ML20042D633
| ML20042D633 | |
| Person / Time | |
|---|---|
| Site: | 05000605 |
| Issue date: | 01/31/1990 |
| From: | Singh J EG&G IDAHO, INC. |
| To: | NRC |
| Shared Package | |
| ML20042D632 | List: |
| References | |
| EGG-AM-8828, NUDOCS 9002070112 | |
| Download: ML20042D633 (20) | |
Text
@, ,
+-. .
ATTACHMENT.1
- EGG-AM-8828
,. January 1990 l*cc c.
INFORMAL REPORT
,,; L
. /daho . ADVANCED BOILING WATER REACTOR STANDARD Nat/onal PLANT SEISMIC DESIGN REVIEW
' Engineering Laboratory M8"898#~ J. N. Singh by the U.S.
' Department ofEnergy
~
94 0 EBsG-o.
Wort performed under DOE Contrect No. ot Aco7 moorsm Prepared for the
' U.S. NUCLEAR REGULATORY COMMISSION
- .. C .
e D}
y-6sr> .
- w
. ABSTRACT EG&G Idaho, Inc., assisted by NCT Engineering, Inc. of Lafayette, ,
California, has evaluated the seismic design of General Electric's advanced boiling water reactor standard plant. In the seismic design of the plant structures, systems, and components, GE has used a total of eight soil profiles to envelope potential sites. SASSI and CLASSI/ASD computer codes are used for soil-structure interaction analysis. The details of the parameters, analyses, and results contained in the standard safety analysis report sections relating to seismic design were reviewed.
On completion, the audit would provide assurance that the parameters are reasonable, the analyses are adequate, and the results are credible, 9
e 4
4 ii
SUF94ARY A review team, consisting of engineers from the Office of the Nuclear Reactor Regulation of the Nuclear Regulatory Commission, the Idaho National Engineering Laboratory, and NCT Engineering, conducted a seismic design audit at the GE offices in San Jose, California from November 28 through November 30, 1989. They reviewed the seismic design sections of the standard safety analysis report and supporting documentation. They also talked to the review support team from GE. The review indicated that the standard design, with respect to the part reviewed, is adequate pending resolution of the open issues.
4 4
+
iii
CONTENTS Abstract................................................................i1 Summary............................................................... 111 Introduction............................................................ 1 Discussion of Audit Findings............................................ 3 .
Seismic Input (Generation of Time History to Envelope Response Spectra).............................................. 3 Modelling of Structures and Supporting Soil Medium................ 5 Validation of. Soil-Structure Interaction Analysis Computer Codes.. 8 Appropriateness of Soil Properties Used in SSI Analysis..........10 Generation of Structure Seismic Loads and Broadened Floor Response Spectra............................................... 11 C o n cl u s i o n s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 References............................................................. 15 e
S iv .
- e
ADVANCED BOILING WATER REACTOR STANDARD PLANT SEISMIC DESIGN REVIEW INTRODUCTION The Nuclear Regulatory Commission (NRC) has been requested by the industry applicants to review the standard design of the Advanced Light Water Reactor (ALWR), General Electric (GE) is one of them. GE'has developed a standard safety analysis report (SSAR) for its advanced boiling water reactor (ABWR) (Ref. 1) standard plant. The SSAR was submitted to the NRC for review and approval. The seismic design of the standard plant structures assumes a set of environmental, geological, seismological, and geotechnical design parameters. On approval, the standard plant is intended to be licensable at any U.S. site provided the site parameters are enveloped by the standard design.
In the seismic design of the ABWR plant structures, GE considered a total of eight generic shear wave velocity profiles for the soil. In addition, it also considered four different soil depths and three different ground water tables for the site with the lowest shear wave velocity. The design ground motion is specified at the plant finished grade. Soil-structure interaction analysis is performed using SASSI computer code for all sites. Further, CLASSI/ASD is used for three of the sites. Seismic response envelopes have been developed for the reactor building. They include shear and moment in structural members and broadened floor response spectra at chosen locations.
