ML20246N319

Forwards Addl Info Re Plant Soil Structure Interaction Analysis/Deconvolution Issue,Per 890213 Request.Results Presented in Encl Supplemental Rept Demonstrate Overall Significant Level of Conservatism in Seismic Design Basis

ML20246N319
Person / Time
Site:	Satsop
Issue date:	08/30/1989
From:	Sorensen G WASHINGTON PUBLIC POWER SUPPLY SYSTEM
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
GO3-89-148, NUDOCS 8909080087
Download: ML20246N319 (36)

Text

..

---m ~ -

N

.p .

.g[_

f .__

l%l ]

J

'4 ' WASHINGTON PUBLIC POWER SUPPLY SYSTEM j

'l

' P.O. Box 968

• 3000 George Washington Way

• Richland, Washington 99352 y -

August 30,:1989' -.

. Docket-No.'50-508 4 'G03-89-148 I

U.S. Nuclear. Regulatory Commission Attn: Document Control Desk Maii ' Station P1-137 Washington, DC 20555 ,

Gent 1emen: 'l

Subject:

RESPONSE ' TO NRCL REQUEST FOR _ ADDITIONAL INFORMATION ON WNP-3 '

SSI ANALYSIS / DECONVOLUTION ISSUE

References:

~1)' Letter,'G. S. Vissing (NRC) to D. W. Mazur (Supply System),

WNP-3 SSI Analysis / Deconvolution Issue, docket 50-508, dated February 13, 1989.  ;

2) Letter, .G. C. Sorensen (Supply System) to U.S. .NRC, i

' Resolution of Key Licensing Issues, dated August 29, 1988 (G03-88-235).

.3) Letter, G. C. Sorensen (Supply System) to U.S. NRC, Preparation of. Draft Safety Evaluation Report, dated May 1

~-25, 1989 (G03-89-084).. j I On : January 12, 1989, the. Supply System presented to the NRC responses to .

L specific, Brookhaven National Laboratory questions -on the subject of  !

soil-structure interaction (SSI) analysis / deconvolution for WNP-3. The l SASSI model results, which constituted the majority of the question  !

responses, were . favorably received by the staff. However, development of supplemental information centering on model parametric studies and comparisons of SASSI results with available empirical seismic data was requested via the Request for Additional Information (RAI) transmitted by

' Reference 1. In accordance with this request, the Supply System is hereby submitting the attached supplemental report in complete response to the  ;!

RAI.  !

By way of background, Reference 2 forwarded the results 'of an alternate methodology that verified WNP-3's soil-structure interaction analyses based l on tP 'esults of the SASSI code, a state-of-the-art analysis tool for SSI.  !

Thes >ASSI results substantiated the deconvolution effect and revealed  !

wide .prgins of conservatism existing in the present (i.e., original) WNP-3 l seismi . design basis.

The results presented by the attached supplemental report again demonstrate the overall significant level of conservatism existing in WNP-3 seismic design basis. The Supply System draws this conclusion primarily from three major facets of the results. First, comparable site seismic data shows a strong trend of significant seismic motion attenuation for depths ,

comparable to the deeply embedded WNP-3 basemat. Furthermore, in benchmark I 8909080087 890830 g6 PDR ADOCK0500g{j9 j

.M \

{- a

U.S. Nuclear Regulatory Commission Page 2 August 30, 1989 RESPONSE 10 NRC REQUEST FOR ADDITIONAL INFORMATION ON WNP-3 SSI ANALYSIS / DECONVOLUTION ISSUE studies utilizing the actual recorded ground moti on , the SASSI code conservatively underpredicted the measured attenuation with depth.

Secondly, comparisons of the basemat and surface control motion show that the basemat design basis spectral motion largely envelopes the surface control motion. Lastly, power spectral density (PSD) curves developed for the WNP-3 control motion and basemat show a rich response in all frequencies and a very favorable comparison with the recently proposed target PSDs as presented by Appendix B of NUREG/CR-5347. These factors, as well as other elements contained in the response to the RAI, supplement the results of Reference 2 to affirm that the seismic deconvolution methodology is sufficiently conservative to assure that the public health and safety will be protected during and following a safe shutdown earthquake.

In the same Reference 1 letter, the staff also expressed the desire to conduct an audit of the computer generated calculations that have been made using the SASSI code in order to independently assess validation and correct implementation. While the methodology and implementation of the SASSI code has previously been used on at least one other NRC licensed plant, the desire for audit by the latest staff reviewers is understandable. The Supply System is concerned, however, that the uncommitted schedule for this audit ("at a later date") continues to make the closure of this long outstanding review even more difficult due to the loss of technical continuity and consistency of the assigned staff reviewers. Alternatively, the staff is encouraged to revisit the 1986 Technical Evaluation Report, the resultant Safety Evaluation, and the supportive documentation that has been made part of the record for the San Onofre plant regarding the application of the Impell version of the SASSI code for their Seismic Reevaluation program. By review of the San Onofre docket, the staff should find the code methodologies and implementation (similarly executed for the WNP-3 evaluation) to have already been assessed to the satisfaction of the Commission. However, if the staff still desires i to perform an independent audit, we request that it be completed in time to i support the issuance of the SER un the deconvolution issue before the end of calendar year 1989.

As you are aware, the Supply System has committed to the preservation of the WNP-3 plant physical assets, including the technical and design related documentation. Questions on SSI Analysis / Deconvolution are of particular importance to this project because of the originally perceived potential design impact and because the NRC concerns have been cited as potential impediments to plant operation by regional planning groups. Thus, we are particularly anxious to close out these longstanding open items. The Supply l

System has responded to all staff requests for additional information and l the results of our studies have confirmed the adequacy of the methodology I

utilized to develop the exi sting seismic design. We ask the NRC to

! l l

L- - __ _

L.,

l j ..

( .

U.S. Nuclear Regulatory Commission

.Page 3 August 30, 1989 .

RESPONSE TO NRC ' REQUEST FOR ADDITIONAL INFORMATION ON WNP-3 SSI ANALYSIS / DECONVOLUTION ISSUE

, evaluate the information we have submitted and to document the findings'in a SER, as this is an essential element in the preservation of the technical and design related ' documentation.

Based on our most recent contact with the NRC. Project Manager for WNP-3 (August 25,- 1989), we understand that the staff is currently finishing preparation of the draft SER.on the WNP-3 Geosciences Program in accordance with the schedule requested in : Reference 3. The staff has-the opportunity to sustain the efforts and continuity of the few WNP-3. knowledgeable personnel who have been involved in both the Geosciences and the

. Deconvolution ' Programs. We, therefore,- trust that the additional information' being transmitted herein will allow for the completion of the.

Deconvolution review to proceed without any delays.

We are prepared to assist in the final resolution of this issue in any way possible. If there are questions, please contact Mr. Richard Latorre, WNP-3 Project Licensing Manager, at (509) 372-5142.

Very truly yours,

~b G. C. Sorensen, Manager Regulatory Programs (MD 280)

RL/tir Attachment cc: . Mr. G. Bagchi, NRC Mr. R. G. Bailey, Puget Sound Power & Light Co.

