ML20137E075
Text
Q
. [qY U G #'c,
. g; ;; , i n vu c -n n c o o m . . . . .~...... . . . .
2- "* r ADVISORY COMMITTEE ON RE ACTOR SAFEGUARDS W ASHINGTON. D. C. 20555 6
s' s..... [j[
March 24, 1982 MEMORANDUM FOR: ACRS Members FROM: R. Savio, Senior Staff Engineer b
SUBJECT:
ACRS COMMENTS ON DR. PAUL JENNINGS PROPOSAL ON SEISMIC DESIGN CRITERIA I have enclosed a letter from Dr. Paul Jennings to Comissioner Gilinsky on Seismic Design Criteria. Commissioner Gilinsky has asked that ACRS comment on Dr. Jennings proposal. I have at Dr. Okrent's request asked three of our ACRS consultants: .Dr. Trifunac, Dr. White and Dr. Zudans, to comment as to how ACRS should respond to this request. I am expecting three reports .
and will make them available to you before the Thursday, April 1,1982 discussions.
Revisions to NR'C seismic design criteria have been under consideration for some time. A significant effort was put into the development of a revised 10 CFR 100 Appendix A in the 1977-1979 time frame.
The work was abandoned before completion because of the assignment of resources to higher priority projects. I have attached a. copy of SECY-79-300 which summarizes this part of the work. Some changes in NRC Staff practices have evolved since then as the result of specific problems dealt with in licensing cases. Site specific spectra have been developed and used in some cases and the NRC Staff has been drawn towards, particularly in the case of the OBE, probabilistic expressions.of the earthquake hazard. In addition, the work being performed under the Seismic Safety Research Program is continuing.
Enclosure:
As stated cc: M. W. Libarkin G. R. Quittschreiber hgg82g 9 850722 l
BELL 85-363 PDR t
i ess
- y
[ % v......, .-._.
NUCLEAR REGULATORY COMMISSION
- s. , j 2 : ADVISORY COMMITTEE ON REACTOR SAFEGUARDS
( [ W ASHINGTON. D. C. 20555
(
ks' w
[*/ March 24, 1982 MEMORANDUM FOR: ACRS Members FROM: R. Savio, Senior Staff Engineer h
SUBJECT:
ACRS COMMENTS ON DR. PAUL JENNINGS PROPOSAL ON SEISMIC DESIGN CRITERIA I have enclosed a letter from Dr. Paul Jennings to Comissioner Gilinsky on Seismic Design Criteria. Comissioner Gilinsky has asked that ACRS coment on Dr. Jennings proposal . I have at Dr. Okrent's request' asked three of our ACRS consultants: Dr. Trifunac, Dr. White and Dr. Zudans, to comment as to how ACRS should respond to this request. I am expecting three reports .
and will make them available to you before the Thursday, April 1,1982 discussions.
Revisions to NR'C seismic design criteria have been under consideration for some time. A significant effort was put into the development of a revised 10 CFR 100 Appendix A in the 1977-1979 time frame.
The work was abandoned before completion because of the assignment of resources to higher priority projects. I have attached a copy of SECY-79-300 which
( summarizes this part of the work. Some changes in NRC Staff practices have evolved since then as the result of specific problems dealt with in licensing cases. Site specific spectra have been developed and used in some cases and the NRC Staff has been drawn towards, particularly in the case of the OBE, probabilistic expressions of the earthquake hazard. In addition, the work being performed under the Seismic Safety Research Program is continuing.
Enclosure:
As stated cc: M. W. Libarkin G. R. Quittschreiber
/hnenMerr f
4
- UNITED STATEE
( [p.~p necg,%,*.
NUCLEAR REOULATORY COMMISSION E I d S WASHINGTON. D. C. 2:ESS 5 .,,)?l
+* DCT 8 II
% *.V..*
MDICRACI FCR: C:I:=tissioner Mason ,
FRCF.: Rcbert B. Minogue, Director Office of Standards Developsent TERU: Executive Director for Operations SUL a_: EE REIATICSSHIP BEIWEE:4 SAFE SEDIDCEN EAic.V.AKES Atm CPERATING BASIS EAR 5CCAKIS In resgense to ycur rec:est this mlercrandtn su=marizes for your information and dat of the other C::=issioners the technical issues involved in the i .terrelationship between the safe shutdown ea'rthquake (SSE) and
(. cperating hasis eardq;ake (CEE) in their deter =ination and application to design.
NRC regulations (Appe. Mix A to Part 100) identify two levels of earthcuake severity t: he a: plied to reactor seistic design. These are called the safe shutd:c erthqcake (SSE) and the operating basis earthquake (CBE).
These earthq2akes can be and are regarded and defined as either geologic events or as engineering design requirements. These two perspectives -
are often difficult to relate to one another. Both Icints of view are explicit in Appe.M ix A. .
Viewed geeleeicallv, the SSE is the most severe earthgaake which can affect :ne size. The GE is the reost severe earthquake which is reasonably likely t: occur during the operating lifetime of t!w plant.
Appe. Mix A defines in detail the elements of the geologic and seismologic investigation of propcsed sites. Deterministic procedures are given te establish de safe shutdown earthquake. These procedures regaire consideratic' of (a) the seismolcgy of the region in which the site is located, (b) the regional and local geology, and (c) the nature of the materials underlying the site. If the structural geology of an area is understoed, Appendix A procedu es for determining the SSE usually have the effect of placing greater esp. asis en structural geology than on historic seismicity, largely
( bec.ause cf the limited historic record.
- Centact:
V)Y~7 of T Robert B. Mi .ogue
- ^ -y
,d u 6 >~ uwv.;rg i _ 443-6914 Enclosure "F"
t
- e
( 1N6 03 g Cw...issic .er l'ascn he intet of these precedures is to identify the maxi =um earthquake
. whid :r. affect the site. Here is, of course, sa::e pro'cability of exc ed2; t'.e S5E. 2is p::blem is because of the uncertainties arising f:c c= 1hited unders anding of the frecuency and severity of large g sede ertsca'<e phen =ana and the limitations of geologic and geophysical
=vestiga :.:ns, rader than lack of validity of the concept of a limiting eart:T i<e. In a sense, den, the likelihood of exceeding the SSE is a =easure of the state-cf-the-art of geolccic and seis:cicgic understandine; it p:::a:ly is en the c:de: of perhaps 10-6 ni any given year for any site, but w_ 2 a very wide er:c: be.d.
