ML20151Q224

Forwards Info Re Postulated Cascadia Subduction Zone Seismicity & Earthquake Hazards.State of or DOE Determined That Plant Can Withstand Postulated Earthquake W/No Danger to Public

ML20151Q224
Person / Time
Site:	Trojan File:Portland General Electric icon.png
Issue date:	08/01/1988
From:	Cockfield D PORTLAND GENERAL ELECTRIC CO.
To:	NRC OFFICE OF ADMINISTRATION & RESOURCES MANAGEMENT (ARM)
References
NUDOCS 8808110076
Download: ML20151Q224 (56)

Text

{{#Wiki_filter:f. A D Portland General ElectricCompany ~ David W. Cockfield Vice President, Nuclear August 1, 1988 Trojan Nuclear Plant Docket 50-344 License NPF-1 U.S. Nuclear Regulatory Commission Document Control Desk Washington DC 20555

Dear Sir:

Pacific Northwest Seismicity As part of a continuing activity, PGE has b?en monitoring the results of research concerning Pacific Northwest seismicity and earthquake hazards associated with the Cascadis. Subduction Zone. Information presented at a workshop and symposium earlier this year was submitted with our letter of June 15, 1988. As described in that letter, PGE has responded to a number of questions from the Oregon Department of Energy (ODOE) concerning the status of Cascadia subduction Zone research, and seismic capability of the Trojan Nuclear Plant with respect to the possibility of a large magnitude subduction zone osrthquake. Copies of the ODOE questions and PGE responses, submitted to ODOE on June 6, 1988 are provided in the enclosure. Attachments B and C to our June 6 letter to ODOE, consisting of Trojan FSAR Sections 2.5 and 3.7, respectively, are not included in the enclosure since the FSAR is readily available at NRC offices. On June 16, 1988, PCE, the intervenor, and OBOE met with a subcommittee of the Energy Facility Siting Co mcil (EFSC) for the State of Oregon. Infor-mation provided in the enclosed question responses was sumnarized in a presentation by PGE, followed by presentations by the intervener and ODOE. Following that meeting, we received ODOE's conclusion that Trojan can do0I 8808110076 80080! in PDR ADOCK 05000344 P PNV g{ 121 S W Sa:rron S::eet. Powd. Oregon 97204

4-f 4 W M M C Ovi @ rf 1 l Document Control Desk August 1, 1988 Page 2 withstand the largest earthquakes now being postulated without danger to the public. That conclusion sas accepted at a full meeting of the EFSC on June 30 1988. We will keep you informed of significant developments on this topic. Sincerely, / Attachments c: Mr. John B. Martin Regional Administrator, Region V U.S. Nuclear Regulatory Commission Mr. Bill Dixon State of Oregon Department of Energy Mr. R. C. Bare NRC Resident Inspector Trojan Nuclear Plant i i

______----- - ~ - - ~ ~ - - - - - n <3 June 6, 1988 PORTLAND GENERAL ELECTRIC COMPANY (PGE) RESPONSE TO OREGON DEPARTMENT OF ENERGY (ODOE) QUESTIONS CONCERNING THE POSTULATED CASCADIA SUBDUCTION ZONE EARTHOUAKE ODOE Ouestion 1 Describe what is meant by the Qascadia Subduction Zone. PGE Response The Cascadia Subduction Zone is the region of interface where the Juan de Fuca plate is being overridden by the North American plate Discussion The Juan de Fuca plate is a geologically young (8 old) oceanic plate originacing from a mid-ocean spreading zone ofto 10 million years upwelling magma off the coasts of southwestern Canada,(molten rock from beneath the earth's cr located and northwestern California (Figure 1-1). western Washington and Oregon, As newly formed oceanic crustal material (lithosphere) spreads away from the Juan de Fuca Ridge, it gradually cools and thickens. The Juan de Fuca plate is converging with the continental North American plate in an east-northeasterly direction at a rate that been estimated at has convergence rate is unknown.three to four centimeters per year; the exact de Fuca and North American platesThe cor. vergence zone between the Juan begins approximately 100 km (62 miles)(referred to as the trench axis) offshore at the base of the continental slope, and it in north-south orientation. extends approximately 1,000 km (620 miles) In the convergence zone, below and being overridden by the North American platethe Juan de Fuca p known as subduction. in a process The convergence zone extends to the east the Juan de Fuca plate dips below the North American plate at a where shallow angle. Approximately 250 km (155 miles) inland, the downward plate eventually assimilatesangle of subduction increases significantly, and the J (by melting) into the mantle beneath the continental crust of area of the convergence ofthe North American plate (Figure 1-2). The the Juan de Fuca plate with the North American plate, to the assimilation ofand extending from the base of the continental slope the Juan de Fuca plate, is generally referred to as the Cascadia Subduction Zone. Study of subduction zones worldwide has shown that earthquakes typically occur in two places subduction zone (see Figure 1-2):

1) the "plate interface" (ie, the boundary) between the subducting and overriding plates, and 2) the "intra-slab" region composed of mater-ial within the subducting plate.

The distinction between these _ - - _ _ - _ _ - _. -

( <4 possiblo subduction zone earthquake sources is important to the -Cascadia Subduction Zone discussion provided in the response to Question 3. f - 2- -~

t CASCADIA SUBDUCTION ZONE I %OuttH & CHARLOTTE Queen LS.l 3 g M te I Wilson !sult " Knolls } \\q BRITISH Brooks 1 A COLUMBIA ,,,N A l Ekpluer \\ %% c.,,.., a ,e CQ Flere,A e, samey /g \\ -.g-..- - - - NNtka t ' Victoria \\ s Fault ~ ~ " \\ f* \\ WASHINGTON -NORTH \\ g [ AMERICAN ,y l / PLATE \\ OO 4 Juan de f-Fuca Plate _ 1 O O ass l A k s l i t %.4 I ^8 I l CREGON I TRENCH AXIS. g O BEGINNING g OF PLATE f, g N,'lg INTERACTIONS g

---

g -- !

0 \\ cores &o \\ a g south 1 l Plate \\ ) } l Csee g P5CIFIC ~

• * '? * "-

t'0' lPLATE CALIFORNIA j i. \\ San Andress " Foull \\, Nn gancisco e , (Adapted from Taber & Smith, 1983) Figure 1-1 ... - - -.... -. _ -..... - -, ~ -.. -.. - -. _ -....

1 REPRESENTATIVE SUllDUCTION ZONE MODEL SPREADING ZONE OCEAN RIDGE OCEAN TRENCH Q :.. OCEAN COAST CONTINENT f -.......\\... + +\\ W TH i ~ P> arse rtare. Ry['"A,'.\\ My" +- + -+ . ~[..~~. ..A =a - + -+ + -- * ), j ~~ f' LOW-VELOCITY LAYER y I8 N MAGMA PLATE N \\ j INTERFACE h & I ) }4}ll RISING MAGMA 1N INTRA-SLAB, SUBDUCTION N ZONE V i Figure 1-2

i '4 ODOE Ouestion 2 How does the Cascadia Subduction Zone compare to other subduction zones around the world? PGE Response In general, the Cascadia Subduction Zone is unlike most duction zones. other sub-oceanic plates in the world andThe Juan de Fuca plate is one of the smallest is one of the youngest known subduc-ting slabs. ison with other subduction zones.The rates of convergence are also very low in com In addition, of earthquakes observed during recorded history (the lastthe number and size has been low compared to other zones. 150 years) Despite these differences, between the postulated characteristics ofresearchers have made several comparisons Zone and those of other subduction zones aroundthe Cascadia Subduction the world. Early stages of research on the subject have provided a variety of viewpoints. Discussion the oceanic crustal materialsIn a comparative analysis of subduction zones con in the subducting plate, the rate of convergence between the plates, and historic earthquakes associated with the subduction zones, it has been proposed that the Cascadia Subduction Zone may have the potential for generation of large earthquakes (Heaton and Kanamori, 1984). Subsequent studies have proposed that tion zones in southern Chile,the Cascadia Subduction Zone may be similar to subduc southwestern Japan and Columbia which have strong seismic coupling (which can result in large earthquake events) and have generated large historic earthquakes

Hartzell, 1986).

(Heaton and Another researcher (Davis, 1988) concludes that comparison with the southwestern Japan, southern Chile, southwestern Columbia, Rivera and northern Cocos (in Mexico) Cascadia Subduction Zone may be very dissimilarsubduction zones suggests that the in seismogenic potential. The Juan de Fuca plate is characterized as a much ) younger topographically smooth segment of octanic lithosphere. The postulated convergence rate of 3 to 4 relatively slower than that of most centimeters per year, which is an anomalously large accumulation of other subduction zones, and that { temperature and high fluid pressure occurs attrench sediment with high the Cascadia Subduc-tion Zone convergence boundary. (such sediment is also present in the southern Chile Subduction Zone, but These sediments have been postulated the volume is much smaller.) to provide a lubricating medium to allow the converging plates to interact without large scale buildup of plate interface resistance. This may provide for a region of weak seismic coupling, ie, occurring mainly in the form of aseismicinterplate motion could then be (not subject to earth-quakes) creep. If the interplate coupling is weak, a seismogenic __

1 .s interface betwson tha Jucn de Fuca and North American plates may be very narrow and incapable of generating earthquakes of the magni-tudes exhibited by other subduction zones. At this time, researchers have not reached consensus on what the 1 subduction zone comparisons mean with respect generation potential of the cascadia Subduction Zone.to the earthquake j 1 ) l 6- ... -.. - ~ _.

ODOE Outstion 3 (s Describe the recorded history of significant and Washington. earthquakes in Oregon associated with the Cascadia subduction Zone. Indicate the earthquakes that PGE Response The recorded history of earthquakes in the Pacific Northwest that most events have occurred offshore. The significant earth-shows quakes that have been recorded from onshore events were generally centered in the Puget Sound area. Deep earthquakes occurred within the subducting slab, not at plate interface. the Juan de Fuca plate-North American quake has been postulated (The very large Cascadia Subduction Zone earth-to occur at the interface.) The largest earthquakes associated with the Cascadia Subduction Zone are those that occurred in 1949 (MB 7.1*) in Punet Sound and in 1965 (MB 6.5) in the Seattle-Tacoma area. Figure 3-1 shows the pattern of earthquake epicenters in the Pacific Northwest. Discussion The majority of the earthquakes in the Pacific Northwest have occur-red offshore and are associated with geologic structural features other than the Cascadia Subduction Zone. These features include / oceanic transform faults such as the Blanco and Mendocino fracture zones, the Queen Charlotte Fault, the San Andreas Fault, internal deformation within the Explorer and Gorda plates, and scattered shallow crustal seismicity. The pattern of onshore earthquakes shown in Figure 3-2 varies significantly in both north-south orientation and in depth. majority of earthquake epicenters are located The in the general area of Puget Sound. Proceeding south from Puget Sound, the number and size of recorded earthquakes diminishes; infrequent earthquakes that have been located in this southern region have recorded or estimated magnitudes in the low to moderate range. Cross-sections through the Puget Sound region (Figures 3-3 show that and 3-4) the earthquakes in this region are occurring in two distinct groups: (1) Shallow crustal earthquakes within the North American plate appear in the projection at depths of approximately 30 km or less. Body Wave Magnitude (M ) is approximately the same as Richter B Magnitude. Refer to the Glossary of Terms, Attachment A. ,7 - 1

(2) A socond rcgion of earthquakes occurs within the subducted Juan do Fuca plate at depths of approximately 30 km or more. t The deeper set of earthquakes are "intra-slab" events and can be referred to as Cascadia Subduction Zone events. The more shallow set of earthquakes appears to be related to shallow faults in the Puget Sound region. The relationship of these more shallow events to the subduction process *< not clear. The largest earthquakes associated with the Cascadia Sub(action Zone are those that occurred in 1949 (MB 7.1) in Puget Sound end in 1965 (11B 6.5) in the Seattle-Tacoma area. Modeling of these events has confirmed that they occurred within the Juan de Fuca slab and not at the plate interface. No recorded earthquakes as low as magnitudes

• of 3 (on the Richter Scale) are known to have occurred on the plate interface.

