ML20029B344
| ML20029B344 | |
| Person / Time | |
|---|---|
| Site: | Satsop |
| Issue date: | 02/28/1991 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20029B342 | List: |
| References | |
| NUDOCS 9103070077 | |
| Download: ML20029B344 (5) | |
Text
.
/pm ** ouq),
UMTED STATES
[
i NUCLEAR REGULATORY COMMISSION n AsmNGTON, D. C 20b5b
,.... +,
DRAFTSAFETYEVALUATig BY THE OFFICE OF NUCLEAR REACTOR REGULATION RELATING TO THE S0ll-STRUCTURE INTERACTION (SSI)
ANALYSIS /DECONYOLUTION ISSUE t
WASHINGTON PUBLIC POWER SUPPLY SYSTEM WASHINGTON NUCLEAR PROJECT NO. 3 (WNP-3)
@CKETNO.50-508 T'
s draf t Safety Evaluation is on WNP-3 the soil-structure interaction (SSI) deconvolution issue, Section 3.7.2 of the Final Safety Analysis Report (FSAR).
3.7.2.
S0ll-STRUCTURE INTERACTION 8ACKGROUND During the initial review of the WNP-3 FSAR in 1983, the NRC staff had expressed a concern that the Washington Public Power Supply System (the applicant) had allowed a reduction <>f over 35 percent at 10 hertz for horizont61 safe shutdown earthquake (SSE) accelerations at the basemat compared to the free field design response spectrum (Ref. 1).
Since the suppotting medium at the WNP-3 site is rock, the staff'!. position was that such reductions in ground motions at the basemat level wob u not be expected. This was communicated to the applicant as item No. 220.13 in the Request for Additional information in May 1983, and again as the Audit finding No. 1 in November 1983 after a staff sudit of the appli-cant's design calculations (Ref. 2). The applicant was recuired to follow the l
seismic analysis methods described in NOREG-0800, Revision 1 Section 3.7.2.
l Specifically, the applicant was asked to confirm its finite element analysis by performing an elastic half-space analysis, or to treat the site as a rock site-l and perform a fixed-base analysis, in any case the applicant was asked to input the full velue of the SSE/ operating basis earthquake at the basemat without l
reduction (Ref. 2),
in response to the staff's Audit Finding No. 1, the applicant performed an l
elastic half space analysis to demonstrate the validity of the original design basis (finite element analysis), and submitted its analysis results in June 1984 (Ref. 3), and additional calculations in July 1986 (Ref. 4).
The NRC staff and its Brookhaven National Laboratory (BNL) consultants reviewed the applicant's half-space analysis and raised several questions regarding the assumptions made in that analysis (Ref. 5). The applicant's responses to these questions "ere 9103070077 910220 PDR ADOCK 05000508 A
i discussed at a meeting held between the staff and the applicent in December 1987 when the staff pointed out that the applicant's analysis was based on certain i
i unacceptable assumptions.
In August 1988 the applicant submitted a new analysis prepared by a different consultant using a computer program (SASSI) based on the state-of-the-art analysis technique (Ref. 6).
The results presented in Reference 6 were discussed at a meeting in January 1989 when the staff generally accepted the new analysis procedure but wanted additional information including empirical field data to support the theoretical analysis results.
The applicant furnished the requested additional infortnation in August 1989 (Ref. 7).
The staff's evaluation of the applicant's new analysis is described below.
EVALUATION The Category I structures inchding the Shield Building, Steel Containment Vessel, Internal Structures and the Reector Auxiliary Building are supported on a 9 foot inick cuncrete mat embedded 64 feet below the plant grade.
The rock overburden (sandstone) was removed from plant grade level at elevation 390 feet to l
the bottom of mat at elevation 326 feet.
The mat is founded on rock which is fresh or we6thered sandstone having a shear wave velocity ranging from 3000 feet per second (fps) to 4300 fps.
i The significant aspect of the applicant's submittal that was evaluated is the deconvolution of ground motion frcm plant grade level to the basemat level l
thraugh a rock medium. As implied in the staff's Audit finding No. 1 (Ref. 2),
the staff's position in 1983 regarding this issue was that in such stiff founda-tion meterial as described above, there would be no attenuation of ground motion with depth.
The staff reviewed the FSAR using NUREG-0800 Revision 1 (1981)
Standard Review Plan (SRP) Section 3.7.2 (Ref. 8), which required that the design response spectra should be prepared by envelopin methods; however, NUREG-0800 Revision 2 (1989) (Ref 9)g the results from two has eliminated the enveloping requirement.