Se c;erati.g basis erdq.ake, although an infrequent and major earthouake,
~
is a =uch =cre likely geclegic event. The recurrence interval frecuentiy cons W -=" applicable t: de GE is in the range of 300 to 1,000 years.
Althcush in engdeering practice the GE is usually established as a fracti n cf de SEE, in any parts of the country there is a sufficient record cf hist:ri: seie-4: rients to previde a basis for a p:cbabilistic assess:en: cf the 3I if these events are considered in light of the regional strue: ural ga:1cgy.
[ '
Appendix A cces no: prescribe specific geologic procedures for establishing l the GE (aldcugh i does specify a mini =t:1 level). An ANSI standard * !
has be dra'ted describing a method of p cbabilistic assescent of the GE. Prc:abilisti: r.alysis was used for establishing the GE for the Koshkenceg W cle r ?:ver Plant. He agplication for a construction permit has been retiewed and a::epted by ACRS.
Bis app::ach of two levels of severity of a natural phenmenon is not unique t: earthgi<as. Here is a good analogy between the concept of ,
the S5E a .d GE r.d that of the Probable ?'.axi:m=r Flood (P!7) and Standard l Projec: Fl:cd (5??) sed by t.% U. S. Ar=y Corps of Emineers (and NEC).
Se le;el cf likeli' ccd of these two floed dh.c+mege levels is substantially the see as de S5E and m, respectively.
Viewed as a lied :: e cineerine desien, the design basis for the SSE 1 is spec.f:.e: :.n Appe c:.x A to assure tnat the plant design adecuately protects de pdli: health and safety in the event of an extre=e earthecake.
The GE is es.ablished as the most sefere earthquake following which the pir.: cr. safely be egerated without special inspections.
- "Geidelines for Ee:ermining the Vibratory Ground Motion for the Design Eartge.a.<e fer Nuclear Facilities", ANS 2.1 Working Group Draft, January 1, 1976.
(
Enclosure "F"
- - - . - - ~
( .
Ccanissioner Mason # l d
. Engineering ccdes and design practice apply these two earthouake levels differently. Category I structures, systa=s, and ccr==onents*, must maintai their safe:y function for earthcuake levels up to and including the SSE.
). . Rat is, although the plant, as a power generating facility, may be severel.
damaged in this extrece event, it must go through the earthcuake without undue hazard to the public and following the earths:ake the reactor must be capable of being shut down and kept in a safe shutdown condition.
On the other hand, the engineering design objective with the CBE is that the plant is capable of being safe in operation after experiencing an i event less than or ecual to the GE.
l As an ex7 e 1 of how these events are applied to design, in use of the l
ASME Eoiler and Pressure Vessel Code, the SSE is normally applied as a Fad.ted C: x'itien, meaning that stress levels allowed by the Code which would result in per anent general deformation are pecnitted except when defo=atics wculd-leed to less of safety function. Cn the other hand ,
'/ the GE is censiderm! as an Opset Condition or Design Condition in applicat:
\ of the Code and is used in cenjunction with lower allowable stress levels l
'- at s.ich ne general deformation would occur. In addition to differences of allcwahle stresses, there are other differences in design analysis methods in the application of Faulted and Coset Confitions. Many seismic designers see the SSE as being the basic seisnic design basis with the i GE playing scre the role of a cross-check basis using different analysis i procedures and different limits to assure the adecuacy of the margin I
provided by the SSI design over a wide range. Viewed fran this perspective
- the GE is more of an engineering safety factor applied to design analysis
!- rather than being seen as a seismic event.
l !
n
- ,' 2b provide such a safety factor, Appendix A to Part 100 centains the statemant that the " Maximi.
- n vibratory ground acceleration of the Cperating
- Bases Earths:akes shall be at least ene-half the maximum vibratory ground acceleration of the Safe Shutdown Eartheuake." Bis is a samed.at arbitrary relatic. =Mp which assures that the stresses associated with design load ccrnbinatiens plus the GE loading on systems essential to safe operation ,
l will net result in general yielding of the materials (i.e. will reain i-in the elastic regime). i 1
02 the other hand, if viewed as a seismic event the acceleration associated with the GE may range frcru as low as 1/10 the SSE to as high as 80 or 1 l 90%. In parts of the country where the structural geology is not well ,
urderstood, current practice would lead to CBE's typically about 1/2 l l
i
(.s the SSE; that is, roughly one Intensity unit icwer. i
)
I i
i !
- Defined and listed in Regulatory Guide 1.29, "Seismir Cesign Classificatic
. _ . _ . .. i _ _ _ _ __ ___ _ _ . _. _ h a _ __
t 1
I OCT a WE
- cannissi:ne: Mas:n -
4-i 2e shaking asscciated with actual earthgakes is a very emplex vibratory metien wi:h wide variations in frequency content, amplitude, and duration depersii g en the ty:e of initiating crustal merenent and the trans:nission of Sa :::icn through the earth. Nocnal engineering practice is to define the vih:1:::y =ction input as it affects the facility with a cam:csite
) respense spect:= based en a number of earthcuake records. Usually.this resp:nse spect in is reduced to a timmtion record (a synthetic seismogra:n) for application to design. Bis engineering definition of vibratory motien input is gite cacplex and contains a number of alments of arbitrary conserta:is=. A 9::up of Regulatory Guides
- has been issued which provides a c _ late definition cf the vibratory motion input for reactor facilities.
Se enly distin :icn between the S5E and the GE in the application of these :sspense srectra to design is a scaling by the canparative accelerations.
Sere are a n=her of specific issues and problen areas that are related tc t;.e de:e=i ati:n a-4 application of these two events which are identified a ,6 hriefly des::ibed bel:w. Attac.=ent I is a detailed staff analysis' dis::ssi::q de subject at greater lorsth, prepared with the assistance of the 22. Divisi= cf site Safety and Envircraental Analysis.
- 1. Is the GI a sa's:y-related event in the strictest sense?
Many pecple do net regard the GE as a safety-related event. Design of the plant to wids a.d de SSE without uniue public hazard is felt to meet the safety need. Se decision *. ether or not to centinue operation of the plant afts: an earthgake is seen as a decisica of the utility.
.If an er hpake s'.ctC! occur, the safety of the plant for centinued operation could be es-^'ished by a sui *=Me, possibly quite extensive, inspecti:n p:ogra: widcut regiring design to an GE level. As originally published for c= rent Appeniix A reflected this perspective by making the ests.clishne.: of an 35 optional. Perhaps the best argtcent for regardi=; the GE as a safety-related gestion is a recognition that in the a'ter. ath of a cajor earth;uake needs for gewer would be significant.