South of the Puget Sound region in northern and central Oregon, the shallow crustal earthquake activity decreases dramatically. In addition, recorded earthquakes within the subducted plate, which is estimated to be deeper in this region (Crosson-Owens, 1987; Wea ve r-Ba ke r, 1988), are also very sparse. It has been proposed that the greiter seismicity of the Puget Sound area, as compared with the Willamette Valley region to the south, is due to arching of the subducted Juan de Fuca plate beneath western Washington. The subducted plate bcneath the Willamette Valley is more planar in its configuration and much steeper in its eastward dip.

• Refer to the Glossary of Terms. Attachment A.

l l ) SFISMICITf AND PLATE TECTONICS OF TH:' PACIFIC NORTHWEST s .e, g u, m. 1% t ~) c.a s a... t. o rf. /;.ma s o 4. ) .s 3 e g. ,) \\ ..L.a...- 4 h. 3 o,p;. \\. t e ,s 1 /. \\ m. .a T,~l..~ s%, 'y woe.ngton

• ., y" :
• g 9

.... y gs.. A g ,,,,.,e ,u,. o u 5,.,,. V. g 1 \\ *..

• I.. t =:i..

i.. 8 . s. ,f Ij./ Ihs.... ...,, ' _ T fQ j l ./ \\ 1 .b \\ p ~ s...,ves -l \\ .u v a n,,;on + j 1,%.. e, ,. f w ,,...e-l 4 { c. .g ,s.. u..- t 1 / g:::. I.

A

,_. x.* :., :...,,.. j t 1 e' j.,; . e i \\ e se e t l c ,t e \\ .s. e ~ 3 4 l 6 s g. .,._",...w ..aaa a'=. I [,a,,,n.a! \\. n-m \\ r1,. es/... ).:. 4,s-s K e.. l s s l s M3rd _'e ~ .d L .e. --~.--.< y... 7,,.=p v::: : f j y-t. .;e.. ..,.., '...,r .,y. 1 '" :. e *. i "~ N y,;;*, 6,a, ' M :.. e c. v '(Q,..4... y..,. ~- S 2 u wa.~= = n... .n e a o - e o we,s.ium

44. o,%,,,n,,,,,,,,,,,,.,,

. so-seu.r.m 4..... w., coa. w so u -ao war'wa 1 9 1 1 0 300 g. -from Heaton-Hartzell. 1988 Figure 3-1

t ~ ~ ~c m mse,s.a.en.>.. M-m m w u ~)(m.< ion m m im i i Wh"'1=ri,% \\! E e g g N i + + y \\ 0, ,oo o s N j g *., s o \\ tb $% e L s e wa Y 4.W, a.AT8 8 = e ix x w M.~:.W%%"% % y or a' /, W. 1% ~efd2t* bf d e o or . NN%@Mr( \\,

• j$c m.

I N.* e. 8 M Mpeh .e +* ( o c.1 3

j e'a.a g a6 1 e* M

~ = v \\. op;o i.[M8eD N6% " e-c

.1 I** jo*' "&,$ $

_i fyioDW.t ic487 *,o y', ur j ,[ [.'U*Ik MN

• eo

.r ..MNTv Y n, ne v L W.W s'/ +/- f*.* oowo o n_ / O o. o t / a g '.5

g. to

• /

/

• /

/ s, g / / a. e I % o 'p' o o .o \\ l \\ ,1 .il / /. l 1

• [

[ fE k 5 i 3 Depth (in km): Io 40 50 so Trojan Nuclear Plant W scar o . o O,. o -,, sca.! tit 7ecoco T88888 O " ' O = "" uxs cr s g,, o m un iMm ul.!'.".,'..* m*m netu.. ut :1? *a . o ai. a= = , y y a g,,,,, 2 =.i. i u. an .i. m - w O " *

• o M ien O

"=e 3 o M tem

==s t. m um -from , i, m i,3 FIGURE 3-2 Geomatrix Consultant 3, Inc. -10

u.am w m.emum. mrA ciem-i ) o era a iem ara) ac omms ac=f wi I = = =r o e a e i t,d 8 _' t i W 4f f ? f 1

• 1 ? I I f f f f f 1 i f f f f f f f f f f f f f f f f f f f f f i f f f ' f f a f fi'A f) d f f f f f I I g

g e i I# ,+g. , J +4'b *t En ,j Y f -o-e !=3 7 p,'r-o i _.9 +-

a 5:

4 a: i in i i i... i i s i i iii ii ii i ii i i ii ii i i ii i ii i i ii ii i i i i i i i ii iii iiiiii i i i i s ti s 8 5 i i i i-a n 5 g V E I E E VI I UNM 7 WA N57RuuDng. DATA (tSW1986) & (DQP PRE 1970 QATA) ,g casnew e aa ese g -c a e G B e i G !is a t-it l o f 1 ? ! 11 f f f f f f f f f f f f f f f f f f I f t f f t e q f f t i l 1 f 1 f t 11) 111 L11 ) a -+- tiff Ifffff g y .d v e. ~ 9 I n: .t I + 1

n

+, \\ != : =+ i : .@..@+ _m -n \\ 2 ac ...iiiiiiI..IiaiieiiiI6in6iiiii Ii4iaiiIi tis'iiiiiieIiiIiiiiii11 ? i n n u 9 9 5 5 5 0 El WI'UC"f W %A tp.NDCA. SATA (19W1984) & (DGP PRE 1970 QA!A) y cas see os e m a es u.2 es a C8 T r. g 8 d 8 1 E t e i i t ffIfffffffffIffff1Iqtttttfe 111111111tfif1f ?fffffttffffffff g e U Iqp!d) , S." g

• yS f

n g.i

n do 1

Y + o 7 +. '. A 6 !";E t W4t k4 .+ E" = n-

n.

IR I .i ...i:iiiiiiiiisasiiiiiiiiiiiiiiiiiiiiiiiiiii I iii.,. i... .. i{10C E I E 5 5 5 p I I I I I I II - f r n :1 FIGURE 3-3 Germatrix Consultants, Inc.

~ Ut4R57Y CF WA DNM OMA (19/O=19 6) Q (DLEP Pea tw/W ueAs E auss scenaw serw c aw estAe4 CC $$d 1 NY d N I d g f -**tiIt!tIttItttttttiIeeitItiItittttttij bt t t t t t_i ) t_1 i t 1 i t e t t t t t t t t t i i n d "..Wg 9. /~ ui II.'9t 9 %,* d.. in i a w w t* 4 e..f 9-'4; L e. t,"a f +. E. t 5 9 71Q j ) .' n. j t=, M ~ i .i.4n.: 4 4 4. 18 6 4 i l 4 4 i s 6 i s n t L Il s t i l l -tm i 16 line16 sillisissalegis44 sus # h h h 4i4s h h lht.CtSTY & WA N$TRVIJCNTAL CATA (1970-1986) & (CEIP PftC 1970 DATA) U *I oms ums wem ao a j m,, a E87C' .s A E 2 a t i t t i l i t t t f it f I f f r f f f Lf) t i l l f l t i l l f i t I t Lf r ! L! ! f t i l f r t t I t f f f r f f f 9 T 'T 'f g ~ p c. 6 67 b u: r5 A.,vg+I*ih*+.._ o

a 3 :

i 8 i :3 s n + i e +4 ~m s t-g :- cas a secox. emmr L Ft31S.1644 4317 te47 *S4 Mt1SLS espe 71 trei q SAS H W ~

== 1 AM 28.9943 thaa 8647.00 wi21400 onom 47 tre seg SJam hat W& AM 1116401943 es47.tr et 21930 esem _4-

7t q

m = = ..eiaiiis444&le444444444lesa444464tili6itil4Il414tillill44ilillil444ilai =gg 41 h h h N pe, cts.m y wr ms;RuutNTA DATA (1970-1954) & (DCIP PftE 1970 CATA) f*1 OtMS SE "De UCSt 30 lod wm pa m ION e a e e t a B G \\ a g 'fIe!vfffIItttttf1eIItttfiti 11111111 IIIfftit1Ltttt iliIJtf11ffelf11tt 4 H Y O

• 3

+ I .t..l_ h. 6 *

• 4

} i 3 jd'-3

• h + %

E" q tyxs or sPccw. smarr , t to titles ca $7 i.anu m:le:S espe 71== g taa H w

== 1 AMt af.tyt3 f t.ta l*47.400 m21400 esem $7 Ir# atet M Mt W.= 1 am 1A19e41914 ees? 130 sim espes 66664IIIi6is4ile4ill4illi3le44( l46tilLliinlisi41: 434il3I1434 14456I 100. -from FIGURE 3-Geoma*.rix Consultants. Inc.

i OpOE Ouestion 4 What are currently proposed r I tectonic models for the Cascadia Subduc-tion Zone? What with these models?is the postulated earthquake potential associated t PGE Response Cascadia Subduction Zone research has resulted in models that from a postulated completely ' locked' range seismic potential of unspecified plate interface with large recurrence intervals (repeat times), to speculation that the plate interface is ' unlocked' which results in continuous, complex theories propose thatessentially aseismic, movement. Other more mented across its north-south extentthe Cascadia Subduction Zone is seg-modes of deformation in terms of configuration, dip and rate ofand may havei convergence, and hence varying seismogenic potential. Cascadia Subduction Zone could rupture in a single eventS re the Chilean earthquake of 1960 which was the largest similar to recorded and had a seismic moment magnitude

• of Mw 9.5earthquake ever Hartzell, 1987).

There are different (Heaton and size of earthquakes (ie, Richter *, Seismic Moment Magnitude, Modi-scales used to measure the fled Mercali", etc). Theso various scales cannot i lated to each othor. magnitude of 9.5. This Mw value is not be exactly corre-equal to a Richter At present fundamental question of Cascadia Subduction Zone seismic A large body of evidence from the geoscientific record subject to consistent interpretation, will be needed to approach a , which is consensus of expert opinion. Discussion Figure 4-1 illustrates the prominent features 'of the oceanic plate l subduction process and the region of interest for discussions concerning the seismogenic potential for large thrust earthquakes. This region, referred to as the seismogenic interface, is the source of thrust earthquakes observed at subduction zones. The seismogenic interface begins in a the weaker accreted materials and the more competentregion of transition between continental crust in the convergence zone, (very strong) reg.on where the slab becomes more ductile (ie,and extends downward to a potential for earthquake occurrence relates to the width ofless brittle). The seismogenic interface, the relative movement the rate and character (gradual or abrupt) of between the plates along the interface, strength of the interface rock formations. and the Refer to the Glossary of Terms, Attachment A..

Invostigators in many areas of the geosciences are continuing their assessment of the seismogenic potential of the Cascadia Subduction Zone. Much of the geologic and seismologic data needed to further develop and calibrate various tectonic models of the Cascadia Sub-duction Zone remain to be obtained and consistently interpreted. consensus of expert A Subduction Zone is thus not yetopinion on a seismogenic model for the Cascadia available and will require many additional studies. This response, therefcre, does not address all of the Cascadia Subduction Zone tectonic models presently under development, but will instead characterize the range of models have been published. that The locked plate theory contends that the interface between the subducting Juan de Fuca plate and the overriding North American plate is strongly coupled, such that plates is resisted. relative motion between the elastic strain energy across the interface.This resistance leads to the accumulation Eventual rupture of the interface plate may result in abrupt energy release in the form of large earthquakes. 1 As previously noted, Subduction Zone with other subduction zones aroundsome researchers have compare the world, particularly those in southern Chile, southwestern Japan, and Columbia, which have exhibited strong seismic coupling and historic earthquakes. large Parameters such as the age of the subducting oceanic plates at the base of the continental slope (trench), the rate of convergence between the underriding oceanic plates and overriding continental plates, the and the magnitudes of historic earth-3 quakes associated with these other subduction zones were selected for comparisons (Heaton and Kanamori, 1984). Based on the age of the Juan de Fuca plate at the interface of the Cascadia Subduction Zone, estimated to be 8 to 10 million years, and a postulated rate of convergence of 3 to 4 centimeters per year, the earthquske potential of the Cascadia subduction Zone has been char-acterized (Heaton-Xanamori, 1984) to be on the order of seismic moment magnitude Mw 8 to Mw 8.5. Figure 4-2 indicates a rela-tionship between maximum energy magnitude Mw, convergence rate, and age of the subducted lithosphere. Speculation has included the possibility that Cascadia Subduction Zone could rupture in a single event, essentially the er a the Chilean earthquake of 1960 which was the largest earthquake ever simile co recorded and had a seismic moment magnitude of Mw 9.5 Heaton and Hartzell, 1987). Simulated ground motions were reopose(d that occur during large subduction zone thrust could earth.tukes. This t approach was required were available for earthquakes with seismic moment magnitudesbecause greater than 7.5. However, special very large Chilean earthquake (a convergence rate 3 geologic conditions favoring the times that of 14

Juan do Fuca/ North America, a very narrow accretionary wedge

• which equates to a very wide seismogenic plate interface) have not been observed in the Pacific Northwest.