Similarly, the Revision 1 SRP Section 3.7.2 (Ref. 8) did not allow the reduction of control motion with depth whereas the Revision 2 SRP Section 3.7.2 (Ref. 9) now permits such reduction limited to a maximum of 40 percent from the surface motion with appropriate verification.
Specifically, the current Revision 2 SRP Section 3.7.2 indicates that the spectral amplitude of the acceleration response spectra (horizontel component of motion) in the free field at the foundation depth shall be not less than 60 percent of the corresponding design response spectra at the finished grade in the free field.
Also, the Revision 2 SRP Section 3.7.1 (Ref. 9) permits the specification of seismic input motion at the surface or et rock outcrop rather than at the foundation level in the free field.
The applicant's submittal in 1989 (Ref. 7) provided detailed responses to the staff's last set of questions on this issue (Ref. 10).
The applicant used the provisions in Revision 2 SRP Sections 3.7.1 and 3.7.2, and performed a revised SSI analysis using a state-of-the-art computer code, SASSI, which has been developed by the University of California, Berkeley, and modified and adapted by the nuclear indr.iry consultants for resolving several licensing issues, e.g., the operating license for the Watts Bar Nuclear Plant (Ref. 11).
The SASSI code permits the specification of the free field ground motion at the plant grade level and provides the response spectro et the basemat level.
Unlike the previous finite element analysis
codes (e.g. FLUSH), the SAS$1 code has the capability to model the half space condition, it also eliminates the need to perform a stparate deconvolution analysis to obtain the rigid base level motion as the free field motion is applied at the grade level in the SASSI code.
Using the results of the revised SSI analysis by SAS$1 code, the applicant has shown in Ref. 7 that the original design basis response spectra are generally adequate when compared to the results obtained by the SAS$1 code.
The design basis spectrum envelopes the SASSI results confirming the sdequacy of the design basis results.
However, the applicant obtained these results by varying the shear modulus of the rock material b and-15%only,whereastheReYision2SRPSection3.7.2(Ref.9)y+15%
requires a much larger variation.
Reference 9 requires that, unless the site is well investigated, the variation in soil properties should be considered by performing SSI onelyses using)three sets of sheer modulus values, namely),
!)
the average (or best estimate value, 2) twice the average value, and the average value of the low strain shear modulus (Gmax defined at 10'g half percent peak shear strain),
in view of the fact that the subsurface material at this site consists of weathered rock in addition to sound rock, it is important that the applicant perform additional SSI analyses varying the rock properties as required by the Revision 2 SRP Section 3.7.2.
Furthermore, in response to a staff question in Ref. 10, the applicant has provided some empirical field measurement data in rock as a part of the justifi-cation of the deconvolution of ground motion in rock foundation material. The field data have been obtained by downhole array measurements in Japan and the United States of America (USA).
The applicant also claims that the SASSI code has conservatively predicted the deemplification of ground motion with depth (i.e. it predicts less reduction than what was actually recorded). The field data included in Ref. 7 pertain to three sites at Choshi and Iwaki in Japan, and McGee Creek in USA, where the shear wave characteristics of the rock foundation material are generally Comparable to those at WNP-3 site.
Swever, the staff has not completely reviewed the WNP-3 specific, SAS$1 calculations for the appropriateness of the elastic parameters, the adequacy of the model, etc.,
The l
staff will review these calculations after completion of parametric analyses on, shear modulus es discussd above.
l l
The main concerns in such an audit would include:
(1)theRevision2SRP Section 3.7.2 states that, for structures supported on rock or rock like material, a fixed base assumption is accepteble; such materials are defined by a shear wave velocity of 3500 fps or greater (Ref. 9).
Since the applicant has chosen to perform an SSI analysis of the structures at WNP-3 site where the shear wave velocity of the foundatio.. rock ranges from 3000 fps to 4,300 fps, it is necessary to examine closely the properties of the rock material (such as the elastic modulus, damping, and Poisson's ratio) and the extent to which their variabilities are considered in the SASSI analysis, (2) the acceptability of the empirical ground motion data in relation to the WNP-3 site should also be evaluated and (3) the validation of the SASSI code has been made on a site-l specific basis only in Ref.11, and the practical application of this code to the WNP-3 plant needs to be ex6 mined in detail. After conducting such an audit the staff will be in a position to determine if the applicant's original-designs l
l
based on deconvolution (attenuation of ground motion with depth) in rock are fully acceptable, or if the applicant should perforr a fixed base analysis os required in Audit finding No. 1.
