ML20153F404
| ML20153F404 | |
| Person / Time | |
|---|---|
| Site: | Satsop |
| Issue date: | 08/29/1988 |
| From: | Sorensen G WASHINGTON PUBLIC POWER SUPPLY SYSTEM |
| To: | NRC OFFICE OF ADMINISTRATION & RESOURCES MANAGEMENT (ARM) |
| References | |
| GO3-88-235, NUDOCS 8809070272 | |
| Download: ML20153F404 (44) | |
Text
_ _ _ , , ,
-N ' ._. --
g 1 EA WASHINGTON PUBLIC POWER SUPPLY SYSTEM P.O. Box 968 e 3000 George Washington Way
- Rkhland, Washington 99352 Docket No: 50-508 i
l' G03-88-235 August 29, 1988 Document Control Desk U.S. Nuclear Regulatory Comission P1-137 Washington, D. C. 20555
Subject:
NUCLEAR PROJECT N0. 3 RESOLUTION OF KEY LICENSING ISSUES FINAL SUBMITTAL 0F S0IL STRUCTURE INTERACTION INFORMATION J
References:
- 1) Letter, GC Sorensen (Supply System) to GW Knighton (NRC),
dated June 27, 1984 (G03-84-410)
- 2) Letter, GC Sorensen (Supply System) to GW Knighton (NRC),
dated July 14, 1986 (G03-86-399)
- 3) Letter, GC Sorensen (Supply System) to Document Control Desk, dated July 31, 1987 (G03-87-236)
The Supply System has recently had Impe11 Corporation perform a SASSI analysis of WNP-3. The results of the SASSI analysis and further clarification of the elastic half space analysis that had previously been submitted by Reference 1, 2, & 3 are presented for your review in the attached document. The combined results of the SASSI and elastic half 4
space analysis provide confirmation that the finite element methodology used for the WNP-3 design basis is sufficiently conservative to assure that public health and safety will be protected during and following a safe shutdown earthquake.
With ,this submittal, sufficient information has been provided for the kRC Staff and their contracto. Brookhaven National Lab, to prepare a safety evaluation or technical eva.uation report. We will be pleased to respond to further questions of clarification that you may have.
A l
8809070272 880829 gDR ADOCK 05000508 PNU J
g,o I 4
y\
.r- i 1
l Document Control Desk ,
WNP-3 Resolution of ~ Key' Licensing Issues Page 2 If you have any questions, please contact Mr. D. W. Coleman, WNP-3 Project Licensing Manager at (509) 372-5304.
Very truly ours, !
a 0 du G.~C. orensen, Manager Regulatory Programs
- Attachtrent
! cc: G Vissing/NRC JR Lewis /BPA NS Reynolds/BCP&R EBASC0/New York M Manrique/impell i JB Martin /NRC RV WC Brauer/PP&L lit. Bryan /WWP C Goodwin/PGE :
WJ Finnegan/PSPL
! i i
f i
i i .
l
1 h I,
WASHINGTON PUBLIC POWER SUPPLY SYSTEM
. WNP-3 SEISk.IC DESIGN BASIS MODEL VALIDATION EFFORT AUGUST 1988 PURPOSE The purpose of this submittal is to provide additional WNP-3 seismic analysis results such tnat NRC question 220.13 and Structural Audit Finding Number 1 can be resolved. The principal new results provided by this response are the soi', structure interaction analy:es completed by Impell Corporation utilizing the SASSI code (Reference 1). The results of the SASSI analyses are also utilized to justify, or assess, analysis methods implemented in the alternate elastic half space studies completed by the A/E in evaluation of the WNP-3 seismic design basis model results.
t CONCLUSIONS The WNP-3 SASSI analysis results demonstrate that the WNP-3 seismic design basis model results contain significant margin of safety. Specifically, the WNP-3 in-structure design basis spectra envelop the SASSI results by wide margins at the spectral peaks, and are bounding over the SASSI results at all significant building or internal structure vibration modes.
The SASSI results demonstrate that the simplified methods utilized in the A/C's elastic half space seismic studies are satisfactory, since the half space models yield generally conservative results. Local exceedances of the design basis spectra that occur with the half space predictions are limited and are acceptable as indicated by subsystem design margins.
The Supply System concludes that the SASSI analysis results presented with this submittal are sufficient to demonstrate that the WNP-3 seismic design basis model yields conservative and appropriate seismic loadings for the safe design of all plant structures, as well as the associated seismic qualifica-tion of plant subsystems. The Supply System further concludes that suf ficient alternate analyses have been completed to satisfy the requirements of NuReg-0800 and that concerns relative to the application of deconvolution are resolved.
t s I
ORGANIZATION The balance of this response is organized into a review of the chronology, or genesis, of the basic deconvolution issue and the progression of responses leading to this submittal; a discussion of the open issues associated with the seismic half space study; and finally, closing remarks on the lack of sub-stantiating evidence supporting the basis of Audit Finding Number 1. The chronologic review is provided mainly for orientation and summation purposes, however, brief coments on the basis for the Supply System's position and actions are included in this discussion.
Following the chronologic review is a brief section outlining the application of the SASSI code to the WNP-3 soil-structure interaction problem as performed by Impell Corporation. The major SASSI analysis steps are identified as well as the scope of the WNP-3 modeling effort.
Having established the primary SASSI analysis steps the following section addresses open issues (additional information) relative to the A/E's half space model inputs and results. In addition, this section also presents the important SASSI model 3nalysis results, including comparisons to the design basis model ,esults, for WNP-3. The discussion of open half space model issues relies heavily on the SASSI model results as the basis for conclusive assessments.
The final section discusses the premise of the audit finding and proposes that the NRC concern regarding deconvolution at rock sites should be resolved since the available theoretic and empirical evidence shows that seismic attenuation does occur in stiff soils ("rock") comparable to the WNP-3 founding media. '
a s s WNP-3 SEISMIC DECONVOLUTION ISSUE CHRONOLOGY Because of the lengthy period of time over which the deconvolution issue has been conducted with the NRC, it is worthwhile to review the chronology of the issue for purposes of orientation. Reviewing the major events associated with deconvolution will establish not only the issue's orig 1n, but also the bases for the Supply System's responses, and the subsequent progression of work com-pleted in response to open NRC technical concerns. The WNP-3 deconvolution issue history is summarized by the following highliahts.
o The NRC first raised the deconvolution question in a request for addi-tional information, item 220.13,. dated May 3, 1983.
o NRC Audit Finding No. 1 (AF-1), questioning the application of deconvolu-tion techniques to a rock site, resulted f rom the NRC Structural Audit conducted in the A/E's (Ebasco) New York offices in September 1983.
During this audit twenty five findings were established by the NRC review staff. Eight findings were closed prior to issuance of the NRC's November 2, 1983 report, and sixteen remaining finding responses were later submitted to the NRC via Supply System letter dated June 25, 1984.
