ML20212L953
| ML20212L953 | |
| Person / Time | |
|---|---|
| Site: | 05000000, Diablo Canyon |
| Issue date: | 04/30/1978 |
| From: | Trifunac M SOUTHERN CALIFORNIA, UNIV. OF, LOS ANGELES, CA |
| To: | |
| Shared Package | |
| ML20150F500 | List:
|
| References | |
| FOIA-86-391 NUDOCS 8608250313 | |
| Download: ML20212L953 (7) | |
Text
9
(
t CC STENTS ON SEIS:.11C 11ESIGN I.EVEl.S F0tt DJ Alli.0 CANYON.SITI: 1:: CAI.ilonNIA by 1-1.D. Trifumic h
i April, 197S 8608200313 860801 PDR FOIA HOUCHS6-301 PDR l
-gg 9;to-yI.-f
- -114 it
e
- +
(
(-
i.
t
)
8.
t The following conunents deal with seismic design criteria for the Diablo Canyon site in Cali fornia and. represent a brief suc::r.ary o f my observations and preliminary conclusions which are based on miscellaneous written :natoria'l and on a number of meetings during the period starting in the summer of 1977 and ending in April of 1978.
Infornation which I had on certain aspect s of this effort may be ineceplete. The general picture and thu sunziatics of the current status of this project nevertheless seem adequate for the following concients and reconunendaticns.
Fluch has been written about detailed aspects of scismic design criteria for the Diablo Canyen site and it would be impractical to W
address again numerous points in detail and completely.
Rather, I will attenpt to present an overall surr: nary of what 1 belicyc to be unresolved prehtens at present, and what might be possible avenues to resolve then.
MJJr w General Cemments en t he Current Inputs and Criteria for Seitmic Design 1.
Cencra11y arcepted assumption appenrs to be that the SSl! on llosgri fault opposite the plant site should be in !!= 7.5 carth-quake. This magnitude, recocunended by USGS, has been deternined mainly on the basis of the possible length of faulting on the llosgri fault system.
2.
Since ela 7.5 at a distance of 5 - 10 L:n from the site leads to large peak neccleration'(about 1-g) considerable effort has hven devoted to t he an.nlyses which a'ro desi'gned t o show that s
s
?
_. =. -,... _., - -,_,_.
~
(
(
2 j
these Jarne an.plitudes can and may be reduced through considera-tion of the following phenomena:
a) Scattering and diffraction of high frequency waves from the foundations of different plant structures has been proposed as a vehicle to justify reduction of high frequency spect ral aiapli tudes (T effect)'. The manner in s.hich this reduction has been affected requires unrealistic assuuptions, for ex-ample, that foundation is rigid. 'the manner in which this assumption is introduced into analysis if often one-sided and considers mainly only those con.cquer.ces of the physical phenomena which lead to reductien of spectral amplitudes.
Other consequences of this phenomencn, for crampic, torsional and rocking excitations of fcundation which may amplify the 7Adi."
structural responr.c have been, so far, either overlooked or o
treated inadequately.
This has been achieved by uti atien of dynamic models for analysis which are so defined thdt only an incomplete physics of the l'roblem, i.e.,
scismic excita-tion and he structural response, can be considered.
b) The term " effective peak acceleration" has been int roduced suggesting that the st ructure will "sec" something smaller than actual peak accelcration. Though such approach may be useful for cart hquake resist ent design of ordinary structures by means of the response spectrum technique, the term "effcc-tive peak acceleration" has not been defined in a way that would enable the derivat ion of consistent re: nil t :. h)t severa l s... r -
,...----:,,_.,.+,------,,,
(.
(
2 different experts in the field.
Since the procedures for scaling Regulatory Guide 1.60 spectia are based on "inaximum vibratory ground acceleration" (as defined in Appendix A) this departure from routine design practices males it di f fi-cult to evaluate the number and the nature of the consequencer which could result from such an approach.
c) Ilypocentral rather than distance closest to the fault has been used to evaluate peak and effective peak acceleration.
This assunption implies certain angles of approach of scistaic wave energy. These angles of approach should then be con-sistent with the extent to which "i effect" is allowed to influence the spectral ar plitudes.
1.ittle or no at;ention scens to have been given to mutual corsistency of inese
~
assumptions and in some cases, inconsistent assumptions have been utili:cd.
For exampic, deep hypocent.cr would sacrease the distance at tthich peak acceleration is evaluated, thus reducing the estimate of peak acceleration amplitudes. This would, however, aise imply that the unves arrive towards the loundation almost vertically.
In consideration of "I effect" houceer, hori: ental dimenr. ions of foun lat ions appear to have been used implying horizontal incidence of
- t. aves, d) The large damping equal to 7*. has been adopted for dynamic response calculatio s.
