ML20247M276
| ML20247M276 | |
| Person / Time | |
|---|---|
| Site: | 05200003 |
| Issue date: | 11/30/1996 |
| From: | Costantino C, Hossain Q COSTANTINO, MILLER & ASSOCIATES, INC., LAWRENCE LIVERMORE NATIONAL LABORATORY |
| To: | NRC (Affiliation Not Assigned) |
| Shared Package | |
| ML20247M262 | List: |
| References | |
| NUDOCS 9805260229 | |
| Download: ML20247M276 (103) | |
Text
{{#Wiki_filter:- i I Confirmatory Soil-Structure Interaction Analysis of the Westinghouse AP600 Nuclear Island Structures l by Carl Costantino Costantino, Miller and Associates i 4 Rockingham Road SpringValley,NewYork 10977 and ! Quazi Hossain , Lawrence Livermore National Laboratory 7000 East Avenue Livermore, California 94550 l Prepared for: , Civil Engineering and Geosciences Branch ! Ofsce of Nuclear Reactor Regulation l U.S. Nuclear Regulatory Commission - l Washington, D.C. l l November 1996 i 9805260229 DR 980409 0 1 h"ITs96-494NHhan ADOCK 05200003 Enclosure 1
l l Table of Contents 1.0 Introduction 3 2.0 Scope 5 3.0 Task 1: Evaluation of Ranges of Soil Conditions Considered 6 4.0 Task 2: Evaluation of the Effects of Soit Layer Thickness Between 8 Basemat and Bedrock l 5.0 Task 3: Evaluation of the Effects of Seismic Input Directions 10 6.0 Task 4: Evaluation of Mesh Size Effects 11 7.0 Task 5: Evaluation of the " Sixty Percent" SRP Requirement 12 8.0 Task 6: Evaluation of 3 D Design Basis Cases Considered 16 9.0 Task 7: Evaluation of Lateral Earth Pressures on Below Grade Exterior Walls 17 10.0 Summary and Conclusions 19 Appendix A: Parametric Studies using 2 D SASSI Analyses A.1 f Appendix B: Evaluation of Soil Pressure on Below Ground Walls B.1 Appendix C: Parametric Studies using CARES Computer Code C.1 , l Appendix D: Evaluation of Response Spectra at Basemat Level D.1
]
I Appendix E: 3-D SASSI Analysis E.1 I 2 NITS 96-4944Hhan
1 -1 1
~
1.0 Introduction In revision 2 of AP600 Standard Safety Analysis Repon (SSAR), Westinghouse Electric Corporation (Westinghouse), the AP600 License Applicant, used two three-dimensional (3-D) soil structure interaction (SSI) analysis cases and one fixed-base 3-D analysis case to calculate the design forces and moments and floor response spectra for the Nuclear Island Structures. These 3-D analysis cases were selected on the basis of a number of two dimensional (2-D) SSI parametric analyses in which several soil and foundation condition parameters (e.g. soil depth, soil stiffness propenies, depth of water table, etc.) were varied. The Nuclear Regulatory Commission (NRC) staff and its consultants (hereafter called the Staff Review Team) had reviewed these SSI analysis results (both 3-D design-bases analyses and 2-D parametric analyses). Even though the applicant's overall approach of determining seismic design loads and responses was found to be appropriate, the staff had major concems regarding the following seven issues: (i) the ranges of soil conditions covered in the parametric 2-D SSI cases (ii) the effects of soil layer thickness between basemat and bedrock l (iii) the effects of variations in assumed seismic input ducctions i (iv) the effects of soil element mesh size on the seismic response (v) the adequacy of the response spectra at the foundation with respect to NRC Standard Review Plan (SRP) requirements (hereafter called the Sixty-Percent SRP requirements) (vi) the adequacy of applicant's 3-D design basis seismic analyses, and (vii) the lateral earth pressure on below-grade exterior walls of the nuclear island To funher assess and resolve these concems, the staff had issued several Requests for AdditionalInformation (RAls). While Westinghouse has been performing additional SSI analyses to respond to staff's RAIs, it was decided by the staff to undenake some independent SSI analyses in order to ensure that AP-600 design basis seismie loads and response spectra are adequate. These independent analyses and their results are presented in this report. l l l 3- NTFs%494/QHhan
[Since these independent analyses have been undenaken, Westinghouse has l submitted an updated version of AP600 SSAR (Revision 9, dated August 9,1996) in which, Westinghouse, based on the results of staff's independent analyses (that l are being reported here), has increased the number of 3-D design basis SSI analysis cases from two to three. The conclusions presented in Section 10 of this repon are based on a cecaparison of staff's independent analysis results reported here and the results presented in Revision 9 of AP600 SSAR] 4 NTFS9994/Qllhan
2.0 Scope The overall scope of work reported here is to verify if the seismic soil-structure interaction (SSI) analyses performed by the applicant to determine the design basis forces and moments and in-structure or floor response spectra for the Nuclear Island are adequate in accordance with SRP requirements and practices. For the purpose of this verification, the work scope was divided by the Staff Review Team into seven taskslisted below: Task 1: Evaluation of the adequacy of the ranges of soil conditions covemd in applicants 2D parametric SSI analyses. Task 2: Evaluation of the effects of soil layer thickness between basemat and bedrock. Task 3: Evaluation of the effects of variations in assumed seismic input directions. Task 4: Evaluation of the effects of soil element mesh sizes on the seismic response. Task 5: Evaluation of the adequacy of SSI response spectra at the foundation level I in the context of sixty-percent SRP requirement. { Task 6: Evaluation of the adequacy of applicant's 3-D design basis seismic analyses. i Task 7: Evaluation of the adequacy of design basis lateral earth pressure on below-grade exterior walls of the Nuclear Island. These seven evaluations and their results are presented and discussed in the
- following sections.
I j 5 NTFS96494/QH/ san
3.0 Task 1: Evaluation of Ranges of Soil Conditions Considered The applicant had originally performed two 3-D SSI analyses and one fixed-base analysis to determine the design basis forces, moments, and in-structure spectra. Based on a number of 2-D SSI analysis case results, the applicant claimed that thee.e three 3-D design basis analysis cases (two SSI and one fixed-base) adequately-cover the entire range of sites for which the license is being sought. To verify this claim, the Staff Review Team selected and analyzed the site conditions listed in i Table 3-1 using the 2-D SASSI computer program. A summary of these analysis cases are presented below; the details are provided in Appendix A. The 2-D structural model of the Nuclear Island used in analyzing the cases listed in
)
Table 3-1 was provided by the applicant. The free-field seismic input motion acceleration time histories were also provided by the applicant. In order to evaluate if the applicants 2-D parametric analysis cases are adequate for the specified range of soil sites, the in-structure response spectra at certain Nuclear Island elevations resulting from the Table 3-1 parametric cases were compared with the enveloping spectra at the corresponding locations from the applicant's parametric analysis i cases. The spectra were compared at the following locations: (i) Shield Building Roof (Elevation 307.25 ft) (ii) Polar Crane (Elevation 205.33 ft) (iii) Containment Intemal Structure Operating Floor (Elevation 135.25 ft) l The comparison of response spectra at these locations showed that applicant's enveloping spectra enveloped most of those generated from Table 3-1 parametric cases. However, a close inspection of the results indicated that the soft rock case (i.e. Case H) produced in structure spectra at some locations that, at certain frequency ranges, exceeded applicant's enveloping spectra. The Staff Review l Teamjudged that the significance of thi:: finding can be high and so it was decided that this finding should be confirmed through a 3-D SASSI analysis so that the conclusions are not contentious. Accordingly,in Task 6, the 3-D SSI analysis case used case H soft rock properties. l NTFS96-494/QHhan
4 Table 3-1 List of 2-D SASSI SSI Analysis Cases
- Soil / Rock Assumed Low Soil Depth Case Type Strain Shear Wave Subcase Moisture tom Velocity Condition Bedrock (feet)
A Soft to 1000 fps at grade Al Dry 80 Medium varyinglinearly to 2400 A2 Dry 120 Stiff fps at 240' (Seed-Idriss A3 Wet 120 degradation pc-des) A4 Sa:urated 120 B StiffRock 11,000 fps B N/A N/A C Medium 8,000 fps C N/A N/A Rock D Soft to 1000 fps at grade D Dry 240 Mediurn varying parabolically to Stiff 2400 fps at 240' (Seed-Idriss degradation properties) F Soft to 1000 fps at grade F Dry 240 Mediurn varying parabolically to Stiff 2400 fps at 240' varying parabolically (lower bound degradation puperdes) G Uniform 1500 fps uniform soil G Dry N/A Site site with Seed-Idriss degradation properties H Soft Rock 3,500 fps H N/A N/A I Soft Site 1000 fps uniform soil 11 Dry 40 site with Seed-Idriss 12 Dry 45 . degradation properties J Specific Savannah Riverdeep Dry N/A Deep soil site with variable Site velocity profile and lower bound degradation supesies K Soft to Same as Case A: K1 Dry 120 Medium . VenicalWaveIncidence K2 Dry 120 Stiff 30* wave Incidence Notes: (1) For cases A,D,F,I, and K, bedrock shear wave velocity of 8,000 fps was used (2) Depth of bedrock from finished grade icvel. Foundation depth is 40ft. NTFS96494/QHAan
r-4 4.0 Task 2: Evaluation of the Effects of Soit Layer Thickness Between Basemat and Bedrock The effects of variation of soil layer thickness between basemat and bedmck was evaluated first by performing four p.r metric 2-D SASSI analysis cases and then by performing seven 3-D parametric cases using the computer code CARES (Computer Analysis for Rapid Evaluation of Structures). The four cases analyzed by the 2-D SASSI computer program assumed soft soil
- sites (shear wave velocity = 1000 FPS) and are characterized as follows
(i) Case 11-40: Depth of soil fonn the finished grade level to the top of the bedmck is 40 ft. Thus, the thickness of soil medium between the bottom of the basemat and the bedmck is zero, and the basemat rests directly on the bedrock. He seismic input motion is applied at the finished grade level I f (ii) Case 11-45: Same as 11-40, except that the depth of soil to bedmck is 45 ft. and the thickness of soil medium between the bottom of the basemat and the bedrock is 5 ft. l (iii) Case 12-40: Same as 11-40, except that the seismic input motion is applied at the top of the bedrock. l (iv) Case 12-45: Same as 11-45., except that the seismic input motion is applied at l the top of the bedmck. The resul(fmm the four parametric cases were compared with the applicant's . ki ;; ;- - ants the results from the other parametric cases listed in Table 3-1 (see Appendix A). The comparison showed that the building response is ! significantly affected by the variation of the thickness (t) of soil medium between i the basemat and the bedmck, especially when the input motion is applied at the top ; i of bedrock and when 't' is small. l i The Staff Review Teamjudged that the finding noted above may increase the design j basis seismic loads significantly, and so this finding should be confirmed by additional analyses. As such, seven more confumatory analyses were performed in which the 3-D computer program CARES was used and the thickness 't' of the soil i
.s. ms96 494mw. ;
J 1.
f l ) i medium between the basemat and the bedrock was varied fmm 10 ft to 100 ft (at 10 ft. intervals). The results of these analyses showed that, as 't' decreases, the floor response spectra peaks shift and accentuates the lower frequency structural modes. It was further observed that, when 't' is less than 20 ft., the lower frequency peak of the response spectra increases markedly (see Appendix C). Thus,it was concluded that either AP600 design basis must include such shallow soil sites and recompute the seismic responses (forces, moments, and floor response spectra) or exclude such sites from the licensing basis.
+ NM%eWWun
L , 5.0 Task 3: Evaluation of the Effects of Seismic Input Directions
'Ihe effects of seismic input motion ducctions on both the in-plane and out-of-plane response of structures were evaluated first by performing two parametric 2-D SASSI analyses and then by performing several 3-D CARES analyses. The 2-D l SASSI analysis consisted of cases K1 and K2 in Table 3-1 (see Appendix A).
I Case K1 used vertical incidence waves and case K2 used 30* inclined (fmm vertical) waves. Comparison of the responses resulting from cases K1 and K2 ) I showed that the effects of angle ofincidence are insignificant. But, since the 2-D SASSI analysis cannot account for the coupling between two orthogonal direction motions and out-of-plane structural behavior, it was decided to perform additional 3-D CARES analyses. These 3-D CARES analyses also, however, showed that the l effects of angle ofincidence on the out-of-plane orin-plane response of the l structure are negligible (see Appendix C). l 1 l l l
. l
( ( 10 NrFS96494/QiUsan {
6.0 Task 4: Evaluation of Mesh Size Effects
. De effects of finite element mesh size representing the soil medium adjacent to the below grade exterior walls were evaluated to determine if the mesh size used by the applicant to calculate the floor response spectra are adequate. For this purpose, thme 2-D SASSI analyses were performed for a soft soil site using three different mesh sizes to represent the side soil. De first analysis was the base case analysis that used a fine mesh with 8 layers of elements to represent the side soil; the.second case used 5 layers, and the third case used a coarse mesh with only 3 layers of elements (see Appendix A).