Variations in convergence rate along the Cascadia Subduction Zone (slower in Oregon), dip of the subducted Juan de Fuca plate (steeper in Oregon), sible reasons why the entire Cascadia Zoneand thickness of sediments being sub (even if seismogenic) would not rupture as a single segment. The physical division of the subducted plate into a central arched slab beneath the Puget Sound area and neighboring, more planar slabs to the north and south, provides an additional basis for segmenting the Cascadia Subduction Zone in terms of its seismogenic behavior. All Pacific subduction zones show such segmentation and no single plate boundary such as the Juan de Fuca/ North American boundary has ever been ruptured in a great earthquake along its entire length. Despite its great length of rupture (approximately 1000 km), the 1960 Chilean earthquake broke only about one-fifth of the plate boundary along which it occurred. In contrast to the locked plate hypothesis. other researchers (Davis, 1988; Byrne, et al. 1987; Sykes, et al, 1987) have suggested that the Cascadia Subduction Zone is unlike the known seismogenic subduction zones to which it has been compared. The Juan de Fuca plate is e celatively young oceanic plate with a slower (relative to other subduction zones) convergence rate. Also, there is a large accumulation of trench sediment at the Cascadia Subduction Zone convergence boundary. These sediments are up to 2 km thick with temperatures estimated to be 200*C at their base, and with high internal fluid pressure; they are believed to enhance stable sliding behavior such that plate convergence can occur without large accumu-lation of strain across the interface. In this model, tude energy releases in the form of earthquakes should notlarge magni-be expected, and the model indicates that generalized comparisons of the Cascadia Subduction Zone to other subduction zones are not valid. For example, comparison has been made to the subduction zone in southern Chile, which was the location of a magnitude My 9.5 earthquake in 1960. However, several differences are apparent: The 1000-km-long rupture initiated in older oceanic crust (30-35 million years as compared tc 8-10 million years for the Juan de Fuca plate). The plate convergence rates for southern Chile are much higher (9 cm/ year vs 3 cm/ year). Despite the fact that the southern Chile zone has a sediment-filled trench, the Cenozoic accretionary prism is very narrow. Crystalline rocks crop out on the sea floor within 100 km of the trench, suggesting subduction erosion, a very close-in back-stop, and a very wide seismogenic interface.

• Refer to the "Glossary of Terms", Attachment A.

15

i Other investigators have suggested that stress buildup in the Cas-cadia Subduction Zone is very complex, and is dominated by the interaction of the Juan de Fuca plate with neighboring plates as [ well as extensional "slab-pull" stresses within the plate itself. 1 Spence (1988) for example, believes that the Juan de Fuca plate provides significantthe youth and buoyancy of resistance to subduction. particularly at north-south compressive stresses. parts of the subduction zone perpendicular to the { This model implies that various parts of the subduction zone may have different seismogenic potential, i l Several studies have shown that subduction has slowed considerably (aboutthe rate of Juan de Fuca plate 60%) during the period from 6.5 million to about 500,000 years ago. There is no direct data on the convergence rate over the past i 500.000 years, but the The Gorda and Explorer plate remnants may have ceas altogether. The Juan de Fuca plate subduction beneath Oregon is considered have greatly slowed and seismicity in this region is very sparseto Analysis shows that in areas to the north.the rate of convergence is lower in Oregon than It has been speculated that the convergence zone resistance to continued subduction and progressive weakening of the subducted slab extensional forces may be evidence of long-term cessation of Juan de Fuca plate subduction. Offshore Holocene deformation confirms that occurring across the Cascadia zone. convergence is still However, if the rate of conver-gence is very slow, occurrence should be very low,then the likelihood or frequency of earthquake particularly in the southern Cascadia Subduction Zone off the coast of Oregon. i 15 -

TYPICAL SUBDUCTION ZONE \\ l t Volcante Arc Contf rU{Mntal UIsa t,,,, s N -880W% N\\\\ N s lllfj- , ACCRETIONARY \\ \\ \\ WEDGE g\\ g8AliRth41 Crvst ') \\ \\ \\s\\ BACKSTOP \\ sg, Oceanft Crust h/c _ BRITTLE DUC M ] 1 \\ 1 ) i (PGE, itBB) FIGURE 4-1

.g I I l h8.5 l l C. Chile 9.5 82 o S. Chile 10 NE. Japon 8.5 9.O 8.2 g pe, o_ g o o T 8 7.1 C. Arnerico 8.8 a5 Jovo Aleu'tions O 7 h* 38b y @w \\7.2

• Ker dec b"*. 9 o

z^ Izu-Bonin 8 o z o 6 8o 9.2 I $0 7.8 RyuEyus MU ko ya~

• New Zealand gg g

y. cu 7.2 -S o Morianas 8.o O ll' Z g M -{ y 8 O' 4 -* 8.6 5 7,9 SW. Jopan o [a[ G 7.5 3 New Hebrides O o c Juan de Fuco 5EE 2 Caribbean ,o s;a, aS" Es! I l\\ l a=a iso u i I I i =ae 14 0 12 0 10 0 80 so 40 20 o " US Age m.y. ( AGE OF PLATE) 1 -from llea r on-Ka namor t. 1984 FIGURE 4-2

ODOE Ouestion 5 What evidence in there from the geologic record that has been inter-preted to support any of these models, and what is the current status of the scientific inquiry? PGE Response Recent geologic investigations have identified evidence for recurrent subsidence of coastal estuary deposits (tidewater marsh burial events that appear to have recurred at irregular intervals) in southwestern Washington and northwestern Ore'gon. "jerky" subsidence has been cited as evidence thatThis evidence of abrupt or large earthquakes may have occurred several times in the Cascadia Subduction Zone dur-ing the past several thousand years (Atwater, Peterson, Grant 1987-1988). Sequences of intertidal bay mud layers overlaying layers of peat-like material that range in depth from 0.5 to 2.0 meters, with relatively sharp interface features, possibility of rapid subsidence sugesst the events. Radiocarbon dating of these sequences has been interpreted to indicate that submergence events occurred in the vicinity of 300, 1,000, 1,500, 1,700, 2,500, 2,800, and 3,500 years ago. Sudden downdrops creating these burial sequences are speculated to have been generated by large, prehistoric Cascadia Subduction-Zone i earthquakes. The presence of sand layers overlying some of the buried low land formations in these sequences has been interpreted as evidence that tsunamis may have resulted from the same events that caused the subsidences. These subsidences may be evidence of paleoseismic earthquake occur-i rences, but they remain to be spatially correlated along the coast and event-correlated from site-to-site. It remains to be demonstra- { ted that these data are consistent with other evidence that be exhibited in the geologic record should Zone earthquakes have occurred. if large Cascadia Subduction For example, additional evidenc6 that would be expected to be exhibited in the geologic record includes evidence for paleoliquifaction such as event-correlated ~ vented sands underlying exposed subsided wetlands, subsided or uplifted wavecut benches along rocky headlands, sea floor and land-ward uplifts, and landslides for which rainfall and non-Cascadia Subduction Zone earthquakes can be excluded as probable causes. Until such observations are made, the evidence for subsidence is not convincing evidence for large-magnitude Cascadia Subduction Zone events. Research efforts in this area are continuing. See the response to Question 6. 2 19

ODOE Ouestion 6 What in progress or planned by the scientific community? additional res I PCR_ Response Much of the scientific inquiry concerning the Cascadia Subduction Zone is being funded as part of a five-year program (which began in 19P7) by the Ur.ited States Geological Survey (USGS) National Earthquake Hazards Reduction Program (NEHRP).through the program is a regional earthquake hazards assessment Part of that for the Pacific Northwest. Their draft work plan for fiscal years 1987 to 1989 intended to defina research guidelines and to identify responsibil - is ities for further activity. Beginning with the recently concluded April 1986 workshop, a series of five annual workshops (1988 through 1992) are to be held in the Puget Sound area to review accomplish-ments and program directions, and to foster knowledge-base develop-ment and application. Proceedings from the annual workshops will be published as USGS Open-File reports. Typically, projects under the NEHRP Program are funded annually and additional investigators may be involved in the future. As described in the response to Question 11, PGE is monitoring the d evelo pme rit t from, and results of, this program. On-going research presented at the 1988 workshop included the fellowing: Affiliation Research Description Dr. Gary Rogers Geological Survey Pacific Northwest of Canada Earthquake Potential Dr. Brian Atwater USGS Coastal evidence of large prehistoric Pacific Northwest earthquakes. ) Al Rogers USGS i Factors affecting i ground motion in Pacific Northwest i earthquakes. Ken King USGS Puget Sound and Portland urban areas earthquake hazards assessment program. Tom Urban USGS Investigation of under-water evidence of geo-logic faulting in Puget Sound. _..

Affiliation Research Description Dr. Curt Peterson Oregon State University Investigation of subsi-and Alan Nelson USGS ded tidewater marshes on the coast of Oregon. Dr. William Spence USGS, National Earthquake Continuing investigation Information Center of the seismicity and tectonics of the Cascadia Plate System. Robert S. Crosson University of Washington Investigation of sub-coastal earthquakes in the Puget Sound region. S. T. Algermissen USGS Estimating anticipated ground motion due to earthquakes in the Pacific Northwest. Dr. John Tinsley USGS Investigation of geo-logic factors that influence propagation of earth motion. quake ground Dr. Robert Schuster USGS Investigation of earth-quake-induced ground f ailures 11. Western Washington. Margaret Hopper USGS Investigation of the effects of past earth-quakes in the Puget Sound area. Tom Heaton USGS Investigation of earth-quake hazards on the Cascadia Subduction Zone. A number of ongoing research activities were reported at a symposium held at ium was coordinated by Dr. Brian Atwaterthe University of Washington 1988. The r.ympos. coastal evidence for the past occurrence o(USGS), whose report on Dr. Atwater's research. earthquakes was published in Science Magazine in M shorelines on the Washington coast, investigating subsided marshes and uplifted is continuing. In addition, coastal evidence of past Gary Carver (Humboldt State University), earthquake events is being investigated by Harvey Kelsey (Weatern Washington University), and Mary Reinhart i (University of hashington). l 4 - 21

Lotor this year. Atwater and Nelson will be investigating the geo-logic record of large magnitude earthquakes in southern Chile. Ongoing research reported at the May symposium included the follow-ing activities, primarily involving review of geologic and other physical records for evidence supporting the theory that

large, possibly subduction zone, earthquakes may have occurred in the Pacific Northwest in the geologic past, most recently about 300 and 17 0 '.) years ago.