Mr. W. L. Bryan, Washington Water Power Co.

Mr. C. Goodwin, Portland General Electric Co.

Mr. J. R. Lewis, Bonneville Power Administration (399)

Mr. T. A. Lockhart, Pacific Power & Light Co.

Mr. J. B. Martin, Region V NRC Mr. L. Reiter, NRC Mr. N. S. Reynolds, Bishop, Cook, Purcell & Reynolds Mr. R. Rothman, NRC Mr. L. Rubenstein, NRC Ms. R. M. Taylor, Ebasco (Elma)

Ebasco (New York) i

4 " c. . _. ;

l .. .

~

. 1 RESPONSES TO USNRC REQUEST FOR ADDITIONAL INFORMATION-DATED FEBRUARY 13, 1989 HNP-3 Soll Structure Interaction Analyses Completed In Response To Structural Audit finding On Deconvolution Washington Public Power Supply System August 31, 1989

r_ _-. . . __ -_

y ,,c i

b,# ' ITEM 1-Compare- the original ground motion 1nputs at the plant grade and basemat levels with those used in the SASSI analyses. Provide a discussion of the differences.between the two sets of input and bases thereof.

l RESPONSE-In' the A/E generated design basis seismic model the ground motion enveloping 4 the required Regulatory Guide 1.60 spectral shape, anchored at 0.329 (SSE) for I i

~NNP-3, was devolved with the aid of the SHAKE code. The SHAKE code yielded a

. subsurface motioer which, when applied at. a depth of 570 feet to a finite ele-ment model of the HNP-3 sandstone foundation monolith, yielded the required ,

grade level control motion. The classic lumped mass and beam element repre- '!

sentation ~ of the WNP-3 structures was then combined into the finite . element l foundation monolith (accounting for the depth of building embedment)- and the )

combined seismic model executed with the devolved control motion input at the 1 570 foot below grade depth. In brief, this procedure constitutes the A/E's finite element / finite boundaries deconvolution seismic design basis model as l

. described fully by the HNP-3 FSAR. j Both wave mechanics and empir? cal downl. ole seismic recordings demonstrate that 'l reductions in motion (e.g. acceleration) occur with depth in rock structures comparable to the HNP-3 foundation media (see data provided with this submit- l tal). Attenuation of seismic motion occurs largely at the surface such that I at the elevation of the deeply embedded WNP-3 basemat (60 feet) significant i reduction in motion has occurred in the free-field. Regardless of the observed attenuation of motion with depth, comparisons of the grade level control motion (R.G. 1.60) with calculated seismic responses at the basemat are required to affirm, in part, that a conservative seismic design basis has been developed for NNP-3.

In fulfillment of the required comparisons both the horizontal and vertical (SSE) " design basis" basemat spectra are provided and compared with the corresponding surface, or control motion spectra. Additionally, with each  :

spectral comparison the SASSI predicted center of basemat, and the SASSI i predicted motion in the free-field at the basemat elevation are provided. As a point of clarity, it is underscored that the applied grade level control motion is identical between the original A/E design basis seismic analyses and that used in All SASSI analyses completed for HNP-3. The comparative SASSI results illustrate two trends, first, the predicted attenuation of seismic response with depth, and secondly, the " differential" response between basemat and free-field motion at depth resulting from the SSI effect.

Additional comparisons of SHAKE and SASSI predicted responses at the basemat elevation in the fre_e__f_1.e.M are also provided. These comparisons are provided 1 to demonstrate that both codes are consistent, that '.s. separate results from the codes, using identical inputs, can be meaningfully compared. And secondly, the comparison assures that SHAKE generated profiles (strain compatible proper-ties) as applied to the SASSI SSI phase of the analysis pr >cedure are valid.

Figure 1.1 contains four plots (all at 2% damping), as follows:

1

___ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ . -. J

., A.- The design basis free-field input ground motion response spectrum. This response spectrum envelops the R.G. 1.60 response spectrum shape anchored at 0.329 peak ground acceleration.

B. The SASSI predicted response spectrum corresponding to the motion obtained at the basemat location and considering the effect of soil-structure interaction.

C. The SASSI predicted response spectrum corresponding to the motion obtained away from the structure (in. the " free-field") but at the elevation corresponding to that of the basemat.

D. The original (Design Basis) broadened response spectrum corresponding to the motion obtained at basemat center location. This spectrum was obtained from the design basis finite element analysis of the soil-structure system.

Some discussion of each of the above curves follows:

Curve (a)'is the " control motion". It represents the input motion to both the original design basis analysis and the current validation analysis. Curve (b) is the motion obtained at the basemat center location calculatcd by ~ SASSI.

Thus, the difference between Curve (a) and Curve (b) represents the effect of SSI including both, inertial and kinematic interaction effects. As expected, the SSI effect is most pronounced in the range of the structural and site fre-quencies. Curve (c) is the motion obtained by SASSI at depth in the free-field, at an elevation corresponding to the center- of the basemat structure.

Comparison of Curve (c) with Curve (a) show the effect of vertical variation of the ground motion in the free-field, without any consideration of the presence of the WNP-3 structures. Comparison between Curves (b) and (c) show the modifying effect of the at depth free-field motion due to-the presence of the structure. Note that in all three cases the peak acceleration (2pa) values are about the same, without significant reduction.

Curve (d), the design-basis curve at basemat center location is also plotted for purposes of comparison with the SASSI results. As observed, the design-basis curve envelops the SASSI results by a significant margin (with a very minor exception around 7Hz). Also observe that the design basis basemat 1 spectra largely envelopes the surface control motion, thus demonstrating a satisfactory level of conservatism in the A/E's design basis model and procedure.

, Figure 1.2 show similar results as figure 1.1 but applicable to the vertical direction.

Figure 1.3 shows the response spectra corresponding to the motion at the base-mat elevation away from the structure (in the free-field) as obtained from SASSI and SHAKE. It is seen from this Figure that both the SASSI and SHAKE results are very close, practically identical. This figure demonstrates the compatibility between the SHAKE and SASSI models as they both predict the same response in the free-field. This comparison serves also to demonstrate that the SHAKE and SASSI modeling of the HNP-3 site is consistent, and that the SASSI profile as derived from SHAKE is adequate for SASSI SSI evaluations.

Figure 1.4 shows similar comparison as figure 1.3 for the vertical direction.

f .....

Thi ; l i

l r ,. ..

1

)

s t

E 1

i g

!- -3 if ei ! 3 C

.f E 4

1. O W

E: L E

-

• m D

$"J E'

~o*

J w >=

r o m C 8a Z Ch>

I s. W L J k , % - . ).. - , m6w

~C . (*

4 W

o. g .s7 j -

- g yg>

A E CL t%= .o 4 g

to .