A comparison of the resulting floor responte spectra showed that the use of 3 soil layers of elements to represent the side soil is unconservative compared to the base case that uses 8 layers; the 3 layer case results in a significant reduction of the spectral peak at a frequency of about 10 fps. Even though the applicant used 8 layers in its 2-D parametric analyses, AP600 design basis floor spectra and member forces and moments (except those for the design of exterior walls) are based either on fixed-base analyses or on 3-D SSI analyses that use only 3 layers of elements to represent the side soil. As such, the applicant must demonstrate that,ifits 3-D SSI analyses had used 8 layers instead of 3 layers of soil elements, the resulting floor spectra and member forces and moment in the Nuclear Island would still be enveloped by the AP600 design-basis values. i NTFS96-494NHhan
7.0 Task 5: Evaluation of the " Sixty-Percent" SRP Requirement In Section 3.7.2-11.4 of NRC's Standard Review Plan (NUREG-0800), it is stated that,"the spectral amplitude of the acceleration response spectra (horizontal l component of motion) in the free field at the foundation depth shall not be less than 60 percent of the corresponding design response spectra at the finished grade in the free field." It further states that for a given site under consideration, if three SSI analyses are performed (the first one using the best-estimate site-specific solid shear moduli (G), the second one using 2G, and the third one using 0.5G), then the l above 60 percent limitation may be satisfied using an envelope of the three spectra corresponding to the three soil yivyerdes. l The applicant has performed several 2-D SSI parametric analyses and only two 3-D SSI analyses (subsequent to the study reported here, the applicant performed a third 3-D SSI analysis). But, in none of these analyses, the compliance with the " sixty-percent" SRP requirement has been demonstrated on art individual case basis. In its interaction with the Staff Review Team on this issue, the applicant indicated that, since the envelope of the 2-D and 3-D case free-field spectra at the foundation level envelops the 60 percent of the design spectra at the finished grade, this may be considered equivalent to satisfying the requirement on the basis of three SSI analyses in which the soil yivyerdes are varied as discussed above. This justification may not be appropriate for the following two reasons: (i) Since the applicant did not vary the soil properties for all cases in accordance with the SRP guidelines discussed above,if a particular site under consideration has soil shear modules value close to the lowest value considered by the applicant (i.e., shear wave velocity of 1000 fps), the site cannot be considered to have complied with the " sixty percent" SRP requirement, because for that site the 0.5G case has not been considered. (ii) De SRP states,"the 60 percent limitation may be satisfied USING AN ENVELOPE of the three spectra....". He use of enveloped spectra at the foundation level eliminates the possibility of underestimating the in-structure ! response that may result from various combinations of soil column frequencies and dominant structural frequencies. Since the applicant did not actually use i
" enveloped" spectra at the foundation level, it is not possible to rule out with NTFS96-494/QH/ san
certainty that the in-structure response have not been underestimated at some frequencies. His is explained below with an illustrated example; additional explanation is provided in Appendix D. I l let Figure 7-I show the free-field spectra for a soil site as labeled. Iet f3, if , and f3 be the fnaAmental soil column frequencies conesponding to best estimate soil property (G),0.5G, and 2G, respectively at which the free-field spectrum at the foundation level dips below the 60% spectrum (i.e., below Spectrum No. 3 in Figure 7-1). Iet Spectrum No. 7 in Figure 7-2 be the enveloped spectra at the foundation level that results from the three different soil properties. If, acwdhig to l the SRP guidelines, this enveloped spectrum is used as input for the SSI analysis, the " sixty-percent" requirement is considered satisfied, because at every frequency point, the input motion is more than 60% of the grade level motion. l let a structure on this soil site have three dominant modes with frequencies fi , f,, and f3. Then,if the SSI analysis uses, as input, a motion compatible with the enveloping Spectrum No. 7, structural response will be based on acceleration responses corresponding to frequenciesif 3, f , and3 f of the enveloping sgu um, i.e., accelerations Ai , A 3, and A 3, respectively. But,if three separate SSI analyses are performed for the three soil cases, the structural response will be based on accelerations ai, 2 a , and a3 . Hus, even mough the envelope of the three foundation level spectra results in a spectrum that exceeds the 50% spectrum, unless this enveloped spectrum is actually used in the SSI analysis, or unless the ir.put motion used for the individual soil pupty cases are perturbed to satisfy the " sixty-percent" requirement, the in-structure responses from the SSI analysis will be underestimated. Sina the applicant performed neither,it is concluded that the l " sixty-percent" SRP requirement has not been satisfied, and the in-structure responses (both response spectra and member forces and moments) from the SSI I analyses may have been underestimated. Even though the 3-D fixed base case dominates structural response at many locations and for many components, unless the effects of the above noncompliance is ascertained quantitatively,it is not l possible to evaluate if the applicant's design basis structural responses (at all frequencies ofinterest) are conservative (when wuW to those from SSI cases where the sixty percent SRP requirement has been complied with). 1
.is. xirsw.e4aw n
N N N l N e v e
/( , N N
ir 3 L i n o t a d _ n u
- 3 o
) f F G .
t a y t r n e 6 = o _ l) p s eo i r s- r a eP s p Ll i
/ m no o
D
)
s ioSe C i s t a t A s, a a d a t r B nm c v ui e is n g ots FE) t
^ S p
e asG)t 2 f d D e 5. G i e tl ee mB u ( (0 2 ( y l F-gvt r l l l
\[ c n
M _ r aLevvve ceee e e e T peee 4 u q F r ( esLLL l d e f e vr aldnnneiii ooo r F o LeGit Faaa t t l e p eaee ddd t nnn uuu m dmra r uFooo a x NN/ Gtr d FFF _ E aec unaaa t t t t /
/ 1 pommm / 7 msBuuu
_ 1 y f e udrt r r r rl t t \ r u ci eeccc t g eFweee _ p- oppp \ 5 i F _ seLsss dr e(ddd l l eF 1 leee f l _ i i i f _ if
- a l
oFFF - - - / eee eu%eee et r c0r r r p /
/
FA6FFF 123456 _ _ E 8
s s N
/ '
3
- 3 f
m g
= u r
t _ c _ e p
)
1
~ S 7 d _
e e p _ r _ u ~ o - g - l e _ Fi v _ e n _ e E _ s 2 f _ ( f o - 6 e y d 5 4 t n a a r c e p w -
/
4 c n e u q e r F U l E 2-7 s e p m a x S _ e
/
s p
' r o !
u g l e f 1 i F v , f . n \ E \ t a # h t
'7# 5 m
u r t c S 7 e p
, [
[ 3 2 1 2 j A A A a a a h 8z! llllllIll1
l d This page intentionally left blank' x l l 1 e i ? L---------------_------- 1
i 1 l i i l l i
)
b 1 This page intentionally left blank I i i O r E___ __ _ _ _ _ _
i 8.0 Task 6: Evaluation of 3-D Design Basis Cases Considered In Task 1 (see Secdon 3.0), the adequacy of the ranges of soil p%-des considered by the applicant was evaluated by perfonning 2-D SASSI analysis for the cases hsted in Table 3-1. It was concluded from these analysis cases that the enveloping spectra generated from applicant's analysis cases reported in Revision 2 of AP600 SSAR do not envelop those from the soft rock cases (i.e. Case H) listed in Table 3-1. To wiu%u this finding, this soft rock case was reanalyzed using a SASSI 3-D model. i nis 3-D SASSI analysis has been described in detailin Ap-adir E. This analysis used a Nuclear Island structural model generated by the applicant which, even though very similar, is not identical to the structural model used in Table 3-1 analysis cases; however, the model is identical to the later design-basis analytical , nxxiel that the applicant used in Revision 9 of AP600 SSAR. In-structure response spectra at few representative locations that resulted from this 3-D SASSI analysis case were compared with the conesponding design basis ! enveloping spectra from appiicants 3-D S ASSI and fixed-base analysis cases as reponed in Revision 2 of AP600s SAR. De couwwison confirmed the findings of the 2-D SASSI analysis case H (see Table 3-1), i.e., the in-structure response at some locations fmm the 3-D SASSI analysis of soil case H exceeds, a: certain frequencies, the responses fmm applicant's earlier design basis enveloping spectra (i.e., spectra provided in Revision 2 of AP600 SSAR). As such,it was concluded that the soil property range covered in applicant's 3-D SSI analysis cases is inadequate without the soft rock case H. [ Subsequently, die applicant updated AP600 SAR incorporating one such case in SAR Revision 9]. l 16 h"ITS96-494/QH/ san
9.0 Task 7: Evaluation of Lateral Earth Pressures on Below-Grade Exterior Walls ne lateral canh pressures on below-grade exterior walls of the Nuclear Island was evaluated from the following four considerations (see also Appendix B): a) ne effect of mesh size (representing the side soil) on the lateral canh pressure determined by the 3-D SASSI analyses. b) ne effect of lateral pressure from rock on the wall design. c) The effect of ground water on the lateral canh pressure during a seismic event. d) De effect of adjacent buildings (Turbine Buildings, Radwaste Building, and Annex Buildings) on the lateral earth pressure. To determine the effect of mesh size on lateral pressure, the Staff Review Team performed three 2-D parametric analyses as described in Section 6.0 by varying the sizes of the elements that represent the side soil. A comparison of the results of these three analyses showed that the coarse representation of the side soil that uses ) 3 layers is inadequate, and a finer representation using 8 layers in necessary. If, however, from cost considerations,3-D S ASSI analysis cases use only 3 soil
. layers (as was done by the applicant), it would be preferable to compute the lateral design pressure from the 2-D SASSI analyses that use 8 soillayers. But,in such cases, the additional loads (i.e. loads in addition to those predicted by the 2-D l SASSI analysis) on exterior walls near the corners because of" Box Effects" must be considered.
