cps),
as shown in the attachment.
Host of the response spectra exceed the Newmark design spectrum for the frequencies lower than 0.5 cps (2 seconds) which is of no interest in the design of nuclear power plants.
17.
For vertical response spectra of the Imperial Valley Earthquake, only three (Stations 6,
7 and El Centro Differential Array) out of ten exceed the Newmark vertical design spectrum in.the frequency range of interest.
Stations 6 and 7 are within 1
km distance from the El Centro fault.
Stations 6 and 7, therefore, are not directly comparable with the Diablo Canyon site which is located approximately 5
km from the Hosgri Fault.
Furthermore, Station 6 is located between two faults in the Mesquite
- 6asin, which is a
down thrown block.
The geological implication from the location of this station is sufficiently complex that the motions recorded at Station 6 are not applicable to the Diablo Canyon site.
18.
The only record which exceeds the Newmark vertical design spectrum and is within the same distance range is from the El Centro Differential Array Station.
However, the representative character of this record is in doubt for the following reasons:
a.
The El Centro Differential Array Station is between Station 8 which is 4 km from the fault and Station 10 which is 9
km from the fau'lt.
A vertical time history record was not available at Station 8, but the vertical peak acceleration at Station 8 was 0.55g which is far lower than 0.93g recorded at the El Centro Differential Array Station.
The vertical response spectrum from Station 10 is also far below the Newmark vertical design
- spectrum, as shown in Attachment 2.
b.
The peak acceleration values used by the Joint Intervenors'n their Motion to Reopen of March 28, 1980 were examined.
The peak
acceleration values they used are those which were scaled from the analog records without regard being given to the frequency of the signal.
The peak acceleration values reported by the U.S.G.S.
as obtained from the digitized and corrected accelograms are signifi-cantly lower.
For example, the peak vertical acceleration for the El Centro Differential Array as read from the analog record is reported as 0.93g while the value from the component after correction is reported as 0.66g by the U.S.G.S.
Nuclear power plant structures will not'e affected by high frequency accelerations of this type.
19.
The above factors raise doubt about the representative character of the records at the El C'entro Differential Array Station.
Thus, the records from the 1979 Imperial Valley Earthquake do not serve as an adequate basis
='for altering the previous conclusion that the Newmark design response spectra be used in design reanalysis of the Diablo Canyon Nuclear Plant.
==
Conclusion:==
20.
The Joint Intervenors'otion to Reopen stated on Page 4 that "earthquakes of N7.5 are expected to produce even higher peak accelerations."
The staff is in basic agreement with the statement providing that a site under consideration is some distance away from the earthquake source.
- However, for sites which are very close to the source such as the Diablo Canyon
- site, the maximum acceleration is practically independent of the magnitude.8/
During the presentation to ACRS on June 14, 1978, Or.
Newmark elaborated on the rationale why the Newmark design spectra anchored at an acceleration of 0.75g for the postulated M7.5 earthquake are adequate for the Diablo
Canyon site.
The following is an excerpt from the ACRS transcript (Pages 105-108):
"Another point that is of extremely great importance that has been raised by several of the ACRS consultants is the statement that if the magnitude of 7-1/2 really occurred, we would get 50 percent more acceleration than for a 6-1/2 earthquake corresponding to the Pacoima Dam record which I used for my estimate of the motions at Diablo
- Canyon, but the data indicate in general from Ancona, Nelendi Ranch as we accept it, and a few other cases, that the maximum acceleration is practically independent of the magnitude, very close the source.
It depends only on the nature of the breaking strength of the rock and its ambient stress
- before, and the magnitude itself has little influence.
When you go out some distance then magnitude has an important effect because it defines how much energy is focused to the point where we are concerned with placing a structure or responding element."
... "in general, the velocity is a better measure of what happens to the response
- spectrum, because of the fact that it takes account of the energies that are available, rather than the peak acceleration, which might occur at a very high frequency."
"In general, the response of structures as inferred by observed
- damage, damage to equipment as well as to structural elements them-
- selves, damage from blast or other kinds of ground motion, all indicate that the response is indeed governed by something less than the peak instrumental recorded acceleration, when one is close to the source of motions."
21.
As pointed out by Newmark, the peak ground acceleration often occurs at a very high frequency.
This was indeed the case with the vertical record at the El Centro Differential Array Station. during the Imperial Valley Earthquake of October 15, 1979.