The Supply System response to AF-1 was submitted separately to the NRC via letter dated June 27, 1984.
o NuReg-0800, Standard Review Plan, Section 3.7.2, Revision 1 was issued in July 1981 requiring validation of plant seismic design basis results by application of alternate analysis techniques. In addition, 10 CFR 50.34(g) requires that WNP-3 must include a seismic evaluation in com-pliance with NuReg-0800 since the project was docketed after May 17, 1982. As a result, and in the spirit of NuReg-0800 Section 3.7.2, the NRC staf f concluded that an appropriate resolution to the deconvolution issue of AF-1 would be to conduct a validation analysis utilizing an alternate plant seismic analysis technique.
o The Supply System and Ebasco presented the results of the WNP-3 elastic half space seismic analyses in a meeting with the NRC staf f on July 10, 1984. The half space analysis included lumped foundation springs repre-senting both the embedment and half space ef fects. The seismic criteria motion was applied directly to the plant basemat in the half space study and comparisons with the design basis model f requencies, mode shapes, accelera tions , and floor spectra were presented. Where local floor spectra exceedances (amplitude or frequency) occurred, results of sub-system analyses were presented to demonstrate maintenance of design qual-l ification. On the basis of this ef fort, the Supply System concluded that I
AF-1 had been resolved by the half space study. The NRC staf f concluded that the technical program was fundamentally sound but questioned the boundary conditions applied to the finite element foundation model which was used to generate t he embedment (sidewall) spring rates. The NRC staff concluded *5at chey would utilize an independent consultant to evaluate the tech w al merit of the half space study.
0 4
o The Supply System, on July 14, 1986, submitted to the NRC further evalua-tions of the model utilized by Ebasco to develop the sidewall foundation springs. The supplemental analyses showed that fixing the basemat nodes was a reasonable means to isolate the embedment stif fness when unit loads were app 1',ed to the foundation sidewall. The premise of the Supply System :nalyses was that the sidewall unit loao deformation in the
. unconstrained model must equal (or approximate) the superposition of the slastic half space deformation (at the basemat level) combined with the sidewall embedment deformation when the basemat nodes are fixed. This is si. gly maintmnce of compatibility between the total solution and the incremental solution in which one component is the deformation due to the sidewall stiffness, o The NRC enlisted the support of Brookhaven National Laboratory (BNL) to review both the WNP-3 elastic half space study and the supplemental study provided by the Supply System which tested the effectiveness of the half space foundation model boundary conditions. BNL commenced their review and on April 7,1987 the NRC submitted to the Supply System a request for additional information needed by BNL to complete their review. The request was divided into three areas comprising a total of thirteen separate line item requests for additional information or analyses, o The Supply System submitted its response to the BNL request for addi-tional information on July 31, 1987.
o BNL comple?.ed its review of the Supply System's July 31 submittal and documented its findings in an October 15 letter to the NRC (Reference 2). As documented by Reference 2 the Supply System responses were acceptable except for Items 2d, 2e, 2f, 2g, and 3a. Revised responses to each of these open items is provided with this submittal (see section entitled: Brookhaven Additional Information Issues).
o On December 15, 1987 a meeting with the NRC, BNL, Ebasco, and the Supply System was held to discuss the open additional information items requested by BNL. BNL strongly recommended util1zation of techniques such as those published by Beredugo et. al. (Reference 3), to formulate the sidewall (embedment) springs for the half space study.
o Ouring January and February of 1988. Ebasco devised an alternate means to develop the sidewall embedment springs f rom what was termed an "energy approach". The energy approach utilized unit load compliances and deformations established f rom a finite element half space model and a finite element model of the embedment. No local fixity of basemat nodes were used in these new finite element models. The finite element results were utilized in an energy equation which balanced the total strain energy with the pure hcif space and embedment strain energies. An out-line of the method was informally forwarded to the NRC and was later dis-cussed with BNL via telecon. During the telecon, BNL citing the results of J. E. Luco's paper (Reference 4), concluded that a zero stif fness value would result (i.e., a rigid body displacement) from a
1 %
static unit horizontal tractien load applied to a half space, and thus, the A/E's finite element model half space results could not be applied and were not correct. As a result, BNL could not endorse the approach and again urged the direct application of the raethods of Beredugo (Reference 3). Further revjew of Luco's paper by the Supply System found that the horizontal compliance (inverse of the real part stiffness) has a finite value for all frequencies. And specifically,for the zero f re-quency (infinite period), the horizontal compliance for the half space is given by Luco as:
CHH(0) = [2.0 - v]/[(8)(G)(r)].
Where v is Poisson's ratio, G the shear modulus, and r is the footing radius. Figure 4 of Reference 4 also plots the complex horizontal com-pliance (normalized to CHH(0)) as a function of frequency. And ' as would be physically reasonabic, Figure 4 (Reference 4) shows that for the zero frequency the real normalized compliance component is unity (1) while the imaginary component goes to zero. Thus, it is concluded that the zero frequency result of Reference 4 is not a basis for dismissing the energy approach as proposed by Ebasco. Further, it is judged that the static loads applied to the half space finite element models used in the ent gy approach (with fixed boundaries at "infinity") will yield, or approximate, mean value unit load influence coef ficients (i.e., stiff-ness) as desired.
o The Supply System and Ebasco could not reconcile implementation of the Beredugo methods (and others that were found in the literature) since all such solutions are based on assumed rigid embedtled sidewalls. Previous NALTRAN analyses for WNP-3 had demonstrated that the reactor auxiliary building sidewalls are flexible, not rigid as required by available solu-tier.s found in the literature. As a result, the Supply System concluded that a limited scope study utilizing a state-of-the-art tool such as SASSI was required. SASSI includes analysis features specifically designed to handle the building embedment af fects and flexibility. In addition. SASSI results have been accepted by the NRC for the seismic analysis, and reanalyses, of several nuclear projects.
o On May 11, 1988 the Supply System contr.?cted with Impell Corporation to complete a SASSI analysis for WNP-3 su.h that the conservatism of the A/E's original design basis seismic mocei (i.e., the finite element /
finite boundaries, deconvolution analysis model) could be demonstrated.
Additional other elements were to be extracted f rom the model to assess the BNL concerns relative to the half space stuuies. This Supply System submittal to the NRC sumarizes this body of work and responds to each of the open BNL concerns, 1
i l
l l
f %
5 WNP-3 SASSI STUDY SASSI (A System for Analysis of Soil-Structure interaction) is a prominent soil-structure interaction analysis code which can be used to solve a wide range of SSI problems including WNP-3's deeply embedded, rock founded, nuclear island structure. The SASSI code has passed numerous validation problems and ^
most notably was successfully implemented to predict seismic responses asso-ciated with EPRI's Lotung Experiment. Additionally, SASSI code results have been accepted by the NRC at several nuclear projects. In sum, the SASSI code is a state-of-the-art tool which has wide industry acceptance and is well suited to the WNP-3 seismic analysis task.
Impell Corporation was contracted to complete a plant specific seismic analy-sis utilizing the SASSI code ss a result of the Supply System's need to bench-mark the WNP-3 seismic design basis model, and assess modeling techniques employed in the A/E's half space model analyses. Impell's experience with the SASSI code is extensive. Recent major ef forts inclutte SASSI based requalifi-Cation analyses for the San Onfre nuclear station and participation in EPRI's Lotung Experiment. Most importantly, Impell's SASSI results, published prior to access to test data, showed f avorable agreement with the measured Lotung selsmic responses. In sum, Impell Cor.noration retains significant experience and credentials in the field of soil-structure interaction, including the application of specialized SSI analysis tools such as SASSI. It is further noted that Impell is a third party contractor having no previous involvement with the WNP-3 seismic design analysis effort.
Impell's ef fort encompassed a complete SASSI model execution for WNP-3. The basic modeling steps include:
- 1) Generation of the strain-compatible shear modulus and damping '
characteristics from the site geophysical properties and input c motion using the SHAKE program.