[ -g

'-* 4' N

' ~

' N

_Ny

.e 2 N

~

"u. Eag y@ g ,

.r.,, .- -

e zo o

e t . >. Z e.

n. -

2 t..g~ -

g . _.

e

.s .

m '

Wpc E

=2O 'm"" Q m.~5..

m@ m.,

-r w >- W -

=

C Em m -E ' . E*

. e ." '

r"~ E OE m kO - gL E' if Q # s

> mmm v b ~6 L 3, ' m

.cf o E{ ._

jW

. m p 8' y E

_ W g . C e L&. ' ,

.9 E 9 .9 oe ea E W k ZE y g- EEo oe w s-20E m- CD s""

.~ w D .

c6 Cllc J u W o

O _m kmm a t C/) u) *iT>

i L

A gy+

W e $ 2 g3 0

m

.s(<O WWW .

M o *c

,t g Il 11 11 11 , ,- E pe ~~n~ t sr> w WW 8083 _ I, mO a

g . .

' E m

o tr>

o L m

e . t . I i P 3 e o e* as

&I & o e (D) NDI16H y336

?I I

m Q

i Em. R R

T d

m .

C E

P I

n fn;, -

S N E

O I

~

S T d

f N R O V

'4 P E T

S N L LN E ECE e

s I R R -

n l T

o L E P R t

p ,

s

! I R M e S t E 2 R n i . P l S o - -

i C i A d it ' ( RB e a 9 E G t

ic v e

y' Y V N D d , t C L e lE N ,IE r

P I 2 P t a E S R MF I

- i U 1 S t aes m - - Q E

I S D' E e r

Sg ,[ Y E u An m a , R L ,R g Sip seB i

s. F R F i

N R F

s.m at v Da Bad RE+ V n f o ie l

L o

i% n r

. A=E C ot eF t I a

o2 T R M n, in

es .. N N O F o

ce e r is E R loirt MCFa B MT U I

t c e- . E S C ne cI I n ., L or i aSSg P RE+

CD f rSSs i

- P I el uAAe -

U T caic a SSSD

\

- I i.

S D R

==== M f t r r ue ))))

abcd k ~

I .

I S S E

SV ((((

I S R I t B

,I

. S S

P

_ ~ - - . - - - / s. P

. l I '

W 2 zo Q 5t_w h y

n .

r o 1 i I t m a -

v e a l

E t a -

a c m R e R s T a C B t E

m P d S l N e ~ E O i - I F N S T N

e O TR e P NV EE r S F .

E CL I e t

R RE

, t E g ,P I

.)

n t e Z R i S 2N I

p . 3 .P t

R {

m C S a

D t

e

(

O=R H

s CB Y N i% I s.C ,ID s2 N y

l , E S- NE PL U I S RI I

an Q no -.

E Y OF Ai L t R ,E A

E c 3. F N URE I

K e E A r I K Hi S A RF SD S H L E .

A S A

'. t S ss I

a

. T=E.K

= N NR ve = E OH W d N I - d e St i h E IS T S s l s L CD A a o a P EN SE S D P RR U I S DI I .

.t I S I .

S S S R 1

S t

t S

t

. S P

~ . . - . - - .

P U

e

. .. l G~ zo_>gEtoEguE

,T

m )

< r 1 n f w.

o i lt t R a .

v e

l .

E i

m R R

t T a . C m E e P s i S

a B

E N S O d I l N e O TRT i P F S N V.

E EE CL e B

. R RE e .

r B L E F R PT C 2MR i I

, S I g t

v. P E i

n C

(

T=S R

E GBR i

p m Y V a N

s. C I

sD N

,I D 4 i PL s% E ME y2 I U S RI 1 I

l Q S DF a , E nn R Y ,E Ao . F L R RRE I i 3

i F Et N EF K c E R V A e I K ,

H r S A Si S H L=E A

D A S I T K sl S

=

N NR OH va = E IS c d MT Ii d e E CD St i h L EN S r l s P RR A e o a P I SV S D U DI I

l

. S S I

. S I

. S R I

S S t

I I

. S S

P

, ~ . - - . - .

s. P

s. .. I W

,, s t '

c G~ zawWaowa.w ua L,

.o : .c L ,

ITEM 2 Provide approp* late empirical measurement data .in rock as a part of the justi-

!. fication of ' the deconvolution of ground motion in rock foundation material.

E (Compare the subsurface traterial properties and reduction in motion with depth for the WNP-3 site with those measured at other sites where field data have been obtained.)

RESPONSE

The fundamental premise of. USNRC Audit finding No.1 is that a reduction in--

. motion between grade level and the HNP-3 basemat is not expected since the foundation media is rock and will behave in an essentially elastic manner.

The Supply System could not concur with this premise since wave propagation in-an elastic media would predict attenuation with depth, and thus the A/E's design basis SHAKE results were consistent with theory. More recently "down-hole" seismic recordings from comparable rock sites have substantiated that significant seismic; attenuation does occur in the range of the WNP-3 embedment depth.

Downhole array sites on rock are limited in numbers on a world-wide basis.

The Japanese constitute the main source of downhole test sites in rock.

Hansen .[1] was perhaps the first to summarize that testing at a rock founded Japanese power plant demonstrated th&t the peak seismic motion at grade (rock) was about twice as large as that measured at depth. A summary of the

. pertinent test sites known to the Supply System are summarized as follows:

DQynhole Test Site Reference Shear Have Velocity (ft/sec)

Choshi- [2] [3] 4580 Higashi-Matsuyama [2] [3] 2450 Iwaki [4] 4200 McGee Creek (USA) [5] 4300 Shuzenji [2] [3] 2300

-Tateyama [2] [3] 2615 Tomioka [4] 2300 The cited shear wave velocity in the above table is a mean value for the rock layers within the first 200 meters of the surface. For reference purposes the mean shear wave velocity for NNP-3's near uniform (i.e., homogeneous) sandstone foundation media is about 3800 ft/sec. Results found in the literature show that seismic motion attenuation with depth occurs, and is prominent for all of the noted rock sites. However, from the above list it is seen that Choshi, Iwaki, and McGee Creek have shear wave characteristics most comparable to the HNP-3 site. Thus, appropriate recorded seismic data is extracted for Iwakt and McGee Creek with this response. Other than gross results, no detailed data is available for Choshi. It is noted that Tomioka data, although not extracted here, is provided with the Iwaki data in NUReg/CR-3805 (Volume 3).

Comparisons of these two sites show that seismic motion attenuation with depth at Tomioka is consistent and quite similar to the Iwakt site recordings.

.- .q In.. orde r to complete our question response, reference to the SASSI results provided with Item 1'of this submittal will be made to illustrate the seismic attenuation -with depth for the HNP-3 site, But first, to assure that the

. SASSI code properly and conservatively predicts rock site responses the McGee Creek site was modeled and then executed with the recorded ground motion event applied at the grade level elevation. The SASSI prcdicted horizontal component response with depth is then compared with the actual McGee Creek downhole seismic recordings.

The McGee Creek array is located in the Hammoth Lake area of California and consists of three-component strong motion accelerometers located at the surface and at- depths of 35.0 meters and 166.0 meters below ground surface.

