ML19345D012

Seismic Evaluation of Vallecitos Site:Basis of Earthquake Ground Motion Design Criteria

ML19345D012
Person / Time
Site:	07000754
Issue date:	09/29/1980
From:	Hall W, Newmark N NATHAN M. NEWMARK CONSULTING ENGINEERING SERVICES
To:
Shared Package
ML19345D004	List: ML19345D003 ML19345D009 ML19345D012
References
CON-NRC-03-78-150, CON-NRC-3-78-150 17584, NUDOCS 8012080779
Download: ML19345D012 (11)

Text

_.

7 n c 6curuneis t <.

(./

~

4 NATHAN M.

NEWMARK 1211 CIVIL ENGINEERING BUILDING CONSULTING ENGINEERING SERVICES URBANA. ILLINCIS 61801 29 September 1980 i

SEISMIC EVALUATIO?l 0F VALLECITOS SITE --

BASIS OF EARTHQUAKE GROUtID HOTION OFSIGil CRITERIA by W. J. Hall and fl. M. tiewmark flathan M. flewmark Consulting Engineering Services 1211 Civil Engineering Building Urbana, Illinoit 61801 IflTR00UCTI0tl The purpose of this background memorandum is to describe in more detail the basis of our reasoning employed in arriving at the recommended seismic review criteria for the Vallecitos site as presented in our memorandum of 14 April 1980. Our recommendations remain unchanged.

It is important to note that the earthquake ground motion design criteria constitute but one part of the design criteria employed in the design or review of a nuclear i

facility.

The most important seismic effects for a nuclear plant arising from earthquake-induced motions at the site normally would be expected to occur from earthquakes having a source close to the site rather than from more distant earthquakes on the various fault systems. This observation has oeen taken into account-in arriving at the recommendations for the Vallecitos facility since the sources are considered to be possibly very close to, or under, the site.

Most of the information available about earthquake motions is probabilistic in nature but, in the absence of definite probabilistic approaches and established criteria for estimating such effects, the 4-m.-,-*

801g9.g9yyg

2 i

4 writers have chosen to use a partly deterministic approach, based on expected maximum values of-ground motion. There is reason to believe that the approaches adopted by the writers are consistent with probabilistic

' approaches now under development. At this time we would be reluctant to.

use on all probabilistic approach.

The current. approach used by the writers to define the design motion i

criteria involves selection, for the appropriate seismic hazard, of earth-l quake design (or " effective" in the sense of applicability to design) ground i

motion parameters, particularly peak values of acceleration, velocity and displac&,.n that are to be used to develop design response spectra. These spectra can be employed directly for design purposes as a basis for evaluating time histories to be used in the analyses. The peak acceleration used to anchor the near-source SSE spectrum generally will be somewhat smaller than the maximum ground acceleration recorded by an array of near-field instruments.

The reduction from the maximum free-field acceleration, as recorded on instruments, to a so-called " effective" or " design" seismic coefficient (acceleration) to which the spectrum is anchored, is generally negligible for earthquake sources that are more than some 30 to 50 km distant by virtue (l

of _the smaller and more regular motions, and the reduced high frequency content, 'as governed largely by surface motion effects. On the other hand thera is a rather well defined and growing body of data which suggests that I

j the response of structures and equipment located close to an earthquake j

source correspond's_ to effects associated with daller ground accelerations than the peak values recorded instrumentally in the near free-field.

The; foregoing observations are not new and have provided in part the

-basis for the coefficients employed in-building codes.

Specifically, the

+.

t 3~

near-field effects (as? educed from measurements and observations) as d

affected by the type and geometry of the structure, by rail-structure interaction and feedback, by the incoherent and complex seismic wave field, and by damping and energy dissipation mechanisms, on motions transmitted to the structure, typically have led to " design or " effective" (acceleration) coefficients in the lower levels of buildings that are less than the peak near free-field instrumental values. Recent unpublished studies by the TERA Corporation suggest that at least a 20 percent reduction in motion is indicated when data on buildings and free-field data are both available.

Because of the foundation conditions (structural mat and a relatively rigid structure) there is probably a more significant reduction for reactor structures; the relatively large and rigid foundation mat responds to some average acceleration value associated with the travel time of the seismic l

waves. An analogy of some help in visualizing this interaction effect is to consider the motions transmitted to a small boat and an ocean liner in rough seas.

The sitaation in the case of the Vallecitos General Electric Test Reactor is somewhat, hot not generally, different from that just described.

i In this case we are attempting to define the ground motion parameters to be employed as criteria in the review of a facility located near or over a fault region.