- 2) Conversion of the WNP-3 NASTRAN building models into SASSI com-patible elements and format. Including completion of frequency and mode shape comparisons between the WNP-3 NASTRAN and SASSI building models to ensure equivalence. i i
- 3) Development of an integrated foundation and building model. This l step includes solution of the layered free-field site response, development of the loading response for each layer, and construction of the excavated soil and building structure finite element mass and stif tness matrices which are needed in the analysis and solution phases of the SASSI procedure. Particular care is taken in these steps to ensure proper layering and nodalization such that all f re-quencies of interest are passed during execution of the analysis.
1 l
J
f b
- 4) Perform the two-dimensional SASSI response analysis and generate two percent damping response spectra at the top elevation of each struc-ture and c2 the center of the basemat. Both horizontal and vertical spectra are generated and plotted with the WNP-3 design basis spectra for ease of comparison.
- 5) Extraction of the f requency dependent impedance functions from the SASSI program results at selected foundation locations.
The SASSI solution was completed for the WNP4 East-West horizontal building model and the vertical building model. Only one horizontal direction was examined due to the nuclear island's high degree of symmetry resulting in very similar responses between horizontal directions. (Note: the issue of WNP-3 building symetry (i.e., ignorable out-of-plane horizontal responses) was addressed by the Supply System's response to Audit Finding 17.) For the SASSI analysis the criteria input uotion was applied at grade in the free-field.
The analyses are documented and verified in compliance with Impell Corporation's nuclear quality assurance program (Reference 5).
i l
1 l
t
-t t BROOKHAVEN ADDITIONAL INFORMATION ISSUES ,
i Attachment I of Reference 2 outlines BNL's assessment of the additional infor-mation responses provided by the Supply System via its July 31, 1987 sub-mittal. In sum, the Supply System responses to Items 2d, 2e, 2f, 2g, and 3a did not resolve BNL's concerns. Each of these items will now be' discussed aided by the SASSI results obtained from the study completed by Impell ;
Corporation. '
Items 2d and 2a 2d: "Provide analyses to demonstrate that proper sidewall interac'clon springs can be obtained by fixing nodes along the basemat."
29: "Provide comparisons of sidewall interactions coef ficients with available analytical approaches.' '
Items 2d and 2g are closely related issues basically involving the validity of the sidewall springs used in the half space model. As will be discussed in the responses to Items 2e and 2f. comparisons with the SASSI generated damping and spring rates dif fer significantly from the values used in the half space analyses. Yet, as will be presented in the response to Item 3a (i .e. , the salient spectral comparisons), it will be shown that the half space results are conservative in nearly all cases as compared to the SASSI results.
In light of the SASSI results being bounded by the WNP-3 design basis spectral results, the Supply System elects to designate Items 2d and 29 as moot issues. SASSI is a verified code (including recent benchmarks with EPRI's Lotung Experiment) with suitable analysis features to accurately assess the WNP-3 embedment and soil (rock) damping effects. The SASSI analysis was .'
executed utilizing standard and accepted methods, and therefore, represents an
., excellent independent test, or measure, of the conservatism contained within j the original WNP-3 seismic design basis model results.
1 !
The Supply System also concludes that the SASSI results corroborate the A/E's !
seismic half space model results. In spite of significant dif ferences in j damping and foundation impedances the half space results exhibit (in general) boundedness over the SASSI results with comparable f requency content in most cases. Thus, based on the favorable comparison with the SASSI generated responses, the A/E's methods and parameter input choices for the half space model were, in retrospect, founded on adequate engineering judgments. In other terms, the validity of judgments such as the means for developing static i
sidewall interaction coef ficients are resolved, or at least put into per- l spective, by comparison to the spectral results obtained f rom the independent l SASSI solution.
l
) The response to Item 3a provides thewompa risons between SASSI, the half !
space, and the WNP-3 design basis spectra. Item 2f includes remarks relative !
, to the apparent lack of model sensitivity to foundation spring rate and l frequency dependence. -
Item 2e:
"Provide justification for the sidewall damping magnitudes which was used."
The reviewers further stated (via Attachment I of Reference 2) that the use of "shear" material damping was inappropriate for the simulation of sidewall "compressive" damping.
In the original seismic half space study submitted by the Supply System (June 1984), damping on the sidewalls was conservatively assumed to consist of the material damping only (i.e., no radiation da.nping was considered or included). Since the WNP-3 sidewalls are flexible (as determined by NASTRAN results) both compressive strains and shear strains will result in the rock foundation. Sidewall shearing will also be significant in the vertical mode and shearing contributions will result in the rocking mode as well. There-fore, the assumption of material damping f rom flexure, or shear responses, is acceptable since significant shear loads are present at the sidewall. In addition, the magnitude of the material damping is significantly less than the radiation damping component which was ignored at the sidewalls in the half space studies. The lower damping values used in the half space analyses could result in some minor shifting of the building f requency response, but the primary effect will be conservative over predictions of maximum spectral amplitudes.
Again, review of the spectral comparisons provided in response to Item 3a do '
show the character of the half space results to over predict the SASSI responses, and at some local frequencies exceed the design basis spectral amplitudes. Albeit approximate, a quick means for assessing the conservatism in ignoring the sidewall radiation damping component is found by simply com- ,
paring the half space model input with Beredugo's damping prediction. Appli- !
cation of the methodology of Reference 3 would yield a 25 percent of critical l damping value for the embedment sidewall, as compared to a 2.4 percent '
material damping value that was used in the half space model.
- Item 2f "Provide justification for the use of static analyses in place of f requency dependent sidewall interaction coefficients."
i Figures 2 f -1 through 2f-14 provide the SASSI results for both the real and imaginary parts of the complex f requency dependent impedances. Cases are i included for lateral, vertical, and rocking modes. Basemat results are also J included with these figures. As can be seen f rom these SASSI results, the 1
) real part lateral impedances are basically invariant with respect to fre-quency, and thus static sidewall interaction coefficients are quite appropriate.
1 1 Review of the real part impedances for the vertical and rocking basemat cases (Figures 2f-6 and 2f-7) show a strong dependence with f requency. Taken as an isolated result it would appear from the results of Figures 2f-6 and 2f-7 that utilization of static foundation impedances ar6 not appropriate for the i
, l
)>
- i. -
~
.i Units are in kips, ft. per foot 400000. - Real part of impedance Unit load at Node 13 - Horizontal translation DESIGN VERIFICATION wo 13-curt;T_arftis l Jonrio._o w roc cAtciraos no.Q2,,g 300000. -
By: 4 g,7g. g,
- \ CHK6:On j oggg.g ,, y ._\.
m ,
to -
E
~ -
4 C"
I
? 8200000. -
z .
E O ~
LA.S g Q_ .
~
E w I 100000.
I I ' '
' g i i 0 0.5 1. 2. 5. 10. 20. .
0.1 0.2 FREQUENCY (CPS) ,
UPPSS SSI IMPEDANCES , DIRECTION = HORIZONTRL ELEV =389.00FT , NODE 13 Figure 2f-l !
/ %,
~
i Units are in kips, ft. per foot 400000. - Real part of impedance .
Unit
- oad at Node 14 - llorizontal translation DESIGN VElliFICATION curra_ettss i' 14
.s:e no _o_}yonpcs ca.er.rt-n'x no WJ -es 300000. -
.iYI3E~ DATE */iffE l
ut.xo c^ o OATE sQ a .
E na -
E
~
I
~
'd 200000. -
z ~
E Q
na (L
r -
100000. -
0 l 0.1 0.2 0.5 1. 2. S. 10. 20.
FREQUENCY (CPS) .