Pro =t centinued reac:c: operation might be a catcunity requircent.
Another ar;=en scretimes advanced is that the widespread shaking of the en-Spake affects reactor systsas in cmplex interactive mcdes which are not easily fereseeable. Bis =akes it appropriate to recuire sa:e le:e1 of general earthquake design within the elastic respense range; but in fact application of CBE design to structures, systems, and canponents 4
nc: aisc c:vered by 53E design is quite rare.
i
- Regulaterf Guides 1.60, " Design Response Spectra for Seisnic Design of Nu:les: Pcwer Plants"; 1.61, "Danping values for Seissic Design of
( R:cles: Pc a. Plants"; 1.92, "Ca.bining Modal Respcases and Spatial Cocponer.:s in Seismic Response Analysis"; and 1.122, " Floor Design
, ~
~
Perpense Spc.tra Development for Seisnic Design of Flocr-Supported Equipent or Cc=prents".
Enclosure "F" vv-..- ..m-__,,.,,. - -%_..y_ . - -
,y.,, , m--,_,_,,.-_,7,y- ,ww, - , , . - ,,y. --w-, ~_,.___-__.__,,____.m_-___---- - - - - - .
l I
1
. j
. . .o ;
( -
Cerraissicner Mason -5 -
OCT 3 M
- 2. Should the mE be ests.blished as a seisnic event or as an emineering safety factor?
p,. A widely-held engineering view is that the analysis of a seismic load in the rar:ge of frem 1/3 to 1/2 the SSE provides the best emineering verificat of seismic design, and that the CBE sheuld be esublished in this range at a value such that the seisnic design is determined by the SSE, not the CBE, since the GE is seen as a design check. Note that other non-seisnic.
factors =ay centrol the design everall. Cn the other hand, if deternined probabilistically as a. geologic event, the GE would not necessarily be within this range, and may or may not dete=nine seisnic design.
- 3. What vibratory motion characteristics should be assumed for the CBE?
2e c=ple: multi-freq;ency shaking of an earthgeake is nocnally represented for design p=peses by a highly censervative stoothed response spectrum.
'Ihe same respense spect:ra shape is noc ally used for both the SSE and
( GE design, adjusted cnly as to acceleration. Since the GE is an earthquakt of Icwer inansity and likely shorter duration, a less conservative response spectru= =ight well be app cpriate.
- 4. How shculd isolated acceleratien peaks be treated?
It is not unusual 2n an earthcuake to have a high amplitude acceleration 7
peak of li-ited duration and little impact en the resp:nse of a cernplex struct= e. Fer exa= le, such a peak (1.25 g) .as measured at the Pacoima l Dam in the San Fernanio earthquake in 1971. Current practice is to base I
the engineering design en a respense spectrum which implicitly assumes
. sustained s'akfr.g (and in effect disregards isolated peaks) based on the general engineering censensus that isolated peaks do not have any significant.
effect en st::ctures. .
- 5. What i spections should be carried out following an earthgaake exceeding tra N ?
Appendix A dces not provide guidance on this : natter. It is very difficult before the fact to identify in a generic way exactly what inspection progren would be appropriate after an earthgaake. Cor.versely, following an earthecaks cbservahle effects on the facility could reasonably be expected to indicate the areas recuiring inspection and the types of inspection needed. Se staff pcsition has been that the level and extent of inspection followim l
( an earthquake should be based on observed damage and a canperison between the snocched respense spectrum used as a design basis and the ressense i
spectru: corresponding to the motion actually experie::r-s3 by t.% facility.
Enclosure "F" 1 .
q$
~
( .
DU 8 M Cannissicner Mason
'D:e actual shaking can be detecnined by seisnic instrtzuentation required on all ::: clear pew.r plants to measure earth =uake input. An A}EI conmittee wh' '- dereleped a s:rdard defining seisnic instrtnantation regairements to =easure input
- is new well along on a standard 161ch defines a basis P. for assessing "exceedance" of the design response soectru:n in a real event **,
to precife a basis for decision on level and type of inspection based on the actual facility res;cnse.
/I 4
! Bobert B. Minogue, Director Office of Standards Develegnant At + ent:
Ce: ailed Staff Analysis ,
]
4 cc: Chai=r. ? wder.
C.m=issicner GilinsW Cz:Lssicner Es-.eiy' (s . Office of the Secra ary
- REI N13.,5 "Earthgaake Instrunentation Criteria for M2 clear Pow.r Plants",
a:dersed a-4 amplified in Regulatory Guide 1.12, " Instrumentation for
- Earth @ ss".
- ANSI N643 (R3 2.10), " Guidelines for Retrieval, Review, Processing, and Evaluari:n for Pecords Obtained fran Seisnic Instrunentation".'
(
Enclosure "F"
.- -..,,.. . -. ,_--.._.,__._-___._.____,-.--.,.._____m__, . . . _ . , , . . _ . _ _ . . _ . _ . _ , , . , . - - . . . . - - . . . . , _ . . . . . - , _ , _ - , -r
~ '
, . .o
(- MTACE*.ENT I STAFF P2 PORT ON BASES FOR SELEX' TION CF SAFE
~
SE.=N EA3CECUAIG5 (SSE'S) AND OPERATING BASIS EA.-CECCAKES (GI'S) AND THE REII.!!CKS~ DIP EETdEEN THE 'MO
.:. . ::::E7 SITUATON WI""i PSGARD 'IO THE BASES FOR SEI.ECTICN OF SSE '
Jt. .
The Safe Shutd wn Eardqua'<e (SSE) is that ear'thecake which is based en an calcation of the'=aximu:a earthquake potential in the region of tne nc lea: ;cwer plant site. The evaluation is based on a censideratien of the regional and local geology and seismology and en the speci'ic c'-=-=--= 'stics of the material on which the power pir.: will he 1 cated.
Fc: sale =i:n cf the SSI, Part 100 Appendix A describes two distinct geci:gi:z.~.'se' -1:qizal situatiens within which sa:auhat different p:ccedres are re=i ei f:: determining the SSE: (1) Wien the seismicity can he re'a:ef : red:gi: struct=es a .d/or capable faults and (2) when se'*-icity :z . :: :e =' =-=* to geologi: structures and/cr capable faults.
Situati:n One: When Earth: ake Generating Structure Can Be
( A.