The results of 2-D SSI cases for rock foundation, i.e., cases B, C, and H (see j l Table 3-1) were inspected to evaluate if, in determining the wall design loads, these cases shoukt be considered. De inspection revealed that, at some locations of the
- exterior walls, rock cases may produce peak lateral pressures that are higher than ;
soil cases. However, since these pressures are localized, their effect on the design l is uncenain and geometry Wa t. nDese pressures may or may not control l design shear or moments, but, it was clear that rock cases should also be l considered in determining the critical design basis lateral loads on walls. l 1 l-i 17 N'ITS96-4944H/ san l l
Determining the effect of ground water on the lateral pressure is difficult because, ( SSI computer codes like SASSI, that are otherwise sophiWa'd have some limitations to properly account for the two-phase interaction between solids and fluids. For example, SASSI solutions tend to be unstable for Poisson ratio values close to 0.5. Thus, instead of explicit or parametric determination of the ground l water effect on the lateral wall pressure, the Staff Review Team studied the effect approximately by using a simple one dimensional constrained rod analysis assuming that wall pressure is influenced only by compressional waves moving through the solid and water fractions of the soil. Even though the results of this simplified study was not conclusive, it was judged that, if SASSI analysis model ( uses a Poisson's ratio of about 0.48, the resulting wall pressure is likely to be l conservative, amp ~4 ally at or near the ground surface where computed pressures are l ~ the largest. l l- 'Ihe effect of adjacent buildings on the lateral pressure was considered only qualitatively, and it was concluded that the applicant should evaluate structure-to-I snucture interaction through the soil to determine lateral pressures on exterior below-grade NuclearIsland walls. I f l l . (' 1 l 1 2 I
-18 NITS 96-494NHhan L--_____-_________-_______________-._._____________ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - - _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
10.0 Summary and Conclusions A. In order to evaluate the adequacy of Westinghouse Electric Corporation's seismic soil structure Interaction analyses and determination of design basis seismic forces, moments, and floor response spectra for AP600 Nuclear Island Structures, the NRC Staff Review Team, consisting of NRC staff and its consultants, reviewed revision 2 of AP600 SSAR. His review raised corwrns on seven technicalissues (see Sections I and 2). To further assess and resolve these issues, the NRC Staff Review Team performed iPnt 2-D and 3-D SSI analyses using computer programs SASSI and CARES De results of these analyses were studied to draw the following conclusions that are considered pertinent and significant for AP600 license review: a) The ranges of soil and rock pivyenies considered by the applican; in its 2-D parametric analysis cases and 3-D design basis cases are not adequate, and the SSI effects from a soft rock case (shear wave velocity equal to about 3500 fps) must be included in determining the design basis forces, moments, and floor response spectra (see also the discussion in Paragraph B of this section). b) De SSI effects of a thin soil layer (less than 20 ft) between the basemat and the bedrock can be significant, and the ranges of site conditions considered by the applicant do not include such sites. As such, either such sites must be excluded from the license basis, or AP600 design basis must include such sites and recompute the design basis seismic forces, moments, and floor response spectra. c) De use of only 3 layers of finite elements to represent the side soil may not be adequate for generating the floor response spectra for the 3-D SSI analysis cases. De applicant must demonstrate that the design basis floor response spectra, that were generated by enveloping the fixed-base case and the SSI cases that used only 3 layers of side soil elements, would not be excMd even if the SSI cases had used more (say about 8) layers of side soil elements. I d) De applicant has not explicitly demonstrated that AP600 seismic design complies with the " Sixty Percent" SRP requirement on the free-field seismic motion at the basemat level (see also Paragraph C of this section). 19- NTFS9649WQH/ san
l e) The design basis seismic forces and moments for the below-grade walls must be dewudned either from the 2-D SSI analysis cases that use about 8 side soil elements and considers the 3-D " Box Effects" properly, or from 3-D SSI analysis cases that use about 8 side soil elements. Also, these design basis forces and moments should include the effects of localized peak pressures on the walls from the rock cases, and the effects of through-the-soil structure-to-structure interaction from the buiklings adjacent to the Nuclear Island. B. In response to conclusion "a" above, the applicant has since revised the AP600 seismic design basis and has included an additional soft rock 3-D SSI analysis case and has reported the results in revision 9 of AP600 SSAR. 'Ihe revised enveloping floor spectra at a few representative locations were vnnpered with those from the Staff Review Team's analysis. The comparison showed that the revision 9 SSAR spectra satisfactorily envelope the SSI effects of the soft rock case. C. In response to conclusion "d" above, the applicant has since submitted the results of an additional evaluation in which it has been claimed / demonstrated that the free-field soil column frequencies (at which " dips" occur at the free-field basemat level spectra) are not close to the dominant structural frequencies of Nuclear Island, and so the likelihood of underestimating the structural responses due to the non-compliance of the " Sixty l I I I l l' NTFS96-4944H/ san
I Appendix A Parametric Studies using 2-D SASSI Analyses ! l
. )
i 1 1 l I l l A.1 i i i
A.I. INTRODUCTION ' In performing the seismic assessment of the AP600 nuclearisland (NI), the Applicant used the appmach of performing three dimensional (3-D) seismic response analyses of the NI for a few critical site conditions, nese critical site conditions were, in turn, defined on the basis of the results obtained from a series of simpler two dimensional (2-D) calculations performed for a wider variation in assumed site conditions. He site conditions which led to the critical or bounding response conditions in the 2-D calculations were used to define the critical 3-D site calculational models. Dese 3-D models were then used for design evaluation of the structure and equipment of the NI, defined in terms of both response spectra at critical points of the NI and stresses in structural elements. To evaluate the adequacy of this design approach, a series of two dimensional l (2-D) seismic response calculations were undertaken by the Staff Review Team, with the objective of assessing the completeness of the SSI (soil-structure interaction) c*n1* ions pehed by the Applicant and the degree of conservatism inherent in their design calculations. He 2-D structural model of the NI used in this evaluation was provided by the Applicant j
- and consists of a rigid box embedded to a depth of 39.5' below the ground surface. He model is based on an average cross-section through the short ducction of the NI (YZ direction of Figure A1)
- and has a width of appmximately 127'. Bree lumped mass structural sticks, mmi.esting the
! coupled shield and auxiliary building, the steel containment vessel and the containment intemal I structure, are attached to the buried bax by means of stiff links as indicated in Figure A2. He j seismic response calculations were performed using the SASSI (System for Analysis of Soil-Structure Interaction) Computer Code, as were those performed by the Applicant. ne free-field seismic input motion used in all cases for these 2 D calculatioris is defined by a time history generated by the Applicant and labeled as the H2 motion. His time history envelopes the Reg.
- j. Guide 1.60 spectra scaled to 0.3 gs and is amplified somewhat in the frequency range from 9 hz to 25 hz to account for cunently described Eastem United States (EUS) potential seismic motions.
l' l De time history provided for these evaluations is 20 seconds long, is digitized at 0.010 seconds and has a Nyquist frequency of 50 hz. The 5% damnM response spectrum for this free-field notion is shown in Figure A3 along with the Reg. Guide 1.60 and AP600 amplified target spectra. In general, the conclusions reached from these 2-D SASSI numencal studies have been based upon a comparison of the in-stmeture spectra developed at several key locations on the structural model from the various SSI cases investigated, ne particular node locations at which much output was generated are Node 3016 (the top of the Shield Building), Node 3116 (the polar A.2 b
1 i t I e L crane), Node 3204 (the operating deck), and Node 3110 (the polar crane support). ,. A2. SITE CONDITIONS EVALUATED l A number of SASSI calculations were psfume by the Staff Review Team to hi.e ( s the sensitivity _of the computed responses to the various parametens of the assumed site conditions l and to determine if enough cases were considered by the Applicant in determining their critical site conditions used for defining the more complete 3-D model evaluations. He site conditions l considered in the Staff's evaluations are listed in Tables Al and A2 and include variations in both soil and rock site conditions. He letters assigned to these site cases bear no relation to the site conditions assumed but are merely pmblem idendfiers. De primary parameters considered in the evaluadon consist of the depth of soil ovetburden to bedrock (which defines the depth of rock to l the basemat of the NI), the stiffness of the soil overburden, the stiffness of bedrock, the format of the soil degradation models assumed to obtain degraded soil evyerdes and the location of the free-
' field ground motion input with respect to the NI. A number of additional runs were conducted to
! investigate the influence of element sizes used in the SASSI computations on the accuracy of the computed responses. l The soil site conditions considered are presented in Table A1. nree of the site conditions
- j. (labeled Cases A, D and F) assume values oflow strain shear wave velocity ranging from 1000 fps at the ground surface and increasing with depth to a value of 2400 fps at a depth of 240 feet below the ground surface. Such a soil profile was initially selected by the Applicant from studies of a number of existing nuclear plant sites and is labeled as " soft to medium stiff" soil. Several variations on their configuration were used in this evaluation, including the assumed depdi to ,
l bedrock, the assumed form of the variation of shear wave-velocity with depth, assumed soil ! degradation properties and soil density. Another site (Case G) uses a low strain shear wave
. velocity of 1500 fps held constant thmughout the depth of soil, with no bedmck considered in this case. One site is labeled as a soft soil site (Case I) and has an assumed low strain shear wave l velocity of 1000 fps throughout the soil overburden. De final soil case (Case J) has a low strain ;
~
shear wave velocity taken fmm the data available to represent the Savannah River site located in ]
' South Carolina. His site is a deep soil site with a shear wave velocity which varies with depth in j the range of 1200 fps to 1500 fps. 'Ihe variation of the low strain or initial shear wave velocity is
- shown in Figure'A4 for these soil cases.
In each of these soil cases, nonlinear soil effects were accounted for in the free-field convolution studies, leading to final oriterated values of shear wave velocity, which are then used j
- as input to the SASSI calculations. These final velocities are somewhat lower than the low strain L - .
4
l shear wave velocities. The soft soil cases (Case I) had final shear wave velocities as low as 500 fps. This site configuration clearly falls into the class of soft soil site as defined in the SRP and requires special consideration. ! For the three rock cases (Cases B, C and H) considered, the NI was considered fully embedded within the rock halfspace, which extends to the gmund surface. 'Ihe purpose of these cases was to determine the influence of rock shear wave velocity on site response since the l Applicant used only a single mck velocity of 8,000 fps in all it: calculations. It should also be l l l noted that a comparison of results from Cases B, C, H and G also allows for an evaluation of the l adequacy of the fixed base evaluations performed by the Applicant.A plot of the shear wave velocity with depth for these cases in the upper 300' of the pmfile is shown in Figure AS. A3. TYPICAL RESULTS From the output generated from these various site conditions, horizontal response spectra were generated at the key output points selected (see Section Al, above) for the purpose of reaching at some allowing for geneml conclusions. Some typical results are presented in this report to ! I illustrate these conclusions. For example, Figure A6 presents a compaison of spectra developed at l l the top of the Shield Building (Node 3016) for the cases with relatively deep layers of soft to j medium stiff soil above bedroc.k, together with an envelope or bounding spectrum generated fmm the results of all the 2-D SASSI runs. This envelope spectrum results from the soft soil, soft to l medium stiff soil and rock cases. The results indicate that Case A using a linear variation in shear velocity with depth and the original Seed-Idriss degradation properties, as originally postulated by the Applicant, almat never govems the soil cases. Using a parabolic variation with depth together i j with more modern approaches to soil degradation (Cases F) lead to higher responses in the frequency range fmm 2 hz to about 7 hz. In addition, using a deeper depth to bedrock of 240' again increases the peak spectral values. However, these soil cases govem (yield enveloping spectral values) in only a narrow frequency range. i , Figure A7, on the other hand, presents the results of the soft thin soil layer above bedrock l (Cases I), together with the envelope spectrum. Cases 11 (11-40 and 11-45) use the H2 criteria motion input at the gmund surface. Case 11-40 has the bedmck situated directly under the hacemar of the NI, while 11-45 has the bedmck located 45' below the ground surface, or only 5' below the , basemat. On the other hand, Cases 12 use the H2 motion input at the top of bedrock, as recommended by the SRP for soft, thin soil sites. Again,12-40 and 12-45 situate the bedrock either directly under the basemat or only 5' below the basemat. As can be no;ed from Figure A7, the open dots associated with Cases Il do not generally reach the envelope values. The NI essentially i A.4 l
moves with the bedmck which is assumed to be anached directly to the basemet. The free-field ground motions at the depth of the bottom of the basemat are generally significantly lower than the surface criteria motions, as indicated in Figure A8. Some minor ex-dam in free-field motions occur at the higher frequency range above about 6 hz. However, defining the criteria ground motion as an outcrop motion at the top of bedrock has a major influence on response of the NI and leads to the largest values throughout the frequency range ofinterest. A comparison of the free-field motions at the level of the h==ammt l which tends to control plant response, is shown in Figurr A 8. He ground motions associated with i the top of rock are significantly higher than the deconvolved free-field motions. In addition, placing a thin soft soil zone only 5' thick below the basemat (Cases 12-45 of Figure A7) produces an additional peak in the low frequency range (around 2 hz). Thus, the treatment of the soft thin soil case can dominate the plant response calculation and is considered a controlling configuration l for plant design. Figure A9 presents a similar comparison of the envelope spectrinn with the results from the uniform site cases. As can be noted, the sites with shear wave velocities exceeding 3,500 fps produce essentially uniform ra-maa and can be considered as the lower bound of the definition of rock for the AP600 plant. Similarly, Figure A10 presents a comparison of the calculated envelope response with the original ymposed design spectrum for this node as well as the results obtained fmm the fixed base calculations. As expected, in the high frequency range, the fixed base results exceed even the results from the hard rock cases, although they are significantly below the envelope and the design spectra in the lower frequency range. A4. SENSITIVITY OF RESULTS TO MESH SIZES - I In addition to the various site conditions, a number of variations in mesh sizes used in the j finite element representation of the SASSI model were evaluated. In the 2-D model, the primary element discretization made use of 8 elements through the depth of the NI model, as indicated ! schematically in Figti A2, together with 26 elements along the length of the h==ammt In addition, most of the 2-D SASSI models made use of soil elements beneath the concrete basemat and to the sides of the NI to determine the contact pressures developed between the soil and the NI. The question of sensitivity of the computed responses to element sizes became significant since it was known that the 3-D models which were to be used to compute design responses would use larger element sizes to keep the size of the 3-D clement meshes within reasonable bounds. Dree additional calculations were run with the 2-D SASSI model using the same general
site conditions (Case 11-40), with the criteria motion input at the foundation depth as an outemp motion, with different soil layer discretizations to the side of the NI. One pmblem was run with only 3 soil layers provided (coarse layer mesh), one with 5 layers (median mesh) and the third using 8 soillayers (fine mesh) to the side of the NI. This configuration was evaluated since 3 soil layers were indicated to be used in the Applicant's 3-D SASSI calculations. Two of the meshes used are indicated schematically in Figure All. He number of elements along the basemat was similarly coarsened for these runs. In addition, SASSI runs were made with elements under the basemat either included in the model or removed, with the NI then resting directly on the rock halfspace. It was found that this latter variation did not play a significant mie on the computed responses. However, it was found that the mesh size used for the side walls had a more significant impact on computed response of the NI. Figures A12 and A13 present typical effects of sidewall-clement number on response spectra. As can be seen, significant differences can occur, particularly at the peaks of the computed spectra. Such behavior indicates that the coarser mesh generates more effective damping leading to reduced peaks of the nodal responses. The 3 layer element discretization for the soft soil site is at the limit of the wave transmission guidance provided for the SASSI calculations at about 10 hz. Some additional responses are noted to be generated at the higher frequencies associated with the second mode response of the structural model. This may indicate that the coarser meshes are somewhat stiffer inducing response at the higher frequencies. In addition, the wall pressures computed fmm the coarser meshes are significantly influenced by element size effects. AS. CONCLUSIONS Based on a review of the numerical responses evaluated, several conclusions have been reached conceming the completeness of the design evaluation performed by the Applicant. h should be noted that these conclusions have been arrived at by wn.yarkg the computed response spectra at several locations on the stick models. De conclusions reached can be summarized as follows.