EG&G Idaho, Inc. (EG&G) is to provide technical assistance to the NRC in the review of standard design for all the applicants. NCT Engineering is assisting EG&G in reviewing the adequacy of the seismic design of the ABWR. A team, consisting of engineers from NRC, NCT, and EG&G visited GE offices in San Jose, California from November 28 through November 30, 1989. The team attended a presentation by GE staff, reviewed pertinent sections of the SSAR, and discussed issues with the applicant. This report is to document the findings. The review covered the following five seismic design areas-I 1
.. iB
- 1. Seismic input (generation of time history to envelope response
' spectra).
I 2. Modelling of structures and supporting soil medium.
- 3. Validation of soil-structure interaction (SSI) analysis computer codes.
.4. Appropriateness of the soil properties used in the analysis, specifically the soil damping values.
- 5. Generation of structural seismic loads and broadened floor .
resoonse spectra.
The following is an abridged version of the technical letter report prepared by Tom N. C. Tsai of NCT Engineering.
0 4
2
w.
DISCUSSION OF AUDIT FINDINGS
-The NRC' staff and one represen.tative each from EG&G and NCT Engineer-
-ing (referred hereafter as the " team") participated in the audit. The attendees list for both the e.ntrance and exit meetings are in Appendix A to this ieport. The a dU. 1.0:rted with an overview presentation by GE of the seismic design methodology of the ABWR reactor building and was followed by review of Section 3.7, Appendix 3A and Appendix 3G of the SSAR. The review specifically covered information related to the five seismic design areas listed above. The basis for the review is Revision 2 of the NRC Standard Review Plan (SRP). The review findings are discussed in this section. '
SEISMIC INPUT (GENERATION OF TIME HISTORY TO ENVELOPE RESPONSE SPECTRA)
The design response spectrum for the safe shutdown earthquake (SSE) is the standardized response spectrum'specified in Regulatory Guide (R.G.)
1.60 with the peak ground acceleration anchored to 0.30g for both the ,
horizontal and vertical components. Section 2.5 of the SSAR specified the operating basis earthquake (OBE) to be one-third of the SSE. However, according to SSAR Section 3.7.1, the seismic design of the ABWR is based on an OBE that is equal to one-half of the SSE. As such, the term OBE in all subsequent discussions is associated with a 0.159 peak ground acceleration.
The design ground motion time history consists of two horizontal components (H1 and H2) and one vertical component (V). They are adopted from GESSAR (Ref. 2) and were generated using GE computer code SIMQK-01 (Ref.3). The team reviewed the adequacy of the artificial time histories '
against the response spectrum enveloping and power :nectral density function (PSDF) matching requirements as specified in SRP Section 3.7.1.
Review finding is as follows:
.. i 3
1 (a) 'GE-computed response spectra of the OBE time histories forl1, 2,.
3,-4, 7, and 10 percent damping. :For each damping, the response spectrum was' computed at the seventy-five frequency points suggested in SRP Table 3.7.1-1, covering the frequency range _of 0.2 to 34 Hz. According to the SRP, no response spectrum should fall below the target design spectrum at more than five frequency poitits. Also, the amount by which a response spectrum falls below the target design spectrum should not exceed 10%. The team found that all three time histories, H1, H2, and V, satisfy the response :pectrum enveloping requirement for 1, 2, 3, and 4 percent damping but not 7 and 10 percent damping. According to-GE, the spectrum enveloping requirement for 7 and 10 percent damping is not applicable because only OBE analyses were performed and the response for SSE condition was obtained by doubling the OBE response. The OBE analyses used structural damping less than 7 percent (see SSAR Table 3.7-1). The team therefore concluoe: that the artificial time. histories are acceptable provided that they are used in-the analysis with structural damping less than 7 percent.