UPPSS SSI IMPEDRNCES , DIRECTION = HORIZONTRL ELEV =361. SOFT , N00E 14 ~
Figure 2f-2 -
~ ^ I.
s
-[. .
s -
Units are in kips, ft. per foot q00000, _ Real part of impedance Unit Load at Node 135 - Horizontal translation o5
~300000. - '2.E_ SIGN Vme!
M1ON
~~ ~
CtJENT_(4ffg g L J)!! NO __67Y 6 - t og '
1 - I C*.LCIPROn . .
g }ur:.cru - - -._n& n'.'
r --
l '
g CHKD:(aq) pi_ . . ..f. .
7 g W 20000G. -
z .
E o
w a.
r
~
100000. - m
' ' ' ' ' I- -- 1- -
0 5. 10. 20.
0.1 0.2 0.5 I. 2.
FREQUENCY (CPat UPPSS SSI IMPEDRNCES , DIRECTION = tt0RIZONTAL ELEV 23d3.0f,F T , NODE 13S Figure 2f-3 ~k
t t
n Units are in kips, ft. per foot 400000. - Real part cf impedance Unit Load at Node 136 - llorizontal translatioa
~
136 o-1000GO. -
DESIGN VERIFICATIONl CUENT,8d([L .
- JOO NO ,_o110- f e 4 -
8 *j_t
+--
w CALC'P908 BY: dAf NO _6pf oATE: tun (I;fft.
i' 5 , coxo- ru) oAre: t/.fygj W 200000. -
l
= l cr i o
w
'L r
100000. -
0 O.1 0.2 0.5 1. 2. 5. 10. 20.
FREQUENCY (CP31
_ _ _ _ _ . ps UPP S S31 IMPEDANCES . DIRECT!GN= HORIZONTRI.
ELEV =3Gl. SOFT , NODE 136 ,
l Figure 2f-4 ,
I - r i k
" .i i .
- Units are in kips, ft. per foot 4100000. -
Real part of impedance
. Unit Load at Node E? - Horizontal translation
'300000. -
OU3tGN VERIFICATION 3 -
7:IEHi._W_Sh_ -
w .scis no.__o bo
ar cAtr.erAOO NO kWfirol,
~
ev. e nt oArB Wifr W 200000. - i cHxo.c d WE. Qukg*
/.
? =
T a
w Q.
r -
~
2 .
100000. -
0 O.1 0.2 0.5 1. 2. 5. 10. 20.
FREQUEtlCY (CPSI UPPSS SSI IMPEDANCES . DIRECT ION = HORIZONTril ELEV =330. SOFT , NODE 62 -
Figure 2f-5
3 L
Di' i
i e
Units are in kips, ft. per foot l Real part of impedance 200000~ -
LHit Load at Node 62 - Vertical translation
~
d
~ ll ;b DELIGN tLT_ M(VER!FICAT!ON
- 0* 0 o. tog ;
120000. - Ic.C6f40 o 62 Jg_4f*/e.torstou OATL647 ran I_pury7 t CHr.re.I W , nATE.c[7,[y 4
K
.' W 40000. -
Y' 5 w
a z
G ~
O -40000. -
- n. -
r
~
I l -120000. -
-20000n 2. Ib. 2. N
- 0. o '2 0.5 1. 5.
FREQUEfJCY (CPZ)
L. - '
V( '
EL 0 SbFT , D 6h ,
Figur f-6
~
F Units are in kips, ft. per foot 5000000000. -
Real part of impedance Unit Load at Node 62 - Rotation
. l 62 m l l-a~ t 4000000000. N, DESIGtJ VERIFICATION
{ CliEtn_Mh$ _
g Jos tio E6 _ toc
[
- CALctsmoB sta su 1 ,- or he im -
J b3000000000.'- -
I CM& M W E. ' i/d t
? 5
~
, a U -
Z tr O200000C000.
O l
i 1000000000.'-
1 I
L t I I I ' ~
20.
~
0 10.
0.T- 0.2 0.5 1. 2. S. Tl FREQUENCY (CPJ! Q l
i . _ _ _ _ _ _ _ _ - - - - -
o OPPsS Sji 1MPEDGNCE3 , OIRECI10N = R0TRT10NRL g ELEV4 JJG.3 F' , NODE 62 -
i . . _ . . . _ _ - -
Figare 2f-7
1- ms we e e .
l' z Md 9
H
? -
%.m a w 4 e O
- 6 i.
w O E 3 l7 E O 5 O
, w r e- z
> 3
~'
Z 7l@(C ew
$ F. k O y a e gl 6
O g w5 % J .. i N
.O,. ClF.C.>;w%
e
= JlC y %= w%
4 r=
W Q
L u
-. a C
Ab 'M'" .- 2 C eJ o
e QN 8 -
- N
.=
Lvb @ g aC C a
"a=
.et -
=
3 4 Q. M
% 6M m.
-- m u
w e-a O. 0 % "W w O - * >== Q CO
.g se n N > i L u" y o, %
N C m9 g 6
-h4 o >, ,
m _, O L.
"3 O 3 L Le o nm) ca . e LL1 g
- m me - % ,O
+J e4 '
cec w-w m.
C LLJ A U CD m Zn
. C it C C tu >
CL LLJ IJ
-W b M o
W th A
CL.
I t t , - U L I - t i 1 . . . , ,
- . . . . . , , o C C O O O O O c c C O O O O O o o O O O O o o o o C O O O O o o o O C C C Q C Q c G N @ f 7 m N -
I ABBNIQUWI) 33NUO3dWI l
I l
(
}
^
- l
~
. s Units are in kips, ft. per foot 000000. - Imaginary part of impedance Unit Load at Mode 14 - llorizontal translation 700000. -
600000. - "f ca: -
G DESIGN VERIFICATIOtC
~500000. -
crean y g U JOH NO._41 Y e .7 g ,
E SiLCIEA08 no,g3.of 4
= 's 00000. -
jfy,3 ,$ 0 9gg; w -
J U
z
@300000.
u G -
E
~200000. -
100000. -
0 O.: 0.2 0.5 1. 2. S. 10. 20. P FREQUENCY.(CPS)
UFPSS SSI IMPEDRNCES , DIRECTION = HORIZONTRL ELEV =3ct.50FT , NODE 14 Figure 2f-9 l 5
e
- II Units are in kips, ft. per foot l
800000. - Imaginary part of impedance Unit 1.oad at Node 135 - 1;arizontal t4anslation oh 700000. -
600000. -
.M,s: SIGN VERIFICATION g ,,,,.
1Mni m
! , wo_c2h -194 L9fctopoct'o.tah? *
- 500000. -
" 3GI # "
, o : car.n gm' , o4t e c i,g e
U180D000. -
su ,
u .
2 E303000.
g 1 U
=
I.
~200000. -
I i
j 100G00. -
1 -
8 , , ,
n n n 0.1 c.S 1. 2.
0.2 FREQUENCY (CPS)
- 5. 10. 20. I Q -
WPPSS SSI iMPEDRNCES , DIRECTION = HORIZONTAL ELE.V =3:19. 00FT , NODE 135
- ^ - ~ - ~ ~ ~ 1 ,
Figure 2f-10
- . . _ _ . = . - _ - -_ - _ . .
m
. g 1
r -- - - - - _ _ _ _ _ . .