I s..: : e:
":e first si:catien, in which seisticity can be related to
. gecl=gic st :::=es a.d/or capable faults, is more typical cf .he western Onited States, which is a region of cenplex.
ge-:1:; and h gh seisaicity. The following four steps represent c=:e . s"'" ; : cad =e for establishing maxi =u:n SSE acceleration 4- --'s si:ca :.:::
- 1. 2,et. 3.a icn cf tectenic structures. An assessnent of the eart quaxe tu.s cry of the area m wnich a nuclear power plant will be l=:ated is mm!e. This assessment includes the i frecuer_ of ec:urrer.:e, and the maximu= earthquake of i rec 5:5. "Ca.snic structures in the area that have associated l
.seis-i:i:y are identified. By definition tectonic structures are larga scale dislocations c: distortions of the earth's crust and may or may not encccpass capable faults. For exa=:le, the San Andreas fault zone of California and the Cincinna:i Arch are both censidered to be tectenic structures.
One represents a rupture of the earth's crust, and the ot'.e:
a large flex =e. If the maximt:n historic earthquake occurs en a tactenic structure either near or upon which the power riant will be situated, the eartheuake is treated-as though in took place at the proposed power plant site.
t
( \-
Enclosure "F"
(
A" TACE 2Nr I
- 2. Co e=-ins:icn of Ca:able Faults. By means of historic records an: secioc:.cs.;. Investigations, capable faults which could c:nceivably affect the site are identified. Tahle I fran Sec icn IV of Appendix A to 10 CFR 100 presents a minit :n fault
-le.g:h versus distance frem site, for censideration of the I fault in establishing the SSE. For c? hie faults, the These characteristic:
characteris-ics of the fault must be known.
are deter =ined frc historic records and by geologic investicatic:
they in:1cde the largth of the fault, acount ard nature of displacacant on the fault, physical properties of rock arx! soil ass:ciated w' h the fault, a.ad infocnation en past mcveents cf the fault.
~
- 1. Ce e=inati n of maxi. :n earthouake. Since a fault does not ir- = y rup:=e along its entire length, an effective rupture leng-2 rus: he establisned. Current staff practice is to ass == -= approxicately 40% of the total fault largth will be i-r:lved in any single event.
Given this rupture length, an earthcuake c:mgnitude is
( dete=ined for the fault. B:pirical relationships between ruptre largth, earthpake magnitude, and displacenent have 2:e relationship bee. faval ped and are used for this purpose.
desnl: pef en AEC contract by Bonilla ard Buchanan (1970) f:: this purpose is most widely used, although others (Alger=iss:
(iiii a=5 A:hraseys and Tchale-Jco (1963)) are sometimes a:ploye:.
This zti== earthgake has al,ays been larger than the maxi =c:
his:::i: ea:thgake associated with the fault; however, the caxi== historic earthgake would be used if it were the larger of -le :w.
2.e maximin
- 4. Dece "-stica of maximum accaleration at site.
ear =:;ua.<.e intenst y is assc::es to occur on the portion cf the fault or tectonic structure closest to the plant site.
Give . One earthcuake and the distance to the site, the a::sleration at the site is deter:nined using attentatien Cther relatic ships developed by Schnabel armi Seed (1973).
relationships (Hof= ann (1974), Ecusner (1965), tenevan (1973))
are also scsetimes used. The largest acceleration resulting a: :he site from the earthcuakes on the various capable faults is then used as the maximum vibratory acceleration for the SEE. Appendix A regires that this acceleration be at least one-tenth the acceleration of gravity (0.1g).
. Enclosure "F"
1 i .
- (
A D .CE: M I J
- 3. S ' '"*' i ~ Two
- When the Cause of Historic Earthcuakes Can=ct be c.eia:ed to Known Geolecic Structures
?.e sec=nd situatien, in which seis:nicity cannot be related to R geci:gic stru:ture and/or capable faults, is typical of the eastern United States, where it has gsnerally not yet been possible j t: relate seici=ity to tectonic structures or capable faults in l
=any areas. In this case, tectonic provinces are used in the
' establishment of the SSE. Se f:llcwing steps present current staff pr:cedure in this situatien.
- 1. Ceteri. .arien of tectonic crovince. A tectonic province is las ced.nnd in Appe.c1x a A) a region of the North American c:n.i.nent characterized by a relative censistency of the gec1:=ic s::::: ural features. There is no definitive gener="y ac:epted identification of such provinces in the
\, Units! : ='=. Several province maps exist (by . King (1969),
j Ea-diey (352), Hadley-Devine (1974)) and are used for
( general guidance, but the basic detectination of such a p:cvin:e, if recuired, is made on a case-by-case basis.
- 2. Dece=ination of maxin_n earth uake. Since in the second situa:.:n we are concerned witn a region in which seismicity is n:t related to known gecicgic structures, the maximu:n his :i: earthquake of the region is treated as though it could oc=ur r.ywhere in the tectenic province (i.e. , at the plant site). Geclogical evidence, a high level of seismicity or a shc:t historical record, may dictate the use of an eardgake intensity greater than that of the maxi:m:n historic ear 2 quake of the tectonic province.
! When an adjacent tectonic province has experienced an earthquakt l
greater than those of the tectenic province in which the power
- pla .
- is to be located, the maximra recorded earthquake of the adja:ent province is treated as though it occurred on the border of the two provinces at the point closest to the power plan: site. De effect of such an earthquake on the plant site is de: ermined as described below.
- 3. Detecination of maxim:n acceleratien at site. Most historic
~eartn=uases in the eastern I.;nited States are recorded in terms of M:dified Mercalli intensity (Itc!), and the maximum earthquak for the site *s also specified in Is. A =axi:m:n acceleration
( - is derived frem this intensity. A nu=ber of correlations of l acceleratien and intensity have been developed b~y various auth::ities. Cocmonly used is the relationship developed by Trifuna: and Brady (1975). A relationship frcm thumann (195;)
is also scuetimes used.
w ..
' AmCCE.TI For maxi = earhtge.akes presced to occur at the border of afjacen: tectonic provinces, the intensity of that ear 27.:a'<e is co:: erted to a Richter magnitude ustg relaticnships developed by P.ichter (1958) or Nattli (1974).
Given this .ag .itude, the maxi == acceleration is calc 9mM in ce sane =anner as for earthcuakes associated with krawn I- - faults, usi.g the sana attenuation , relationships.