From the available earthquake records obtained close to fault zones, such as the 1979 Coyote Lake (f1 5.9) and the 1979 Imperial L

Valley (ft 6.9) earthquakes it is apparent that there is considerable 3

scatter in the values of. recorded free-field acceleration in a given region.

Moreover, just as in the case of the noted recent earthquakes, it has been observed for many years that the dama%e associated with' earthquakes having T

P

~* +

-=e6-,.g c

g

,,g

4 high frequency peaks of acceleration have not been consistent with the force levels one might infer from such accelerations. Other examples include the 1971 San Fernando (M 6.4) earthquake (1.15 g)* with the

gh acceleration at Pacoima Dam which did not produce damage to the dam nor to a caretaker's house located at the base of the dam. Similarly in the case of the 1972 Ancona, Italy (M 4 to 4.5) earthquakes (0.4 to 0.6 g), the authors are aware that structures in the region, which were subjected to rather a vere shaking, came througn in excellent condition.

Likewise in the case of the 1966 Parkfield (M 5.5) earthquake (0.5 g) and the 1972 Melendy Ranch (M 5.1) earthquake (0.7 g) the damage noted was not commensurate with the high acceleration peaks that were recorded. These documented cases strongly suggest that a single peak of lar; amplitude, short duration acceleration does not contribute significantly to the damage po.tential of an earthquake.

In short, the instru-mentally recorded peak acceleration (especially in the near-field) is a poor indicator of the severity (damaging potential) of the motion.

Studies of the effects of harmonic and transient excitations on the response of simple systems have been carried out for years and have been reported in the literature. The significant amplifications in motion, as reflected in response spectra, are associated generally with repetitive motions in the frequency ranges of consideration. For these reasons then, a single short duration spike of acceleration would not be expected to constitute a majur factor in causing damage. And this observation seems to be borne out by field surveys following earthquakes, by the results of military blast studies, as well as 1 om theorc;ical considerations.

This

• The acceleration values in parentheses are approximate peak va:ues.

S subject, namely the response arising fe:m transient excitation, is under continuing investigation in many laboratories; and through research and field observations in a few years one anticipates that additional understanding of these phenomena will be forthcoming, especially as it affects the response of structures and equipment.

With realization that structures and equipment systems in a nuclear reactor would not be permitted to be strained inelastically to a degree which would permit significant amcunts of deformation, it would seem unwise to permit major reductions in motion based on ductility demand as is the case in standard building codes. Similarly the damping values employed in nuclear plants are normally less than those wnich might be expected in structures and systems undergoing significant motion and some degree of nonlinear behavicr. Althougn these items have not been explicitly taken into account by the authors in developing the recommended criteria, we feel that the effects noted add a significant degree of conservatism to the evaluations in this particular case.

The above observations and concepts were used in an implicit manner by the authors in judgmentally arri ing at tneir recommendations. Even further details pertaining to seismic resistance and structural damage have been presented recently by Newmark (1980).

BASIS FOR RECOMMENDATIONS In that which follows, essentially the same procedure was employed in arriving at the criteria applicable to motions arising from excitation on the Calaveras fault system as for motions arising from effects associated with the Verona system. The slight differences in philosophy of approach are noted below.

{

l

~

6 Initially, and as only one step in the process, we considered the earthquake hazard as reflected by the magnitude of the earthquake expected on the respective fault; actually we gave consideration to a range of magnitudes.

It should be recognized that the magnitude is normally f

conservatively selected, and thus is not to be considered a "mean" magnitude in 'ter=s of definition of the earthquake. Thereupon a series of parametric calculations were made employing the attenuation relationship of the general type proposed by Dr.

N.' Donovan in his papers appearing over the past decade, using for conservatism 20 km instead of 25 km as the residual term in the expression for estimating distance effects. The distance values employed for both sets of close-in estimates corresponded to a horizontal distance of about 5 km and a depth of about 10 km in the final analysis. Our experience in recent years in making estimates of this type suggests that values ia this general range lead to reasonable estimates of acceleration when coupled with -

the appropriate magnitudes for geologies of the types being considered here.

For earthquakes of magnitude 6.5 to 7 the mean accelerations were found to be 0.35 to 0.4 g and for magnitudes 5.5 to 6 the mean accelerations were found to be about 0.2 to 0.25 g.

The mean plus standard deviation values (16 per-cent exceedance), which' we believe to be the desired values to be used for comparative purposes were estimated to be about 1.6 times the values just noted,' or 0.6 g in the first case, and about 0.4' g in the second case.