Units are in kips, ft. per foot 800000. -
Imaginary part of impedance Unit Load at Mode 136 - llorizont&1 translation 136 700000.
f-a-l L
I
_600000 g- DESIGN VER!FICATIOt3 g [
- CalENT_M[M _
g -
JOB NO._07.Y0:_til I z '
G500000.f "C'"
ov: .&a,"0MO odiT4% "T fj '
m OME:f r[gg CHKO.' Es<)
to r o
- 8400000. r-t a w (
8 o ;
z j g?00000.i m
n- r
- C *
'~200000. -
100000.'- -
e W
0 3 8 -- -- 8 -- '- ' ' ' '
O.1 0.2 0.5 1. 2. 5. 10. 20. T-)
FRE0t'ENCY ( C P'.21 sJ "N WPP33 SCI IMPEDRNCES , DIRtCT ION = If0RIZ0llTHl.
ELEY =361.5GFT , Nt!Dt 136 , ,
Figure 2f-Il
m s -l l
I Units are in kips, ft. per foot t
000000. - IHaginary part of impedance Ur:it Load at Node 62 - Horizontal translation 700000. -
62 I.
h i I
- 600000. r DESIGN VERIFtOATION a- t CLsLNT 10t(G a -
Joe no . 6He 26C
.;500000.I- cAtciPROd DOJtf) _
o BY: Wh DATE4 ~
E f CHKD' M OA10 6 U400000. r-
=
e m
u 2 1 E 300000. -
' sa
- n. t E .
~ 200000. -
e 1000co. -
- 0 O .' t 0.2 0.5 1. 2. 5. 10. 20.
FREDUENCY (CPhi
.)
WPPSS 531 IMPEDANCES , DIRECTION = HORIZONTAL ELEV =JJ0. SOFT , N00E 62 Figure 2f-12 !
. -. . _ - .-= .- _ _ . . - . . - _ , . . -. -- . =-. - , . - ., - - - - - .-
^
l i
s
~~
\ .
Units are in kips, f t. per foot
~J00C000. -
Imaginary part of impedance Unit Load at Node 62 = Vertical translation 1 - l 2:400000.h I.
g .
x .
G z -
G Its00000. - '
.I .
r
~ .
DESIGN VERIFICATION e'o -
CLIENT ,_{3Pf$f '
'.# us .
. Joe w Jivo_e g z I200000- -
a ar. .testf UU-ci'.a o '
~
comtisu> QTV 4j'l n.
r 600000,l-F r
0 ' ' ' ' ' --
O.1 0.2 0.5 1. 2. 5. 10 20.
FREQUENCY (CPLs } .
G LN WPPaS 351 IMPEDhNCEJ , DIRECTICN = VERTICAL \
ELEV=~330.5 FT , Nn0E 62 Figure 2f-13 a
e . .
e D
g .g e
Co,e a
e o**s s u T. T,, :<e E
w WddM gz
> c.
2 8
l fk 59 0% ..
y *2 . *
~- -
? 8Rb u.l 5 d~
C _j ,
I Q*
Z o
4 c ..
o -
8
% 9
. 3 f.bd gco C Q. a: %
. t Il N
9 t!) g
% N L
- W u Z d% e
- C "y
C. o 9 . o N u g cg9 Z W
- EN W !E , Z ghv l 3 "'"
e O b l- 6 6 g
- g o
',g , Qy p N
l - E e
- m- IA. *p b j 4 Cn M
-m- :s i
- Cn i C 3'. c G3 i
"D . W $
l es E',
a c O p ,
l d
l gJ i
W t "1
" N. t!"t ,
C
- i
)
C 1 L !
t, ..
1; . . g _t,. 4 W, . ...I, ,.i1 . . ", "3 O O O C C . C C C C C C C C C C C C C C C C C C C C C C C C C Cs C O O C C C C O C C C C C C C C C C C C C C C C C Ln C D C C
& N N w - V4 (ABBNIQSWil 33NUO3dW1 T
i vertical and rocking modes. However, if the model responses (Item 3a) are examined between the f requency domain solution--(SASSI) and the static half space solution, then it can be inferred that the use of static vertical and rocking foundation impedances provides satisfactory results since the char-acter of the responses are similar.
The apparent lack of model sensitivity to the f requency variation in vertical and rocking foundation impedances is judged to result f rom two primary analy-sis factors. First, in the SASSI solution the complex f rcquency dependent i stif fness matrix, included in the equation of motion for the complete solu-2 tion, is of the form:
g - E - w25 where E is the complex stif fness matrix, w the f requency of vibration and B the mass matrix. Even though the complex stif fness matrix 5 may have a strong functional dependence on frequency, the total frequency dependent stif fness matrix C is probably cominated over most of the f requency range by the f re-quency-squared times mass term. This is especially true as 'requency l
, increases and, in our case, Ereal diminishes. Thus, for the specific case of WNP-3, the use of static, or mean value, foundation stiffnesses could be applied (for a simplified study) since the mass stiffness is apparently controlling. Also, it is recognized that a significant mass damping term exists in the SA5SI stiffness formulation. I Although the SASS! rocking impedant is functionally dependent on frequency it is seen that all of the lateral fo. idation impedances are basically invariant with frequency. Since the rocking node is coupled with the lateral sidewall -
impedances the rocking response is therefore not simply governed by the rotational rocking impedance. For the Auxiliary Building (at least) it is reasonable to assume that fairly strong couples would be formed by the fre-l quency independent sidewall impedances. Thus, these static (frequency inde-pendent) impedance couples will act to temper the rocking response with respect to vibration f requency. It is also noted f rom a review of the SASSI impedance verses f requency figures that the rocking impedance is diminishing with increasing frequency, with a corresponding substantial increase in 1 imaginary impedance (i.e. component damping). These results also tend to '
j infer that the rocking component, as comptred to the sidewall lateral j impedances, becomes less dominant and more highly damped over much of the f requency range. Finally, it is worthwhile to observe that the same general impedance trends can be observed in Beredugo's formulations for the rigid sidewall ,:ase as presented by Reference 3. t i
! In the half space modeling approach it was the intent, following the recom- l mendations of Reference 6, that limiting the available model damping would J penni t a compromise in the foundation impedance accuracy (i.e., the use of 3 static value foundation springs). The low damping input raises model responses and tends to broaden tha f requency range where response amplitudes 1 are of interest. Indeed the results provided in response to Item 3a show this trend of the half space model results enveloping the SASSI results.
1
}
l 1
. , t i
To quantify, or isolate, the effect of individual parameters (e.g., mass affects, impedance values, etc.) on model sensitivity is not the Supply System's objective.
i Rather, the issue for WNP-3 is: are the design basis '
spectra conservative when compared to results f rom an acceptable alternate soluttor.? And secondly, given the half space model assumptions and analysis '
methodology are the response results reasonable and consistent? The benchmark results provided by the SASSI study resolves each of these issues. Spe- !
cifically, the WNP-3 seismic design basis spectra contain a significant level i of conservatism, and the seismic half space s6udy retains a good deal of merit as a comparative solution.
- Finally, as measured by the SASSI response results (Item 3a) the Supply System concludes that for the specific case of WNP-3's soil conditions it is justi-fled (for a comparative study) to assume static sidewall interaction coef- !
ficients as derived f rom a consistent finite element foundation model and unit 3
influence load application methodology.
Item 3a l
"Provide a comparison of the fini e element finite boundary calc o ations with the half space approach." :
i The intent of this BNL concern was to seek an explanation, or justification, !
of the dif ferences between the design basis model and the half space study ;
floor spectra results. Since a mutually agreed approach for generating the
, half space model foundation springs was not established, a SASSI study incor- !
porating all aspects cJ the foundation media and building model was completed as an alternative to the half space approach. Floor spectra results f rom the WNP-3 SASSI study (Reference 5) are plotted in Figures Ja-1 through 3a-10.
f For comp 6rison purposes, the corresponding design basis floor spectra is .
plotted with each of these figures. Bt.th the SASSI and design basis floor
- spectra which are presented by these figures are SSE responses' at two percent I j damping.