The letter alternative (i.e. the need for a fixed base analysis) deserves to be looked into carefully because of the SRP 3.7.2 recomendation that, for structures supported on rock or rock-like materiel (defined by a shear wave velocity of 3500 fps or greater), a fixed base assumption is acceptable for modeling the sup Mrting soil in the seismic dynamic analysis.
CONCLUSION Based on a review of the applicant's submittals (Ref. 6&7) by the staff and its RNL consultant (Ref. 12), the staff hot concluded that the applicant's new methodology using the state-of-the-art analytical technique for resolving the deconvolution issue is acceptable.
Howeser, beveral open items specific to the WNp-3 application of the methodology remain.
Specifically, the applicent has responded to the staff's questions on the deconvolution issue by submitting the results of its revised SSI analysis, and by citing field measurement data in partial justification of the deconvolution of ground motion in rock foundation material similar to thet et WNP-3 site (Ref. 7).
Although the revised analysis generally conforms to the provisions of the SRP Sections 3.7.1 end 3.7.2 in NUREG-0800, Revision 2 (1989), it does not setisfy a Lignificant requirement in the SRP Section 3.7.2 related to the variation of rock foundation n.aterial properties as described earlier.
Therefore, the applicant should perform additional SSI analyses using ecceptable variation of rock propec ties and demonstrate that the design basis spectro still envelop the new sEcctra as claimed by the applicant in Ref. 7.
Then the staff will check the applicant's SS' calculations through a confirma-tory cudit as stated in Reference 10 and detailed above.
Principol Contributor:
R. Nchumani Dated:
February 28, 1991
_.. ~. _ _ _ _. _ _...
o 1
i.-
5-
REFERENCES:
1.
Memo from G. Lear, NRC, to G. W. Knighton, NRC, dated March 3,1983, l
Subject:
WNP-3 Seismic Input l-
- 2. -Memo from G. W. Knighton, NRC, to D. W. Mazur, WPPSS, dated November 2,
)
1983,
Subject:
WNP-3 Structural Design Audit Report of findings 3.
Letter from G. C. Sorensen, to NRC, dated June 27, 1984, S'Jbject:
Nuclear Project 3, NRC Structural Design Audit-(Attachment 1 contains the response to Audit finding No. 1) 4.
WPPSS WNP 3 Supplemental Response to Audit Finding No. 1, " Validation of Ebasco's Foundation Spring Modeling for Elastic Half-Space Seismic 2-Model",-June 1986 L
5 '. Letter from R. S. Lee, NRC to D. W. Mazur, WPPSS, dated April 7, 1987,
Subject:
Request for Additional Information - WNP-3 Soil Structure Interaction issue i
6.
Letter from G. C. Sorensen, WPPSS, to NRC dated August 29, 1988,
Subject:
Nuclear Project No. 3 - Resolution of Key Licensing Issues - Final Submittel of Soil Structure Interaction Information.
7.
Letter from G. C. Sorensen, WPPSS, to NRC, dated August-30. 1989,
Subject:
Response to NRC Request for Additional Information on WNP 3 SSI Analysis / Deconvolution issue 8.
NUREG-0800, Revision 1,' dated July, 1981 9.
NUREG-0800, Revision 2, dated August, 1989
' 10. Letter from Guy S. Vissing, NRC to D.-W. Mazur, WPPSS, dated
- February 13, 1989,
Subject:
WNP-3 SSI-Analysis / Deconvolution Issue 11.-Letter from S. Black, NRC, to 0. 0 ingsley, 'Jr., TVA, dated October 31, L
1989,
Subject:
Watts Bar Nuclear Plant, Unit 1 - Validation of SASSI
- computer code for Soil-Structure Interaction Analysis
- 12. Letter from A. Philippacopoulos, BNL, to R. Pichumani, NRC, dated January
. 10, 1989,
Subject:
1 FIN A3850 - WNP-3 SSI Analysis Review f
r-
.E-.-=r
-. - mv. ~ m h ww < m er,e,w.em-w+,.
eni.-w-. -e v-we,
w ine- +m v e w
.,iw g-ene.,r,,,+-mmr e-wete
- r-v=,*ier
-r-
= = r w e y, we,.aom y
.+y
'