In de case of the Charlesten SC 1886 and New Madrid PO 1311-1312 earecuakes, the recorded intensity contours are used _c dete=ine the intensity at the site, and the sita accelerati:n is detectined fren relationships such as Trifunac f
a-d 3:ady (1575) or Nem.n (1954).
I' . 55I IS C ??_5:O'5I SPECTRA
~ '="#-= Se !!I precisely for engineering design purposes, the maxi:m=1 h::i= ..a1 g::u-i a:: ale:ation associated with the S5E for a given nu: es: p es: plan: si:n is established by the a;plicant and a;croved by the 2C. ~.f:e: ag:eeing en the maxi. um ground acceleration for a given si a, ?.agLa. :y Guide 1.60, " Design Response Spectra for Seismic
( Design cf Nu laa: ?cwer Plants," scaled to the SSE acceleration, is
$ no: sally used f:: es.ah11shing the free field ground vibratorf motion associated with :he 55I f:r the ' site. T..e applicant also has the option of usi .; site-fa;andant design respense spectra in place of those given in Reg-da_::7 Gnde 1.60.
- It is ;ciated c= in ?.agulatory Guide 1.60 that the acceptable design res;cnse s,.ec== p::cedure for nuclea: ;cwer plants is a procedure deva _1. pad as a result of two statistical studies of response spectra f:s: as: ear h::akas. A stumary of the two studies and the rec =d..andai design procedure is contained in the paper entitled " seismic Design 5 ectra f:: Unclear Power Plants" by Nathan M. Newnark, Jch: A. El=e a-d Kanwar K. Kapur (.EI, Journal of the Power Division, Neve=e: 1972) . In a study by John A. 31t=e, a total of 33 different
.eartir.:ar:e rec::ds were considered, with the peak ground accelerations f:: :..:sa en:2;;akes ra. w ing fran 0.11g to 0.51g. A total of 23 records were used in a s::dy by Nathan M. Newtark, with the maxi == ground a::e' stations in de horizontal direction ra.ging f:cm 0.03g to 1.25g.
Resp:nse spectra were calculated for each record for varying degrees of dz pi..g and mean spectra were derived frcm statistical analyses. The resi_s f the two studies were cccbine:1 and a si.gle spect:t=t was rec = a-ded fs: design purposes, usi : a nean plusAne standard deviation as t.h.e. de_si.gn s ect:t=t probability l.e. vel.
h
( \
Enclosure "F" !
.._)
4
. l i( .
ATIACEm I It can he seen therefore that the characteristics of the actual ea:27:a' <e :e::::ds used by 31t:me a-d Newnark, such as their frequency t
c:ntent and the d::ation of their metiens, are inharent in the design spectra f:rd in ?.egulator.y Guide 1.60. Since a nunber of conditions 1 i
a a e. ialeped in 7.egulate:y Guide 1.60, the detelog:ent of c:cothed sita dependent spectra f::s a single earthgaake record would be less c:nsa- ative be:ause it :cvers only one cordition. ,
It is i..,_ rtant to e-lauize that the peak ground accelerations of the eard? cakes used in the Blu:e and Newra:k studies are maximum motions ej:: served i~Jepe . dent &the duratioq_qf. creund__s_haking_at those levels. ~
Ois :.n::: duces a degree of conse:vatism into design spectra, ~since it is .all establ'*ed in earthcuake engineering that the most damaging g::r.d reticas are tycically those levels of motion (fregaentIy of lower a p'i =de "= . the pea'<s) that are maintained for longer periods of time.
I!!. ??J53 f 5_ ... OF %IIE REGARD 70 '5E BASES EUR SELECTION OF THE CBE
-'he C; era _i~; 3 asis Ea-dq2ake, as defined in 10 CFR Part 100, Appendix A, is da: en-27.:ake .ich, considering the regional and lccal geology ard seir:1::;7 a-4 s;e:ifi: characteristics of local subsurface material,
.(.s c: ld :sas:nahl he expe:ted to affect the plant site during the operati~. g l .
- is that earthq ake which produces the vibratory life =f de pla.:.
grou.i :::i:n f:: whi:h those features of the nuclear power plant necessary fo: c:ntinued :; ara:icn without undue risk to the health ard safety of the puh'_i: are designed to reain functienal.
De mini un value cf de acceleration level for the CBE is currently spec _-fied in Appe.-din A to 10 CFR Part 100 as at least one-half the acce' ara:icn de:erri .e5 f:r the 55E and this is the value normally s,.ecifiad. T:r si:n,s ..ot in highly se:i.snic__ areas, thineaguinsment In these areas, earthcuake o' is.c:. :11in: in .he selection of CBE.
~
~
accela a icns st.i:iTah reasoriably be expected to occur in the life of the plant will usually be less than one-half the SSE acceleration.
Fcr si as in hi:hly seismic regions (mainly in the western United States' a c::tpleta description of the CBE is developed. Geologic structure
- ca.sh?e
- faults :: tectonic previnces with which historical earthgaake ac:i.-i:7 has hes. associatti are considered as possible source mechanisns.
P "-i "y calculations such as those described by Algermissen and Perkins (197 , 1976) can be used to help estimate the acceleration leia_1 that can reas nahiy be expected to affect the plant during its operating life (ap;: nisataly forty years). Bis acceleration may be greater than one-hal th-s 35I a--=' =-= Mon for sane sites.
k I
Enclosure "F"
i 1
! =
j~ s l(
i s 1
j ATmC.Trzr I
- To define the C5E precisely for engineering design purposes, the !
i
/ sa=e res;c..se spec::= as that for the S5E is used, scaled only as to acOS1armtiOn. ,
l
'2.e ; i'.:s:;hy behird establishing the GE to be not less than one-half g 55I is as follcws: ,
! 1. '2.e stress level in 'the safety-related struct=es, systems, ard
, c _* ents is all:wed to reach yield level when the plant is subjected 4
t= a 53E in c=3ination with other applicable loads, provided the i ne:essary safety f=ctions are maintained. For the CBE, all structures, i
syste=s, a-3 :=;=nents necessary for continued operation without
=due :isk t: the health and safety of the public are designed to
- - = ' : f=cti:nal a-4 vell within elastic limits when subjected to the CEI in c==f..a '-'. with nor=al operathg loads. Se usual range of
=~ *:.a:1e elas i: stress is from 0.43 to 0.6 of the yield stress.