(b) The PSDF matching requirement applies to the horizontal time histories. GE initially computed the PSDF for the two horizontal time histories'at a frequency interval of 0.024 Hz. It was then smoothed using the 3-point moving average method recommended by the draft Revision 2 of the'SRP'(Ref 4). SSAR Figs. 3.7-24 and 3.7-25 compare the smoothed PSDF of the horizontal time histories to the Kanai-Tajimi target PSDF as recommended. The comparison shows that.both horizontal time histories do not meet the PSOF matching requirement at frequencies above 10 Hz. However, tha PSDF matching requirement recommended in Ref. 4 has been revised. The revised requirement is contained in Appendix A to SRP Section 3.7.1. It allows the PSDF to be smoothed by averaging over a frequency band width of 20%. It also recommends that the smoothed PSDF envelopes the 80% level of a 4
3
r
' newly defined target PSDF over the frequency range of 0.3 to 24 Hz'. Based on preliminary information provided by GE during the
! audit', it appears that both horizontal time histories would meet the revised PSDF matching requirement. However, GE is to finalize the PSDF calculation according to the revised requirement. The calculation is to be based on the time histories- being normalized to the 0.15g OBE' ground acceleration.
! It would demonstrate the adequacy of the time histories.
MoDELLING OF STaucTunes AND $UPPORTING SOIL MEDIUM The SSAR has presented the seismic response of only the' reactor building. The control building, in the SSI analysis, was included for-the purpose of studying the effect of adjacent buildings. Review finding for the adequacy of the fixed-base structural models and SSI analysis models are discussed in the following:
(a) Fixed-Base Structural Model - The reactor Milding is represented by a two-dimensional -(2-D) multiple-stick 'sumped mass model in
-each horizontal direction as shown in SSAR Fig. 3.7-31. One stick each represents the enclosure structure, reinforced concrete containment vessel (RCCV), and reactor pressure vessel (RPV) pedestal. A 2-D lumped-mass stick model for the RPV and internals, as shown'in SSAR Fig. 3.7-32, was included in the-model through structural coupling with the' RPV pedestal and the RCCV. Vertical flexibility of the floors is accounted for in the model by spring element representing the funda' mental mode of the floor. To justify the adequacy of using a 2-D structural model, GE performed a parametric study comparing the modal frequencies and participating factors between a 3-D and 2-D model of the reactor building in which the RPV pedestal, RPV, and internals
. were omitted. Two 3-D models were considered. One includes the actual eccentricity between center of mass and center of 5
t
" rigidity. The largest eccentricity is about 1.7m. This is only
. about 3% of the shorter horizontal dimension of the enclosure '
structure.. The other has a uniform eccentricity of 0.5m. The team reviewed GE calculation DRF A00 02990 (Ref 5). The difference in modal frequencies and participating factors is negligibly small between the 2-D and 3-D models. The use of the 2-D. stick model for the reactor building is therefore ,
reasonable. The modelling of the RPV and internals follows the standard GE practice (Ref. 2) and is acceptable.
The control building is represented by a 1-0 lumped mass stick model: for analysis in the O to 180-degree direction (i.e., along the reactor building-control building-turbine building axis).
Vertical flo.or flexibility is omitted. As mentioned before, the control building was used only in the study of the structure-to--
structure interaction effect. For this particular purpose, the control building model is sufficient.
(b) SSI Analysis Model - GE used computer code SASSI in a 2-D SSI analysis-of the reactor building for fourteen generic site conditions as listed in SSAR Table 3A.3-6. These generic sites comprise eight generic shear wave velocity profiles of soil. GE also applied the SASSI code in various parametric studies including comparison of 2-D vs. 3-D SSI modelling, effect of adjacent buildings (control-building and turbine building), and effect of revised control building configuration. SASSI is a linear finite element analysis code that performs seismic analysis in the frequency domain. It was originally developed at the University of California at Berkeley (Ref. 6). The GE version of the code is SASSI015 (Ref. 7). As a study, GE also applied CLASSI/ASD code to a 3-D SSI analysis of the reactor building for three generic site conditions. These cases were run with and without adjacent buildings. CLASSI/ASD is a proprietary 6
~
u 7
' computer code of ASD International, Inc. It analyzes the problem with-the 3-D continuum impedance approach in the frequency domain (Ref. 8).- Since the main computer code is SASSI and the SRP does not mandate the use of two different SSI methods', it is sufficient to focus the review on the SASSI analysis method and results only.