Site characterization and analytical work on results from this array are described in Reference 5. Reference 5 describes the site as consisting of three layers: glacial till from om to 14m with a shear wave velocity of 330 m/sec, glacial till from 14m to 30m with a shear wave velocity of 620 m/sec, and hornfels from 30m to 166m with a shear wave velocity of 1320 m/sec -(4300 ft/sec). Figure 2.1 taken from R?ference 5, shows the site profile and instrumented locations. The recorded acceleration time history applied at the surface corresponds to the Round 'ulley, California earthquake event of November 23, 1984.

The McGee Creek SASSI model consists of 23 layers overlying a half-space which is modeled within SASSI. The layering is appropriate for frequencies up to approximately 30 Hz. Material damping is assumed to be l'/. for all layers and only vertically propagating shear waves are considered in the SASSI evaluation.

1he ground motion is applied at the surface and corresponds to surface record-ing of the above mentioned Round Valley earthquake event.

Figure 2.2 shows the comparison of response spectra obtained from the actually recorded response time histories and the SASSI-generated responses. Curve labeled number 1 in the plot is the response spectrum corresponding to the recorded motion at 166m below the surface. Curve 2 is the. recorded response spectrum at 35m below the surface and Curve 3 is the response spectrum corres-ponding to the recorded motion at the surface level. As mentioned above Curve 3 represents the input to the SASSI analysis. Comparison of these three curves indicates that significant reduction is obtained in the first 35 meters (in the glacial till layers) with a lesser magnitude reduction between 35m and 166m.

Curves lA and 2A in Figure 2.2 correspond to the SASSI predicted responses at depths of 166m and 35m, respectively. It is seen that SASSI conservatively l predicts the denmplification of the motion with depth, i.e., it predicts less l reduction than what is actually recorded. The peaks are not quite damped out; l

however, consistent trends are observed between actually recorded and SASSI predicted response, i.e., the peaks and valleys are in good agreement.

L Based on this limited study, we conclude that: (1) the SASSI approach does conservatively predict reduction in motion with depth, (ii) the SASSI pre-l dicted reduction is less than what is actually recorded and (iii) that reduc-tion of motion with depth is feasible even for quite stiff sites such as the McGee Creek site.

• o.g

. i The Iwakt site is essentially an all sandstone site very similar to the WNP-3 site. At Iwaki the shear wave velocity at grade is a minimum of 2750 ft/sec, increasing to 3860 ft/sec in the first 20 meters (65 ft) below grade elevation.

A ready comparison of the attenuation of recorded seismic motion with depth at Iwaki is provided by the spectral ratio plots developed by Omote [3), and t reproduced in Figure 2.3. The figure 2.3 plots are the average results for ,

some 28 separate seismic events recorded at Iwakt. In these plots all spectral responses are normalized to the responses recorded at a depth of 330 meters.

It is easily seen that the attenuation in the first 20 meters is significant, ,

t especially for frequencies above about 2 Hz. In this frequency range the recorded motion at 20 meters is typically 40 to 70 percent of the surface ,

motion magnitude.

l A SASSI model of the Iwaki site was developed in the hope that actual ground motion time history data could be secured. Unfortunately, after numerous i inquiries it is concluded that these records are not available in the United States. Even though the frequency content and magnitudes will not match, the HNP-3 synthetic time history was applied to the Iwakt site SASSI model for, at least, heuristic purposes. Figure 2.4 plots the subsurface attenuation magni-tude (percent in the first 20 meters) as predicted by SASSI for the Iwaki site using the WNP-3 control motion input, and compares this result with the cor- '

responding percentage attenuation as derived from Omote's recorded data.

These results suggest, from the expectation of attenuation observed in actual recorded seismic motions, that SASSI again conservatively under predicts the i attenuation with depth. However, since actual digitized seismic records are not available for Iwaki, the McGee Creek results presented above serve as a verifiable benchmark of the SASSI capability and conservatism in rock site modeling.

Finally, the HNP-3 predicted reductions in motion with depth (i .e. , at the basemat elevation) can be seen by inspection of Figures 1.1 and 1.2 provided  !

with Item 1 of this response. In the cited figures, Curve (a) is the surface control motion, and depending on whether the basemat or free-field response is of interest, either Curve (b) or (c) may be compared with Curve (a). For the i basemat response the SASSI predicted attenuation (Figure 1.1) reaches a maximum of near 50 percent at 10 Hz for the horizontal case. As seen from the Iwaki data presented above this is a realistic, or even conservative, attenuation magnitude at this frequency. It is also observed that the SASSI results  !

conservatively converge to the ZPA, whereas the Iwaki recordings show signif-icantly attenuated high frequency components (Figure 2.3). As a last observa-tion, the inherent conservatism in the HNP-3 seismic design basis is again  !

demonstrated from figures 1.1 and 1.2 in that the A/E's design basis spectra l for the basemat, envelop, or nearly envelop, the surface control motion over the entire frequency range.

In summary, it has been shown from both empirical results and solutions taken  !

from analysis coces based on, or emulating, wave propagation theory (SASSI/ ,

SHAKE), that seismic attenuation is present and significant in rock at depths l comparable to the deeply embedded HNP-3 foundation basemat. Pertinent down- j hole seismic response data from the McGee Creek and Iwaki sites have been pre-  !

sented and citations to other significant rock site test data provided. A

• ,.  : O g O i benchmark analysis against the McGee Creek recordings demonstrated the SASSI code's capability - to conservatively replicate measured responses.

A SASSI analysis of the Iwaki site, using the WNP-3 control motion. . was completed to assure 'that consistent and conservative results (as compared to recorded values) are obtained on a rock site similar to the HNP-3 site. Comparisons of

' the SASSI predicted reduction in motion for the HNP-3 site (basemat) were completed showing' results . that are consistent yet conservative when compared to measured downhole responses. It is also observed that the SASSI results at depth all conservatively converge to the. surface motion magnitude for frequen-cles approaching the ZPA. . Recorded responses (Iwaki) show strong deamplifica-tion at high frequencies. Lastly, the actual design basis spectral responses at .the basemat were shown to be conservative in that they nearly match, and

- often exceed the surface control motion by a wide margin.

l l

l i

i e 1 1

1

\. - _ _ - _ - . _ _ - -. _ _ _ - _ _ _ _ _ _ _ _ _ _A

V ,

l McGEE CREEK RECORDING SITE

~ '

RECORDING INSTRUMENTS g

AND ACCELEROMETER -

I- PACKAGE p

E VELOCITY TRANSOUCER -d PACKAGE ]<

S

-~~8

. as.1 --

ACCELEROMETER-VELOCITY TRANSDUCER PACKAGE uncaos

' ~

ACCELEROMETER =

VELOOITY

' ' '~ TRANSOUCER PACKAGE The _ _ and geslagy as McGee Cseek. The fism 30.5 m me glacal till At espahs asum 30.3 m to le6A a Ibs moduun a homials. The mammasum aus locumut as 0 m.35 mi and us m.

Figure 2.1 (from Reference 5) r 1

- w d

e r s u ,t' s

a f i, e

' 'f .