Affiliation - Research Description Alan G. Hull University of California, Radiocarbon dating of Santa Barbara buried soil layers on the Washington coast. David K. Yamaguchi University of Colorado Identification, by dendrochronology, of the time when cedar trees died, possibly as a result of salt water intrusion when the land elevation changed rela-tive to sea level. Wendy C. Grant USGS Investigation of tidal marshes for evidence of i earthquake-related sub-sidence along the Salmon and Nehalem Rivers in northern

oregon, Michael Lisowski USGS i

Investigation of geo. logic strain accumula-tion in western Washington and south-t western British Columbia. Paul Vincent University of Oregon Resea,ching elevation surveys for evidence of uplift and subsidence along the Oregon coast. Albert A. Eggers University of Puget Sound Investigation of defor-mation of Pleistocene sediments in the Tacoma Narrows, Washington area. 4 __

ODOE Ouestion ? What is the seismic design basis for the Trojan Plant? PGE Response The Trojan Nuclear Plant. seismic design basis is represented by the design response spectra provided as Figures 7-1 and 7-2. These spectra were established in accordance with Title 10 of the Code of Federal Regulations Part 100, Appendix A, "Seismic and Geologic Siting Criteria for Nuclear Power Plants". These spectra, which establish the ground motions to be withstood as a function of period and damping ratio peak ground acceleration represent the design basis for Trojan, not the single-point peak ground acceleration levels of the Safe Shutdown Earthquake (SSE) and Operating Basis Earthquake (OBE). Discussion Investigations performed pursuant to establish Trojan's seismic design basisto 10 CFR 100 Appendix A criteria included: (1) Determination of the lithologic, stratigraphic, hydrologic, and structural geologic conditions of the site and the region surrounding the site, including its geologic history. (2) Identification and evaluation of tectonic structures underlying the site and the region surrounding the site, the surface. whether buried or expressed at (3) Evaluation of physical evidence concerning the behavior during prior earthquakes of the surficial geologic materials and the substrata underlying the site from the lithologic, strati-graphic, and structural geologic studies. (4) Determination of the static and dynamic engineering properties of the materials underlying the site. (5) Listing of all historically reported earthquakes which have affected or which could reasonably be expected to have affected the site, including the date of occurrence and the following measured or estimated data: magnitude or highest and a plot of the epicenter or location of highest iniensity,* intensity. (6) Correlation of epicenters or locations of highest intensity of historically reported earthquakes, where possible, with tectonic structures, any part of which is located within 200 miles of the site. Epicenters or locationc of highest intensity which be reasonably correlated with tectonic structures were cannot identified with tectonic provinces, to be any part of which is located within 200 miles of the site.

• Refer to the Glossary of Terms, Attachment A.

- 23 i

1 (7) For faults, any part of which is within 200 miles of the site and which may be of significance in establishing the Safe Shut-( down Earthquake (SSE), determination of whether these faults are to be considered as capable faults. A "capable fault" is a fault which has exhibited one or more of the following characteristics: Movement at or near the ground surface at least once within the past 35,000 years, or movement of a recurring nature within the past 500,000 years. . Macro-seismicity instrumentally determined with records of sufficient precision to demonstrate a direct relationship with a fault. . A structural relationship to a capable fault according to the above characteristics such that movement on one could be reasonably expected to be accompanied by movement on the other. Desailed information developed from these siting studies is presented in Sections 2.5 and 3.7 of the Trojan Final Safety Analysis Report (FSAR), provided as Attachments B and C, respectively. In summary, the Trojan Nuclear Plant is located in an area that experiences moderate to low seismic activity, and safety-related structures are located on bedrock of ancient volcanic origin (see the response to j Question 12). Significant historic earthquake epicenters and their intensities within 150 miles of the site are shown in Figure 7-3. Records show that the maximum intensity reported at Rainier, Oregon, four miles north of i intensity occurred on alluvial deposits,the site, was Modified Mercali Vill. Since this it the Trojan site the intensity for this same event did notis probable that on rock at 1 V11. eration of approximately 0.12 g.Ir. tensity V11 correlates with a peak horizo exceed The Safe Shutdown Earthquake (SSE), also referred to as the Design Basis Earthquake, is based on an evaluation of the maximum earth-quake potential considering regional and local geology and seis-mology and specific characteristics of local subsurface material. The Operating Basis Earthquake (OBE) is that earthquake which, considering the regional and local geology and seismology and 1 specific characteristics of local subsurface material, could reason-ably be expected to affect the Plant of the Plant. site during the operating life For the Safe Shutdown Earthquake (SSE), a conservative intensity of Modified Mercali Vill was selected, since it is probable that 1 intensity of Vill has never been experienced at the Trojan sitean during recorded history. horizontal ground acceleration of 0.25 g.An intensity Vill corresponds to a peak Therefore, a horizontal acceleration of 0.25 g was used for the design of the Trojan Nuclear Plant for the SSE. Although intensity Vll correlates with a peak horizontal ground acceleration of 0.12 g, a more conservative value of 0.15 g was used for the Trojan OBE. - 24

D0Volop2cnt of tho soisaic design basis for Trojan was reviewed by the United States Geological Survey (USGS), serving as a consultant to the Nuclear Regttlatory Commission (NRC). Their findings were l, provided to the Advisory Committee for Reactor Safeguards (ACHS), an advisory group for the NRC Commissioners composed of recognized experts in various engineering and technical fields. The ACRS con-cluded that the 0.15 g OBE and 0.25 g SSE are adequate, to be used in conjunction with conserva-tively derived response spectra. The Trojan seismic design response spectra are shown in Figures 7-1 and 7-2 for the OBE and SSE, respectively. Conservatism applied in developing the ground accelerations for the OBE and the SSE and in developing the seismic design response spectra provides significant design margin. These spectra are developed for close-in, medium range and distant earthquakes, and thus are "broad band" to account for a wide range of ground motions. These spectra, which establish the ground motions to be withstood as a function of period and damping ratio, represent the design basis for Trojan, not the single-point peak ground acceleration levels of the SSE and OBE. i

O '8 DESIGN RESPONSE SPECTRA OPERATING BASIS EARTHQUAKE c0 s / A-A n t w w - v s h/v a n a va w omo v/evana vem v -/rJ//X /t A X NNYs / \\ /f/ /Y/iA XKXW / 's /y'// Y /', Aj X fx T A /l 80 jfgj g fgg yyw gfyj g fx A x ygg yfffg fx g y ygu,% t 60 't <# O{P/\\ M 'NA{ A /\\ /# \\f\\ VN i S e N / \\ X N:XWRV/Y X/\\ X XW2W/Y X A\\ X XWa 60 ..# NA /VNX$h /N/ /V\\X3 PERCENT CRITICAL.N7/h ggg y\\X gg,ggmNG 4oIM_ _My c MM NO@MNMib,\\ g,y ' 2@ xx>swxxxvmeA8mx>NoR e w m v J,. Y %MNNW4QM N v' N N k ,N N' A x 'N / x' NN' 8,x / N N yfy, j y n g gg7w t gf,y,fgfs j7g fy _ uf, f A fx g y sg s Y U 6 AWWW!W Y\\W!N! ! W y bWAYLX/\\ LXNH2MffX f\\h XDDWAYLX /\\ i& ~,, N !$$>(/VN /vNYRF//Yb/V/vW65R</K/N 'hs 4 5 W 'XY '>$8W8A f,4X M8i$@W X>QM6'i (, MEIk(MN/MM[ex&xxa:mw, DYMXMk@ a wxmwavran $la = WlW)x m m w xl m} W =,v >M WW uxm 2 ,. xovi x n x nww. u,v7 x /x x wmp ^ N ^ '6 FW/x'tX /\\ 'M//WXXX//Y!X /\\ XM6 4cY' tX AN iXiVON E E$$(/h/N MNX@M/v^ { f (G ' W %'N/NXNdKf( 4 c k r, 'e h<mxxaam'mxxw-w 2 ,e \\ .ut .02 04 .06.08 .1 ,2 4 6 .8 L 2 4 6 8 80 PE Rl00 sec 3 4 -f rot: Trojan FSAR, Figure 3.7-1 FIGURE 7-1 .~

DESIGN RSSPONSE SPECTRA SAFE SHUTDOWN EARTHQU \\ l l @[<Y/ xYsN1y$$( ?\\ Y// YYx ^/( ao $(Y\\N A/Y/$y?x(,h* m ^ m-svm x O{ / P' W//> N A A t'>N PERCENT CRITICAL ^ mv 60 e"%l& / '\\ ^ 'XX'WRV/X X/\\X >N3p'g' D A M PIN,G, 7/\\ XN D 60 g 40 TA$88K.hX hs2M M 'XX' XV4MNI:o ' i 40 20 I r xxwasw<nowzeum/ A x"mmmm 20 [MXM7//M3MMENO 8 \\/ ^^ X \\

1. X

\\ \\ / r N/V / X r s A NVNNN i //// / # /X A 'K YNINNJ i 'A N I 6 u 6 A//V I A INA\\V j g W//'>(!X A X W MQV/\\iX'XMX//X'X #N IXl 6 4 E EA $t G @2<!)6tISK XM !x W m1 _W OM k n' 4 i, M wraraawmaseanw a 1 wemaw I '8 ^ t X/N/ X A E YNW i X//y'/ X fx X XYNK' % B \\ V s '6 '7vJ__A IN X N YNNX: ! F//x'tX /XtX%'WW//X'!X /\\ X Y6 'h !O YN O d8)(/N4(/b/\\N/b/\\ /'vf4 # N, v'M X AN!X m NXp 6 '4 E

1. v\\ /hAV'dZ

_X38!$R$83t4 'XWN2C ')fx' M3 ? 4 y j A y m Axx, mAxw + wg ,e 1, ,0 i .02 .04 06.08.I ,2 4 .6 .8 L 2 4 6 8 to PE RIOD, see -from Trojan FSAR, Figure 3.7-2 FIGURE 7-2

N t26' 12 5' 124' 123s f225 l E [* gl 4ee.$_. f. G'Q por t 4 I . y ".'b.-t y d-t ' ' h, h Y 8,,,15 I j h T A f

• E N f y,rg y y,,<

I FF l s r r r f M I T TF G Y j g3Mfy k /, y E( N 'k s r a 4 ./ ,z/ C in g i f i bp TACOMal i,h j

• di i E

I r 4rti - 2 -,7' Q' g y- --f,n - l West 9 E r O F ' g lster 81 s ,yn hv (,' h! Q ' centrate " " ~ ~ I h g -?~~; fl E" N l 1 'y N,\\ \\ h,h l $ W I ] y D Castle R6cm

t., -

t ut st un,, N k' N h. a cetin u A.,, g teagu. 46' Reinieri cati , p ._g__ _ g,3 SITE i / 1 s a k k,j 2 \\.N r 'y ' fe 3,w3 .I y [E gg g idlornest F v ,S r m, N u,P i s 4 2 !y ! n, 3 M -I l q l 7 / E 45-4 m r_ t O \\ hl E 1, g 1/' // / OdLsaw l 'i I g R \\ i, \\' / / o l 5 / F UtJGint 44' ~ i 1 i 12 6 ' 52 5' 12 4 ' 12 3' 12 2 ' % _o io to m 40 v) v a, e. w< e e

sts* SIGNIFICANT EARTHQUAKE EPICENTERS AND INTENSITIES, 1769-1971 I 7 s EXPLANATION r r t i h.u k Y E Roman Numerol indicotes the epicenter of on eorthquake, the t 4 9 epicentrol intensity of which was 3[II on the Moddied Mercotti Scale. F i f g Indicotes eorthquoke was felt but hos no assigned intensity F i T "~' foult doshed where inferred o twAtcMet w \\lshk gERENCES: g 1 Bose mcp is token from Internoteonel Mop of the World N L-10 3 2 Coscode Range g 2ag i 2. Faults shown are from. _ _..Y} 7p._ 7.r, P 'N, v -47' Geologic Mon of Woshington,1961, secle 1:500,000 e 2 7 fir,cu.ao.nej Geol)gic Map of Oregon west of the 121Meridion,t96f, rF Wi 8ader g ggglg g 500,000 m a 3. U S Eerthquake5 1928 - 196B" U.S.C. B G.S. rk "Earthquake Histo,y of the United States" Port 1. U.S.C. 8 G,S g ' o YmW A E.S S A.Hypotenter Octa Cords,1968 - Sept.1969. U.S.C.8 G.S. 8 y i r 'r "BuHetin Seismological Society of Americo" "Abstrccis of Earthqucke Reports' 1967 - Sept.1968. U.S.C.8 G.S. I E g j Jon.1939. Desenptive Cotolog of Earthquokes of the Pacific Coast of the United States 1769-1928. Jort 1953-Washingtort Eorthquokes (930 - 1951. A Jcn 1963 Earthquakes in Oregon 1841-1958. Junel967-Washington State Earthquokes 1840-1965. 46 4 The Tectorne Mop of North Americo,19695 shows o possible I I fouff c!ong the Columbio River between Portiond and Kelsci [~ Evoluoton of this possibility indicctes o fault should not be inferred there. 7 2 g e Y.S 25 g } } [ I g,...e

1. (I

} \\I a l l l f - 4 s-1 1 i i t

g..,..