In the model for the reactor building, the-soil medium is typically represented by soil layers. Soil element representa-
. tion was.used only where necessary, such as between two adjacent buildings. The team found the modelling of the soil medium to be consistent with the SASSI code theory. However, there is concern with the size of the soil layers and/or elements because it dictates the cutoff frequency in the analysis. The size of soil layers / elements used in the GE calculation results in a cutoff frequency of 18 Hz for UB site and 25 Hz or higher for other generic sites (Ref. 5). According to GE, 90% of the structural
. mass has been captured with this cutoff frequency. The criteria of capturing 90% of total structural mass may be sufficient for the response of the massive structures dominating the total mass such as the RCCV and enclosure structure. However,.it may not be sufficient for the response of the basemat and other lighter structures such as the RPV and internals. For these, the contribution from higher frequency modes up to 33 Hz may be significant. 33 Hz is the cutoff frequency for ground motion spectral amplification and typical industry practice for seismic analysis. The adequacy of the size of soil layer-element used in the SASSI model is, therefore, questionable. GE is to justify the adequacy of the SASSI model.
The team raised another concern with the modelling of structural embedment. To model the 85-foot embedment' of the enclosure structure with SASSI, it is necessary to laterally expand the lower portion of the stick model to the full plan dimension to i
7
L A
' allow intteface between the embedded wall and surrounding soil
^
medium. The team found the required expansion in the stick model sufficient for the 2 0 SAS$1 model but questionable for the'3 D model as shown in $$AR Fig. 3A.8 23. For the expansion to be proper for the embedded portion of the enclosure structur.e, a rigid beam should be placed between the following pairs of nodes:
208 and 264, 408 and 464, 708 and 764, 908 and 964. This modelling concern may not affect the global structural response but could affect the local response such as the dynamic soil pressure on the structure wall. GE is also to address the potential of separation of the side soil from the embedded wall ,
and any effect it may have on the structural response.
(c) The seismic analysis results of the control building and radwaste building substructure were not available for review during the audit. Therefore, in addition to addressing the team's concerns with the reactor building SSI model, GE is to complete the seismic analysis of the control building and radwaste building substructure and submit the results to the NRC for review.
VALIDATION or Sott-STaucTunE INTERACTION ANALYSIS COMPUTER CODES The team reviewed the validation manual of SASSI code that contains fourteen generic validation problems (Ref. 9). Seven problems are the more important ones. They encompass four areas of potential applications, i.e., computation of impedances and wave scattering for embedded footings, SSI response solution, and response spectrum computation. Acceptable basis for validation is close form solution or other SSI analysis techniques previously accepted by the NRC for licensing of other nuclear plants. The seven validation problems are summarized in the following table:
e 6
9 8
4
, j
. t Validation -
Prob. No. Summary of Problem 1
A-2 Impedance and SSI response of a building with a surface i founded circular base on a uniform half space and with
, vertically incident wave.
A-3 Impedance and wave scattering for a square surface footing on a uniform half space and subjected to inclined incident wave. .
A5 Impedance of a square surface footing (rigid at center and flexible at edge) on a unifom half space.
l A9 Impedance of a rigid cubical footing fully embedded in a uniform half space.
A 10 Response of two identical nearby buildings having square ,
surface footings on a unifom half space. ,
A 11 Scattering response of a rigid massless cylinder embedded in '
a surface layer and subjected to vertical incident waves. ;
A 14 Response spectrum calculation.
The reactor building is deeply embedded to a depth of 85 feet, which accounts for about 40% of its overall height. A generic site condition where the reactor building is founded on rock and surrounded by soil constitutes a unique situation. The applicability of the generic ,
validation problems to this case is not clear. To increase the confidence in the applicability of SASSI to the SSI analysis of the ABWR, therefore, GE is to provide validation of the code with respect to the SSI response of structures deeply embedded in soil and founded on rock.
Otherwise, the generic validation of SASSI appears to be sufficient.
GE is to submit a copy each of the User's Manual (Ref. 7) and Validation .
Manual (Ref 9) for further review.
9 -
l
. . ---,.,w,- ,w , ..n- - , --,
i i
APPROPRIATENESS OF SoxL PnoPERT:ES Usto IN $$I ANALY313 l Soil properties used in the SSI analysis of the reactor building are strain-dependent. The strain dependency of soil properties were estab- !