- Figures 3a-1 through la-10 present the maximum horizontal and vertical j building / internals responses found at the highest respective structure 1 elevation. In addition, the responses at the center of the basemat are also ;
plotted for the horizontal and vertical directions. Because of the high ,
' degree of symmetry between the North-South and East-West horizontal building l models it is only necessary to examine one horizontal direction for comparison l to the design basis floor spectra results. With each figure the corresponding l 4
fundamental building frequency (ft ) is given for reference purposes. For !
example, Figure 3a-7 cites the fundamental vertical Reactor Auxiliary Building '
l f requency as 17.1 Hertz. Because of its rigidity the basemat "frequency' is -
taken as characteristic of the ZPA, or say greater than 33 Hertz. And thereby !
the response of the basemat is governed by the dynamic interaction of the soil !
and building model elements.
l i
l l
.<5-
- g. -g l
.i i
SASSI vs. Design. East - West Input !
Top of Shield building. 2% damping. E - W corponent ,
Solid = SASSI Dash = Design DESIGN VERIFICATIO:8
~
CLIENT lMT<<
g no _ -e7N 0-IOC cAtcWROS NO.Mfhk fI = 3*5 Hz gy: W H DATEM3'[F8 b
~
CHMD-g @ DATE: <iff s... -
[ ~t*
2 i g flalf Space tiodei
- a l t
, i a s 0" -
I g to a i J l t g
m d
u I
i 8 *
' a .
/ 'g ,-- ,
- ,
- s .
- / l s -
p r.. ~
n.,
~------..--
. 1 =M t t I a e s e e t e e t
- 8. . ..i. . . . . . . . . . .
. a. 3- .. .. .. v. ....n.. m. m. m.
FREQUENCY- (CPSI WPPSS SSI ANALYSIS , TOP OF SHIELD BUILDING DIRECTION = E-U , DAMPINC = 2 PERCENT
~
Figure 'la-1
_ _ _ _ . ~ _ - _ - . -- - _ __- . _. - . _ - . . - _ . . _ _ _ _ . __ . . . _ __ -. _ . _ _ _ _ _ -
/~ m SASSI vs. Design. East - West Input Top of Reactor Aux. Building. 2% damping. E-W component
,,,,_ Solid = SASSI Dash = Design Ifalf Space flodel DESIGN VERIFICATION j _
cuoun1s r- i l fy = 6.9 Hz g m uo o14 e- tec l
I t cAtciPROB NO Wf 7- *1 "
! I I BY: 4 84 DATE: 6/w/PF N ** g CHKD- M DATE- 6fts[g
~
f .---
2 - I o -
1
.-- g a
j 6- '
' e e l' l' .
, m .
< 1 w (
m 7 a l ' Half Space Model s
o ./ N c
f f f i f I f f f f f f I f I f f
- I. . ..e .........i. a. 3. .. . . .. v . .. . . . . m. m. .. ..
FREQUENCY (CPSI !
l,4 UPPSS SSI RNALYSIS , TOP OF REACTOR HUX. BUILDINC DIRECTION = E-u , DAMP!NC = 2-PERCENT , [
1 Figure 3a-2 .
~
?
' ?.
l .o
~
e e
l I
- SASSI vs. Design. East - West Input j Top of Containment Yessel, 2% damping. E-W component Solid = SASSI Dash = Design
( DESIGN VERIFICATION CUENT laffEI m uo. on.-i.c l ~
cAtcPROS NO.W8-
- I
~ Half Space Model av: (r at DATE:& fMlfJP i
. CHED; im DATE:grefft 3 u. -
z .
e a y f3 = 6,0 Hz
~
[
cr I-t I
x -l sa. , t I
6 e.. -
,i o i Half Space Model y 'i t, l \
s
~
^' I _f f I I I f f 8 = _ _^- e -; a f f f t
- m. m. .. w.
.. s. .. v. .. . . . . .
FREQUENCY (CPSI 3SI RNRLYSIS , TOP OF' CONTAINMENT BUILDING
! WPPSS DIRECTION = E-U , DRMPINC = 2 PERCENT l
Figure 3a-3 1
l l ll
?I. 1. e
^.
s N g g O
i 4
0 p
, f s
i t T -
f m A 1.s 6:
g E C c. y e ;
I t
F i . .
r e O go lO
, A T
t IR t E N -
n e V Mw o B o . .
O t t n N G top
.n . ( -
m 4D :
e NuI -
I S E s, C L K E I A y H c D L
CxC a C . .
W % t o S n
E .
E
. si 5 R J
i g ,I,,i i Ii1 I n T i m .
C T t p - U N um pa [I e s8 :E R E
,, T nd n d I
%i g
i o -
t o S C R .
t s L E t2s -
s e e -, t e ,D c ! A P Ws e= a p
e8 o ' t e
. )
S N r S R 2 t_ s. P tt s uh f ' C
(
E=
T sca l - .
NC i
auD a o E r t I
Y I
N
,ts t s C I n ,
N F P 4 gl -
E G M -
i a a snI t
. U P D R 3 N erS des -
Q O e E r sI
.tA nS R T
. u g
v =
t s.F i f ,U z F xt I od H S-E S i S pl 1 I Aoo STS 2 S=
y-
. Y 1 a L
= A N 1
N O I f -
A T I
C S E S R I t
D
. S t
S i e. P e P t
e
. W t
e s.
, . E o.
e .
- o~ zo teed;ouoa - t t A, ?.
rI I llI \, llI \ fil lllll;!l l L
f *
.- . ., e i
e SASSI vs. Design. East - West Input , .
Center of Basemat, 2% damping. E-W component Solid = SASSI Dash = Design
- ~* ~
Half Space Model DESIGN VERIFICATION cueni Men m uo_ oM*- lo s CALCIPROS NO. WAlP3 -8I n V; Ir M OATE-'P' M N U CHKD- (m DATE cl eaf yr
~ 3.. - ZPA Z e . \
0 '
)
g e
-v
- i t
x w
l l \ ;
a
- I l
l b- d'*
o
\. , , ,
e t i
~~ N
~ ~~. _ _
I f f f f f f f f f v i I f l
g,8.3 ..........f. a. 2 .. ..
f I I I
.. 3. .. ..f.. . . .
FREQUENCY (CPS)
l DIRECTION = E-U , ORMPINC = 2 PERCENT Figure 3a-5 l b
[>
S
_ ~ _ _ - .. ._ . . .. . - - - . _ - - _ - .. - - _ . - -_. - -.
r m s
SASSI vs. Design. Vertical Input l
j Top of shield Building. 2% damping. Vertical component -
Solid = SASSI Dash = Design DESIGt! VERIFICATION I ** ' -
cuaa_idDC g ,o .wo- sec otcfPDOB tt0 2*LII*
av-9*
DME /" "
one st'*ln l
. -~3 g gm .
8 1 9 ... -
I Half Space Model-2 f3 = 10.8 Hz e- :
E
/ !
l a g l i 1
\.b 7 J
! w u
... - ,e
\
u T
E . '
, - s
\
I -
l -- -s,
? ,
Tf* I t t t t I 1 9 I I I t t I
- v. . . . . . .