- 1. By comparing spectra at given locations developed fmm Cases B, C, H and O and comparing these to similar results obtained from the fixed base case, it was dAmJr,ed that for rock shear wave velocities of 3,500 fps and greater, the fixed base analysis produces conservative results. At frequencies above 10 hz, the fixed base analysis is very conservative since the SSI effects eliminate the second mode noted in some of the fixed base spectra. The spectra from the fixed base analysis fall below the SSI results at the lower A.6
(
l frequencies conesponding to the SSI modes of the system; that is, SSI causes spectral amplifications at the coupled SSI/ structure frequencies.
- 2. The depth to bedmck was found to be a controlling parameter. For the soft to medium stiff soil case (Case A), the Applicant indicated in the SSAR that they considered two depths to bedrock in their 2-D evaluations,120 feet and 240 feet, with the 120 foot depth found to be more critical. He results of the Staff Review Team evaluation indicates that the results at deeper depths may lead to increased, particularly when combined with more recent degradation propenies and more realistic variations in shear wave velocities.
- 3. For the soft soil site (Case I), the final iterated strains lead to shear wave velocities as low as 500 fps. Fcr the cases considered, with depths to bedrock of either 40 feet or 45 feet, such sites qualify as thin soft soil sites according to the SRP definitions. For such cases, the broad banded RG 1.60 type spectra are considered more appropriate as input as outcrop motions at the top of bedrock. Each of the two cases were then run with the H2 motions input at the ground surface and at the rock interface as an outcrop.
For Case II, with the depth to bedmck at the basemat level, the spectral responses computed for rock outcrop input were not significantly different from those computed from the rock cases, since the motion input at the foundation level for the rock cases is similar to thatinputin the rock outcrop case. De loadings applied by the soil to the sidewalls of the NI, however, can be expected to be significantly higher than those from Case 11 with the motion input at the ground surface, since the peak soil displacements throughout the soil layer above the bedrock are greater than the motions computed with the input at the ground surface. These soilloads can also be expected to be different from the results for the rock cases, although the peak values for all these cases have yet to be evaluated. For Case 12, however, the spectral responses were found to significantly exceed the spectra for the fixed base case as well as the rock cases at the horizontal SSI frequency associated with the thin layer of soil between the basemat and the bedrock. This exceedance will change with both frequency and magnitude as the layer of soil remaining below the ha_wmat gets thicker. Therefore, the allowable depth to bedrock is an imponant parameter controlling response of the NI which is felt to have been incompletely evaluated by the Applicant.
- 4. Three calculations were run with SASSI using the same general site conditions (Case II) with different soillayer discretizations to the side of the NI. The msults of the calculations A.7
I i indicate that the use of 3 soil layers for this soft soil condition leads to a reduction in the spectral peak (as compared to the 8 layer case) of about 25% at a frequency of near 10 hz and an increase in the computed spectra at higher frequencies. The adequacy of the 3D SASSI models to be used by the Applicant for the design basis computations should therefore be checked to ensure that they exceed the wave transmission guidance provided in the SASSI users manual. In addition, the computation of wall loadings from such 3D calculations are notrecommended. l
- 5. For Case K, which is the same site condition as Case A2, two computations were made l using two different angles of incidence of the incoming S-wave, one of 0* (vertical >
incidence) and one of 30' from the venical. It was found that the effects of angle of incidence of the incoming wave had negligible effects on computed spectra.
- 6. The evaluation of the effects of soil saturation on computed responses and on soil pressures acting on the walls of the NI have not been completely evaluated. He Staff Review Team SASSI computations were made assuming dry soil conditions. 'Ihe Applicant's computations assumed pseudo-saturated conditions, that is, the P-wave velocity of the soil was set equal to 5,000 fps (approximately that of water) while the shear wave velocity and soil unit weights were kept at their dry values. Additional 2D computations made by the Applicant using an increased value of soil unit weight (from 120 pcf to 135 pcf) to account for typical saturated soil effects showed negligible effects on computed responses.
- 7. An additional SASSI calculation was performed by the Staff Review Team for the soft to medium stiff soil site (Case A) but using an enhanced H2 free-field ground motion for input j to the site at the ground surface. In this evaluation, a typical soil variation in the free-field deconvolution calculation was performed, varying the shear moduli from 1/2 to 2 times their best estimate values defined by the Case A configuration. De free-field spectrum at the foundation level of 40 feet was then computed for each case and found to fall below the 60% criteria indicated in the SRP over a frequency range of from about 1.5 hz to about 10 hz. The foundation level spectrum in this frequency range was then amplified and convolved to the surface using the best estimate soil column to arrive at an enhanced surface H2 motion which presumably would satisfy the intent of the SRP guidance. This enhanced motion was then used in the SASSI computation to arrive at response spectra which show significant exceedances over the corresponding fixed base spectra typically used for comparison purposes.
- 8. Combining the effects of parabolic variations in shear wave velocity with more modem A.8 1
i
( 1 i degradation models leads to somewhat gmater responses than those soil cases investigated by the Applicant. l l A.9
TABLE A.1
SUMMARY
OF 2-D SASSI ANALYSIS CASES FOR SOIL SITES (ROCK VELOCITY = 8000 FPS) CASE SOIL ASSUMED LOW STRAIN SUBCASE DESCRIP'IOR DEFIH TO TYPE SHEAR WAVE VEIDCITY ROCK (feet) A Soft to 1000 fps to 2400 fps @ 240' Al dry soil 80 Medium varying linewly with depth with A2 dry soil 120 Stiff Seed-Idriss degradation properties A3 wet 120 A4 saturated 120 D Soft to 1000 fps to 2400 fps @ 240' dry soil 240 Medium varying parabolically with depth with Stiff Seed-Idriss degradation pugnies F Soft to 1000 fps to 2400 fps @ 240' dry soil 240 Medium varying parabolically with depth with Sd-X lower bound degradation pivgnies G Uniform 1500 fps uniform soil site with dry soil none Site Seed-Idriss degradation piugrdes I Soft 1000 fps uniform soil site with 11 dry soil 40 Site Seed-Idriss degradation pivgdes 12 dry soil 45 J S@ Savannah River deep soil site with dry soil none Deep variable velocity profile andlower Site bound degradation pugnies K Soft to Same as Case A: Medium VenicalWaveIncidence K1 dry soil 120 Stiff 30' WaveIncidence K2 dry soil 120 l l A.10 l l 1
l TABLE A.2
SUMMARY
OF 2-D SASSI PROBLEMS FOR ROCK SITES CASE ASSUMEDIDW STRAIN ROCK SHEAR WAVEVELOCITY (fps) B 11,000 C 8,000 l H 3,500 I e I I i A.11 ) i
[ l l l 2 127 m
'l ;
I d 91'
^^
l l l p EQUIVAENT j, I i RECTANGE g ; iree f - l FOUNDATION l PIAN 72' R V 3'
/ w U d 63.9 m m 23' m d "Y m, - , -
(West) F-28' 4 > .
, 117
, m, X
r (North) FIGURE A1 AP600 FOUNDATION PLAN 3 l
i l l J k mR l Bevation (ft) Y v 100 l -
^N^ ; ; "m- .; .,
d Building Sidewall senapiragm g,,,,,t,- ; si;ck goa,i,
. g,,,,g, . . ,
f MM ll 3
%^ j ;,
q m side soil Ecznents . , I. - Rigid Unk Bements .
- '. z.:s9 l INbbbl//NN/>7)//M Conaste Basemat Bements % fj ff y g7 ,/ ,f 3.:.k. -
1y 55
#'I"I'l 'N'N #M44'id'W '^^'tUti SoilBa$cBemcats 1412.:'F$1, p'.g..eF;/yng.: b,--l..4 ,
2 127 W FIGUREA2 SASSI NUCLEARISLAND MODEL l IN THE YZ PLANE 1 l h I L____._______ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
I i l I I i 1.25 ., . . , H2 Motion 5% Damping l AP600 Target Specinun l 1 no 1.60 specinsa
- ,s,s .
z ' O % 0.75 -
\, -
g a, -
\. .
\*
g f/3 0.25 - [
./
0 O.1 1 10 100 FREQUENCY (hz) . FIGURE A3 COMPARISON OF H2 MCTTION SPECTRUM WITH TARGET AND RO 1.60 SPECTRA i I i i ___._m___.___-_ ..__m--_____-__m _ _ _ _ _ _ - - - - _ .
1 l i I O i i i
-9 .
&-4
-t 1 M, .
50 - 4 - O - 0--O 0-4 i im - - e
- ?.
T : e 3 b a b'? 12 - 9 - CASE A b -- O OM o..... 9 i
---C CASE D 200 -
....o."* CASEJ QD M-O CASEll i i' '
o-- 250 O 500 1000 1500 2000 2500 IDW STRAIN SHEAR WAVE VEIDCrrY(fps) FIGURE A4 VARIATIONS IN LOW STRAIN SHEAR WAVE VELOCITY FOR SOIL SITE MODELS - 1
!ll Il 0
l
~
0 ; g , ,. , l l !
. e
' i 100 .
l !
'. -l l a
200 - e E s i e g . I CASE B. Vs = 11.000 fps CASE C. Vs = 8,000 fps
# ~ ~
CASE G Vs = 1.500 fps CASEH Vs =3.500 fps g i e 0 5000 10000 15000 IDW STRAIN SHEAR WAVE VEIDCETY (fps) FIGURE AS VARIATIONS IN LOW STRAIN SHEAR WAVE . VELOCITY FOR ROCK SITE MODELS
i 10 i 's ENVEIDPE OF A11 CASES RUN 5% Damping CASE A I O- - CASE D120 7.5 CASE D240
~
i j ---m--- CASE F120 i l z t 0 - - +- -- CASE F240 < >
-- -N(-- - CASE J sev l 5 - ~
.. y .
a , h . .. z$ . ' f ( 0 O.1 1 10 100 FREQUENCY (hz) FIGURE A6 RESPONSESPECTRA ATNODE3016, SHIELD BUILDING ROOF, AT DEEPER SOIL SITES i
)
i 1
)
10 .. i 5% Damping l ENVEIDPE
--- 0 -- CASEll-40 lli <
a CASE ll-45 ll 7.5 - -
---E--- CASE 12-40 O >1 llll f - - +- -- CASE 12-45 i o 5 ,, 11l l l
l {< 5 - - I i 1
" i
,.hi l
m - W*% , is - * - I e _ _ - _ _ . i
%; . a,w\
O' "' i 0.1 1 10 100 l FREQLT.NG' (hz) FIGURE A7 RESPONSE SPECTRA AT NODE 3016, SHIELD - BUILDING ROOF, AT SHALLOW SOIL SITES 1
l 1.5 ., ,, , H2 CIUTERIA MOTION 5% Dam #g 0 OUTPUr 9 45, H2 6 SURFACE
"-*"" OUTPUT @ 40'. H2 6 $URFACE .
OUTPUT @ 4S, H2 9 45 i l 3 , - --- - OUTPUT 847, H2 9 40' , x 9 o a og \ 5 . l&% '. . N g
.;}i A :
J % *
$Rllft ~
k [ Qe <= 4,
/_
03 - - e 5 k < 0 O.1 1 10 100 FREQUENCY (bz) , i FIGURE A8 FREE-FIELD MOTIONS FOR SOFT THIN SOIL LAYER l ATOP BEDROCK 1
I i 10
., ., i G l
~
5% Damping l ENVELOPEOFALLCASES , j
--- O---- ROCK 11.000 FPS F -
^ I ROCK 8.000 FPS l 7.5 - '
- - -gi- - - ROCK. 3.500 FPS !I '
l 2 lI o -.-.m.... soti.1.500 FPS ' I z O i h. e I ll EI - l
? 5 -
c< h.. gl a 19 { l
- c. E l Gl b
15 - - N ~~,au.......... ......; ; 0-s c O.1 1 10 100
. FREQUENCY On)
FIGURE A9 RESPONSE SPECTRA AT NODE 3016, SHIELD j BUILDING ROOF, AT UNIFORM SITES 1 1 1 1
)
1 1 l l l 1 a
115 .i .. ENVELOPE OF ALL 2D CASES 5% Damping 0 DESIGN ENVELOPE h 6 10 - j' -
-o-*** HXED BASE Wmi4% .Q STRUCTURAL DAMP!NG y ~6- FIXED BASE WirH 7% I ~
Z MU D M NG I )
- o 7.5 -
^
b h i : m 9 g .- : J $ -
\ .