Though the apparent damping for the complete soil-st ructure system, subjected to earttuguake excitz. tion may he much larger t han 7t, inadequat e basis lias
.heen presented to justif); 7'. dampityg in st ruc t ural sy s terr..
g
.g y-g ampeg 4 - = * -
4pm m
(
c t
only.
Selection of too large structural damping coupled with only two-dimensional or simpic three-dimensional analysis
~
of soil-structure interaction can lead to unreliabic response estimates.
3.
At least three scismic rish studies have been prepared to cr.timate the probability of exceeding the selected desij;n criteria at the 111ablo Canyon site (Illume, Ann am! Neuuart, Anderson and Trifunac'..
These studies have produced re.eults which, in some cases, differ by as much as two orders of magni tmle. Concurrent. with the com-paris ons of these studics, considerable effort has been devoted to different details in the nethodology caliloyed in these calcula -
tions.
Little or no explicit effort and discussion has been de-voted to the models of scismicity which arc esscatial input into M
such calculations, even though this may reptrsent the most impor-tant contribution to the discrepancies anong the results of different studies.
In some extreme cases (e.g., report by Blume and Kiremidjian) claborate work has been carried out, apparently in vain, to show that a particular metho.d for scaling peak accel-cration (Trifnnac, 1976) supposedly leads to "too large" estimates of peak acceleration irregardless of the fact that those results of Trifunac (1976) have never heen used and do not represent a basis for the derivation of scismic risk it.cdels by Anderson and Tri funac.
In the report by Ang and Nevanart, substantially smalle'r than average scismicity has been assoned near t he site.
This saay lead to an underestimat e of actual ri:.k.
t:
..-,,,.,-s.'.~-.----.,,.
(
(
\\
l.
s
?.
l(ccorMienda t ions A.
Ground flotion:
1.
Deterministic approach based on the assumption that an carthquake of magnitude fl = 7.5 or g: cater will occur oppo-si te the plant sit e shonld he re--cvalua ted.
This magnitu*e might be an indienter of the ext ent of geologic faulting phenomena but it is not necessarily t he most reliabic basis for evaluat ing the nature of :;trong shaking close to the fault. There are nu..creus examples in literature of sig-nificant differences between II and 1.i, for exampic, which g
g are based on-siiort and leng period seismic waves, respectively.
Often r.tudies have shown that larger carthquakes may be t heught of as a sequence of several or many discretc events which can sequentially take place along a long fault.
Fi na lly, a
the largest recorded acceleration, so far, has resulted for
-,'s Il ~ 6.5 only.
For these reasons, and from the design view-
"~
point, I would prefer to adopt :1= 6.5 on liosgri opposite the si te and not i.I = 7. 5.
2.
Near-ficId source theory (not a finite element or finite-difference nodel of the scuree and its surrotuidings) could be used in conjunction with the spectral analysir. of strong motions recorded cir.cwhere to evaluate the amplitudes of renponse spectra indepeinlent of peak acceleration cr.timates or of scismic risk analyses.
- c. -
l.,
6M 6
_r_'
m mme
o f
^
6 t
R.
Ilesponse:
1.
Three-dimensional soil-structure interaction analyils should be carricil out. This should be donc assuming that the free-Lield z er,ponse spectra fer design resul t from incident SII, SV or llayleigh waves.
For S!! and SV cxcitation, horinental, vertical and 45" incidence analysis should be considered.
Thi:
approach would offer the following advantages:
a.
The "T cffect" if present will he necounted for correctly.
b.
Torsional and rocking excitation:; will be included into the analysis correctly.
c.
The proximity of the carthquake source and the fact that the waves must likely arrive horizontally will be account cJ for correctly, h
d.
The radiation damping in the soil will i>c introduced into analysis properly so that the high value of M for struc-t u ret. would not be regteired.
t.,., y.
ls 4
lixecpt for the fact that T', damping is permi:. :ible according to the regulat ine, guide 1.61, thi s high st ructural daml ing reco:cended far the scistiiic analysis at the Diablo Canyon site has not been jus _tt-fled. l'orced vibrat ion t est (available in li.S. and. lap.in) data, where the effect uf soil-st ruet tare int eract inn an.1 di f ferent. mode of energy inl ut i nt o (In :.t r.uc t its, c i'u ri ng.ui e spe r i:.mnt. relative acco6' t ed'*fe r, uny be o f-l'i r t l e
,to, incident ea rtlupitike water, itt1301 n
use 'in' e :t alil'i;.hing t he actua l,ilampi ng i.n st ruc t ures :.n I t hei,r ce postent:.f. r sei smic r. rpopSe.c:t h.i! at innu s
..