\\ d '\\ Sfotnoo W o; I, e e. l~~ I .gg_j tru* sto - - - - - ~ l -from l ~ ~..... FIcugg 7,3 f l ~~~~ ^-----.-.m__

ODOE Ou@stion 8 Compare the Trojan design criteria with design requirements used nonnuclear construction in this area. for PcE Response In general, the Uniform Building Code (UBC for a Seismic Zone 2 is used to design non-nuc) lear structures in this area (Figure 8-1) a r e a'. Trojan is designed are considerably more stringentin accordance with NRC requirements, which in the response to Question 7. than UBC requirements, as described Figure 8-2 compares the seismic design response spectrum for Troj and the corresponding lateral force coefficients from the UBC an earthquake design criteria for the Trojan Plant, The natural period, are a large factor above in all ranges of for conventional structures. the UBC earthquake design requirement Discussion The most dramatic differences between the seismic design basis for the Trojan Plant for other classes of nonnuclear structures such as high rise bu ings, dams, bridges, etc. inelastic earthquake structuralare in the design concepts of elastic vs

response, for the Trojan Plant,In an elastic seismic design,and static versus dynamic analysis methods.

such as that used structures are required respond to the amplitude and frequency content to be considered to motions without significant of earthquake ground would otherwise reduce nonlinear structural performance which the earthquake lateral Stated differently, in an elastic design, load design demand. to resonate at their structural frequenciesstructures are considered in response to the earthquake ground motion. Thus, without effects, the structures experience consideration of 'detuning' earthquake ground motion input. large amplifications over the As illustrated in Figure 8-2, the result is a substantial demand, approaching 0.50 g, increase in the earthquake design load over the input ground motion level of 0.25 g. 1.n the more realistic inelastic seismic design approach is given to the experience based performance capabilities of recognition tures to deform in a nonlinear struc-required structural integrity and thus, fashion while maintaining their because of inelastic per-formance, not The seismic design of conventional structures has tra motions. tionally taken this more realistic approach, the inherent with the result that become "detuned" and dissipate energy based on inelastic str response is acknowledged.

Thus, ural exhibit increased amplitudes.

the UBC curves in Figure 8-2 do not. -.

Another najor difference in these seismic design concepts is the equivalent static lateral load method used in the UBC design i approach versus the dynamic analysis method used in nuclear plant (. design. In the equivalent static load approach, for the particular UBC seismic zone in which the structure is located, earthquake lateral load coefficients are determined from categories of site conditions (rock, stiff soil and medium to soft soil) and categories of lateral load resisting systems (shear wall or braced frame, moment resisting frame, etc). These coefficients are then used to develop equivalent static lateral loads based on the structure weight tributary to the various stories in a building, for example, to determine the earthquake lateral force demands. Structural res-ponse modes higher than the fundamental are only approximately accounted for. This static equivalent lateral load approach is simplified and rational but of earthquake structural demands.provides only a general characterization In the dynamic analysis method, such as used the site seismic response characteristics are firstin nuclear plant design, determined and then translated into foundation models for the structures (fixed-base, elastic springs, non-linear finite element, etc.) to determine their dynamic response parameters. Seismic input motions are then applied eration time-histories (accelerograms)to these complex models either in th 1 or design seismic response spectrum model inputs (spectral accelerations). The seismic loading demands are then determined by structural account for all characteristics of response solutions which structural behav37r of interest (translation, rocking, torsion, etc.) and all inertLa load response modes of significance. The dynamic analysis method used in nuclear plant design provides a much better representation of the actual seismic characteristics of structures. is time consuming and costly,Being a much more datailed analysis method, however, it and may not be practical for certain classes of conventional structures. Even in areas of high seismic exposure, such as Los Angeles and San Francisco, local codes do not require that in the structural design ofsite specific studies and dynamic analysis be performed such structures as high-rise office buildings, except under specific conditions. criteria for these classes of conventional The seismic design mic risk zones is seldom, if ever, required to be as conservative as structures in high seis-it is for the design of nuclear plants. A visual Trojan Nuclear Plantrepresentation of the dramatic differences between the and Uniform Building Code seismic design requirements is illustrated in Figure 8-2. In that figure, the maximum and minimum UBC required earthquake cients for conventional structures located lateral design coeffi-in Seismic Zone 2 are plotted against the Trojan safe shutdown earthquake (SSE) design response spectrum. As shown, the earthquake design criteria for Trojan plant, in all the ranges of natural period, is a large factor above the UBC earthquake design requirement structures. for conventional 30 -

SEISMIC ZONE MAP OF THE UNITED STATES

• s e sw e, r s.,

we av f.Cl ,e w / ~q .~ f 1-x-n ,.e E ~' ~ ~~. { [~. )- - -4 l \\ D ,/s < >. 3 l\\ Q.-..,/ _)m 7_._ M X, y f hT ,.y y I I ) k,,,,,,,, f tL \\ ar \\ h g 2 j ~~ 'i=T f j yE.. '} h. l >g 0 I ~. 2Cett 0.No esmese. ~ 204(E 1. hetner omrnage; disture earthauskes may some comece to strucewee with fursemanes pertoos N e i Y and the Scese. k p 2DNE 2 essenge,eermepende to Weenefy W of _ 20ME 3. tenear g telesensey n and / tome 4. Those areas seem Zane peo. 3 estermhed h the prozam.'ty w eartain snepr taun eyonema, onnesthed tieresel beenemy sense of i331 f f me f ar ) er er w ) -from Uniform Building Code Figure 8-1

l COMPARISON OF TROJAN SEISMIC DESZGN WITH UNZFORM BUZLDING CODE REQUZREMENTS 7 x s e l T k 5 + l 9 H ni d 5, WC bi b \\\\\\ ec ,83 = a ~ / S@ q n \\ O O \\ / m_\\' 1 I d \\ / %3.- I I / a cc 3 W / 1 Y M 7G / / "ZR" / / 9594 / a / SQ$/ 5 o ps o t 24-y esez tt ? 5 1 N 3 N A o \\ \\ \\ N \\ s N s N N \\, N N ^ N 9 e. 9 ~ E o o o 6 o 0 -D S ' NOl1VW37300V 7VW103dS -..

ODOE Ouestion 9 How do the consequences of a postulated subduction zone earthquake differ from the anticipated effect of a Trojan design basis earthquake? PGE Response Figure 9-1 compares the Trojan seismic design basis with the synthe-sized large subduction zone earthquake ground motion spec *c.ra postu-lated by Heaton and Hartzell, 1987. seismic design basis with spectra generated by recorded groundFigure 9-2 motions from actual large subduction zone earthquakes. These comparisons show that the Trojan Plant can accommodate the ground motions resulting from postulated extraordinarily large subduction zone earthquakes. Discussion In a hazard assessment of earthquake ground motion exposure, the proper context of comparison between the postulated occurrence of large Cascadia Subduction Zone earthquakes and the Trojan seismic design basis (or any other seismic design criteria) would be a probabilistic evaluation. At present, however, consensus data on the hazard due to Cascadia Subduction Zone earthquake ground motions at any location in the Pacific Northwest are not available. ) The following, however, is a comparison of the Trojan seismic design basis with (1) synthesized large subduction zone earthquake ground motion spectra, and (2) with the spectra generated from recorded ground motions from large subduction zone earthquakes that have occurred elsewhere in the world. the actual earthquake resistance ofIn these comparisons, note that the Trojan Plant structures and systems above the seismic design basis due to inherent conserva-tisms, which is a significant increase (see the response to Question 10), is not represented. j It should be recognized that Trojan is located on a rock founda-tion. The rock reduces the effect of earthquakes since there is less amplification of the ground motion, as seen when structures are located on alluvium (soil). Several rupture zone models and types of grorni motion from some of the largest subduction zone earthquakes ever recorded (Hyuganda, Tokachi-Oki, and Miyagi-Oki in Japan and St. Elias in Alaska) have been used to synthesize ground motions projected for postulated earthquakes by Heaton-Hartzell (1987). A rupture zone model used in this synthetic earthquake ground motion generation extends 950 km (600 miles) in north-south orientation and 200 km (125 miles) east to west (Figure 9-3). This from extending from about rupture would encompass an area northern California to southern Canada and from the Cascadia Subduction Zone convergence region (approximately 135 km or 85 miles off the coast of Oregon and Washington) (125 miles) inland to about the longitude of Puget Sound (this would 200 km be a rupture zone area of approximately 73,400 square miles). 33 -

Ground motion was simulated for the model by adding the seismic energies of the recorded subduction zone earthquakes described above, to equal the earthquake energy corresponding to that if the Chilean earthquake of 1960 were transposed to the Pacific assumed Northwest. (PGE considers this transposition of the Chilean earth-quake to represent a highly improbable worst-case scenario.) effect of The this assumed energy release, in the form of ground motion response spectra for various size earthquakes, is shown in Figure 9-4. The spectra are represented as average (expected) ground motions (using a 5 percent damping or energy dissipation ratio) as a function of seismic moment magnitude for locations approximately 50 km (30 miles) inland from the coast [the Trojan Nuclear Plant is located approximately 75 km (45 miles) inland from the coast and about 115 km (70 miles) from the assumed center of seismic energy release for this synthetic model). For the largest earthquakes, motions are postulated to be about 25 percent larger than shown in Figure 9-4 at coastal locations and about 67 percent large in the Puget Sound area. as Average peak horizontal ground motion accelerations are characterized as being in the range of 0.60 times the force of gravity (referred to as 0.60 g) sites and 0.26 g for Puget for coastal Sound sites (approximately the same longitude as Trojan). (The Trojan Plant Safe Shutdown Earthquake (SSE) is based on a peak horizontal acceleration of 0.25 (; the Operating Basis Earthquake (OBE)

however, of 0.15 g (with the attendant assumptions)

SSE greater than 0.25 g.)octually controlled the Trojan design and equatesto an The synthesized ground motion spectra adjusted for (67 percent of inland locations Heaton and Hartzell,the spectra shown in Figure 9-4'as suggested by 1987) are plotted in Figure 9-1 together with the Trojan seismic design basis Safe Shutdown Earthquake (SSE) spectrum (5 percent damping ratio). The synthesized spectra have been adjusted for the Trojan rock site response characteristics; not this is conservative since the adjustment would tend to reduce the computed ground motion levels for the synthesized curve. ing Figure 9-1, it is important to note that In analyz-periods of Trojan's safety-related structures,the natural response piping and equipment are typically less than 0.30

seconds, have natural periods approaching 1.0 second.and no structures or systems Thus, the relevant portion of the figure is shown in the boxed area.