I g3.5 ............i. a. t .. s. .. n. n. . ..
l
! FREQUENCY (CPS)
I . L t- f*T WPPSS SSI RNRLYSIS , TOP OF StilELD BUILD'NC ^
DIRECTION = VER , DRMPINC = 2 PERCENT i
P Figure 3a-6
+
3 e
SASSI vs. Design. Vertical Input Top of Reactor Aux. Building. 21 damping. Vertical crwponent Solid = SASSI Dash = Design DESIGtt VERIFICATIOi!
s-- I currn b>ff55 Italf Space flodel '
i m o tsr- a n.- s'G
,- 1 CALCJFROB tiO AP83-!1__
s g . ov. er cati ~Qs/ri .
f3 = 17.1 ifz g CHKD 5A.0 DATE:4lid f1 l 9 a. '
I
- . g 2 ,-.s-~~~,__ -
~s g S
/ - ... i r i
- s
. e w
N a: - ,
y t
to t a '
y ... -
- t y ,
f l f f f I l l f I I I I I I I I I FREQUENCY (CPSI UPPSS SSI ANALYSIS . TOP OF REACTOR AUX. BUILDINC DIRECTION = VER , DAMPINC = 2 PERCENT 3
~
. Figure 3a-7
s
- 3. . .
I O
b
~
SASSI vs. Design, Vertical Input Top of Containment Vessel, 2% damping Vertical component Solid = SASSI Da d = Design g ' DESIGft VERIFICATIOil C11ENT luffit .
J A NO 01' *
- f.E CALCIP%B NOMf]- *., 1
~
BY. (ysu OATE: 6fts/f t
- CHetD: Sy> t DATE:6 help O
~... -
!!alf Space flodel e- - g
,8 ',
~
fg = 17.8 Hz s
t 5* E ,
/ i i
m -
i Half Space Model
- ^j U** -
s o ,
( .
a . .
f I f f f f f f f f 1 9 i f f f f s- v. m m. . .
- I.. .t. . . . . . . . . . . 1. . .. .. .. . . . .
FREQUENCY (CPS) ,,
l I
I .
UPPSS SSI RNALYSIS , TOP OF CONTAINMENT BUILDINC DIRECTION = VER , DAMPINC = 2 PERCENT [-
I Figure 3a-8 i e I
l
(' .
')
{,
m
-t SASSI vs. Design. Vertical Input Top of Internal structures, 2% damping, Vertical component DES Grl VERIFICATIOt2 Solid = SASSI Dash = Design cutur_#tf55 s.e - . xe no. one-nec cAtcienos no t-wP l 8Y. 44 DATL afss/ff I
ChFD' W DATE Clis G Half Space Model h *~* ~
f3 = 21.5 Hz 5 _
H e
d i e
E -
'N ,- _ 3 g
\
s ,
e w
u l.a g ,.-, --
.,f .
r J'
p 4
g
- Se
- f f f f f f f f f f f e e f f f __
f so. ee. ee.
,, E.s e.fe e. f e..e.e s. s. s .. s. e. s. e. e. no. se..
FREQUENCY (CPS) n OLA WPPSS SSI RNRLYSIC , TOP OF INTERNAL STRUCTURES D j
\ DIRECTION = VER , ORMPING = 2 PERCENT
' figure 3a-9 s
I e
.= -
,e 1
4 I
e I SASSI vs. Design. Vertical Input ,
j Center of Basenat. 21 damping. Vertical component 1 Solid = SASSI Bash - Design DESIGN VERIFICATION
- ~ cuun_Mtu Josuo owe- tot 1
_ gas ceaan wo en n at 45 OA1E: ah Half Space ibdel
- CHMD iu_ DATE:sfrs]t a
s.. - ,
z ZPA o ,
e W
. K, u ,
r ,
J ', -
y 9.. -
g , ,,
s ,
\
, s' ,,a' % ^- ,
f f f e i I t
- i t f __1 t t if I f
- s. v. . . . . . . m. m. w. m.
- I.. .I ..........9. 1. 3 .. ..
FREQUENCY (CPSI WPPSS SSI ANRLYSIS , CENTER OF BASEMAT DIRECTION = VER , DAMPINC = 2 PERCENT
\ --- ..
i 9
Figure 3a-10 1
4 1
l
. o l
I A comparative review of the SASSI versus design basis spectral results shows that the SASSI results are bounded at virtually all f requencies of interest.
In general, at the fundamental building f requency of interest, the design basis spectral accelerations are more than 100 percent higher than the SASSI predicted results. Trivial exceedances of the design basis spectra do occur i
with the vertical responses at l ow ( i . e . , less than 2.5 Hertz) f requencies; but the acceleration magnitude dif ferences are small in this regime, and most l importantly, no active building or subsystem f requency response is found at
, these long periods (i.e., the WNP-3 design is stiff, as evidenced by the referenced fundamental building frequencies).
Examination of Figure 3a-8 shows a small exceedance of the design basis spectra at approximately 13 to 15 Hz, but the associated containment funda-mental vertical frequency, although close, is higher (17.8 Hz). Thus, direct building f requency matching at the exceedance does not result, however, con-tainment mounted subsystems active in this f requency regime could experience an approximate 10 percent vertical component seismic load increase. But again, it is noted that the exceedance is small, and further, that the rela-tive acceleration magnitude compared to the horizontal response is small.
Furthermore, WNP-3's total seismic loads are derived as the SRSS combination of the two horizontal and the vertical direction as required by Regulatory Guide 1.92. Thus, a minor exceedance of the vertical component is greatly suppressed by the combination of the larger horizontal responses. Of course, other design conservatisms in terms of design allowables, damping, as well as the basic conservatism of assuming concurrent maximum responses for all three directions of seismic motion (as cited by Newmark, R.G.1.92), all cortpound to form a very large margin of safety.
In sum, the Supply System concludes that the WNP-3 SASSI analysis results demonstrate that the WNP-3 seismic design basis model results contain signifi-cant margins of safety. Specifically, the WNP-3 in-rtructure design basis spectra envelop the SASSI results by wide margins at the spectral peaks, and
, are bounding over the SASSI results at all significant building or internal I structure vibration modes.
I Figures 3a-1 through 3a-10 also plot the two percent damping curves generated l from the A/E's elastic half space study. These curves are plotted in the l range f rom 2.0 to approximately 20.0 Hz. Where the half space curves overlay l the SASSI or design basis curves the half space plotting is suspended for l purposes of clarity. The half space curves are labeled on each figure.
In general terms, the half space results show f requency characteristics sim-ilar to the design basis or SASSI results. And as a further generalization, the half space results are bounded in magnitude between the SASSI and design basis spectral results. Furthermore, the half space results are bounded by the design basis results at each of the respective building fundamental l f requencies, except for the Reactor Auxiliary Building (RAB).
l I
l l
l j .~.
I Figure 3a-2 shows that the peak of the half space response for the RAB is near
] -the fundamental response of this structure (6.9 Hz). Whereas both the SASSI i and design basis models shift the spectra l response of this foundation-i interfaced structure up to near 10 Hz. A stif fening af fect with the sur- r rounding foundation media is judged to cause this shif t. The discrete soil l i
support interfaces (i.e., lumped foundation springs) applied in the half space t i model would in ef fect yield less constraint en the RAB than the continuous !
I finite ele'nent boundaries applied in the SASSI and design basis model tech-
] niques. In any case, the SASSI results at 6.9 Hz are (by a f actor of three) [
l far below either the design basis or half space results (see Figure 3a-2).