- k '
te ( O 15 - D - p u - 0
~ **
O.1 1 10 100 FREQUENCY (bz) FIGURE A10 RESPONSE SPECTRA AT NODE 3016, SHIELD BUILDING ROOF, COMPARISON WITH DESIGN SPECTRUM -
1 l 4 aR l t i
' ~
; "N l l ^6' .;
.. 6 "II '
.: i Building - :
I jh ' StifTDiaphragm _ Stick Models U '"*""
]'
Bements WM )*
?' I l l ; ;
M' i ,
-l Side Soil _ {
~'
s ;
. : Rigid Link Bements
)
l$%%bb/O/%Y)//f/ Concrete Basemat Bements /
% '/Y/f/% / // %!h 1
'///
.:@sy$$i@3Rd:.##:- .X:q M$kye'/ Xid Soil Base Bements lif.f9M9.:$?::0:LW.:X@@::?:M:Qs?
, 127 ,
hR
;4
-c ^N l l 'h' -. .
.- a i
- sj Building Sidewall ' :: l S; .].;
. Stiff Diaphragm _ Stick Models Bemenu ;j S
Bements .#.' )
- .'I NM l l l l M^ 4 .f f .:" Side Soil 'j.
-Rigid Iink Bements -'
! es.:/gggggggs/c.:/aggggg, Soti Bose Beme.a s:gggn:<g/;.- ;:gggey;.:ge .t.:, 127 4 O l RGURE All VARIATIONS IN SASSI RNITE ELEMEbtr DISCRETIZATION l ) l t
l 10 ., .. . 8 ELEMENTS THROUGH DEPTH 5% Damping .
. I l 3 E11MENTS THROUGH DEPTH P
7.5 - F g - k 5 - h
.J j [
1 - i 15 N N v . a 4 0 ' 0.1 1 10 100 FREQUENCY (hz) i FIGURE A12 RESPONSE SPECTRA AT NODE 3016, SHIELD BUILDING ROOF, FOR DIFFERENT ELEMENT SIZES THROUGH DEFTH OF NUCLEAR ISLAND 9
5 i i 8 ELEMENTS THROUGH DEPTH 5% Damping 3 ELEMENTS THROUGH DEFTH l \ 4 -
?
3 5 3 g - f' l d 2 - i p
\,
I '- 4 0 ' ' O.1 1 10 100 FREQUENQ' (hz) FIGURE A13 RESPONSE SPECTRA AT NODE 3204, OPERATING DECK, FOR DIFFERENT ELEMENT SIZES THROUGH DEFTH OF . NUCLEAR ISLAND
l I l i t I Appendix B Evaluation of Soil Pressure on Below Ground W alls i
)
t I 4 I ( f i B.1
l Bl. INTRODUCTION-As part of the seismic assessment of the proposed AP600 advanced reacmr system,
- - investigations were undertaken by the Applicant to estimate the dynamic wall pressure that may develop along sidewalls of the nuclear island (NI) from seismic motions of the plant and
( surroundmg soilhock. Initialpresentations to the Staff Review Team had proposed the use of the l Mononobe-Okabe (MO) method to estimate both magnitude and distribution of wall pressures. 'Ihe Staff Review Team disagreed with this approach since the method is formally appropriate only for l retaining wall structures, could not be shown to be a conservative estimate of wall pressures and l could not easily inccigine the effects of structures adjacent to the NI on dynamic wall pressures. In addition, the current Staff position clearly outlines the pivMess inherent with the MO method and iMac that the procedure to estimate soil pressures must be able to incorporate the potential l l effects of relative dynamic displacement of the NI with respect to side soils as well as the soils below the i==an=t. including potential rocking and rotation of the NI. l i Following these discussions, it was generally agreed that dynamic pressures can best be handled within the framework of the SASSI formulation since these were to be performed anyway
- . to evaluate seismic response ofinternal stmetures and systems within the NI. 'Ihe major drawback to this formulation was the anticipated adequacy of the SASSI computations for the soil pressure computations. As discussed in Appendix A, it was L a "r.si from the many 2D SASSI models performed by both the Applicant and the Staff Review team that relatively fine finite alamant -
meshes are required to obtain reasonable accuracy in computation of dynamic wall pressures. For example,it was foend from the 2-D SASSI computations that about 8 elements were required along the walls of the NI to achieve a reasonably accurate stress computation. Using such a discretization requirement in the 3-D SASSI models would lead to extremely large 3-D computational models that had to be used for the design computations. It was therefere agreed that an altemate approach would make use of the 2-D SASSI stress computations which could then be l modified in a simplified manner to incorporate corrections for potential torsional motias of the NI (anticipated to be small) as well as potential comer effects. In the 2-D S ASSI models, the effects of the light structures adjacent to the NI could be included to estimate any additional pressures that may be inW by differential seismic motions of the NI and the adjacent structures. In this Appendix, we summarize the results of the 2-D SASSI computations conducted by the Staff Review' Team. 'Ihese were done to detennine the influence of the additional site conditions
evaluated on the computed wall loads and to pmvide confirmation of the Applicant's estimates of - wall pressures to ensure that seismic designs of the NI are appropriate. In addition, an area of concem in developing these wall loading conditions has to do with the effect of ground water on induced wall pressures. He plant design condition assumes that the ground water table is at or near the operating deck level, or nominal elevation 100. ne only impact , of this assumption on the almiah of dynamic wall pressures is in the selection of the value of Poisson's ratio of the surrounding clastic soil material. In this eypoich estimates of horizontal shear wave velocity for the various site conditions are combined with an assumed P-wave velocity of approximately 5,000 fps for the soil which matches the wave velocity through water. His leads to a value of Poisson's ratio which is relatively high (approaching a value of 0.5). Howevcc, as Poisson's ratio approaches 0.5, the SASSI computation (as well as most other codes) h=== unstable since it cannot treat this incompressibility condition. attempts were made to address the issue of the inclusion of pore water effects on wall loads and if in fact the elastic SASSI type solutions provide reasonable estimates. Values of Poisson's ratio typically utilized in the SASSI runs varied from 0.45 to 0.48. De question of the adequacy of this assumption should be evaluated. Since no complete computational solution similar to SASSlis currently available to treat the water / soil system as coupled but separate systems, the results of other computations were used to try to evaluate the adequacy of the SASSI approach. Dese results are also described in this Appendix. B2. SITE CONDITIONS EVALUATED ! A number of SASSI calculations were performed by the Staff Review Team to Ludue the sensitivity of the computed responses to the various parameters of the assumed site conditions and these have been described in Appendix A. De site conditions considered in the Staff's evaluations are listed in Tables Al and A2 and consider variations in both soil and rock site j conditions. De soil site condidons considered are presented in Table Al while the rock site assumptions are listed in Table A2. The soft thin soil cases (Case Il and I2) fall into the class of l soft soil site as defined in the SRP and requires special consideration. For the three rock cases l (Cases B, C and H) considered, the NI was considered fully embedded within the rock halfspace, l which extends to the ground surface. The purpose of these cases was to determine the influence of l rock shear wave velocity on site response since the Applicant used only a single rock velocity of l 8,000 fps in all its calculations. It should also be noted that a comparison of results from Cases B, C, H and G also allows for an evaluation of the influence of rock modulus on induced dynamic ; wall pressures. B.3
l B3. TYPICAL RESULTS OF 2 D SASSI ANALYSES From the output generated from these various site conditions, the pressures generated along the walls were generated for several of the site cases. A special purpose pc-giucessor was written which converts the stresses p.d by SASSI in the soil elements adjacent to the NI to obtain peak pressures applied to the wall with depth as well as the total forces generated along the wall with time. From this computation, the marinunn force was determined as well as the pressure distribution along the wall at the time of the peak total horizontal force. Figure BI presents a plot of the peak computed tensile dynamic pressures (in ksf) generated along the wall for the various rock cases while Figure B2 presents the same data for the maximum compressive dynamic pressures. l These pressures do not occur at the same instant of time, but are the maximum at the particular levels shown. For these runs, eight wall elements were used to compute pressures which were assumed to be uniformly ' distributed over the width of the element. In addition, the maximum j induced pressures generated from all the runs (both soil and rock) is also plotted in the figures. h can be noted that the maximum pressures developed along the wall are developed from the hard rock case, Case B, which has a shear wave velocity of 11,000 fps. As the shear wave velocity , decreases, the pressures on the wall similarly decrease. Maximum dynamic pressures are always l induced at the top of the wallindicating the participation of the walls in resisting rocking of the NI. l The values of pressure for the hard rock case are similar in magnitude to the values generated by the Applicant. It should be noted that the static at-rest pressures are also plotted on these figures and it is clear that on the tensile side, the dynamic pressures exceed the static over a major part of the wall, indicating the tendency for separation of the NI from the surrounding rock. If this is is fact true, l the tensile forces developed by SASSI cannot be sustained, and either the dynamic loads are transferred to the compressive side, increasing their value or are transferred through other parts of the NI wall /basemat system to the surrounding rock. This issue has not been previously discussed and should be evaluated by the Applicant.
'Ihe results from the soft soil cases (Cases Il and I2) are shown in Figure B3. Again, the peak pressures occur at the top of the wall. It is clear that the induced wall pressures are generally
'iower than those developed by the stiff rock case, although for the cases with the criteria motion input at bedrock, the induced pressures are significantly higher than the results from the other soil cases. In all cases, however, the dynamic pressures again exczed the at-rest condition, indicating that someload transfer must occur. B.4
It should be noted that the results from the other 2D runs using a coarser finite A-.m mesh through the walllead to signincantly different wall pressure distributions. 'Ihis is obviously the case as the variation in wall pressures is generally so sharp at the top of the wall. Using a larger element would then smear this distribution, leading to potentially incorrect wall design loads. This lends further support for using the process of modifying the 2-D SASSI results mther than using the results from the coarser meshes of the 3D models.
. B 4. INFLUENCE OF PORE WATER ON SEISMIC RESPONSE The effect of pore water on seismic response of the AP600 or any other plant plant is of interest from two different perspectives. The considerations for AP600 were of interest since one ,
of the design conditions was that the ground water table was at or near the ground surface, with the plant exterior walls then being entirely below the ground water table. 'Ihe implications of this j assumption on seismic response can be summarized as follows. l l l *Ihe first issue has to do with the effect of saturation on the dynamic seismic response of structuits. For most ordinary buildings, soil-structure interaction (SSI) has not been an iiufui.nt ; issue in evaluating dynamic response and structural response can be adequately d-m*M by fixed base analyses. However, for those structures where SSI is lup-st (such as the NI , structure), previous studies have been performed for NRC (Ref. B1 and B2) to determine the effect of pore water on SSI interaction coefficients. A special finite element computer code ! (POROSLAM) was developed which performs a frequency domain linear finite element analysis of a two dimensional plane strain clastic medium, similar to the SASSI formulation. 'Ihe Code treats both the solid and fluid fractions ceparately and accounts for their interaction through Biot's two-L phase theory. Coupling between the soil and water takes place by accounting for both peehility l as well as volume compressibility effects. The analysis is similar to considering both intergranular L and pore water in ordinary static soil mechanics effects to arrive at the total stresses developed in the ground or against a wall. l 'Ihe effects of soil saturation on dynamic response can be shown by considering the l response of a rigid footing founded at the surface of an elastic two-phased half space (Figure B4). The forces developed on the footing for unit harmonic footing motions are then the impedance functions that are gered from similar one-phase solutions, such as with SASSI. Figures B5 > and B6inacee the comparison of the rocking and vertical interaction coefficients as functions of frequency for the fully dry and fully saturated soil cases. As you may note, differences can be significant. Fortunately, the differences that exist for the horizontal impedance coefficients are not l B.5
F l as great as for the vertical and rocking cases. Dese differences in impedance functions can have a significant impact on seismic response and critical frequencies of some facilities, particularly when I SSI is an important component in the seismic response. he typical way that water is included in the one-phase analyses, such as a SASSI type l seismic response analysis, assumes that the soil / water act as one material. He approach is to select the shear modulus of the material to match S-wave properties of the soil and the Poisson's ratio to match the P-wave velocity of the water. His often results in selecting values of Poisson's ratio between 0.45 and 0.48. De pmblem then Wmes one of numerical stability as the value of Poisson's ratio approaches 0.5. Therefore, most people select Poisson's ratio no greater than 0.48. Some comparisons (Ref. B2) of the conect fully saturated two-phase analysis with the one-phase approximation using the fully incompressible assumption with Poisson's ratio equal to 0.5 indicates significant differences in impedance functions, particularly at higher frequencies. De iuwinnce of these differences obviously depends on the particular geometries and frequencies of interest or any one project. ney have not been evaluated on any of the advanced reactor programs. De second issue ofimportance involves the computation oflateral soil pressures computed on vertical walls developed by dynamic shaking of the facility. Again, the typical approach is to use a code like SASSI (one-phase approximation) with a Poisson's ratio of about 0.48 to match P-wave velocity through the pore water as described above. We have atwayisi to estimate the j potential magnitude of such effects by using a simple ID constrained rod analysis. De model is I shown in Figure B7, in which the left end of the constrained rod represents the vertical wall that moves horizontally into the saturated soil. Obviously, this model assumes that the wall pressures are only influence by compressional waves moving through the solid and water fractions of the ; soil. A comparison of results is shown in the bottom of Figure B7. In the analysis, a steady state wall movement of 1" maximum magnitude is moved into the saturated soil at a given frequency. He stresses in both the saturated and dry cases developed at the wall have been computed and are j shown in Figure B7. A comparison of stresses is made with the dry solution using different values , of Poisson's ratio. For this case, to match the total stresses developed on the wall, the Poisson's l l ratio used in the dry analysis must reach a value of between 0.49 and 0.50, a value higher than l typically used in SASSI type analyses. For lower values of Poisson's ratio, the computed stresses in the dry case are lower than the saturated results. De impact at regions closer to the free-surface where dissipation of pore pressure is easier and where total p[ressures are higher are expected to be less severe. l B.6
B5. ~ CONCLUSIONS The tentadve conclusion mached support the approaches that are being performed by the Applicant to genente wall pressures on the vertical walls of the NI. Since the discatization requirements are so severe for the 3-D SASSI models, the approach of using the 2-D SASSI results with modifications based upon rational engineering approaches is felt to be more appropriate than wag to extend the capability of the 3-D SASSI model. Secondly, the pressures computed using the one-phase soil-water representation is felt to lead to conservative , estimates of wall loadings, particularly at or nea- the ground surface where computed pressures are largest. I B 6. REFERENCES Bl. C. J. Costantino, A. J. Philippacopoulos, " Influence of Ground Water on Soil-Structure l Interaction", Civil Engineering Dept., City College of New York for Brookhaven National Laboratory , NUREG/CR-4784, October,1987 - B2. C. J. Costantino, " Soil-Structure Interaction: Influence of Ground Water", NUREG/CR-l 4588, Volume 3, Brookhaven National Laboratory, April,1986 l B.7
i i 1 { l l l 100 , , .4 , ,
; i.