Comparison indi-cates that the response spectrum used for the seismic design of Trojan plant will accommodate postulated ground motions the Figure 9-2 shows the relationship of velocity spectrum with response velocity spectrathe Trojan SSE design response (adjusted for distance, as explained below) generated from the largest subduction zone earthquakes for which ground motions have been reliably recor-ded. The subduction zone earthquakes represented had seismic moment magnitudes Mw - 8, which is in the range of postulated magnitudes for large Cascadia Subduction Zone earthquakes. ing locations were classified as rock sites. All of the record-Distances from the subduction zone earthquake rupture surface to the location of ground motion recording stations were as follows: the - -

Valparaiso, Chilo - 39 km l [ - La Villita, Mexico - 19 km La Union, Mexico - 23 km Caleta de Campo, Mexico - 15 km Zihuatenejo, Mexico - 28 km I i The distance between the Trojan site and the hypothesized Cascadia Subduction Zone seismogenic interface is presently characterized as j being 60 km or more, which is greater than the distances at which the ground motions shown in Figure 9-2 were recorded. The recorded spectra shown in Figure 9-2 were adjusted for the difference in distance from earthquake rupture using appropriate ground motion attenuation relationships (Young, et al, 1988). As was the case with Figure 9-1, in analysis of Figure 9-2 it is important to note that the natural response periods of Trojan's I safety-related structures, piping, and equipment are typically less than 0.30 seconds, periods approaching 1.0 second.and no such structures or systems have natural As shown in Figure 9-2, the Trojan seismic design basis spectrum envelopes the spectra based on actual measured subduction zonc earthquake data from Chile and Mexico for magnitude Mg 8 events. This comparison again indicates that the Trojan seismic design can accommodate ground motions resulting from very large subduction zone earthquakes. i In Figure 9-1, the comparisons show very good agreement except motions at for periods greater than about 0.70 seconds. In this long-period range, the divergence in shape between the Trojan SSE spec-trum and the hypothesized subduction zone earthquake spectra 1 is mostly attributable to the effect of soil site rosponse, which tends to amplify the longer period motions, ponse where such amplifications do notas compared to rock site res-appear. As can be seen in Figure 9-2, where the spectra were developed from earthquake motions recorded on rock sites, amplified long-period motions do not ally result. gener-A discussion on the significant design margins available above the Trojan seismic design basis capability is presented to Question 10. in the response - - -

COMPARZSON OF THE TROJAN SEISMIC DESIGN BASZS WZTH SYNTHESIZED LARGE SUBDUOTION ZONE EARTHQUAKE GROUND MOTION SPECTRA (- t f.s-MS SEISMIC RESPONSE OF TROJAN x,, x, zm,. - 9f"N SAFETY-RELATED SYSTEMS, MT f,'f ,5 go f y f s r"////V STRUCTURES, AND COMPONENTS WANN/// A / 6 /i W NXN p

Qq///*K M'fX e/ 7,/., T.

/g 6XdA j gd '&'f .X. /' 0 ~ k W 9,4 ~ / Y 0,4j

y%'spkTV. _ :&, C,Q'*

~ %jf d r ^ ~ ;. q u y igg 7 N M : ![\\ / I M g i L' s. I J ,- W 44M' '-N,,M. A' %"" lM _/ 2 - -.'. T - , /,,v < n n. we i m m rr m v.,~nn_ ~ ,+s 3' ' g,' ' / / /; v% %N {v'/<//E'/tyX 4 era N W'/, A' A 7 /\\ X WhXh%//EV/AMC MisAn/ene4 ZN X'N y MYA/N/\\ N W N 'r#dh.PC@fMMM'N 7AN 4 /0# 0 MM d\\ )db'3N'/MMMMI /": D@d W A 3 e g g/ ,,s ,s s 2 \\ \\g / / 5, 2 /y(pygggA pAggp m f />s <.\\ g /X MA fYufHE612696@UWP a WPTIAJS (EAfou 4 > 0& 'l $$$$Q Y 'f;-Q^fff& HACuiLL) UNAQ)ll$7ED \\ y //// V A A rs'vs W//// A A A x n" f0R T(t,0 JAM PM s*~s st~> x rx Ar nw ss/v r r rx x WX{ gifg, pggpgfg4, n h','/ t X, /Y_N NA'7xW / /:X /\\'N XXhu _. ..s . g. t xv.' /o SM /1 /N% WNA /N /N N ')'221W' /. /\\ ^x. GW 't 'M s ~ n ~ " ' ^ ' N iS$ Rh( / N! D G 99 f g h%[ M.)N [ d r $ M 's h 4 % ' 3 61 y K, s. .-') N '$) f, ,s .,, ',/,h 0.1 N e ' !' /\\ A. _\\ i '\\. '., N 'h, u ./i s .x,. .,/. pu

• ?'Y Y~: WrS $ $ ll Y-l-lx\\'M'&Il'

.am-s ~x'~ ,l N O ! t* f// ' A ' A A WNw's//// A' A A NNNXxV ' ,K. <.' q / x 6 w X X W X / / e. y rs. q N y Nx} N *' N' ' A x VsinN n [' ~ /N /N N 'df6&M- /\\ \\ \\ XNiN's\\ _4_i v ,_c. W. o, 'S MM [N [ 1 h '%gt ' d d \\ \\ \\\\ 'Y \\ M- .'/ \\ .h 's '\\ ., ',s '8 %s. s ',s s.N \\ i s' o s,.. 0 0' 'os ^ s .o. io to o PERIOD SEC (PGE, 1988) Figure 9-1 i -. -. - - -, - - -. - ~,., - - -. -. _,., - - - - -. -, - - - -. - l

[ COMPARISON OF THE TROJAN SEISMIC DESIGN BASIS I SPECTRUM ($% DAMPED) WITH SPECTRA GENERATED FROM RECORDED GROUND MOTIONS FROM LARGE SUBDUCTION ZONE EARTHQUAKES.

• LARGEST SUBDUCTION ZONE EARTHQUAKES EVEI; RELIABLY RECORDED 1

' RECORDED ON ROCK SITES \\ 200, l I i i 1 .00 ) 1 t L 50 h J 3, e 3 i

)YN.h,,,t T

I \\i s 20 M \\. .n 8 !'d% % *, \\ '- ~/ ik;t' q t h 4 \\

lll (., b':/.)1p' 'J &,,,.h\\

f '0 h-n. .:L. t , 5,1 p, e ...eV E . / Jtd.e). ri ( '~ f , f. Y \\ /,,,)- Treien ssE ;esign More, '/ a fl m i ((g jReco ced Magnits:e 5 veders(M) ', p, f, f,'<,,. 6 ' ',, S:c'ee to 60 k-Diste :e 2 F l Vcocce:.0a e 1955 ,.' r [b [*;' g t.c W,.c. Wei:: 1985 i { / - - La woe,vei : 1985 .' '[/M.;[', ceet: ce ce co, w re: me [ e - + - I:n e.e e;e, vnto 1985 ,7 .02 .05 2 .5 2 5 Pe*:c:: 'se:) P -from Geomatrix Consultants, Inc. fGURE 9-2

t SCHEMATIC DRAWING OF THE SIMULATION OF AN EARTHQUAYJ SIMILAR TO THE 22 MAY 1960 CHILEAN EARTHQUAKE (M 9.5) BY THE SUPERPOSITION OF 1978 MIYAGI-OKI EARTHQUAKES (M 7.5) (.. l l 950 km --- Tench axie = r ~' r - Rupture from ,0 ' ~ ~?'~ ~ ~ ~,l'- -,$...l.'.17 _}y

a...:

?- (s c.aw . t.n... '---)-- sfgd., y r --- -- at:if / //, /(K'_ e %m

~..

tows.na, Cascade O q / GrouM motfobs from volca A 4 A A, A A each bubfault are simu ground moUarts recorded durf"9 2 M, 7.5 Aftyeg5.ou gg i -from Heaton-Hartzell. 1988 FIGURE 9-3

ESTIMATES OF THE VARIATION IN AVERAGE HORIZONTAL GROUND MOTION RESPONSE SPECTRA (5% DAMPING) AS A FUNCTION OF ENERGY MAGNITUDE FOR SITES LOCATED 50 km INLAND FROM THE COAST 'o b. 4,,O A e n. // s b hMa i W

1. )

\\0 {"o S 1-g Sai 9 i w*: .5 N-sn 2, O 9. g"' [/ 6. m. m ~ >. o 7.9 t t F-e. U*:. O O* a.." 7.25 \\_? w n. n. t w. o i o 10* i. 5 i ii.iii'10" i i 4 5itii'10' i 5 4 5itii'10' i 5 iiihii10' PERIOD (SEC) > -from Heaton-Hartzell. 1988 Figure 9-4

4 ODOE Ouention 1_0 Put aside questions of whether an event of the postulated. magnitude could occur and theifrequency of its recurrence. What margins exist in the seismic design of the Trojan Plant that would provide for safety at ground motion levels significantly greater than those represented by the seismic design basis? PGE Response There are numerous conservatisms in the seismic design of the Trojan Nuclear Plant. The conservatisms result in an additional margin of safety for large earthquakes even beyond that shown in the response to Question 9, which show that Trojan can withstand the ground motions resulting from postulated large subduction zone earthquakes. These margins provide Trojan with the apability to withstand ground motions two or more times the design basis ground motions. The increased seismic capability from these conservatisms is shown in Figure 10-1. Discussion There are numerous different conservatisms inherent in the seismic (earthquake) design of Trojan that result in substantial margins for structures, components and equipment. Trojaa structures, components and equipment can withstand ground motion levels significantly greater than the seismic design basis earthquake without endangering the public health and safety. Independent studies concerning earth-quake margins and conservatisms are described on the following pages. Seismic margins in nuclear plant design and construction have been investigated in the nuclear industry for the past decade. Seismic margin research programs include the Seismic Safety Margin Research 3 Program sponsored by the Nuclear Regulatory Comn.ission (NRC), cut-rent research by the Seismic Qualification Utility Group (an indus-try group representing essentially all nuclear utilities organized to verify seismic qualifications of nuclear plant equipment), on-going seismic margin identification programs sponsored by the Electric Power Research Institute, and probabilistic risk assessments plant-specific seismic 20 such assessments performed). (there have been approximately The conclusion of these detailed programs and studies is that there are substantial margins with respect nuclear plants in the United States. to the seismic design basis of This uniform industry exper-ience demonstrates that similar margins exist at the Trojan Plant. Seismic margins specific to Trojan provide further confidence that the Trojan Plant has margins similar to those observed at other plants where these detailed margin studies have been conducted. of One these characteristics of Trojan is its relatively recent design vintage. (As described in the response to Question 7, seismic design criteria for the Trojan Plant are in accordance Uith 10 CFR 100, Appendix A. Prior to Appendix A, regulations with respect to seismic criteria were not clearly defined.) i.

l There are other specific conservatisms in the Trojan design criteria, such as lower-than-usual damping (energy absorption) values and the l relatively high ratio (0.60) of Operating Basis Earthquake (OBE) to [ Safe Shutdown Earthquake (SSE), that create greater-than-usual margins with respect to the SSE. In addition, factors that studies to be of the most have been identified in early margins significance in several of the lower sels-mic margin nuclear plants (such as mechanical and electrical equip-ment anchorages) have been the subject of conscientious seismic evaluations at the Trojan Plant in the years since it was originally licensed. All of these factors support the conclusion that industrywide seis-mic margin evaluation experience and Trojan-specific factors provide assurance that the Trojan Plant could safely sustain ground motions two or more times the seismic design basis ground motions. This mergin above the design basis seismic event provides additional assurance that Trojan would safely withstand the subduction zone event postulated in Question 9 without and safety. endangering the public health The seismic margin assessments that have been conducted by the NRC, the Electric Power Research Institute, and by individual utilities are based on three approaches: engineering analysis, testing of structures and equipment to beyond their design basis levels, and the actual performance of structures. piping, and equipment in conventional power plants that were subjected to very strong ground motions. These bases and the resulting conservatisms and margins are explained in further detail below. 1. Conservatisms in Desion The engineering design process uses assumptions and meth<dology which contain conservatisms with respect to the actual perfor-mance of the structures and systems. Examples of these areas of conservatism are: Actual Versus Desion Minimum Material Properties - In the orig-inal design, before actual material properties are known, con-servative, minimum material strengths are assumed. As construc-tion progresses, samples are tested to determine representative actual material strengths. these tests are typically 10 to 30 percentMean value material strengths from greater than those in the design. As an example of this, review of represen-used tative concrete test cylinder data for the Trojan Containment s concrete (47 samples) showed that, for a concrete design compressive strength of 5000 psi, the mean value from the test cylinders was 6876 psi, about 38 percent greater than the design value. In another comparison for concrete in the Fuel Building (60 samples), which had a design compressive strength of 3000 psi, the mean value from the test cylinders was 5094 psi, about 70 percent greator than the design value. Representative i test samples for the reinforcing steel used in the Containment 41 -