.I l Other exceedances (outside of the building frequencies of interest) do occur l J with the half space model. Specifically, Figures 3a-8 and 3a-9 show frequency i shif ts and magnitude increases which are significant. These exceedance are
] mitigated largely in that they are relatively low magnitude vertical responses ;
j (i.e., SRSS combination ef fects), and secondly, subsystem evaluations (June
- 1984 submittal) show substantial design margins of safety. i i
- In sum, the Supply System concludes that the half space model results are a
! basic and sound first approximation to the design basis model spectral i t
results, however, the isolated spectral exceedances which result must be [
i rationalized against other inherent and applied design conservatisms which t j exist in the WNP-3 seismic design basis, i
1 l I f
l !
1 l
t 4
i
\
i 4 !
i !
)
i E
I l l 1
! 37 I
j 4 I
o a .
PREMISE OF AUDIT FINDING NUMBER 1 4 The premise of AF-1 is summarized by the reviewer's contention that a reduc-tion in seismic motion with depth is not expected because the rock foundation is an elastic media. Specifically, the reviewer states: " ..Since the rock is essentially elastic in the range of interest, a reduction in motion at the basemat from that postulated at the surface would not be expected". The
' Supply System does not agree with this premise either f rom a theoretic or empirical point of view.
THe WNP-3 nuclear island is a deeply embedded structure essentially socketed i int,o rock down to a depth of over 60 feet. For this depth, even the simplis- '
3 tic assumption of wav'n propagation in an elastic medium would predict wave L attenuation, pa rti'cula rly at higher frequencies (see for example i Reference 7). In other words, it is theoretically as well as physically Impossible for a motion to have identical characteristics between the surface I
and at any depth below the surface. This is particularly true for surface
, motions which are broad banded, rich in all frequencies in the range of It:terest, such as in the case for the WNP-3 seismic design. .
Figures ( AF-1)-1 and ( AF-1)-2 show a comparison of the SASSI spectral results obtained at the basemat foundation elevation with the free ' field or control r motion input to the SSI analysis. These figures depict the ef fects due to SSI. As expected, it is seen that some reduction does occur as compared with j the free-field motion. Also as expected, the interaction ef fects are shown to be more significant in 'the range of the structural frequencies (e.g., approxi- i mately 3 to 20 Hz, depending on motion direction, with reductions as large as '
1 45 percent). At very high f requencies, the f ree-field trotion and the inter-action motion are basically identical. This particular ef fect demonstrates the conservatism of the SASSI calculations.
In addition, the SASSI methodology has been shown to conservatively predict reductions of motion over I
4 the embedment depth as compared to actual recorded responses. This was demonstrated, for example, by the L.otung experiment where SASSI predicted reductions of a lessor magnitude than those actually recorded, i -
Empi rica *, results from downhole arrays also substantiate that rapid seismic attenuation with depth occurs in rock (see References 8, 9,' and 10). Seismic data f rom Tateyama and Choshi test sites (Ref erence 8), as well as Iwaki
! (Reference 10) are judged to be "spresentative of the WNP-3 sandstone site. ,
t
- And each of these test sites sho. rapid measured seismic motion attenuations i
1 l o
4 4
y l
< l 'I 9 : .
i
, ~
,s O
s T
v OND
' ~ REL TCE
\ . CRI
/
\ .
EEF 1
PPE I
S E S 2R
.i .
. P D F t
,I C L=
(
'\ . E = 1
,t i IG )
i s Y FNi t 1
. C EIS
's N EPR F
/ E RMD A
(
y\
s . U FG I .Q D er E .
s R S , g u
'I i ,F V T i RR F
/ TON RHE N S e
E=A
, S B RN BO=
!t IS i
D i
i i
i
- - . - - - I, ,
- e. e. s s s .
- _9 2o..- ege1J()1 A L 8 a
O 1
~
9 s. -
z o
t-G ,s e \ r sti J \p%
u2 <
f..
%#'wt g i, e ,,\
,8.s I
.........,f.
f f f i 1
1 2
1 1 1
s.
i t
v.
I t i i
a..
t
- 2. ...
t J.s FRE01;ENCY ( C P .,1
__ -- _. -_ -. _ Q UNP3 BASENAT V5. FREEFIELP SPECIRP #
DIRECTION = VEh , OAMPING = 2 PERCENI SOLIO = BASENHT 00;11 = FREEFIELD Figure (AF-2)-2
oo i
i l
with depth. Moreover, the measured results .are consistent with the WNP-3 analysis results which show a 35 percent reduction of the surface motion at the basemat elevation. Reference 11 also presents SHAKE analysis comparisons with empirical data f rom sites which are comparable with WNP-3. The SHAKE results are consistent with the empiricai data and show significant attenua-tion with depth.
In conclusion, the Supply System finds that deconvolution in a rock media is '
substantiated by theory and published seismic data for comparable rock sites.
Thus, the Supply System reconsnends that the NRC's concern as identified in '
AF-1 be closed.
Finally, in the spirit of the requirements of NuReg-0800 (SRP 3.7.2), the Supply System is obliged to validate the conservatism of the WNP-3 seismic design basis model by application of alternate and independent analysis tech-niques. In fulfillment .o f this SRP requirement, the comparative SASSI analysis spectral results, as completed by Imoell Corporation (Reference 5),
are submitted with this response.
t i
l P
i i
f r
r 41
- - . . - - , -- , - ,,-.~,y --- . - - -- c+ ~-,-- -
~
REFERENCES
- 1) Impell Standard Program SASSI -
"A Computer Program for Dynamic Soll-Structure Interaction Analysis". Version 2.0, March 1985. Users Manual Revision 0,
- 2) letter No. 2115, Brookhaven National Laboratory, A. Philippacopoulos to R. Pichumani (NRC), ' Status of WNP-3 Review", dated October 15, 1987.
- 3) Y. O. Beredugo, et. al., "Coupled Horizontal and Rocking Vibration of Embedded Footings", Canadian Geotechnical Journal, Volume 2, 1972.
- 4) J. E. Luco, et. al., "Dynamic Response of Circular Footings", Journal of Engineering Mechanics, ASCE, October 1971.
- 5) Letter No. 0740-106-001, Impe11 Corporation, M. A. Manrique to D. M. Bosi (Supply System), "WNP3 2-0 SASSI Analyses", dated June 21, 1988, with enclosure: Impell Calculation No. WNP3-01, "WNP3 -
20 SASSI SSI Analysis", Rev. O, dated June 21, 1988.
- 6) R. C. Murray, et al., "Seismic Review of the Oyster Creek Nuclear Power Plant as Part of the Systematic Evaluation Program *, U.S. NRC, NuReg/CR-1981 UCRL-53018,
- 7) W. Flugge, "Handbook of Engineering Mechanics", Chapter 64, "Propagation of Elastic Waves', First Edition, 1962.
- 8) O. Syun'itiro, et. al., "Recently Developed Strong Motion Earthquake Instruments Array in Japan', Research Committee of Strong-Motion E'rthquake Instrumented Arrays on Rock Sites", Circa. 1983.
- 9) C. Y. Chang, et, al., "Engineering Characterization of Ground Motion",
NuRt1/CR-3805, Volume 3, February 1986,
- 10) R. J. Hanson, "Saismic Design for Nuclear Power Plants", The M. I.T.
Press, Library of ( cess No. 79-110237.
- 11) C. Y. Chang, Draft Paper: "Use of Observational Data in Evaluating Theoretical Models of Ground Response", Geomatrix Consultants, Inc., San Francisco, California.