l of o ; ! p'
- 1 t i n, o < >
! 9 . 90 i = -
/ l
;e :
'< ld MAX VALUES FOR A11 CASES
! I' i m. n . ! "*- T' C CASE B. I1,030 FPS < l g a v e o j< > I Z '
, O CASEC.8,000 FPS O e p 80 -e <W> -
i a
.e e i S ; i --4-- CASE H,3.500 FPS ,
i 4 e m< > l l C I ======* STATIC. AT REST l 2: ': , -
.e ; e i l 5 m> *< > .
l l
', e ; <
' e ! 70 <>:: - l e
.a i el e i
1
.W> I #h e e
l } e4 > e e , e g $I > f f i f l 0 5 10 15 20 25 i l l i ! PEAK DYNAMIC WALLTENSIONS (ksf) l - FIGURE B1 COMPUTED DYNAMIC WALL TENSIONS FROM 2D SASSI PROBLEM SET FOR UNIFORM ROCK SITES . 1 i t o--_ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
100 \ i e f e~
, i 2:
l
. O d '
li
.g e------ ~ ::
'ta. (:.
m bh *
'i -
MAX VALUES FOR ALL CASES
~ <>I i! - - + - CASE B. I1.000 FPS .
q-1 9 l l ****4**** CASEC 8,000 FPS 5 . 4 o
^
! ---*--- CASE H. 3.500 FPS p< 80 - 1,; 4p H > -
. *e ======* STATIC, AT REST t ; alo ti:
. e.l s e
- e: a
.e:
70 k-{ ~
.1 4.;.
s o .
. g
- e. .
e: . Id> *v> l! e. 0 5 10 15 20 25 PEAK DYNAMIC WALLCOMPRESSIONS (ksf) FIOURE B2 COMPUTED DYNAMIC WALL COMPRESSIONS FROM 2D SASSI PROBLEM SET FOR UNIFORM ROCK SITES
100 y
.p..
M $ 1
- 1 : i
- asti .;.= : : i
- l i:
. : 40 + 8 90
- I
;4Il Al H i!
4
- gh , MAX VALUES FOR All. CASES m
hr35 9 *8 SMMC. ATREST 6 ---A--- CASE 11. H2 @ SURFAG, D = 45 z d,,[d !;E 80 - d- ---D- CASE 11R, H2 @ SURFACE D = 40'
~
' 40 Q l
.e -*e - -*A - -* CASE 12. H2 @ ROCK. D = 49 I l
- e. 1 s ---E--- CASE 12R, H2 @ ROCK D = 40'
-N '1q .
.i. , ,
8
$, Y, :. A. !
as
!!i' a
V 4 +,;
- i -
60 6 A l:i . . . i 0 5 10 15 20 25 PEAK DYNAhDC WALL COMPRESSIONS (ksf) FIGURE B3 COMPUTED DYNAMIC WALL COMPRESSIONS FROM 2D SASSI PROBLEM SET FOR SOFT SOIL CASES
1 f k UNIT FREQUENCY DEPENDENT FORCING FUNCHONS i I m I & -+
- RIGID IUOUNG 1
< t HA15-100 TING WIDIH,a i +
o l l \ __ J ' FULLY SATURATED _ UNIFORM SOIL I l Q l l FIGURE B4 RIGID FOOTING ATOP AN ELASTIC, FULLY , i l l SATURATED SOILIN A TWO DIMENSIONAL PLANE STRAIN CONFIGURATION l 1 l l l l l l l : 1 l l I l i l l l l l l _ _ . . _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - .>
l i m 1.0 - dry Soil - m ma 2
, ~ ,, -
'~..,P-..,,,,,,...,,,,,,,,...:
j g '- b 0 l o . \N l % . s . l p: - 1.0 Saturated l Soll 1 - 2.0 - . . 1 0 1 2 3 4 5 DIMENSIONLESS FREQUENCY a O i Y, l l 7 . . . . . . . . . 6 - O Z
- 5 - -
l l 1 A 4 . . O E l 3 - - 1 Dry Soil l Saturated S i1 2 -
,,,...v..
.. ~-
...... - i
----~~~~~',,-
0 - - - - - - 0 1 2 3 4 5 DIMENSIONLESS FREQUENCY a OIV3 FIGURE B5 COMPARISON OF ROCKING IMPEDANCES FOR DRY AND FULLY SKIURATED CASES l l l 1
L 2 . . . . . . . . .
........o.................................,..............
. 0 Dry Soit -
m m na -4 - . Z A m . Saturated ' l @ Soll - d -8 - . O _{2 . . j .. l -16 . . . . . . . . - l 0 1 2 3 4 5 1 DIMENSIONLESS FREQUENCY a O lV, l l 25 . . . . . . . . . i l l 20 - - l l l O ~ iC - - fA 15 g Saturated o Soil 10 - -
~
Dry Soil 5 -
.,.....P.....""
O ~~Y'' . . . . . . - 0 1 2 3 4 5 DIMENSIONLESS FREQUENCY a O I V, FIGURE B6 COMPARISON OF VERTICAL IMP 5 DANCES FOR DRY AND FULLY SATURATED CASES 1 I l 4
)
1 I u = U exp(iwt) U = [1",0]
\
One Duneasional Constraned Bastic Rod
\
1250 . . iao - 1.o ma.e S.H u. e] 1= One Phase Sou Model ; C; 500 - I 250 ( , . 0 ' S.25 03 035 0.4 0.45 0.5 POISSON'S RATIO e 4 FIGURE B7 COMPARISON OF DRY AND SATURATED SOLUTIONS FOR ONE DIMENSIONAL CONSTRAINED SOIL
, _ _ . - _ . _ _ _ _ . _ _ - - _ - _ - _ - _ . - _ - _ - - - - _ - - - _ . - - - - - _ - - . _ - _ _ . . . - _ - - . _ . _ _ _ _ _ _ _ _ _ E
Appendix C Parametric Studies using CARES Computer Code 1 I I l i l c.1 1 ---------- _ _- -
I l C 1. INTRODUCTION As part of the seismic assessment of the pmposed AP600 advanced reactor system, several j efforts were undertaken by the Staff Review Team to investigate additional problems associated l .with soil-structure interaction (SSI) aspects of the seismic review. One of these areas of investigation was concemed with the critical depth of soil overburden to bedmck since it was found from the 2D SASSI M-..;-etional set that this p- i was a significant player in the seismic response of the nuclearisland (NI). As described in Appendix A, the Applicant performed l 2D SASSI computations which had a limited variation in the mck depth. However, it was found fmm the SASSI analyses perfonned by the Staff Review Team that when the rock interface l approached the bottom of the NI ha-mar. critical response could be calculated which may exceed the design response spectra, particularly at low frequencies. To evaluate this impact using further SASSI computations would requite significant additional time and expense. A second issue that arose during the review of the Applicant's computations had to do with the amount of out-of-plane motion that could be developed at critical locations on the NI, and the impact of variations in assumed ducction of seismic input motions on these responses. 'Ihe primary variational study performed by the Applicant made use of 2-D SASSI models which obviously cannot be used to evaluate this issues. Only the 3-D SASSI model can be used to evaluate this issue and since so few D pmblems were to be evaluated, the issue could not be properly evaluated. For both these issues, the Staff Review Team performed several simplified analyses, making use of the CARES Computer Code (Ref. Cl) to perform approximate confirmatory studies to indicate if potential problems may exist. CARES is a simplified computer code developed for the NRC to treat in a simplified fashion the SSI pmblem of interest for plant response analyses. 'Ihe approach c.yims the primary effects of SSI but cannot be used for design evaluations except for simplified site and stmetural configurations. If the calculated responses fmm CARES for either issue indicated potential problems in NI responses, the more exact S ASSI studies wculd have to be undertaken by the Applicant to further evaluate these issues. 'Ihe following paragraphs attempt to summarize these additional studies. C.2 l
. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ _]
C2. INFLUENCE OF DEPTH TO BEDROCK CARES is designed to operate in a relatively quick and cheap manner to detemrine both free-field site and structural responses. In order to do this, however, it makes use of approdmate interaction coefficients which are available fmm the open literature for simplified configurations of both the structure (circular or rectangular plan areas) and the site (uniform site conditions to bedrock). For the problem of interest herein, the plan area of the NI (see Figure A1) is not rectangular although it was felt that this deviation was not significant and the equivalent rectangle shown in Figure AI was used in the CARES computations. However, the ima~ia- function data W into CARES and available in the literature are appropriate for the case of relatively deep depths to bedrock. For the shallow rock depths of concem for the AP600 problem of interest herein, the data on imp ~iae functions had to be extended to make it appropriate for the shallow depths ofinterest. , Approximate imaai== functions were developed by performing the following finite element studies. First, static FE analyses were performed with the ANSYS Computer Code as indicated in Figure C1. A number of meshes were developed varying the depth, d, the width of the mesh, L, and the FE size, keeping the parameters (L, W) of the embedded plant constant. From each computer run, the sidewall and basemat stiffness functions were determined. De impact of these parameters on the computed stiffnesses was determined to obtain a reasonable assessment of adequacy in the computation as indica'~i in Figures C2 and C3. From these results, the variation of stiffness with depth below the basemat was estimated as indica'~i in Figure C4, with the estimated parameters obtained from the computer output. De corresponding values of damping impedances was determined as shown in Figure C4, using the available results for the deeper depths to bedrock. De conversion of this data to treat the two layer (soft over stiff site) conditions ofinterest was made using the superposition argument shown schematically in Figure C5. Dese approximations :o the impedance functions were then treated as frequency independent par-T~~.s
. and input to the CARES Code for use in the NI response analyses.