structure also showed c usen tensile strength value of 104,400' psi, which is about a 16 percent increase over the required minimum value of 90,000 psi. In typical nuclear plant structures, determination.mean strength is appropriate for actual capacity Seismic Analysis Method - The response spectrum method of analysis was used in the seismic design of most of the Trojan Plant structures, systems, and components. This method of i analysis, in comparison to the time-history and finite element modeling methods that could be used in a seismic analysis, has been recognized to provide a margin on the order of 10 percent or more. Operatina Basis Earthouake Versus Safe Shutdown Earthouake Design Criteria - For Trojan, the OBE is larger than the typical one-half the SSE value for nuclear plants. The Trojan Plant OBE peak horizontal ground acceleration (0.15 g) is 60 percent of i that for the SSE (0.25 g). Because of the low damping ratios and allowable stresres associated with the OBE design criteria, the design for most structural elements is determined by the OBE requirements. This results in an additional margin beyond the design requirement for the SSE of 20 percent or more. Conservative Dampine Ratios - Damping is a measure of energy dissipation in a structure or system as it responds to vibratory motion. As more energy is dissipated, the overall response (vibration) of the structure or system is reduced. Thus, the higher the damping ratio is, the lower the response becomes. Duch lowered response in turn reduces the potential for damage i to structures ind systems. Very conservative damping values, as ) low as 0.5 percent for piping analysis, 2 percent for OBE, and i 5 percent for SSE structural analysis have been used in the j Trojan Plant design. Much higher damping values, ie, 5 percent for piping and 10 percent for structures, can realistically be expected. Thus, a design margin that can range from 25 percent { to more than 100 percent can result due to the difference between the conservative design and realistically expected damp-l ing values. Broadened Floor Response Spectra - For Trojan, response spectra that are used to define the seismic input for a component or structure were broadened to account for potential variations in structural properties (conservatively assumed in the original design stage) that could affect the location of resonance repre-sented by the spectral peaks. The spectra peaks are broadened by amounts of plus and minus 10 percent or (usually) more. A percentage of conservatism around 15 percent attributable to broadened spectra is usually reported in the literature. Desian Criteria - Loading combinations used in nuclear plant design are very conservative. In actuality, the combined loads are very unlikely to exist simultaneously. Mcre importantly. l the design criteria parformance is one where there is essenti-ally no yielding allowed during, or permanent deformation 1 allowed after, a design basis earthquake. This is a very large [ conservatism since in actual performance, one of the single j important is its ability to yield, characteristic of a structure in resisting earthquakes l decouple from large structurbi resoninces, and absorb energy. Such limited yielding and limitad permanent deformation does not impair the functionality of a rtructure. In the design of conventional structures, these characteristics are implicitly allowed in the design process, but they are not allowed to be accounted in the design for Trojan's structures, systems or components. For Troian, the full effect of the earthquake was designed to be resisted without reliance on this very large reserve energy capacity. In the unlikely event of an earthquake exceeding the design basis, this reserve capacity is available even though it was not relied upon during the original design. There have been a number of studies performed for nuclear plant structures at other sites (such as probabilistic risk assessments for the Zion and Indian Point Nuclear Plants) that evaluates the margin attributable to the inherent ability of structures to perform inelastically, and it ranges from approximately 200 percent 500 percent (a factor of to two to five times the design ground motions) depending upon the type of design and construction. 2. Conservatisms Demonstrated From Tests. The major safety-related electrical equipment installed in the Trojan Nuclear Plant which must remain functional during and after an earthquake includes: motor control centers, 480-V meta)-clad switchgear, transformers, control panels, and bat-teries and battery racks. This equipment has been seismically qualified by actual testing on a shake table. The expected largest seismic motion at the location of the equipment in the plant is simulated, and the equipment is tested for its cap-ability to perform required safety-related functions during and after the design basis earthquake event. This equipment is typically very similar or identical to equip-ment installed in most nuclear power plants where it was seis-mically qualified 'or earthquake levels specific to those plants. These items have, therefore, different levels of vibratory motion. been tested for many At several other nuclear plants, the same equipment has been qualified to levels higher than those that result from the Trojan Plant seismic design criteria. These successful test. results indicate that electri-cal equipment is very rugged and capable of withstanding earth-quake motions larger than the Trojan SSE. The Electrical Power Research Institute (EPRI) has compiled industry-wide test and experience-based information in the form of generic equipment response spectra, which provide a data bank - 43 m

of the seismic capabilities of such equipment. strates that This data demon-seismic demand and equipment capacities.there is substantial margin ava In addition to the EPRI data, the nuclear industry has made an extensive effort assess the seismic capabilities of generic types of equipment. to These studies, to-date, in the seismic capacity of generic equipmenthave also shown large margins in the Trojan Plant. types similar to those used In terms of structural dynamic response of cable tray systems, there is a significant difference between design criteria and expected performance. The most dramatic difference is betwoon damping ratios (energy dissipation) used in design criteria, which are usually low, and demonstrated by cable tray system dthe high values which have been actual seismic response experience.ynamic tests and evidenced in

systems, For Trojan cable tray which included the SSE,the design damping ratio used for load combinations, was 5 percent.

ing ratios demonstrated by generic tests are on the order ofCable tray system da 10 percent to 30 percent. In terms of seismic response, the reduct. ion in structural demand between 5 and 20 percent damping can approach a margin of system 100 percent or more. Therefore, cable trays at Trojan have significant margin above that sented by the seismic design bases. repre-3. Experience-Based Marcin Demonstration The safety-related structures at Trojan have an inherent to resist those associated with the seismic design basis. earthquake grou s Such margins can be deduced by reviewing actual building performance in res-ponse to ground motions experienced during the earthquake with the seismic design basis ground motion. I This comparison demon-strates conventional structures can survive ground motions well above their design basis with only minor, Since the structures at Trojan are designed to criteria signif-acceptable damage. icantly more conservative than conventional structures, be expected to survive earthquake motions significantly above 4 they can their design basis, also with only minor, acceptable damage and no loss of function. The performance of has been observed for centuries, structures subjected to strong ground motions but only recently have detailed reviews of this performance been conducted and documented. performance of The structures and systems designed to seismic criteria less than those used at Trojan have been reported recent in a and Industrial Facilities andpublication entitled "The Effects of Earthquakes on P i i the Implications for Nuclear Power Plant Design" published by the American Society of Civil Engi-neers (ASCE), 1987. quakes in the world, This publication covers 15 major earth-earthquake in 1952. starting with the Kern County, California uakes included conventional power plants,The facilities affected by these earth substations, transmis-sion and distribution facilities, industrial plants, paper mills 44 -

and rofinery facilities. Thsso fccilities include structures and systems similar in function, but not design and quality, to those at Trojan. The conventional power plants and other indus-trial facilities were designed to commercial standards. These standards are far less conservative than those that would be used for a nuclear plant. An example of inherent reserve strength is the El Centre Steam Plant. As stated in the ASCE report. the structural frame and equipment was designed to accommodate specified loads corres-ponding to a peak ground acceleration range of 0.10 to 0.20 g. During the Imperial Valley earthquake (1979), the El Centre Steam Plant experienced ground motion considerably above 0.20 g acceleration; less than 1 km from the plant the measured ground. acceleration was 0.35 g in one horizontal direction and 0.49 g in the other. No significant structural damage was observed, even though the peak ground acceleration experienced by the Plant was at least twice that assumed for its design basis. These observations at the El Centre Steam Plont are typical and are borne out by the general observations following the 15 earthquakes reported in this publication, which are: i i Power plants experienced only local damage in earthquakes. Engineered structures and facilities designed and constructed to withstand seismic loads survive seismic events beyond their i design levels. ( As another example of experience-based demonstrations, during the past few years the Seismic Qualification Utilities Group carried out a comprehensive survey of non-nuclear facilities which have experienced strong motion earthquakes with peak ground acceleration as high as 0.60 g. The facilities surveyed had equipment similar to that used in Trojan. The results of the survey indicated that such equipment properly performed its functions during and after the earthquakes provided their anchorages were properly designed. In nuclear plant design, special attention is focused on provision of reliable equipment anchorages. These findings from non-nuclear facilities provido confidence that equipment and piping used in nuclear plants, with their more rigorous design requirements, would also properly function following earthquakes significantly more severe than their design basis. A fourth basis for PGE's conclusions regarding conservatisms in Trojan's design comes from knowledge gained since performance of original design and from additional conservatisms introduced by the Plant upgrades. A detailed review and strengthening of the Control. - 45 ~

Auxiliary and Fuel Building complox* and review and modifications equipment and piping supports are examples of such increased know-to ledge and added conservatisms. l, Summary For Trojan Plant structures and systems, taking the aggregate of the margin factors described above, and including additional conserva-tisms and knowledge of the seismic response o' Trojan, the struc-tures and systems can be expected to withstand earthquake ground motions two or more times larger than those associated with the seismic design basis without safety-related systems. impairing the safety function of any The Atomic Safety and Licensing Board which reviewed.the Control Building matter concluded that even though 30 percent to 50 percent of the intended margin was lacking, sufficient capability remained to withstand an event at leas: 50 percent greater than the 0.25 SSE. The intended margins were all by the strengthening the structures. restored by PGE in 1981 46 - -.. ~ - - _ _ _.. _ _

TROJAN SETSMIC CAPABZLITY AS A FUNCTION OF DESIGN MARGIN \\ ,l ? f.s ' ", % SEISMIC RESPONSE OF TROJAN N "'..I ff 2U' Y SAFETY-RELATED SYSTEMS, , frn M-t-Y y-s;h 4 k o " -[N = S fx STRUCTURES AND COMPONENTS j/ V.] j7%. o, f c, . '/ _ g.b_ y ~ , ij i s: dil V4 7 (# 4 /sy' ~ -QN] [ fUp'_ (s gfd of,V . ;.yj)(, MARGI k __,f o - i tr s._ .x-d I. /,Di .'s 'V .3 It. srr.v wr > ,__r i t/ ~ f B i ' -t _ ' f iT. L. ' 7m / ~ _x. _/ / / fyr f r /\\ 7\\\\ Y\\ \\% \\tfor'Y f f/ W % '.ff f'/ ' ' FT.M'. r% G-&) .<sA .A f[//,' ' A A v1 W G% /M'/< // VA n//// A' / s ' s% \\ s5 Q t, /NtX h sN 9Kr/ E V # M sMILb f t y W/i /NI /\\ NWMhM .w i Y lx 10 4 3 2 /o, s M 7/56^2MM*'\\ ? d' L a SEM A M4WM'/AX 4%6MXM / hPM 't Mk AW h M, M W N., W,, slbV$l w p g I p /h, $ W V,,A O A Y W W f s 2- .,.i.,,, y s s/ /* X U / fYufHE117Ji9 dlf.'UNP VpfladS (MATou 4 _-- A'Ah A $NY ~l'd b'l' '<N?: 0 $- W IELL) UNhON STSO ^ \\ nma rr u m /~r,r1 A /#vvWWer/ A< A A x m rzm str.o.is.ue m

,o*'%

y /N v K'//!X mNs' 3;Tp,,, gpyp;fgowe,, MX) NAMW/ /W. rsia x v ! /o J NNN.m 4 ' ~ /i A%NWSW/l /"N A N WW' ' /\\ \\N_._1>'M 't ',q gqve < xrxmgg fs; gk M k z [ N [ [ p u d t I l M A M b' ), g 'N \\ y- / A. m h .., e. \\, C3 \\ - - f / \\ /A ..e 7. 7. xx/ ' ' r \\ 'A

• / D jN X'Y t h -t~

s sW ' - n., x+ + m. v,' m w w w e A > O d 'i' N ss. s

v. ~.~ -

o -- /s - N' ' NY, ! o* 2 ..'.x fxav.N y m er <,y ~rsn mww '.- N.f(NN ~ / Ai\\ XWJM gi' N v;NNNM...v s s se w v