11shed from the free field site response analysis based on the assumption ;
of vertical propagating waves. The site response analysis was done for the OBE condition using the 10 soil column analysis cod *"AKE. L,imits l
imposed by GE on the strain dependent soil properties .2mply th the SRP l criteria. That is, the shear modulus would not be let th" 40% o'/ the l low strain value and the soil material damping would not + eeed 15%. ;
GE used horizontal time history H1 in the calculation of strain-l dependent soil properties for all generic site conditions. Validity of using only H1 in the site response analysis to generate the soil properties was confirmed through a comparison of the results with those
, [
generated by using the other horizontal time history, H2, for represen-tative site conditions (Ref. 5). The comparison is close and hence the {
use of only H1 in the soil property generation is sufficient.
To account for the effect of ground water, GE maintained a minimum compressive wave velocity equal to 4800 ft/see for soil medium below the L ground water table. This velocity is equal to the speed of sound in l
water. It was achieved in the analysis by adjusting the Poisson's ratio l of soil so long as it remained lower than 0.5. The team found the soil j properties below ground water table reasonable.
The SRP requires that soil property uncertainty be accounted for in ;
the SSI analysis. This requirement has not been explicitly implemented in 1
GE is to provide a clarification in the the seismic design of the ABWR, SSAR to the effect that the generic shear wave velocity profiles represent low-strain soil properties after uncertainty is taken into account. .
l 10
[
. i l
GENERATION OF $7RUCTURAL $EISMIC LOADS AND ,
BROADENED FLo0R RESPONSE $PECTRA ;
GE performed OBE analyses of the reactor building for fourteen generic )
site conditions. They cover eight generic shear wave velocity profiles of l soil. For the UB site, four different soil depths (85 feet, 150 feet, 200 !
feet, and 300 feet) and three different ground water tables ( 2 feet, 40 i feet, and 85 feet) were also considered. In addition, GE studied the sensitivity of the reactor building response to the effect of 2 D vs. 3 D ;
SAS$1 model, the effect of adjacent buildings, the effect of using another
$$1 technique (CLAS$1/ASD), and the effect of revised control building configuration. The OBE design seismic loads (shear and moment in ,
structural members) and floor response spectra were then developed by enveloping the seismic loads and floor response spectra for all individual ;
cases. Specifically, the OBE design floor spectra were obtained in the
{
following manner:
Step 1 - Floor response spectra from all SASSI cases were enveloped at representative locations in each of the three directions. '
Step 2 - Envelope spectra in the two horizontal directions at each location obtained in Step 1 were further enveloped to form one !
bounding horizontal spectrum at each location for use in both !
l horizontal directions. '
Step 3 - The calculated 2 percent damping floor spectra from all CLASSI/ASD runs were enveloped at corresponding locations and ;
directions. ;
Step 4 - By comparing the SASSI and CLASSI/ASD envelope spectra of 2 percent damping, a generic scale factor was developed for >
application to the SASSI spectra so that the scaled spectra would I also envelope the CLASSI/ASD spectra. The generic scale factor is frequency dependent, as listed below:
Freauency Ranae SASSI Enveloce Spectrum- Scale Factor ,
0 to 1 Hz 1.0 1 to 2 Hz 1.35 ,
. 2 to 20 Hz 1.70 Beyond 20 Hz 1.33 ,
11
Step 5 - The scaled SASSI spectra were broadened by + 10% to obtain the
.06E design floor spectra, s
The calculation of the OBE design seismic loads and broadened floor i spectra is acceptable. The use of doubled OBE design seism'ic loads as the ,
SSE seismic loads is also acceptable, but the adequacy of using doubled OBE floor spectra as the SSE floor spectra is questionable. The floor
[*
spectrum peaks are measured not only by magnitude but also frequency location. Since the strain dependent shear modulus of soil will be lower !
under the SSE than under the OBE, frequencies of the SSE floor spectrum peaks are expected to be somewhat lower than those of the OBE spectrum peaks. Thus, using doubled OBE floor spectra as the SSE floor spectra may l be sufficient magnitude wise but may not be frequency wise. This may be ;
especially so for the softest sites such as UB and VP2. This is a concern that requires an investigation by GE.
k e
4 i
f 9
4
CONCLUSIONS i Based on the review, the team reached the following conclusions:
(1) The artificial time histories are acceptable provided that (a) {
they are used in seismic analyses with structural damping less i than 7 percent, and (b) the finalized PSDF calculation base'd on 0.15g OBE ground acceleration satisfies the matching requirement !
specified in Appendix A to SRP Section 3.7.1.. I i
(2) The modelling of structures and soil medium in the $$1 analysis is reasonable for the reactor building provided that the -
following concerns are resolved:
(a) The validity of using a cutoff frequency equal to 18 Hz for UB sites and 25 Hz or higher for the stiffer sites.