l Prior to mnning the variations on bedrock depths, a run was made with the AP600 model for the uniform rock problem to ensure that CARES output pmvides a reasonable estimate of NI responses. Figures C6 through C9 show some comparisons with the more accumte 2D SASSI analysis for this problem. Some differences can be noted, particularly in Figure C8 for the Shield Building Roof node, where the SASSI spectral peak exceeds the CARES result, indicating some differences in impedances controlling rocking response of the NI. However, the results are considered reasonable for assessment of general trends. Figure C9 indicates a comparison with the C.3
i fixed base BSAP results obtained by the Applicant. Using these approximations, the CARES runs were made varying the depth to bedmck below the basemat from @ to 100'. De available data fmm the 2-D SASSI computations L indicated significant differences in response between these depth ranges. The results at two nodes ! of the AP600 stick are shown in Figures C10 and C11 which indicane the general trend in the results. Figure C10 presents spectral output at the Shield Building Roof, Node 3016, for the l various depths to bedrock while Figure C11 presents similar output at the Steel Containment Vessel, Node 3110. It can be noted that as the depth to bedrock decreases, a fundamental shift in character of the response changes, with peaks of the response spectra shifting from the higher mode of the stick response to a lower mode. This shift is caused by the change in effective damping of the iWaam function as the depth to rock decreases. It should be noted that this result is in keeping with the 2D SASSI results obtained for the shallow rock case. These same CARES results have been replotted in Figures C12 and C13 in the form of amplification ratios of the spectra at a particular depth to that at the deeper depth. In both cases, the results indicate that when the depth to bedmck is within 20' of the NI basemat, the lower frequency peak of the response increases significantly and may exceed the design allowables. C3. OUT-OF-PLANE STRUCTURAL RESPONSES The variation in NI response due to a variation of direction of the input horizontal snotions was considered by again using the simplified CARES model, but in this case using a 3-D model of the NI. De purpose of these results was to investigate the potential for significant out-of-plane response at critical locations of the NI,in which case further evaluation of direction of inputs may be ofinterest. Several runs were made using both the H1 (X-direction) and H2 (Y-direction) input motions and determining the nodal responses. A typical result is shown in Figure C14, in which it can be noted that only minor or negligible out-of-plane response is generated. His is a typical result for all the nodes evaluated. It should be noted that similar results were obtained from the one 3-D SASSI model performed by the Staff Review Team. C4. CONCLUSIONS On the basis of these simplified results using the CARES Computer Code, two general conclusions were reached. First,in developing the range of site conditions considered appropriate l for siting the AP600 reactor system, the depth to bedrock is a significant pariurder in defining plant response and must be properly evaluated by the Applicant. De current site descriptions presented in the SSAR are primarily concerned with ranges of shear wave velocity to define C.4 l
- - ____ ______-_ ._ - - _ _ _ _ _ _ _ - _-. . _ _ _ _ _ _ _ - l
} acceptable site conditions. However, the depth of the bedrock below the basemat of the NI may be important, particularly for the softer site soil conditions. It is rex- . + ded that the Applicant specifically consider this parameter and evaluate its po:ential effect on the design. Secondly, it was I determined that out-of-plane responses am generally not a significant concern when evaluating plant responses. C5. REFERENCES 1 C1. Costantino, C. J., C. A. Miller, E. Heymsfield and A. Yang," CARES: Computer Analysis j for Rapid Evaluation of Stmetures" Version 1.2, Draft Report, Civil Engineering Dept., City College of New York, for U.S. Nuclear Regulatory Commission, September 1995 l I i r C.5
l
. .W l' 'l a
Side Soil S1 D 1 r J b Soil S2 d g r, pace ir (a) General Site Configuration of Interest 1 W/2 L i 1 A D 1 r d L f Z < w l l d u$ l l IP i Fixed Bottom Nodes E i (b) ANSYS Finite Element Generic Model l l l 1 i FIGURE C1 CONFIGURATION FOR CARES CONFIRMATORY ANALYSES t.
l l P i 68.32* 60* t............y d ' .... ....12............ 0' ,L ..... .y e.s a 80*
'l I 840*
o it, o ...., ., . . HALF SPACE <
*a~~~*
, HALF SPACE MORE MESH ,
..~,".,
WHOLE SPACE 20 - .',** - l - MUS <
,.~ -
,/ <
40 - f < c - ' .
.so -
s~ - 4 4 g . i ..i.. s . . . . . . . 0 30 100 150 200 248.32 l l DISTANCE FROMTHE CENER LNE(R) i HORIZONTAL NODAL PORCE AI.ONG THE BASE ("+"is the right direction) kvec cs 1 i l I
- t. . .
120* I 60' l j,._.. 68.32'. ..._. _ y._. _. . _. _ ._... . .. . . .,y. ._ _ _ . . _ _ S
. =
e e n. a .
. s I .e
- g i
- n. i. 80' i.
- t. l.
40- .I I t E l 0 I *
. i - . -
i \ . 30
....~...,**. -
\.t l
.x -
\* / \-
\ / . t .
.[ j 30 -
[ l i g 40 - HALF SPAG -
- - - ~ ~ ~ -
HALP SPAG,MORE MESH 50 -- ---- WHOG SPACE DENSER MESH . 4 i .e e i . ..- 0 50 100 150 200 248.32 DISTANG FROM THE GNTER LINE(n) VERTICAL NODAL FORCE ALONG 'IIIE BASE ("+"is the up direction) tGoR.E C$ 1 l 1 l 1 I i
l Stiffness K JL Static ANSYS Results 1 l l Kbme K ocat = K siewat + Khme 7 d* Depth d l Damping C dL
.Cbec Parabolic Variation C totg =[C stewgi + Cbme]
- Cbme 1-h d.
d** Depth d FIGURE C;Z VARIATION OFIMPEDANCE FUNC.TIONS WITH DEFFH 4 TO BOITOM BOUNDARY
)
l I-l Side Soil S1 y Soil Halfspaa S2 l II . Side Soil S1 ' l Soil Halfspace S1 1 L
+
L f l Vj + .
- .. Soil Halfspace S2
/ ,A l Soil Halfspace S1 l
. i h
FIGURE Cf SUPERPOSITION APPROACH TO OBTAIN IMPEDANCE FUNCTIONS FOR CARES CONFIRMATORY ANALYSES
l l l l I I ! e s ; I 2 3116 SASSI g - MN . l
.I \p 3 < ,I i .2 %
I g 4 - e \ \ < > q 4 ' 1 I b I i l 2 - t I l o- ' . 4.1 1 to tot Frequesey (Hz) AP600 RESPONSE SPECTRA (5% DAMPING) ROCK SITE WITH Ves11,000 fps (CASE 8) POLAR CRANE NODE 3116 (X=0.0, el 205.3')
i I l l
.. i 1
l l , 4 '
, I l
( I l l i ! 3 - i ~ , .E C 3110 SAS$1 > 0 l l N
= soc a ns -
g ; l 3 ip , ,
,E l l 2
8 ,i - O i a i M l to 1 , i 01 1 i
*I I le 100
. Freguesey (Hr)
AP600 RESPONSE SPECTRA (5% DAMPlNG) ROCK SITE WITH Vs=11,000 fps (CASE 8) STEEL CONTAINMENT VESSEL NODE 3110 (X=0.0, el=205.3') N Ceu t E C-7 1 l
10 ,
; 3016 SASSI
< ,i , 16 CARES 7.5 -
.E (I 3 '
th
.I i n 1 <
1o
.k g 5 -
Y, - , . 1 1< , l .g '<> l g;,P i j= ' {' 4
'i ,
2.5 - ' i i - 5
, e i
s.1 1 se les Frequacy (Hz) l AP600 RESPONSE SPECTRA (5% DAMPlNG) ROCK SITE WITN Ve=11,000 fps (CASE 8) SHIELD 8UfLDING ROOF NODE 3016 (X= 0.0, el 307.3') kCrobi, C8
9 4 9 4 I. t 12.5 y i a
- 3016 FB4 SAP f
I s 01 0 3016 FB7 SAP d i " 10 ~
, ......... 16 CARES
?= I# t& 7.5 - I ~ u,' h ' k E
- !i u : 1,:
e 5 = 0 4.1 1 38 3" Frequency (eps) AP600 RESPONSE SPECTRA (5% DAMPlNG) ROCK SITE WITH Vs=11,000 fps (CASE B) SHIELO SUILDING ROOF NODE 3016 (X=0.0, el=307.3') bGc0ft,E G3
i i i l l 15 i . I l Dept 10
- Dep& 20
? Dep& 30 5 30 . .
.lw t
- l I
!! ---m--- Deps s0 ll t 8 e . . . . . . . oops ,0 l
4 Deps 100
){i E
I 3 ------- ! l' -
;r 8!:,
s i >
\- g
'~~~.
f l 01 ' '
~
l O.I I 10 100 l - Frequency (eps) l AP600 RESPONSE SPECTRA ($% DAMPING) . SHALLOW SOIL SITE (h=45') WITH Vs=1,000 fps . l H2 MOTION @ SOIL SURFACE (CASE 11) l , SHIELD BUILDING ROOP NODE 3016 (X=0.0, el=307.3') Chus CAO
r i l s . . Depth 10 ! :  A
;d, ;
6 - Depsh30 - g8
.=
8 l l e
- Depth 30 :l 8I l j ---m--- Depth 80~
.g
} ---.+ -... Depth 90 M*
g 4 g . w 4 .
*E
....... Depik 100 (%
2 'l E.
=
! 2 - \. -
\
01 ' ' e.i a io. Frequency (eps) AP600 RESPONSE SPECTRA (5% DAMPING) SHALLOW SOfL SITE (h=45') WITH Vs=I,000 fps - H2 MOTION @ SOIL SURFACE (CASE II) STEEL CONTAINMENT YESSEL NODE 3110 (X=0.0, el=205.3') EW Ek ( . )
4 h 1 l 9 --
- Depth 10 / Depth 100 j 8 * ' '
- DepdG0 / Depskl00 l 7 --- * -- Depth 50 / Depch100 , ,
^
Depth 80 / Depth 100 88 s p . 1 1 1 3 - lj 8 4 4 - b I j2 - 2 - l , . y l I si 4 p he dh t dllREl me s e s q y ,#.
..I i i. ... l Frequesey (eps)
, )
i AP600 RESPONSE SPECTRAL ACCELERATION RATIOS (5% DAMPING) l SHALLOW SOIL SITE (h=45') WITH ' Vs=1.000 fps j H2 MOTION @ SOIL SURFACE (CASE II) J SHIELD BUILDING ROOF NODE 3016 (X=0.6, el=307J') bbEG CCE 1
l 7 l I I I Illli I I ! C Depskl0 / Depthl00 l
~
' O Deptb30 / Depskl00 '
l Y
. . . . .... Depth 50 / Depeh100 \
.t 5 .
^
. j Depsh00 / Depth 100 ha .
? .
l 4 3 - i I e 4 ' Ul k
- 2 -
I i r 1 gg p u u l l l l -
' qqg o Qibu s e.1 1 10 IO0 Frequency (eps) ,
AMet SPECTRAL ACCELERATION RATIOS (5% DAMPING) SHALLOW SOIL SITE (ks45') WITH Vs=1,000 fps H2 MOTION 9 SOIL SURFACE (CASE II) STF.EL CONTAINMENT VESSEL NODE 3110 (X=0.8, el=2055) W O k,t"~ CS
l 1 i I l - II 5 I I g ._ I 1 b 0.8
~
i
? \
8 ; c
.9 e 0.6 -
5
" l z !
E ' s i 3: 0.4 - w " ! I. mu a smuun Respot.se in s. direction i 0.2 - Response in y direedon due to H2 in yestion 0 ' 1 le 100 Frequency (eps) ! AP600 RESPONSE SPECTRA ($% DAMPING) DEEP SOIL SITE (b=I20') WITH , Vs=1000 fps.2400 fps @ 240' VARYING PARABOLICALLY WITH DEPTH (CASE D) SHIELD BUILDING ROOF NODE 16 (3016) (x=9.0, pr0.0, eI=307.3') bQ;0N bN
Appendix D Evaluation of Response Spectra at Basemat Level I 1
)
{ r ( l D.1 l w-____-____-_____-_-__ _ - _ _ _ _ - _ - - _
i l l l l DI. INTRODUCTION L Section 3.7.2 of the SRP (Ref. D1) indicates that when considering variations in amplitude I and frequency content at depth for partially emhaMad structures, the free-field motion at ! l : l fanadeian depth shall not be less than 60% than the conenpaading amplitude of the design motion ! L at the finished grade. The 60% limitation may be satisfied by considering the envelope of the thru foundation level spectra in the free-field corresponding to the three soil cases considered. if the analysis does not include consideration of the rotational E-rt:==ts of the motion at depth, no i reduction is permined. In previous years, when simplified SSI methods of analysis were used (constant SSI spring / dashpot models with simple structural sticks), no reduction in input ground motions was allowed in SSI analyses. Concerns with abilities to predict incoming unmir wave fields, j completeness of the methods of SSI analysis and uncertainties in specifying site properties all played 'a pan in this position. As methods of SSI analysis improved, the concern with the determination of allowable levels of reductions of the free-field motions at depth have evolved since most recorded data at depth did indicate some amount of reduction as compred to the measured surface spectrum. However, uncertainties in the completeness of the mtJuxis of SSI analyses as well as variability in site conditions still remained. Current analysis -thevis are now j considered complete enough such that the concem with rotational input is no longer an issue, i leaving the definition ofincoming wave fields as well as treatment of variability of soil p@cs I in quantifying free-field motions at depth as the prime driver in maintaining the 60% requirement. Some indicate that the 60% requirement is no longer required since potential variability in ground motion at depth is fully captured by the variability in soil column properties (1/2 to 2 times best estimate p@es) considered in the usual SSI analyses. The current revised version of ASCE 4-86 (Ref. D2) has apparently dropped this concem in the Cu --mry section of this guideline. *Ihe SSI procedure reammand~iin ASCE 4-86 indicates that at least 3 analyses be performed, using l lower bound, best. estimate and upper bound soil puysiles together with the best estimate
- structural model.