The peak vertical acceleration of 0.93g was recorded at a frequency above 50 Hz which is of no interest in the design of nuclear plant structures.
High frequency peaks of short duration will contribute little to the response of structures.
Furthermore, damage due to the 1979 Imperial Valley Earthquake at the
El Centro Steam Plant, which was approximately 5
km away from the fault on which the earthquake
- occurred, was minor.
No damage of any type that would have safety significance at a nuclear power plant occurred and the units were back on line in a matter of a few hours.
The systems, components and instrumentation at the El Centro Plant are analogous to those in a nuclear plant and the El Centro Plant had little or no specific seismic design considerations factored into its design.
The performance of the El Centro Steam Plant is entirely consistent with Dr. Newmark's design response spectra anchored at
.75g for the 7.5N postulated earthquake on the Hosgri.
- 22. Contrary to the Joint Intervenors'ontention, the new data set does not provide any evidence that would invalidate the Newmark design response spectra for DCNPP developed in accordance with the procedures and require-ments of 10 CFR Part 100, Appendix A..
References 1.
Brune, J. N., F. Vernon III and R. Simons, (1979), Strong Motion Data Recorded in Mexico for the October 15, 1979 Imperial Valley Earthquake, Abstracts, The Seismological Society of American 1980 Annual Meeting, Earthquake Notes, Vol. 50, No. 4, Page 49.
2.
Helmberger, D. V. (1980),
Personal Communication 16 April 1980.
3.
Helmberger, D. V. and D.
M. Hadley (1979),
A Note on the High Frequency Vertical Strong Motions Recorded in Sedimentary Basins, Abstracts, The Seismological Society of America 1980 Annual Meeting, Earthquake
- Notes, Vol ~ 50, No. 4, Page 50.
4.
Hartzell, Stephen, (1980), Faulting Process of the May 17, 1976 Gazli, U.S.S.R.
Earthquake Pre-print of Paper submitted for publication to the Bulletin of the Seismological Society of America.
5.
- Newmark, N. M., Appendix C to SER Supplement No. 5, Page C 1-3.
6.
- Rajahn, C.
and Ragsdale, J. D.,
A Preliminary Report on the Strong Motion Records from the Imperial County Services Building.
7.
- Pardoen, G. C., Imperial County Services Building, Ambient Vibration Test Results, Research Report Abstract, EERI Newsletter, Volume 14, No. 1, January 1980.
8.
- Newmark, N. M., Presentation to ACRS on June 14, 1978, Transcript Pages 105 through 108.
The above statements and opinions are true and correct to the best of my personal knowledge and belief.
7 I
Pao-Tsin Kuo Subscribed and sworn to before me this
day of May, 1980.
Notary Public I-'/
I~
4
/I Ny Commission Expires:
/
The above statements and opinions are true and correct to the 'best of my personal knowledge and belief.
Robert Rothman Subscribed and sworn to before me this day of May, 1980.
Notary Public I'
My Commission Expires:
(
).
ROBERT L.
ROTHMAN GEOSCI)NCES BRANCH DIVISION OF ENGINEERING U. S.
NUCLEAR REGULATORY COMMISSION i'y name is Robert L. Rothman.
I presently reside at 8409 Stonewall Drive, Vienna, Virginia 22180 and I am employed as a Seismologist in the Geosciences Branch, Division of Engineering, Office of Nuclear Reactor Regulation, washington, D.
C. 20555.
PROFESSIONAL QUALIFICATIONS I received a B.S.
degree in Geology from Brooklyn College and N.S.
and Ph.D.
degrees in Geophysics from the Pennsylvania State University.
I.have been employed by the NRC since October 1979 as a Seismologist in the evaluation of the suitabi lity of nuclear power plant sites.
tly.
areas of expertise include seismicity, rupture mechanics, seismic wave propagation and seismic instrumentation.
From 1975 through
- 1979, I was employed by the U. S. Air Force Technical Applications Center as a seismologist in the nuclear explosion detection program.
I was involved in several projects of this program both as a Technical Project Officer and as a
researcher.
These projects included the detection of and the discrimination between underground explosions and earthquakes, magnitude and yield relationship studies, seismic network detection and location capability studies, regional and teleseismic wave propagation studies and projects to operate seismic instrument arrays and automatic data=processing and coornunications systems.
From 1965 through 1970 I was employed as a seismologist by the U. S.
Coast and Geogetic Survey.