Che ch:=e cf .he 3I to be cae-half SSE is consistent with the ra:i: Of the yield s=ess (alload for SSI) to the allowable elastic s= sss i C =ved'f:: GI) .
- 2. Appe-li A : _. we. cart 100 requires that the nuclear pcur ~ plant be
's sh:: d:wn if the vibratory, ground mction exceeds that of the mE.
tis ragi: sten: 1-di:ates the advisability of an CBE which is large encugh so tha drin; a streng earthecake all the nuclear pcwer pir.:s -in a large geographical area are not shut down and the public lef: vi:ncut ele ==i: ;cuer.
(It shed.d be :::ed :a: another applica:ica of the GE arises when seis=ic effects ara c=sida:ed in cccbination with other natural phencnena. For exa:ple, in dete=izing the design basis for certain structures of the ultimate -
heat si ~<., he CEI is c=sidered in erbination with waterflow based on severe his Ori:al e ents in the region of the structure. 'Ihis applicatien of the CEI is basi:Cy not ge=ane to the issues of SSE/C3E interrelationship discussed in this pape-j'.
IV. . . ::_ 2 5 2. . .c.- CN *T"E REGARD 'IO ENGIh:.:.rdNG DESIGN !E"lCT.CG'l '
. '5: FC2 5EI5:CC OISIG.i CF LWIS Si;3JECTF.D TO SSE AND CEE the a:tze..ed diag:s= :tlines the doc:: tented seismic design methodology being : sed today in the engineering design of nuclear power plants rela:i.e t: t':e 55I and GE.
~
- A;gendin A to 10 C??. Part 100 requires that vibratory motions of the S5I a-2 .3I be defined by res;cnse spectra of the fce:x*ation level o f - .e .s. .t . '2.ese s; metra are cbtained by deconvolving the,three-( di-=-- --= =pectra specified in Regulatory Guide 1.60, which are Enclosure "F" 4
- - - - - - _ __---- - - _ _ _ _ . _ _ , _ . _ , - -- ,,,,-,-,-,.,-..m..- e,,. , . , , - - -. , . , , .r-.
. o
( ,
ATmCn'.4DT. I f:ee field spectra. Fc: these free field spectra Regulatory Guide 1.60 specifies tw: hori =:a1 si:ectra having the sa=e amplification factor el a ' c ti:al spe:t::= hrling amplification factors which range frcct
> =e zare :: 1/3 of the a plification factor used for the horizental spectr.
- :a p:: cess of fe :::1ving, ti=e histeries ccmpatible with the free k fiC: 2:e==a are generated. Engineeri .g analysis of the site is perfecte
- fi-i tne c=.pa:ihle bedrock motions. These bedrock motions are used in' sci ~ s====e interaction to find the plant foundation motions. Se f: -izti:n level =cti:ns (represented by spect:a or time-histories) are in :== csed in de a: ::tural medel of the nuclear power plant to find s= esses a-i displa:ce :s in various elcents of the plant and the flect des .g resp =se spa:=a to be used sdsegaently in the design of floor-
. s-w.::ed egnigen: a-i c=ponents. Sc=etimes structure-to-structure i :tra:t.= s .z".- sis ::st he perfo=ed if two or more structures are in
- ':se ;::nizi:/ Of en= nhe:. This anslysis may further mcdify the i f:c-in i= 1rtel desi 1.put into a nuclear power plant structure.
- 7. ::GTE :55-'Z5 CT.:.r -2 TING SEI.EC*PE OF TEE SSE O.ere is ;a::ity of data for use in detarmining the mag'nitude
(. 7..
ef de +=:=7:ake associated'with a given fault length. certain
(
types cf ez:=g:ske scurce nachaniras (such as overthrust faults s .f r e ,- = - = = 's - ; resulting frem cc:cressional failures) ga e:al'.:- hr e hi:'.er effective stresses across the fault, landi.; :: 7:na:a: energy releases and thus greater magnitudes than :cu~f he expected for a strike-slip or normal dip-slip-fault. -=-= '= s::me feeling that the more frecuently used' expir -=' -v =-" . ships are dcninated by data for the reverse-d
., .; a '= '-e, ~.eading to unduly conservative magnitudes for other i121:s.
'ese a: pirical relationshias contain an additional conservative cias due t: difficulties in determining the rtsture lengths of 5:lts ass:ciated with past earthquakes. Since po:tions of the fau.lt =r,- he h= led beneath laMslides or a deep allu7ial soil -
.rfer, as- s es of rupture lengths associated with past earthgaak:'
. :end :: :s s'.::ter . than the actual rupture lengths. The ra atic .si.i;s develeped using these rupture lengths will then t+1. conse:vative (la:=a) earthquake magnitudes for
. =11-esta:lishec raua.c . Length and postulated rupture length.
- 3. In establishing the maximus acceleration for a site, there
' are differences of opinion among seismology experts regarding
-he naxir_ possible near-fiele (close to the fault) acceleration
( g
- .st cr. he generated. nat is, there is same physical upper limit en near-field acceleration, regardless of earthquake magnitude: the ~ disagreement is on the value for this upper limit. '
4 Enclosure "F"
] .
,1 . -
- ( t ATI: ACE'.c2 I C. Va y hich isolated ceak accelerations have been measured but are not used in scaling and applying Regulatory Guide 1.60 design spectra. These maxima accelerations are usually of short du atica and do not have time to build up dataging energy.
l Near-peak acceleraticas maintained for longer duraticas will y be more danaging. Tne duration of strong notion is particularly significant for- design stresses near the yield point (as is the 2
case with the SSI).
?
D. Cetermination of tectonic provinces is a point of contention between the staff and industry. In general, staff practice is to use relatively larce tectente crovince-; to ensure ecnservative design. However, there is a great ceaTof tas..nical arg=ent frcm industry to use s: aller tectonic pr::L res.
I. e.ar=inatien of =axi=ca acceleration frem intensity creates d:dfinuities in that intensity is a subjective measure of ground n:-i:n. Alsc, direct information about the duration of strong n:. ion is Ics: 2.en using intensity as a measure. Tnia is a
( difficult problem since al=est all earthcuake data for the eastern I Crited States is in terms of Mcdified Mercalli intensity.