(b) The adequacy in modelling the embedded structure wall in the 3-D SASSI model as shown in SSAR Fig. 3A.8 23.
1 (c) The potential of separation between soil foundation and embedded wall, and its effect on the seismic response. i (3) The generic validation of SASSI code is sufficient. To increase the confidence in the applicability of SASSI to the SSI analysis l of ABWR, GE is to provide certain application specific !
validations of the code, for example, the response of structures !
deeply embedded in soil and founded on rock.
f (4) The soil properties used in the SSI analysis are appropriate for the OBE condition. GE is to provide a clarification in the SSAR to the effect that the generic shear wave velocity profiles ?
represent the low strain property after uncertainty is taken into ac:,ount.
13 t
4
(
L ,o I;
'. l (5) The structural seismic loads and broadened floor response spectra for the OBE condition are acceptable. The $$E condition
]
structural seismic loads, as obtained from doubling the OBE i seismic loads, are also acceptable. However, GE is to justify the adequacy of using doubled OBE floor response spectra as the {
SSE floor spectra.
l (6) The standard design considered two categories of generic sites. )
The first without soil deposit and the second with a minimum soil i depth of 85 feet. It did not, however, consider a generic j condition where the soil depth is between 0 and 85 feet. GE is '
to justify that this condition is bounded by the two categories considered.
(7) GE is to provide the results of the analysis of the control building and radwaste building substructure to the NRC for review.
I e
P L
e 14 9
9
~- , - - - -- - - - - - - -
l' t REFERENCES -
t
- 1. General Electric Company ABWR Standard Safety Analysis Report, Section ;
3.7 (Amendment 1), Appendix 3A (Amendment 1), and ADpendix 3G !
(Amendment 4). 1
- 2. General Electric Company BWR/6 238 Standard Safety Analysis Report (GESSAR), Docket No. STN 50 447, November 7, 1975. '
- 3. General Electric Company Computer Code $1MQK-01, "A Computer Program !
for Artificial Motion Generation,' June 1976.
- 4. NUREG 0800, Standard Review Plan, Draft of Revision 2.
- 5. General Elsetric Company Calculation No. DRF A00-02990.
- 6. J. Lysmer, M. Tabatabaie Raissi, F. Tajirian, S. Vahdani, and F. Ostandan, *SASSI - A System for Analysis of Soil-Structure Interaction,' UCB/GT/81 02, University of California, Berkeley, California, April 1981.
- 7. SASS 10lS User's Manual, GE Report No. NEDE 31496, October 1987.
- 8. CLASSI/ASD Computer Program for Three Dimensional Soil / Multiple Foundation Interaction Analysis, User's Manual, Version 2.1, ASD International Inc., San Francisco, ' California, November 1987.
9 .. SASSI Validation Manual, GE Report No. DRF A00 03065.
(
i e
i 1,
15
APPENDIX A A.1 Attendees List for Entrance Meeting :
Name Orosnization A. H. Hsia NRC/NRR ,
Chen P. Tan NRC/NRR Ramon Pichumani NRC/NRR Jag N. Singh EG&G Idaho, Inc.
Tom N. C. Tsai NCT Engineering Jack Fox GE Earl Nichols GE Al-Shen Liu GE H. E. Townsend GE A.2NttendeesListforExitNeeting Name Oraanization A. H. Hsik NPC/NRR Chen P. Tan NRC/NRR Ramon Pichumani NRC/NRR Jag N. Singh EG&G Idaho, Inc.
Tom N. C. Tsai NCT Engineering L Jack Fox GE L Earl Nichols GE Al-Shen L'in GE L
1 A1 i
l