In this position I was involved in studies in the areas of engineering seismology, seismicity and earthquake aftershock sequences.
This v ork was per ormed as part of a,program to investigate seismic hazard in the United States.
From 1959 to 1961 and during 1964-1965 I was an Engineering Geologist with the New York State Department of Public Morks.
In this position, I conducted geophysical field surveys in support of construction projects such as bridges, buildings and highways.
ROBERT L. ROTH!fAN GEOSCIENCES BRANCH DIVISION OF ENGINEERING U. S.
NUCLEAR REGULATORY COt&ilISSION i'y name is Robert L. Rothman.
I presently reside at 8409 Stonewall
- Drive, Vienna, Virginia 22180 and I am employed as a Seismologist in the Geosciences Branch, Division of Engineering, Office of Nuclear Reactor Regulation,
! ashington, D.
C. 20555.
PROFESSIONAL EQUAL I F I CATIONS I received a B.S. degree in Geology from Brooklyn College and M.S.
and Ph.D.
degrees in Geophysics from the Pennsylvania State University.
I.have been employed by the NRC since October 1979 as a Seismologist in the evaluation of the suitability of nuclear power plant sites.
le areas of expertise include seismicity, rupture mechanics, seismic wave propagation and seismic instrumentation.
From 1975 through
- 1979, I vras employed by the U. S. Air Force Technical Applications Center as a seismologist in the nuclear explosion detection program.
I was involved in several projects of this program both as a Technical Project Officer and as a
researcher.
These projects included the detection of and the discrimination between underground explosions and earthquakes, magnitude and yield relationship studies, seismic network detection and location capability studies, regional and teleseismic wave propagation studies and projects to operate seismic instrument arrays and automatic data processing and comounications systems.
From 1965 through 1970 I:,as employed as a seismologist by the U. S.
Coast and Geogetic Survey.
In this position I was involved in studies in the areas of engineering seismology, seismicity and earthquake aftershock sequences.
This work v as performed as part of a program to invedtigate seismic hazard in the United States.
From 1959 to 1961 and during 1964-1965 I was an Engineering Geologist with the New York State Department of Public Works.
In this position, I conducted geophysical field surveys in support of construction projects such as bridges, buildings and highways.
l
I0.0 5
ioo.o
~
~
D 230 140 IO.O VERTICAL qP a
C~/
O.l LI
n
=t P
f
\\
00.0 N0.0 315 to.:
225 VERTICAL l.O
>0.0
II 4'
N00.0 g
eo.o
~
~
~n 225 3l5 4g e
I lK IO.O l
o g0$
~b'
/////
VERTICAL
~a C~
LO I PERIOD-SEC.
I.O
>0.0
I'I ll I
1
$ o>5 5'/~ 9 t.as~
y) 50 Wl7'vik(E pys w CJF /icP gM peak iil vAk4+
is o~~ Ve N00.0 225 315 p4
~
0'Q.O
//////
VERTICAL
~o C~
O. I C,.OI
- 0. I PKRIOO-SEC.
I.O
/
I0.0
'l
~
$0.0 40'30 IO.O VERTICAL O. I 0.0 I O.I PERIQO-SEC.
I 1'
4.
I C
'l
g ioo.o
~
~
l40 VERTICAL
$0.0
////
230 0.0 I PERIOO-SEC.
I.O
1
SPo
]QQ g4o gl csmto A8Kay' kg.. )KK.
N00.0 230 140
<oo.o X
~ERTtCAt.
l0.0
~0+
0'/
LOt.
O.I PERtOO-SEC.
l.0
~ 'I
i>dc7 7p
'0.0 230 g
too.o Ih
~g Qe IO.O VERTICAL
////
io Cg n gA 0.0 I O.I l.0
<0.0
N
gg e461kb'>IXI g5 f. C Ch<t K&HRK iso'pO' lb oc.f 7$
.f
.1 N00.0 rP"
$0.0 140 230
~ge Qr
}0.0 LOI PERCD-SEC.
I.O I0.0
'I 4
Ac cgWI5 jbt4Q i'0 gal QOWWIAHI7f ~5@ijul, 2 ~
go'~4 C taws 4
NQ0.0 0.0 IO.O
//
///
320 50 qEPTICAL O. I LOI O. I PERIOD-SEC.
I.O I0.0
8 I
fl i
oo.o v~
60 270 VERTICAL
'0
A