M A

• 2 , R R

s 5

gi'\g T v s' C 0 l

4 ., E .

I af I i t P 1 S

8

,,' S =

S A ' k' l*

, 1 E M P

S I o S R f.1 'i - i N T O n

l f O N l o P E i

s e, /[/

v'

'r v.

S E C R

l u .

r R S E

a P p ) . I m S Z S 2 o ' A P I S C 1

s. C R R O=S e

s nm o6 m

.I

[ (

Y C

HG=

N

,IH S

p65 m N s1 3O 3 E S P M R 2 e U I R 0 2 Rhhh 3. Q S O ,

ttt E Y O e eppp L ,E r l eeem R u F R R g oddd6 m N 0 U i h 65 R 7 S F wttt1 3 oaaa .

L E 2 R D -

tt 3 A

ddd aa

,eee T=M=

yrrrII N T D d uuuSS E N I usssSS ME L t aaaAA E N O S eeeSS - L O S MMM - P P ,

e - - P M t - - -

i Ah U O E S1 2312 i S C T I

ee I S ev . S G r S E cu , E MC C S c

. S M P

P

~ J - J - J - J ~

. W

' ' s.

3 1

a. 1 -

' o o o '

- 2am Eaujyu" L w' il?

a. ... ,

4 '

N o. 6 ' '(NENd GL- C 3 N o.1

- 1 S a- DN o. 5

" I 8 m- ON o. 4 7s.

f,-130m- ON o. 3

.c -! 8 8 8- ON o, t E

-330m- O N o.1

, . Absamma Location of Downhole Instruments 5-m 7 (E 4.8ml/mHE-33Dal

, s (E e. sal /mua ana) s (E-mawmHE neal

"" 1 -

4 (E-MalthME anal 4 (R lapultunH E aata) at (E atta)#mHE assa)

y.

y ./

e 2~ .

l f 5

[j '

1 .

/> 7-%---b5---

0 .

. ,....i . .

. i ....; i i 's j 0.02 0.05 O.1 0. 2 0. 5  ; 2 5 10 Period (sec) )

- l Average Ratio of Spectral Acceleration - Iwaki Site (Reference 3)

Figure 2.3 l

4

)

i l

I

n 5 i

o ih n 7 ~

.o n  ?

.I 3

t c d mi 3 gS S

n i d n 0t a '

hea2a e nR u t wr c u 5 tS t

c n

iS s

uot n 7

/ o o ut s daas a

d e

^ not t

e e 5 ehensn rt moia f r

e i u 0

' 3 6 e ni'sst5m.R i

d t p t n a i

/ a a eh cn Air

'to n' 5 k 5 e t n.t ui G c) hS oo p d n O rgt i e es dio et r c ow uB e

l I

W o Nk i a

w ti e( p g tt d p i enaa eo r

dit mrdss 8 i m

ra t o i i 2

it0 o 0 et )

ud I So2 c m e6

.e t im o

t z

s Rteh o i t r n k r a p o kl a aa do n ni s al wtro' 20e H et nD e wr e e ati N/t I

It re a (

e cuw6t7 dn Mcr r e epcer ect sobo3 S eo t c c

y c

et i3 n vsP e N d -

eP 5eu m r N' PW 1

q 0 ( e de2t $ r F

I eit tS a i

c i n

• [ s S

S R

' 0 1

dai k o -o S ewtI o ^

r M P o ' 5

/

I .

/

S S

'o A A

~___O S 0 R 0 ! 5 05 0 5 0 5 0 5 0 g 0 8 7 7 ! 5 5 A 4 3 3 2 2 1 1

)

$g >. O cLO Tm1 goe y

'.T

.. ~ x. }

m.

ITEM 3 Justify the modifications made to the original structural model for the SASSI

~

calculations. At the meeting on January 12, 1989 it was noted that the stick model?for the auxiliary building was not taken all the way down to the basemat. ,

.What would be the effect of taking the auxiliary building stick model down to l the basemat?. ' Also, justify the use of one-half of the stiffness of the wall y to account for concrete cracking. j

RESPONSE

The SASSI analysis model presented on January 12, 1989 included the Auxiliary Building modeled as a single stick above grade elevation. Below grade eleva-tion, the building was modeled with two sticks, one on each side along-(and in ,

contact with) the-side soil (see Figure 3.1). The stiffness and' mass proper- I ties for each of these two sticks were taken as one-half the corresponding full  !

stick values, a reasonable approximation given the symmetry of the structure. i In response to the above question, the model presented last January 12 was )

modified, consistent with the NRC's request, as follows: The " stick" repre- i senting : the Auxiliary Building was taken down to the foundation level. At i each major floor elevation (corresponding to the mass point locations) the stick model was connected to the side soll layers by means of rigid horizontal links representing the rigid in-plane behavior of the floor slabs. This model-ing approach is essentially equivalent to the original A/E model as shown in-FSAR Figure 3.7.2-2. The SASSI analysis was re-executed and new response spectra were generated at the basemat level and at the top elevation of the  ;

building. ]

Figures 3.2 and 3.3 show the comparisons of response spectra obtained using the " January 12, 1989" model and the revised model at the basemat and top

~

building elevation, respectively. At the basemat level, minimal differences are observed. At the top of-the structure, the revised model shows a broader  !

peak response but reduced in magnitude, and exhibiting a frequency shift from  !

about 10 Hz to 8.5-9.0 Hz. This is attributed to some differences in rocking response resulting from the - two modeling approaches. Superimposed on these plots is the design basis response spectrum. When compared with the design basis responses however, the above differences resulting from slightly different modeling assumptions, are of no consequence as the resulting responses are fully enveloped by the design basis curves.

In conclusion, the differences in modeling approach do not affect the response ,

of the structure at the lower elevations. The maximum effect is obtained at the top of the structure and consist of a slight shif t in building response frequency and a reduction in the peak amplitudes of the response spectrum ordinates. The design basis spectrum envelops both the January 12, 1989 model i results and the revised results by a significant margin confirming the conser- l vatism of the design basis results. l The modeling approach in our " January 12, 1989" Auxiliary Building model (split the single stick into two sticks for the below grade elevation portion) was developed to properly account, in two-dimensions, for the Auxiliary Building i foundation embedment, and thus is not related to a cracked section.

I l

j

Z (Vertical) '-

Mass Point Q

[

(

r X (East-West) FlexiMe Beam Q Rigid -

Beam .._

Shield Building g., , Containment Yessel cM n b y 4, V4541.

m 87 o N , ric co b b"m no esy.co m.go ,

nt7e oy

'9 s @

__g Mo M Cos.60 s Internals m Reactor 4 Structure

@ #" 4h Auxiliary Bldg 34 o 4 75-ro ~

jnyg.co MD 44*7 4o g g 7, @& g g M.).4D b 4,g .s o '

ri -

e __@ M o' a4r a 4 arn d'M o Wade g 14110 , a ss,400 oo , HS **,y v i' i @ o m ro !? y 4elI112An j' ;

it a i m n.