With regard to the mean values it is instructive to note that the recent Incerial Valley (M 6.9) instrumental data available to us indicate 0.35g mean, s

and a mean plus _one standard deviation value of about 0.55 g for all recordings within 20 km (16 stations) of. the fault trace, and 0.40 g and 0.57 g respec-L

- tively within 13 km (13 stations)'of the fault trace.

In the case of Coyote

..U

s 7

Lake (fi 5.9)'within 8 km of the fault (6 stations) the values were 0.24 g g

and 0.32 g respectively.

It is interesting to note how close the values come to the computed values just described. We are aware of other studies that-have been made recently in which the same general trends have been observed for larger _ groups of close-in data for a range of magnitudes.

A recent study by Donovan (1980) shows data from 115 stations within 20 kilometers of a fault, plotted to show the relation between peak recorded acceleration and distance from the fault in kilometers. iiagnitudes from 3.5 or less up to 7.5 are considered, and the data show no consistent relationship with magnitude. This is an indication of the fact that the high dependence on magnitudes reportet in some attenuation relationships is probably a spurious one. Basic physical principles indicate that close to the source of the earthquake the peak acceleration should be deperident only_ on the local conditions, including the shear strength of tiie rock at the focus and is not a function of the magnitude, although at larger distances' the acceleration is' dependent on magnitude or amount of. energy release. A plot of these data is shown in Plate C-1 taken from Donovan (1980). The linear relationship shown on plate C-1 was drawn to indicate the approximate level of the 84' percentile or median plus one standard deviation value and support the estimates made by us herein. Also recent similar studies by TERA submitted as a part of the-Diablo Canyon ASLAB hearings provide general confirmatory support l.

Over and above the estimated motion values just noted it should be apparent frorn' the foregoing discussion that higher peak values would be -

expected to be interspersed. The USGS in their evaluation state that one -

could expect instrumental peak acceleration values in excess of 1 g.

.8:

While we concur with this ' observation se applicability to design needs to be evaluated in some manner, much as we-have attempted to explain in the foregoing ma'terial. : At the present time based on information available to-us, it is our belief that even in the vicinity of the largest fault on the West Coast, the San Andreas fault,. the use of a design acceleration in the range of 0.8 g for anchoring the Reg. Guide 1.60 spectra or the flUREG 0098 spectra would lead to a' reasonably conservative basis of design.

'/hus, on the basis of considerations of the type noted, we recommended in the 14 April 1980 memorandum that the most reasonable value of acceleration to use for anchoring the spectra'for effects arising from the Calaveras fault would correspond to 0.6 g, but for design or review conservatism we suggested a value of 0.75.

This value reflects the fact that there is some degree of uncertainty in estimating such motions and that the hazard specified by the USGS corresponded to a magnitude 7.5 earthquake. We noted that we did not expect fault motion of significance _to be transferred to the site form activity on the Calaveras fault.

In a similar manner, in the case of the Verona fault, we stated,that it was our opinion that, from the information available, an acceleration value of about 0.40 g was the most reasonable value for anchoring the response spectra, but for conservatism we recommended use of a value of 0.6 g.

The margin between the most likely value and the recommended value here is larger to account for a greater degree of uncertainty and for the fact that the USGS specified the hazard to be that associated with a magnitude 6.5 earthquake. Also we noted that the motion was to be taken as acting simul-taneously. with a fault motion of. not more than 1 meter, interpreted to be the' resultant (net) motion in-any arbitrary direction.

9 Faulting As for faulting, our recomendations are based on a study of existing data as reported in available documents, and from other information available to us generally on faulting. There appears to be no conclusive evidence that past faulting extends into the GETR foundation. Thern is evidence that the fault planes may pass beneath the reactor foundation at some depth. On the basis of that information it is our judgment that one meter of net displace-ment (in any direction) is an acceptable value to use for evaluation purposes when combined with the vibratory motion. Even if faulting occurs at or near the foundation, that does not mean the foundation would be subjected to the peak fault displacement; more likely it will be affected by an average fault displacement which most likely would be on the order of 40 percent of the peak displacement.

Concludina Observation Thus, in summary, it should be apparent that the values recommended were judgmentally determined in a deterministic manner based on available information, and with realization of the effects upon stractures and equipment, and the resistance of such systems to seismic motion. We believe that the recommended values are adequately conservative to assure,

.when cot pled with the other criteria, that GETR will have a system resis-tance to excitation and displacement such that it has a high probability of achieving safe shutdown and renoval of residual heat.