VI. J?~r-20 253~'Is WI'IH RESPECT 'IO ??LATIONSHIP EETdEEN CSE 72:D SSE A. be CEI can he censidered to be not safety-related and therefore apprcpriately chosen by tha applicant as a matter rf econc.ic judgeant. In such a case the guiding factor f:: the applican: to choose a certain g-level for the CBE would
- e the ecs- of shutting down and restarting the plant after an
.=. . cce:rs. If the applicant chcoses too low a leiel for the "EI, he sight incur heavy costs of shutting down and restarting he plant several times during the life of the plant. Further, if the plant generally has an inadequate seismic design basis ha ma/- incur extensive costs due to eartheuake danage to power ganeratica and transnission egaignent. (As an example damage t: utility systens in the San Fernando eerthqaake ancunted to ahcut 5100,000,000).
Regulatorf Guide 1.12 specifies that various seisnic instrtments he used to ascertain whether the cas has been exceeded during an earthganke. When instrinnentation shows that the peak acceleration er the response spectra experienced at the fourdation of the containmet building er in the free field exceed the CBE acceleration level or response spectra, the plant is, required to be shut down pending per:nission to restree operation. To I determine whether or not the plant can safely rescue operations, Enclosure "F"
- _ _ ~ ~ __ * "~ ' :
__, , . .. -_-_-_1
. z 1
3 .
t -
l( ATDCEiENr I l l
l visual field inspection of safety-related items, which may include n ndestructive testing if need is indicated, is i ple en:ei a-4 the measured respenses frem both the peak-j re::rdin: a-2 at:eng-motion accelerographs are co: pared with 1
. :se ass =ei in the design. Ccnsiderable cost and time is j inv:17e5 in dis precess of inspection ard verification, which i; includes such iter.s as (1) inspecting the piping for any mcvement c: rdbi.p, (2Y inspecting the structures and ecuignet for !
a::c-=' displacement, (3) reviewing the recorded data on plant i j
- arating pars eters for any ahner2al operation of egipment during even
- , and (4) cenparir.s the recorded responses with ,
i de design basis. To ccmpare the responses, the earthquake data m:
he re::iered f:= the inst:t=ents, digitized, and used in the ec=p calyses. Us ='17 technical specifications of the licensees are 4
- sed := ac:=;11sh the inspecticn and ccmparison of the responses. .
- 3. Sc== -=- e s of the staff feel that the CBE should be esta
=ased en p:dabilistic methcxis, rather than beim specified as
' = "-=- #-- cf the SSE. Sis cculd be done by chocsing an earthqua: ii f
-r--=- e inte ral to obtain a sufficiently small pecbab l ty o
(
4 ec:eeds :e d::ing the cperating life of the plant, an$ mcdifyi.g i \ er eld =inatin- C .e de recuirement that the CBEisbe inat least to one-half the S5I. difficulty with this aporoach how select an a.r..,.inte re:urrence interval. Also, there is very little
' earthq ar:e data for the eastern United States, creating problems
- in de
- =-ini-: de earthquake seierity for a given r zurrence inte. val. '2.ere is scte feeling that the historic records for l
Cord .'. eri:2 are too short fe: use in predict.ing geologic phenar.!
Se s.aff has a wide range of opinions on how such an approach w '-= -=af. Sere have been cases for which an CBE less
( s.c- n c a-ha*.f 55I has been accepted by the staff as the earthquakt
. ich c c.'d :easonably be expected to affect the plant during its:
1
' :pera ing life.
. Sc=e fee' that the ministm CBE should be It setisscmetere their in th
~
a e :f ene-dird to one-fourth of the SSE.
- e .ti:n dat, with an CBE equal to one-half of th SSE, the i
. J-design :f =cy ccaponents in a nuclear power plant is severned by 1:mi c=hinations involving the CSE rather,than by conbinations While reducing the CBE/SSE ratio might which in:lude the SSE.
i Y.: s=e :ases, reverse the situation, in other cases it will not.-
! What reWy controls the design are the choices of load factors,
! leadi .; embinations, ard allowable stressas used. The CBE is or
- ne of de loadings but in many cases the ecmbination of nonseicn 1:sds'has the controlling effect on the design. To overecme 4 ( '
- his p::blem, the load casination questions and allowable i s ressas would have to be reiised to give different weightings to the loads involved. Differences of opinion exist within t'.a staff on whether or not th'.s should be dene.
Enclosure "F"
( ~- ATIACEM I VII. .Iy?O.NT I5SES WIG T?.E SEISEC DESIGN MESC::OI.CXTl A. Pagulatory Guide 1.60 is based only on limited data.
here is recm for.imprcvement in the shape and a:Plitude of the free field design response spectra as a- function of =agnitude and distance. ,
- 3. Pagulatory Guide 1.61, which gives the da= ping values to be used in the seisaic analysis of the nuclear power plants is based on limited test data and ergineering judgment. Room for impreveent also exists here, because a slight difference in da= ping values can make a substantial difference in the resultire stresses in various ccrnponents of the plant.
C. ?sculatcry Guide 1.60 specifies that the same amplification factors he used for the two horizontal spectra, and gives a separate set of a=plification factors for the vertical spectrtra.
2ere is see feeling that the amplification factors for the seccr4 horizcntal spectrum and the vertical spectrum can be -
less than these given in Regulatory Guide 1.60.
(' s D. ':he shape of the spectra as given in Regulatory Guide 1.60 is net suitable for soil licuification analysis, according to scrse .
views. A site-ccrepatible spectrum is more appropriate for ,
thase analyses.
E. 3 oil is.a nonlinear material when subjected to the dynamic
=ccien cf earthquakes. At present most of the soil-structures interactics analyses and site analyses for deconvolving the i=e-histcry =otions are based on the elastic (linear) properties cf the scil. 'Ihere must be a substantial improvement in the state-of-the-art before nonlinear analyses for soils can be incorporated.into the design methodology.
- 7. In carrying out the structure-to-structure interaction analyses, the major problem encountered is the sa:e as in soil-structure interactica analyses, i.e., the elastic rather than inelastic properties of the soil are used in the analyses.
- 3. 'Jhere are two usual kinds of analyses which are performed on structures subjected to earthquake motions: (1) response spectrtra analysis and (2) time-history analysis.
The major question raised about the response spectrum technicue (x is how to ccrnbine the modal respense and the effects of three ccm;cnents of earthquakes. Regulatory Guide 1.92 specifies procedures based on the present state-of-tFa-art.
Enclosure "F"
. .u ATmCS?.cc I If i=ehistory technique is used and a simultanecus analysis of the three directions is perfor ed, the problems of modal c:rbi .atic and three ccupenents of earthquakes are eliminated.
I:wver, the procedure regaires a eccputer with very large meccry capacity, therefore it is not being widely used at present.