,e du221 m g/ i l<f,y, W40 li 9,q., , / foi \

)! b id,L_.___E _ . A - _ t r a _ L M _ e_ 3 Ih5 _ _ _ _ _ _

j!a7

.i ,,n _.E Primary Structres with Foundation (All elevations in feet)

Figure 3.1

- , ,'

• l  : :! ,;i' .! i i i {  ! i1jliI1i

. . .d

~ .

f e s

I i .

m T

, t R

- m M s E S S A L

- ,. B E

- D s , F TO l , -

O NM e t c\ g '

E da .. CG o m  %

I R RN Me

~

a f E EI e' T ks PD ca it B

j. e e 'l ..

i )

S N

E 2I L Sf e .P C U o w., I 3

C g r ,i1 i

( ,=B in e - I dt n S G Y 2 l

iu e I NX I

s. C 3 BC .u l .* N S IU E Y PR e y g r

U L RR M r n t u

lia i ip ' Q A OO N

g xm u ,- \' E R R:C ,

T i

F AaD -

t

3. F L UEA r

t o% .J RR T

E ac2 .

N U E

Ro en i

,l d

e s l t a

M=N E

E ol s i

N wtc Me d a

- L OD Ni ee r l a oB

(.e i P

P IN TR st

.D i nMng U S C vse gwis ,

EL l

i r ee - RA a I i

n W- OND t

=== . I .

S IN S DIG igs .

i r a )))

abc

- t I OE (((

S R

I S O I

P

' P 1

. W s.

o.

3 e ,,

I o zoCexwswOu

'.= .

l l!

ltlli'l'

Im I

G

. N

~.I m

. I D

g n _~ L I

id l

_- U B

iu .

,- I m

B -

~ .

S y '

X L r \ U E a

lsi l i N R D ex t*' ,

TO R NM du <

oA \i

%. , f O E i

i'i{ l Mro i t 1 T CG kt ...

I

. C EI RN cc ia t

-t

- I R PD Se ' ,t I E L e ',' l' ly I g R '

~

, S R 2I

,_ P nf F=U I

io ld p

~*

<<I o.I !e. I C

( O B iuo l .

P G .- 3 BT -

/ Y O NX 3

s. C T IU

, ~

y g ..

i N

r n PH e r

ia i ip l

xma

! ' , < ' J.- ,-

t E

U Q

M

,RR u g

u -

E S OOT i F

Ar D  %- R I

,C

~ .- S

. F R o% ~ i 3 Y L UE t ~

c2 a ,

J.. AR en N E l

e Ro s wc it d

ol es i -

- R=W E ee MdoBa L N N i

- a r

Ni l - A OD st

.D aMn n g ,

T N IN vse igwis E TA ir ee MC lawn -

OND E EL RA igs t

=== L IN P DI ir a )))

abc P OE (((

I f.

. U G I

I S R I s. O

. I t

t S

- S S

. - ' . a . . ' . ,

. S

. I

- P s . ' P i

U x o s. e1a 2 Uua G- .

' . e'

f 4.

4' q

ITEM 4-

-9 Provide and_ discuss the results of parametric studies varying a) the SSI model dimensions used in SASSI and b)'the dimensions of the SHAKE model to show con-vergence of the analysis results.

RESPONSE.

The SASSI soll ^ model presented at the January 12, 1989 meeting consists of

'three soll (rock) layers overlying a half-space. The material properties specified for.these layers and the half-space are consistent with the level of

. shaking (i .e. , are " strain-compatible") and were obtained separately from a

. free-field analysis of the site using SHAKE. Input to SHAKE were the esti-mated " low-strain" properties (obtained from Figure 2.5.D.10 of the FSAR), the free-field. design basis ground motion time history, and the shear modulus and damping versus shear strain developed specifically for the HNP-3 site as presented in Figures 2.5.121 and 2.5.122 of the FSAR.

The SHAKE model consisted of six layers,- each layer of about 25.0 f t to 30.0 ft. in thickness. The model extends between elevations . 389.0 ft. (grade

' level) and 225.5 f t. giving a layered model depth of 163.5 ft. Below eleva-tion 225.5 a half-space is specified in SHAKE. Note that because the site is essentially uniform, the final " strain-iterated" properties of these layers and the half-space are very similar (less than 21. difference in shear moduli among the layers). Because of this uniformity in material properties and because SASSI is able to model the half-space condition (i.e., no rigid bound-artes . at the base of the model are necessary), the half-space was modeled to begin at the level below the embedment depth. Thus, in SASSI, the site was modeled with three discrete layers followed by a representation of the half-space which, again, .SASSI is able to model. Note that this cr.pability of

'SASSI to model the half-space condition is an improvement over other SSI tech-niques such as the FLUSH approach or other standard finite element programs.

.The FLUSH approach (or other standard finite element methods) requires a rigid base at the base.of the soil model where the motion is specified. Typically the motion at this level is derived from a deconvolution analysis using SHAKE.

Men using this approach, parametric analysis are required in order to locate the appropriate depth of the model and to ensure that the response at the.

basemat level is not affected by the location of the rigid boundary. As a ,

point of reference, these parametric studies were completed by the A/E in his original design basis development work with the SHAKE code.

In SASSI, the control motion is specified at the ground surface and amplitudes I of the motion are computed at each layer interface. The half-space is expit-citly modeled by a system of layers whose total depth is automatically adjusted by the pr'ogram and depend on the particular solution frequency (SASSI works in the frequency domain and, thus, obtains solutions at discrete frequencies and then interpolates in ' between the solution frequencies). The capability of SASSI to model the. half-space conditon eliminates the need to perform model depth parametric analyses. Furthermore, SHAKE deconvolution analysis to obtain the " rigid base" level motion is not necessary as the surface motion is applied in the SASSI evaluations. The only purpose of the SHAKE analysis is, then, to obtain " strain compatible" material properties using the iterative equivalent linear elastic approach implemented in SHAKE.

g 1. : . e . . ;

- From .the above discussion, it is clear that, when the SASSI methodology is used, ne parametric analysis for.model depth determination and convergence of results is'necessary.- Furthermore, as part of our response to Item I we have

-shown that our SHAKE model (six layers, layered depth of 163.4 f t. and half-space below it) and our SASSI model (three layers, layered nodsl. depth of 83.5 f t. and half-space below) are consistent, since both models predict the same -

response. To further confirm that the two models are consistent, a parametric SHAKE analysis was performed to show that the " strain-compatible" soil proper-ties do not change when either a three. layer over a half-space or a six layer over a half space is used. The six-layer SHAKE model was modified to contain three layers from elevation 389.0 (grade level) to elevation 305.5 ft., with a half-space below it. This model is consistent with the SASSI model. Results of this parametric analysis are shown in Table 4.1 and show that the properties obtained from the two SHAKE models (six-layer and three-layer) are practically

-identical.

In conclusion, we have shown the following: (i) the SASSI model used in the

'SSI evaluations is' consistent with the SHAKE model used for determining

" strain-iterated" soil properties, (11) the SASSI approach to SSI models the half-space- condition and does not require artificially placed rigid boundaries at the model base. This eliminates any need to perform model depth parametric evaluations, and (iii) the SASSI model is consistent with the three-layer SHAKE model.