1

e; 10

~

References N. 'C. ' Donovan (1980), " Problems.of an Expanding Field of Applications in Seismic Risk," American Geophysical Union,, Abstract, Spring Meeting,-

Toronto.

N. M. tiewmark '(1980), " Testimony of Nathan M. tiewmark" before the Atomic'

Safety and Licensing Appeal Board in the matter of Pacific Gas and Electric Company (Diablo Canyon Nuclear Power Plant Unit flos. I and 2), Docket flos'.

50-275 0.L., 50-323 0.L.

~

b t

I f

i t

i L'

I-3 31m I

mi u m o e m wwvtz y mm.

Q. @d)TafPL La oja-Y ld 799 g P E A K t.;; E E RI.TI O N ( 01.LS )

5 h

o 7

o o

o O

o o

o o

o o

o o

o x

88G L

n

~

J in em 5

8.3 GG G

a>H e

G GG G Cn GG % GG G

c I

E mm eGGG E G ob G

!G i

I l

l l

~

5CerJB GG G9 l

1, O

I

=

GG m

eGA j

j m

ee l

o n

i 2

! O e ee9'O

/

E e

z i

+

x i

m sn - we o

e t

GGG 5l' H

C

=

5 E

S3GD3 Gol E

I l

8 8

. e w, v I

GG3 G '

c y e e e aaaaa C3E333:i e

19 e

!EE55 I

l f

C C C C

C Gg d5350 1

5" GB

,.. v r i

OO S f:M "4 y y & v

• O A

l 55 o

E so esa :

er.;

a I

x f

I u v v =

G G3.G 'GG d

u I

O n!

I 2

t' Cn l

11 t

S

.c

).

i 0

i Attachment i e'~

N A,T H A N M.

NEWMARK 1211 CIVit. ENGINEERING BUlt. DING CONSULTING ENGINEERING SERVICES URBANA. ILLINCIS 61801 14 April 1980 eo

- g o

o Mr. Chris fielson i

0;:erating Reactors, Branch tio. 4 Division of Operating Reactors Nuclear Regulatory Comission Washington, D. C. 20555 Mr.~W. Burkhardt Fuel Reprocessing and Recycle ~ Branch Divisicn of Fuel Cycling and Material Safety ticclear Regulatory Commission

.Willste Building 7915 Eastern Avenue Silver Spring, Maryland 20910 Re: Seismic Evaluation of Vallecitos Site Contract NRC-03-78-150 Gentlemen:

We are enclosing two copies for each of you of the report on the Seismic Evaluation of Vallecitos Site by N. M. tiewmark and W. J. Hall dated 14 April 1980.

Respectfully submitted, bN. h k

W. J. Hall q %. 9h P

enclosures N. M. flewmark Distribution:

2 - l'r. Chris t'elson, NRC 2 - Mr. W. Burkhardt, NRC 3 - Cr. N. M. flewmark 3 - Cr. W. J. Hall 17531 h,

r

NA'THAN M.

NEWMARK 1211 CIVIL ENGINEERING BUILDING CONSULTING ENGINEERING SERVICES URB ANA. ILLINOIS 61801 14 April 1980 SEISMIC EVALUATIOt 0F VALLECITOS SITE by fl. M. tiewmark & W. J. Hall Nathan M. Newmark Consulting Encineerinc Services 1211 Civil Engineering Building Urbana, Illinois 61801 I.

CC:iDITIO:is CC.ISIDERED AfiD BASIS 0.c EVALUATION The purpose of this report is to define a rational basis for seismic evaluaticn of the General Electric test reactor and other facilities located near Vallecitos, California. The major facilities considered are within three miles of the Calaveras Fault and very close to, or possibly just over, a fault identified as the Verona Fault.

After discussion with a number of persons and a review of reports, documents, and letters from flRC, the U.S. Geological Survey, and the TERA Corporation, studies for Diablo Canyon, and recogni::ing the lack of correlation of damage to structures and equipment in relation to peak accelera-tion (including the 6 August 1979 Coyote Lake earthquake and the 15 October 1979 Imperial Valley earthquake), in the light of our judgment and experience we reco=end the use of the criteria described nerein for the seismic evalua-tion of the site and for the review of structures and equipment in structures It is noted that these recommendations are the writers' sole at the site.

he official views of t!RC or the USGS.

DUPLICATE DOCUMENT rthquake of magnitude 7 to 7.5 might Entire document previously epicenter might be located close to the entered into system under:

ANO k O D (o %. M @

s No. of pages:

(o 9

=-r

_