. m R, m

\\ \\XWs'$\\ _ _ s.ZL ' O 4 4

\\hx 1 - N g )'y 4 M 5 p.x. g- @ NN$'Q).;. 0 0, - o oi og ic ic o PERIOD 5f C (PGE, 1988) Figure 10-1

4 ODOE Ouestion 11 What programs exist at PGE to monitor development of the possibility that a major earthquake in the Pacific Northwest could result from a subduction zone event? PGE Response For many years PGE has had an active program for monitoring activ-ities relating to the seismicity of the Pacific Northwest and the Cascadia Subduction Zone. PGE will continue monitoring the results of ongoing research related to Cascadia Subduction Zone earthquake issues, as well as other geo-science research that may have a bearing on operation of the Trojan Nuclear Plant. This review is being conducted by PGE's technical staff and management, using consultants as appropriate. Representa-tives of PGE have been active in attending seminars, workshops, conferences dealing with seismic issues, and and they routinely review information received from the Earthquake Engineering Research Insti- ) tute. Also, PGE routinely reviews the summary reports of data collected from the seismic monitoring network located in Washington and northern Oregon (see Figure 11-1). This network is jointly sponsored by the U.S. Geological Survey and U.S. Depar'.ent of Energy, and operated by the University of Washington. monitoring stations, twenty-two including two in the Portland area, have been operating in Oregon since approximately 1980. Programs in place at PGE are responsive and adequate for the evalu-ation of seismic research and developments in the Pacif'c Northwest that may potentially affect facilities owned and operated by PGE. 48 -

EEISMOGRAPH STATIONS OPERATING DURING THE lst QUARTER 198 125.00 117.00 49.50 59 5n %+ I

13v

[ + i Pv4 4 4 5 XV NE ) ' oov,0SD M TV Y vev a %s ah ~ 0 gg a wAT,5Av Q , PW 4 anv {N BH +a + EPH+ csn o;s 7g vo nv +' g 3 g. a vic

• pD get 3pveny,MPv,d ss b 4y

?t" .:V ' '

• Y+

"' eTP ,s. + av a a ^* + h ^ * +[T #II.,,. . 4 a a CW +, cL2 VC2 4 ,47 y cac + ' ^ + + t yC2a VTH a vet a Y,, 4 4 + vPo + + G VIP 4T + . A y T + ^ + + + N:o 93.50 F a 125.00 93.50 117.00 OLf ' ' ' ' ' ' ' ' ' 200 KM from University of k'ashington Geophysics Program Quarterly Network Report 88-A January 1 - March 31, 1988 FIGURE 11-1 .~

r.. ODOE Ousstion 12 What is PGE's action plan to obtain data on the geological structure of the areas near Trojan and Trojan? to assess the effects of earthquakes on PGE Response Data on the geologic structure at the Trojan Plant site and in the surrounding area were collected and evaluated during the original plant siting studies. determined from geological and geophysicalProperties of the rock structure which w evaluations (described in Trojan FSAR Section 2.5, Attachment B) are summarized below: The site is underlain by bedrock, which is a part of the Goble series of Upper Eocene age. The bedrock is exposed on the ground surface along a narrow, elongated ridge bordering the west bank of the Columbia River. All Seismic Category I structures and Seismic Category II structures housing Seismic Category I equipment are founded on the rock which forms the ridge. The bedrock is volcanic in origin and consists principally of with lesser amounts of tuffs flow breccias, tuff breccias, agglomerates, and basalt flows. ground surface since they are more resistantBasalt and agglomerate often are ex to erosion than is the tuff: however, tuffs and flow breccias are the predominant in the ridge and they generally provide the structural foundations rock type Thus the strength of the tuffs generally determine the bearing capacity of the foundation rocks. The lowest compressive strength of 41 samples of and highest unconfined tuff which were tested 360 psi (26 ton /ft2) and 2,790 psi (200 tons /ft2). The average are is 1,225 psi or 88 tons /ft2 The geophysical surveys showed the compression wave velocities in the bedrock to be 8,200 to 10,600 fps, which indicates good foundation conditions. velocities of Shear-wave the foundation rock ranged from 4,500 to 5,000 fps. Site response studies pursuant were performed as part of the development ofthe applicable federal regulations to the Trojan seismic design criteria (see the response to Question 7). Trojan structures are founded on very strong rock, as described above, and the seismic design response spectra are broad band to account inputs over a wide range of frequencies. for ground motion Thus, the seismic response characteristics of the Plant site have been evaluated, and were enveloped in the seismic design response spectra. In addition, seismic monitoring instrumentation, installed as part of the orig-inal plant design, ing both ground motions andprovides the capability of recording and analyz-the actual Plant safety-related structures and structural response of Trojan systems (for seismic motions of significance that may be received at the site). The Trojan Plant seismic monitoring system consists of three differ-ent categories of instrumentation: an array of five triaxial strong-motion accelerograph sensors with its initiating trigger and magnetic tape recording system, an array of six peak acceleration recorders, and a 36 channel multielement seismoscope. 50 -

...s s The function of the strong-motion accelerograph system is to provid location of the sensors. triaxial acceleration time-history data of seismic res e / the To obtain a free-field ground motion reference, the initiating trigger for this system (and for one of the sensors) is located on rock adjacent Fuel Building complex. (1 each on the base slab and wall),Two sensors are locatedto the C in the Containment Fuel Building and Control Building. and sensors are located in the seismic trigger is 0.01 g. The threshold setting of the to a central multichannel magnetic tape data recorder locatThe accele main control room. ed in the j rechargeable batteries which provide sufficientThe entire system is powere ) event of a-c power failure. reserve power in the \\ The system is operated in a standby mode until activated by th seismic trigger. Upon activation, e ational status within 0.10 second and continuesthe system achieves fully oper-adjustable range of to operate for an falls below the trigger threshold 10 to 60 seconds after the last seismic shock level. continue to run each time the The system will restart threshold is exceeded. Recorded sels-or mic data on each sensor channel tape playback system. is available through the magnetic i Any channel may be selected individually to be recorded on a paper strip chart. In addition to the triaxial accelerographs, i recorders are installed at various locations. peak acceleration These are on the emergency diesel generator, the component cooling water heat exchan-ger, top of the Control Building, top of the intake structure, and top and bottom ofthe Fuel Building, top of the Containment. The multi-element seismoscope is a 36 channel triaxial response spectrum recorder the three axes)(which is rigidly attached to the Containm 12 peak acceleration response recorders of in each slab. is a mechanical recorder and does not It ent base power source. Following a seismic event, rely on an external spectra for the event instrument also has presetcan be plotted from the recorded data.the ground m j This SSE response accelerations and frequencies which would bindicators ated in the control room if such events were to occure annunci-vides indication to the control room operating <> This pro-initiate the appropriate planned actions in response to th lC which would e event. PGE recently consulted geologists and seismologists to det their collective judgment ermine supplemental geophysical data thatconcerning the need for, a means to obtain, characterization of the Trojan site response would enhance siting studies. monitoring prog) ram beyondOpinions received were that (based on the origina the level of a continuous site installed would be unlikely to provide significantinstrumentation currently of the very low level of results because years of monitoring would regional seismic activity. likely be required Thus, tens of data for evaluation, and a change in the characterization to collect sufficient site response would be unlikely to result. of the 51

"es In summary, geologie and seismologic evaluations of the Trojan Plant site and surrounding area were performed during the original plant siting studies. Instrumentation is currently in place to provide additional data on response characteristics of the site and safety-related structures and systems if a seismic event of significant intensity were to be experienced. considered to be of appreciable value. Additional monitoring programs are not Therefore, PGE does feel there is value added by engaging in supplemental seismic not monitoring or geophysical investigation programs at this time. j Relevant data from existing seismic monitoring networks and ongoing seismic site response studies (Figure 11-1) (King et al) in the Pacific Northwest will continue to be evaluated, i As stated in the responses to Questions 6 and 11, PGE is actively monitoring the progress and results of ongoing geoscience research programs focused on the evaluation of seismic hazards in the Pacific Northwest. 1 i 52 -

Qe* s ODOE Ouestion 13 ~ 7'" What will PGE do to assure independent review and resolution of future developments which might impact Trojan earthquake resistance? s PGE Response The United States Regulatory Commission (NRC) is charged with the responsibility of reviewing issues the Trojan Nuclear Plant. relating to the safe operation of ications, if deemed necessary, andThey have the authority to require modif-to restrict Plant operation if the public health and safety is at risk. In 1984, the NRC reviewed the matter of subduction of the Juan de Fuca plate in the Pacific Northwest and the potential for a great earthquake along the subduc-tion interface. In their safety evaluation for Trojan, the NRC stated that "the United States Geological Survey, National Science Foundation, and seismologic and geologic investigations to assess the possibility of whether or not a great earthquake is likely or even credible. The (NRC) staff is well informed as to the progress of research and will continue to aaintain this awareness.this ongoing We conclude that for Trojan. there is no reason to alter the seismic design basis ating license reviews.". approved during the construction permit and oper-Review of research findings concerning the Cascadia Subduction Zone has been a continuing effort (USGS), which serves as a consultantby the United States Geological Survey to the NRC. Investigators affiliated with the USGS (Atwater, Heaton, Spence, Weaver, et al) are responsible for the m&jority of Cascadia Subduction Zone research currently in progress. The NRC contributes to the funding toward this research. Results of USGS research are published as "Open File Reports" which, if relevant to their activities, are reviewed by the NRC's Structural and Geosciences Branch of Office of Nuclear Reactor Regulation. the i i As requested by the March 1984 letter, PGE Will continue to keep the 4 NRC informed of significant findings relative to the Trojan site, in particular with respect to the Cascadia subduction Zone. In addition, PGE will obtain the services of a consulting firm to review research results concerning the Cascadia subduction Zone (with respect to safe operation of consultant will advise PGE of significantthe Trojan Nuclear Plant). The research developments and new disveveries of evidence concerning Cascadia Subduction 7.ono charac*stictics. 1 - 53

k %O d s ATTACRMENT A GLOSSARY OF TERMS Accretionary Wedce refers to the low-strength sediment deposit subduction process, composed of material sheared off from the oceanic peak during the t of deformations where the stiffness of the lateralElastic Response system remains essentially linear. load-carrying Inelastic Response refers to the response of a structure in the range of deformations where the stiffness of the structure's load-carrying s members occurs,ystem becomes non-linear. Yielding of structural and energy is thereby absorbed. Intensity is a measure of an earthquake's offect on people and structures. well as the duration and depth, Intensity depends on the magnitude of an earthqua as geology, frequency of ground motion, distance from the epicenter, the type and quality of construction, and on accurate observations of damage. usually reported using the Modified Mercali Intensity is earthquake with MM intensity of (MM) scale. An

people, II will be noticed by only a few an earthquake causes destruction of structures,and may cause susp If landslides and observable cracks in the ground, it would be classified as Intensity X.

Macnitude refers to the amount of energy released at the earthquake source, or the "strength" of an earthquake based on seismographic observations. the Richter magnitude scale. Earthquake magnitudes are frequently reported using Generally, an earthquake of Richter magnitude 3 will cause no damage and be noticed by only a few people, while a Richter magnitude of 7 may cause considerable damage to buildings and other structures located close to the zone of energy release. Other magnitude scales are: . Body Wave Macnitude (M ) is derived from the B characteristics of body waves recorded from an earthquake. i l Seismic Moment Macnitude (Mw) is a measure of earthquake energy in terms of rigidity of the rock formation through which the rupture occurs, the area of the rupture surface, and the slip displacement. The largest this century was of Mw = 9.5 in southern Chiledocumented earthquake of (May 22, 1960). Plate - segments of the earth's outer shell (lithosphere) 50-150 km thick which move horizontally across the earth's surface relative to one another. Most { of the earth's major tectonic patterns, including mountain chains, trenches to plate tectonic processe,s and interaccions atand ocean basins, may be directly linke j f plate boundaries, ,.-3,. -m.~.- _.,w g-,mm,, ,,., - y,-,,,,,,.oq,* -}}