.a .; ,

\

i Table 4.1 Original (6-layer) and New (3-layer) SHAKE Analysis 6-Layer Model 3-Layer Model S-Wave S-Wave Layoi Thickness Velocity Damping Layer Thickness Velocity Damping No. (ft) ft/see No. (ft) ft/see 1 27.5 3884 0.010 1 27.5 3884 0.010 2 31.0 3864 0.015 2 31.0 3864 0.015 3 25.0 3847. 0.018 3 25.0 3847 0.018 4 25.0 3836 0.019 5 25.0 3828 0.020 6 30.0 3820 0.021

v4 A ITEM 5 lustify the use of the apparently high value of 0.4 for the Poisson's ratio for the rock medium. Provide .the analysis results for a range of foundation rock properties (e.g., shear modulus and damping) used in the SASSI analysis.

The basis for selecting the range of rock properties should also be discussed.

RESPONSE.

The value of Poisson's ratio (0.4) used in the SASSI calculations is an average value taken from a summary of engineering properties as given by FSAR Tables 2.5.15 and 2.5.16. These tables are based on the results of geophysical tests completed on actual samples of weathered and fresh sandstone obtained from numerous ' core borings conducted at the HNP-3 site. FSAR Tables 2.5.C.8 and

'2.5.C.10' provide more detail by listing the specific core bores (23 total) and the actual tests results obtained in the measurement of Poisson's ratio. The mean value of these 23 tests'is 0.375. Additionally, the classical. theory of elasticity would predict a Poisson's ratio of about 0.38 based on the mean value of shear / compressional wave velocities measured at the. site. Also note that the measured variation of site shear and compression wave velocity is remarkably small (plus/minus 12 percent). Thus, a value of Poisson's ratio of

'0.4 is considered satisfactory.

Additional SASSI parametric analyses were performed by varying the shear modulus of the rock material by plus/minus 151.. The basis for selection of this ran';e is the relatively small variation of the as-measured shear wave velocities for the site as documented in Figure 2.5.D.10 of the FSAR. As

.shown in this Figure, the measured range is generally within 12 percent.

Results obtained from the SASSI parametric analyses are shown in Figures 5.1 to 5.3 and correspond to the center of the basemat, top of the reactor auxi-

'liary building, and top of the containment building. Comparison of- these parametric analysis results with the design basis results shows the adequacy of:the design basis curves, even after consideration of variation of the site rock properties. Variation in damping is not considered since extremely con-servative (low) material - damping is already assumed. Moreover, the SASSI solution radiation damping will, by far, dominate the material damping input.

l

r t

I I

- T

' R I

. M

, . E

',~

S

~

R

.~,

B t

. T a .

F N

~

O E m a..\'\

e s ~

,\

R RLCS a , c s. b . I E EI B , I T

f el , ' ' , '

I N PO .S o ,

• t S E 2 r

,. P C F e 2 I C

t n ,', ,,

( ,=FI e . f ls C e

. Y S GTNS I

d ,g .,, I C I o n  !'. N S PD Mip E Y U L MN A RR im l

o a I

Q S D E N O L

\'

f i

R A ,R L U N f

i%

t S2 I

3

. F f

RI G d , T n n ol s ,. N EI aiit c o is E=R M O f

t o eSr l l a V I

3 E N ,

Si

.DaoilnS SgB i

L on -

P O l

sti v s gf t f s eir oie f i

'( l ' '

P U TF IT CO a t S ES i

nWOSSD gs= = ==

t

.., I R

I ir a)))) S D O E (a(((b c d .

I i.

S I .

I

. S S

k - I

. P P

. W

- ' ~ -

s.

i

' NI 1, g

s.

8-

- Q g _ gg$y sr

t i

G f

. N

_ w I D

g L I

n m

• U i

s ld <

' f B iu ,

B ' -

y ,~ X r g' U ia s R i

l x

,s .

T s' 1 A

u s) t - M '{i 1

i

~ft R E O CS N

t r

o -

a' -

I f

T C RL c I R EI a ,!  !

E POS e t l I. I e

.)

R 2*

R . S v.P F t

f

. F o ,

C F p ,j

• t

(

O=I lseT o ~

r Y P GT

/ i i

C O INS -

d ,g o

e,a, N T PD Mipn '.

1 E

U MN

,AR il m Q o a E S D L S D f

f i%

I 3

.F R 'I S

Y

,R N

S2 t

!.. EUIG d n, RI N E n o ' R D=O ai l is -

t t

c io s a . i .

i

~

f o er S L

. R N l l B Si i laoion -, T IT O st

.D i n S S o~ ,

, N TF vse gt is E M CO f f l

ir oie f a E ES n W- OSSD t

L IR igs t

==== P ir a ))))

abcd I .

P D OE ((((

I f

. U S

I I

s t

S

' S

• S

s. S

. . ' I,

s. '- g P

i P U -

_ u ,en$wJ* -uT Aw'

e iw -

I w

G im. . N I

D

~

L I

g U

in . in.

B d

l

. iu B

t n

m. T N

E MT e N N m  % I n N R ECS ia ~ T N RL a

t ~ .

n ~ i O EI o * ~

C POS C - I )

a S F 2 f

o k's P O=F i

f.

p i ri I

,i C F s To 3 b, I

(

P GT I

le r Y O NS d g, tj

s. C T on I

t' 1 I l N ,P D Mip im /, E U

MN S RR l

I oa - Q D SD -

- E R

I S',L f -

R i%

f t

S2 gI . F 3 Y

L UIN R G d n, NI n ol s A E R aii ceSo t is f

t o B a L=O A N Si r

.Daoion sti l li nS Sg l

Y a.

~

T O

N IT E TF v s gf t f is -

l eir oie f t

M CO a E ES inWOSSD igs = = = =,

t -

L R P I r a)))) P D O E (a(((b c d  :

i U

, S I

. S S

=

S S

s. 'P s

' P t.

5 "- U '

e eO~ ED.ya:wJ0u Ae l i lli

. s .,

L  !

ITEM 6  !

L .

l

-Confirm that' the original design bases of the plant will not change even if  !

the SASSI analysis results are accepted by the staff as justification for the

~ deconvolution of ground motion.

' RESPONSE l

The Supply System confirms that the original HNP-3 seismic design basis will not be changed as a result of the SASSI studies that have been completed in  !

response to Structural Audit finding No. 1. Furthermore, these SASSI gene-rated results, while being significantly lower in magnitude than the design basis ' loads, shall not be implemented or used as a basis, or partial basis, .

for the qualification of any HNP-3 structure, component, or system.