E. Regulatory Guide 1.122 specifies procedures to broaden and sEooth
- b. the ficc: resxnse spectra. The guide also prevides the procede:
to ccr.line the three floor res:ense spectra for a given direction The questicas raised about the guide include the artcunt of broadening aM the procedure to ccr.bine the three spectra for a given direction. The guide is ' cased on the present state-cf-the-art aM can be irproved in the future.
VIII. ?.- -ISCIS 10 -- :=-- '00, Appe..di: A, "Seisnic and Geologic Siting Criteria for Nuclea: Pr * # =--*"
?.eguia :rf C-cide !.12, "I .strt:nentation for Earthgaakes" Regulat: y Guide 1.60, " Design Response Spectra for Seisnic Design of Nuclea: Pcwer Pla ._s"
\
Regulat:rf Guide 1.51, " Camping Values for Seis:uic Design of Nuclear Pcwer Plants" Regula::rf hide 1.92, "O=bining Modal Responses and Spatial Cc=panents in Seiz:i: Respense A.alysis" Regula::rf hide 1.122, "Flcor Design Respense Spectra Development for Seis-i: Cesign of Ilcc: Supported Equi;=ent or Ccmpenents" '
S. T. Alge=isse ., " Studies in Seisnicity and Earthcuake Damage
, Statisti:=," U. S. Coast and Geodetic Survey (1969)
S. T. .:C_;cissen and D. M. Perkins, " Techniques for Seis:nic Zoning:
- 1. C-e.e ri C::siderati:ns and Parameters," Proceedings of the Inter .a :.:na' C:nference on Microzonatien, Seattle (1972)
S. T. Alger:issan a-d D. M. Perkins, "A Prebabilistic Estimate of Maxirnrn
- c: ale:a:i:n i . 7.0ck in the Contiguous United States," U. S. Geological 52: ey pen File Reper: 76-416 (1976)
N. n. A ::ueys a-4 J. Tchalenko, "Documen" tion of Faulting Associated wi:n rar- .;:akes," Part I, Dept. of Civil Engineering, Imperial College
~( \
of 3cien:n, LeM en (1968)
- Enclosure "F"
W
(
AD.CKE?f I M. G. Sonilla a .d J. M. Buchanan, " Interim Report on Worldwida Historic Surface Faulti.vg," U. S. Geological Survey Cpen File Report '
(1970)
N. C. Conovan, "A Statistical Evaluation of Strong Motion Data Including the Fac:cary 9,1971 San Fernando Earthgaske," Precee51ngs, Fifth World Cen'erence on Ear $qJake Engineering, Rc:e (1973)
A. J. Eardley, "Structu'ral Geology of North America," Earper & Row (1962)
R. B. Ho** .n, " State-of-the-Art for Assessing Earthgaake Eazards in the Unitsi States, Report 3, Facters in the Specification of. Ground Motion for resign Ear $q;akes in California," Miscellaneous Paper S-73-1, U. S. Arf Engineers Waterways Experiment Station (1974)
G. W. Ec Iner, "Intensirl of Earthgaake Ground Shaking near the Causative Fault," Proceedings, Third World Conference on EarthTJake Engineerin:, :~ew Zeala .d (1965)
P. S. Ki .g, "2e Tec::nics of North America - A Discussion to Acconpany Tect: .ic :'ap cf Neri. A=arica Scale 1:5,000,000," Professional Paper
(\
628, U. S. Geological Sc:vey (1969)
F. Net =:a:r., "Ea: S ?
- 2ke Intensity an$ Related Ground Motion,"
Universiri of Washin:::n Press (1954)
O. W. Nuc li, " Magni _ de Recurrence Relation for Central Mississippi
, Valley Ea:27 akas," Eulletin seismological Society of America, Vol. 64 (1974) .
C. F. Ri :ar, "Elacr.tary Seismology," Freeman & Co., San Francisco (1958)
P. B. Sch. thel and E. B. Seed, " Accelerations in Rock for Earthgaakes in the Western Unitsi States," Bulletin seismological Society of America, 7cl. 63 (1973) i M.D. ?-if.:.ac and A. G. Brady, "Cn the Ccrrelation of Seismic Intensity Scales W12 Peaks of Recorded Ground Motion," Bulletin Seismological Society of America, Vol. 65 (1975) i 1
4
. ( s Enclosure "F" -
N 4j c %- I g, 0, ;,, n "s 4, -
,J, w ns3 Lou 3 -
. . 3 j g
s,
%t .c a 4 4 9 x d
s J
x 4 %-
h su S%% % ~.M9 1T 57,
~
i #i xd aehn <
4- # .t v0,, U $~ $ y, 9' E' ? 4 2 ea, 4 es
= 5 ,
J <d %
w't q
.i;
=4 o +3 &~-
)
g q%v, ir -5<
4 N .
e g --
, ., , o 4'- . = ~
.b e 4 ~ .,M sr M. ,2 3
= r %%.- r--g<5
.e g m.
4{Mal 4
% , 4, e
<M as
.f,. w*
W- e; -
3 3
T. ,ss.t44 e
4 9 ,,
1 .o, 5 .
g .
.l . 'v -i! T ' ~ 45. 4 42M i : Es 'I.;:: 4
.!A.Dg E 95 id $ 'I 5 t44 hi M 'i 'i S:
i
, a a
'T e , e 4 i
e 4 El 3.
T
" .4 .
% T n n . . ..
.~ '
,$w 'y . *
)M
?, . J. .
~
5 ~4 t >-
w - 1x s 1 45 s * .
2 d
t h[ r -
)4 3 n y .
a 3 49 .
\
3 y q . . -
)
z +e y
a 4 .
? -Q ? n i '
JePtw ,, g 5 %: 4
.v z u:n<
k%n<
. ~s 5 Q. t 51% -
l1%e t;d g% i!- 1 y < va{ N ib S g, .
4 UN99W - 3 -',h
~ -
%w $, h
$ Ag c%@l 4 4 -
8
.g
~!?
3 -
> c L 4g 4
u-.
g w
.j D .m
- o. .
E C
% C % .q O *c 47 3 Is H #7 -
N a'$eo.".: .: 6 3 4-a w e I- .:
o .*. .E. ei wN.
- J f
. . u a. u. .
.? a,g e h e$
C d 4U$
OW -
U 9
~4.a W 4
- kn .s :f N
. e -
W Cl . Cd
.d
)
O
- -