At present, over 70 percent of the HNP-3 construction phase is complete, with i the associated plant design engineering effort being more . than 90 percent  !

complete. Thus, the Supply System views that the plant design is complete and  ;

fixed. Additionally, the Supply System is staying abreast of USNRC sponsor- j ship of various continuing research projects in the area of seismic design for .i nuclear power plant facilities. These programs, at a future time, may result  !

in generic. regulatory requirement revisions that potentially could include l HNP-3. In this circumstance, the Supply System suggests that utilization of 1 state-of-the-art design tools, whether it be SASSI or some other tool, should i be permitted. Of course, any program revision to the HNP-3 seismic design basis would be very closely coordinated with the NRC to assure, prior to imple-mentation, that complete agreement exists on the suitability and limits of application of the new analysis methods. ,

l

)

l l

l 1

l 1

i

- __ _-__ _L

n e-ITEM 7 Provide the power spectral density (PSD) function for the free-field surface ground acceleration time history used in the original plant design. Compare this PSD function to that of the basemat motion obtained from the SSI analysis.

RESPONSE

The requested PSDs, for both vertical and horizontal cases, are provided with the response to Item 8.

,4 ,

h., ,

ITEM 8  ;

Explain as to what extent the SASSI results depend on the choice of the ground motion time history at the surface.

[

RESPONSE

.The SASSI results are dependent on the choice of the time history only to the ~

extent that different histories can be developed to match the same target

. response spectrum and because of this each time history will produce slightly different results commensurate with the differences that each time history's response spectrum is able to match the target response spectrum. The same SASSI inodel analyzed using two time histories that were developed to match the same target spectrum, and that closely match the target spectrum, will give essentially the same results as long as the frequency content and power of the two time histories are similar.

In the case of HNP-3, as part of the original- design basis analysis, time histories were developed to match the R.G.1.60 Response Spectrum shape. The generated time histories actually envelop the R.G. 1.60 requirements with ample margin. To get a better representation of the frequency characteristics of these design basis control motion time histories, the power spectral density (PSD) function of the time histories has been generated and are shown in Figures 8.1 and 8.2. Also, the PSD of the SSI response time histories obtained by SASSI at the basemat level have been generated and are shown in the same figures for the horizontal and vertical direction, respectively. Comparison of the PSD of the above time histories show the following: (1) The PSD cor-responding to the free-field motions show that the time histories are rich in the frequency range of interest, there are no deep valleys in any frequency range, (ii) The PSD of the SSI response time histories show the same trend as the previous response spectra comparisons, i.e., deamplification of the motion due to SSI. and spatial variation with depth. Thus, the differences that we can see between the free-field and basemat level motions are not attributable to the' choice of the time history but are a direct result of the physics of the problem, i.e., SSI effects and spatial variation of motion.

He conclude that the HNP-3 control motion (surface) time histories used for the SASSI evaluations are appropriate, and that because the HNP-3 time his-tories are rich in the entire frequency range, no significant differences are expected by the use of a different time history that will equally match the same target response spectrum.

Figures 8.1 and 8.2 also include plots of the proposed " target" or minimum PSD as recently published (June 1989) in Appendix B of Reference 6. The proposed Appendix B target PSD are scaled to ZPA level accelerations of 0.32g and 0.21g for the HNP-3 SSE horizontal and vertical cases, respectively. Additionally, the target PSDs are converted to g-squared-second units versus frequency (Hertz). The units conversion was completed to attain consistency between Appendix B and the HNP-3 PSD plots which were completed prior to publication of Reference 6. Similarly, the " Function B" time envelope function of Appendix B was used to normalize the PSD time durations for comparison consistency. The

i, control points as provided with Appendix ~ B Equation 11 were followed for nor-

- malizing .the time. durations and plotting, except that in the horizontal case the target PSD was extended to 30 Hz from the 20 Hz plotted in Reference 6.

Inspection of- the Figures'shows basic compliance with the proposed target PSD for both horizontal and vertical cases, The vertical target PSD is completely bounded by the ' WNP-3 . control ' motion PSD. In the horizontal case the. target PSD matches the control motion PSD with only very limited frequency bands where the target PSD magnitude slightly exceeds the control motion PSD.- On an integrated basis (i.e., area under the respective power density curves) it is-seen that the control motion PSD will at least match, or likely exceed, the energy or " power" content of the horizontal target PSD.

I

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . - - _ _ _ _ _ _ _ _ _ _ _ . _ _ .I

w- _,  ;;-.

l

c. (

m p.;

,10 ,

7-m Free field' .

---

Base mat

~

10-3 .  :

s 10 " 7 '.  :

it .

h i v& 10 5 -

'. J i

. j' l

- . *\ . ,

l .

e,d -

l l

Jo i .

5c .

10~' : . i o .

3 .

.ts .

8 .

1. Proposed Target PSD per . _

k 10-7 -

j

. Appendix B. NUREG-5347 (June 1989) j i g .

n. .

i 1

l l 10-8 7 ,

e .  ;

e , .

. e 5 '

. j 10-' 7

. i e  :

.s .

~

I, 10' '

.5 1' 2 5 to 20 50 100 l

.01 .02 .05 .1 .2 i Frequency (Hz) .

1

~ Figure 8.1 Power Spectral Density Function Comparisons -

Horizontal Component of SSE Motion.

T 4-u ++ 'o C 1

~

D p

escWlATptix

' 10~2 . . . . . ....g . . . .

....s . . . .- n

'y

~

Free field .

- ---- Base mat l

10*3 e .

10-* :- -

9, . ..

?

S 10 r

/ i -

')g .

s g

g. -

w .

• b  :

Y104  :

- 1 o .

i.

15 . ' . .

k .

8 l' o  : .

i 8'

V 10-7 :-

I ,

t  :

g ~  :

n. .

l .

- Proposed Target PSD per l -

. Appendix B, NUREG/CR-5347 (June 1989) -

10-s _

-h, { _

- i  :

e .

- t

. 8 e

10 ~' l 7 e .

- e  :

8 .

. 6 i

10~'O

.01 .02 .05 .1 .2 .5 1 2 5 to 20 50 100 Frequency (Hz)

Figure 8.e Power Spectral Density Function Comparisons -

Vertical Component of SSE Motion.

-32

[ <, == o ,

o,

REFERENCES

1. R. J. Hansen, Seismic Design for Nuclear Power Plant Projects, The H.I.T.

Press, Library of Congress No. 79-110237

2. Omote, S., et. al., Recently Developed Strong-Motion Earthquake Instruments Array in Japan, Committee of Strong-Motion Earthquake Instrumented Arravs on Rock Sites, Circa 1983

3. Omote, S., et. al., 'bservation of Earthquake Strong-Motion Hith Deep Boreholes - An Introductory Note for Iwakia and Tamioka Observation Station in Japan, Proceeologs, Eight World Conference on Earthquake Engineering, San Francisco, California, Volume II, 1984

4. NuReg/CR-3805, Volume 3, Engineering Characterization of Ground Motion, February 1986

5. Seale, S. H. and Archuleta, R. J., Site Effects at McGee Creek, California, Proceedings of Speciality Conference on Earthquake Engineering and Soil Dynamics II - Recent Advances in Ground Motion Evaluation, Park City, Utah, July 1988

6. NuReg/CR-5347, Recomrhandations for Resolution of Public Comments on USI A-40, Seismic Design Criteria, June 1989 i