Recent experience with other evaluations of critical facilities, however, indicate that this approach may not be acceptable for those cases where SSI effects are impun nt; that is, where critical responses are strongly influenced by the SSI effects. By uniformly changing the shear D.2
i modulus of the soil column by a constant factor,it has been found that both the foundation level fme-field spectrum as well as the SSI frequencies may all be pivycidonal to the shear moduli.
*Iherefore, if in a lower bound calculation, for example, it is found that the structural response frequency falls "in the valley" of the free-field spectrum, the msponse may always be in that valley 3 I
as the soil column is uniformly stiffened since all parameters contmiling the response are rwGonal to the t, hear modulus. 'Ihis is illustrated schematically in Figures Dl, D2 and D3 in which both the frequency of the free-field and the SSI mode of the rigid structure are found to be l rvyvnional to the site shear wave velocity. Of course,ifit is found that other p--we of the lower bound, best estimate and upper bound analyses control the relative locations of the " valleys" of the in-structure spectra, then the concern with this issue may be lessened. In the AP600 review, this concern has been raised since each assumed site condition has been performed for best estimate values only. 'Ihe counter argument presented by the Applicant is that the free-field spectra at a depth of 40' (bottom of the basemat) from all the cases studied cxceed the 60% requirement. However, whether they have met I the intent of the rule is not clear until the issue of relative spread of the SSI frequencies is also i 1 addressed. It is not clear how a simple procedure can be developed that will work to overcome this j concem. A procedure that may work is to first envelope the free-field motion at depth and use this j l broad-banded motion as the input into a single SSI analysis of the nuclearisland (NI). An ahemate ; would be to use random selection of soil properties with depth such that the shear moduli are not uniformly modified. However, this would lead to an inordinate amount of SSI SASSI l computanons, i D2. AMPLIFIED INPUT MOTIONS In this section we present the implication of an interpretation of the 60% rule on computed NI responses. In this approach, the foundation level motions are maintained at or greater than the 60% surface spectrcl values for all variations in site properties, and this single enhanced motion is used as input to a single SSI response analysis. For the one problem evaluated, the amplified l ground motion was developed to satisfy the requirement of no greater than a 60% reduction in ground motion at the foundation level as described in the Standard Review Plan (SRP). An additional 2D SASSI c+nMon was performed fer the soft to medium stiff soil rite (Case A of Appendix A) but using an enhanced H2 free-field ground motion defined at the ground surface. In this evaluation, a typical soil variation in the free-field deconvolution calculedon was performed, j varying the shear moduli from 1/2 to 2 tirres their best estimate shear wave velocities defined by the Case A configuration. D.3
i 1 ne original free-field seismic input motion used is the H2 time history generated by the Applicant and described in Appendix A.De free-field spectrum at the foundation level depth of 40 feet was then computed for each site condition assumed (best estimate, lower bound and upper bound) and the results at the 40' dcyth are shown in Figure D4. It can be noted that the spectra at depth fall below the 60% criteria spectrum over a frequency range of from about 1.5 hz to about 10 hz. In this problem, the found don level spectrum in this frequency range was then amplified to match the 60% target spectrum as shown in Figure D5, a new time history generated (Figure D6) which was then convolved to the ground surface using the best estimate soil column properties to arrive at an enhanced surface H2 motion, as noted in Figure D7. 'Ihis ground motion can then be used as the input groundynotion to the 2D SASSI model and presumably would satisfy the intent of the SRP guidance. ' Die resulting response spectra are shown in Figures D8 through D11, and show exMances over the corresponding fixed base spectra which have typically controlled the design. D3. CONCLUSIONS It is not clear that the process followed by the Applicant of using the resulting spectra at the 40'c'epth from each site response analysis to show that the design satisfies the intent of the 60% rule at foundation depth. It is recommended that they consider this problem further to indicate that in fact the relationship between the " valleys" of the free-field spectra at depth do not all bear the same relationship to the SSI frequencies in each of the 2-D problems investigated. i D4. REFERENCES D1. " Standard Review Plan", NUREG-0800, Office of Nuclear Reactor Regulation, U.S. Nuclear Regulatory Commission, Revision 2, August,1989 - D2. American Society of Civil Engineers Standard 4-86, " Seismic Analysis of Safety Related l Nuclear Structures and Cnmentary on Standard for Seismic Analysis of Safety Related Nuclear Structures",1986 D.4
i 11llI jI \j I I p p g F m F u tr l e
- 3 S pve l3 V V y S S 1x s E m
u r l I S S tclh e uhe r ct ud r S c o K c o " S Y L A t I c e t r Su o n C S S S p n F g N el s
;wv e I
A E
..le S S
S
-e n N n nl
, O I
t S P L _ J
, p S
E 3 I I C R I S J S L L A N K R E N E G v1E
= R
< I U G
nmu t r I F c am e du se r i c t eS l' t p I F V 1 Ope d pu F 1 4 eS le i i n ct F a u f p le t un Si wv Fl e %0 6 n m u
' MlC o S
S
, miS Fo l
N L J 4
, I11ll
i 1 1 r l l l l
~
aL S a _ Broad-Banded Surface Criteriainput V __
-_\
vv l g 60% Surface ' I Cri:eria FF E FF
>f f
BE pp J f f U FF E FF
= 0.71 = 1.41 g g E W E n l
Interpretation: Envelope of Input Spectra at Foundation Level Is Greater Than 60% of Surface Criteria Spectrum 9 FIGURE 2 FREE-FIELD SPECTRA AT FOUNDATION LEVEL l I l l l l L-_--_---------- )
l l A a SSI Amplification FF Deamplification l g Surface F Criteria 60% Surface Criteria f E SSI yf f W W t l f f l W SSI UB SSI
- 0.71 = 1.41 ,
f ! f E SSI E SSI j l
- Envelope (Widened
- 15%) ofInstmeture Spectra i 1
i Defined as Design Spectrum . l
- Question of Similar Approach to Determine Design l Structural I;xwis and Element Forces i
! 9 l FIGURE 3 INSTRUCTURESPECTRA AT FOUNDATIONIIVEL L
O 2.40 i i 2.20 - G BEST ESh'fE -
=
~
o nnwensom f1x 14 l 1.40 .-- -- o .asoi 0 _ l
< 1.20 - -
l 1.00 .,-, l A / J
< e.
A-W 0.80 -
/ \( .
'a-m,w , .
o.60 _ N . .
** $ , ,e t 'ed.,t, 0.40 - ',,.. , ;
\/<* ...',y ' _
,_20 .
eseos em
'?,d' .. .f ,
0.00 I 10 100 NQUENCY(cys) AP600: CASE A - RESPONSE SPECTRA AT40 MM H2 MOTION AT SURFACE i
= m G F3 ,#{
l
l l
)
1.2 . m mm 5% DAMPING
~ ~
i REVISED H2 MODON 1 o < b
$ o.: - -
A a: 4 7 0.6 - - Q
< 4 a r. <
l o.4 - - IVii* f . '.4. iv - o.2 - -
^ ' ' ' ' ' ' ' ' ' ' '
0.0 ' O.1 1 to too FREQUSCY(9s) AP600: CASE A RESPONSE SPECTRA @40*FOR . DIE REVISED H2 MOTION & MINIMUM RESPONSE SPECTRA ENVELOPEFUNCTION
~
F 6 7 1
I ! 1.2 . l . ! GENERATED 5% DAMPING 1.0 .......... TARGET ! 5 - l @ 0.s' - ! A ' O.6 I l W e
~
- +
e o.2 - 1
, e **
I
^^ ^' ' ' ^
0.0 O.1 1 to 100 PREQUINCY(cps) AP600: CASE A TARGET AND GENERA *IED SPECTRA @40'OF
... THE REVISED H2 MODON i
l emmann & e I~q i 6
l l , ORIGINAL H2 @ SURIACE l ENVELOPE AT40'
- --- - I REVISED SURFACE 140 .
., ., . l
, a < ! 12 - A
. il8 5% DAMPING $
2.00 -
!s ,
isI b g 1.80 - W ;! - I'" - h w I ,I i; - 1.40 . -
./
6 1.20 -
! \ -
,.# i
. L .
1.00 - 8 ii j - ! i ,
. i. 1
,/I%
. l 0.80 - t
,e 5 l . ,
o t
% ~~
Q ' O.20 - o,og .....e . . . . . . . . . . . 0.1 1 10 100 FREQUENCY (cps) AP600: REVISED CASE A e s.$7 0
i I l 1 l
.h 12.5 .
..>.~
Revised H2 Motles 10 . n
. Fixed Base 45 Structural y Damplag h 7.5 -
.......... Ftsed Base
~
E 75 Structural f . ' f+. d Daapuis p. u . u :0
< : 3 5 - : 8 -
a : t I sc b k..
- M, .* .
A 2.5 -
' :Y.
i
\
0 O.1 1 10 100 FREQUENCY (cys) .,,., AP600 RESPONSE SPECTRA (5% DAMPING) COMPARISON BETWEEN H2 MOTION AND REVISED H2 MOTION FOR CASE A SHIELD BUILDING ROOF ,, ,ht. NODE. 3016 (r=0.6, et.=307.3') .- .:
,. .?.t .
t
.../. .. . .
~ egg
- - ' 17 ,.
.. ~ ThCG&~
C.. W
.! -; VetX
I I j l l 1 i i i W W W W W WW W W W W V W WW W W W W W W WW Revised H2 Metles e b 3 - Fized Base - 4% Structural k . $ Damplag l E I
.......... Ftsed Base !
)
75 Structural i 2 - Damplag . < 1 l .J - 4 1
~
\ ~
m \ l
'~.... ;
^ ^^^' ' " ^ ^ ^ ' ' ^
O i a.1 , 1 10 100 FREQUENCY (cys) AP600 RESPONSE SPECTRA (5% DAMPING) COMPARISON BETWEEN H2 MOTION AND REVISED H2 MOT 10N FOR CASE A STEEL CONTAINMENT VESSEL NODE 3110 (ame.4, et.=20S.3*)
~
e vw 3
f l l l 15 . . Revised R2 Motles T l Fixed Base 45 Structural~ 10 ! p Damplag l < at g ........~ Fixed Base os 75 Structural O , Damplag
< : 1 m
a 5 - l1 - U I o 1 as 1 5 \ 0 O.1 1 10 100 FREQUENCY (eps) AP600 RESPONSE SPECTRA (5% DAMPING) COMPARISON BETWEEN H2 MOTION AND REVISED H2 MOTION FOR CASE A POLAR CRANE NODE 3116 (z=0.6, et.=245.3') l
- m 1
\
l I N 'l
%+3 1
1 1 ,
I 1-4 T w W w y y w y W W W M W WW W W W W W w w w Revised R2 Metles ^ 5 -
.S W Fhed Base g 4% Structural C 4 -
Damptog - E
* .......... Fhed Base 75 Structural e g3
_ D.. ,i., r J I
< l E
l 2 - e 1 . - O
^ ' ' ' ' ' ' ' " " ' ' "
0.1 1 10 100 , FREQUENCY (eps) AP600 RESPONSE SPECTRA (5% DAMPlNG) COMPARISON BETWEEN H2 MOTION AND REVISED H2 MOTION FOR CASE A CONTAINMENT INTERNAL STUCTURE NODE 3204 (x=14.5, el.=135.3')
- =
l 90 Eg T-
l l Appendix E 3-D SASSI Analysis i
)
E.1
{ l El. INTRODUCTION Following the completion of the 2-D SASSI computations performed by the Staff Review Team, discussions were held with the Applicant concerning the .yymph set of 3-D SASSI analyses that were required to be performed to provide +E'=e in the adequacy of the design responses for the nuclearisland (NI). On the basis of the Staff's parametric 2-D calculations, and ] discussions among the Staff Review Team and with the Applicant, it was duwd that one j additional confirmatory 3-D SASSI computation should be performed, using an assumed site
]
condition having a uniform shear wave velocity of 3,500 fps. 'Ihis site category can be cw=L44 j a soft rock case and would fall between the soft to medium soil and hard rock cases proposed by the Applicant. E2. RESULTS OF 3-D SASSI ANALYSIS Using the new 3-D structural model provided by the Applicant, which contained some l minor variations from the models previously used in the studies, together with the soft rock site I condition, three separate accesses were made to the 3-D SASSI model, one for each duection of motion using the H1, H2 and V criteria motions defined at the ground surface. From these runs, response spectra were generated at several key locations mutually agreed upon with the Applicant, these being Nodes 3004 (Control Room @ elevation 117.5'), 3016 (Shield Building Roof @ elevation 307'), 3110 (Polar Crane Support @ elevation 205') and 3115 (Top of Containment Vessel @ elevation 256'). Both 2% and 5% response spectra were generated at these locations and cuiwM with the design spectra proposed by the Applicant in the SSAR which were described as being suitably amplified and widened. Comparison of some typical results are shown in Figures El through E3. It cen be noted that the computed spectra generated from the output for the soft rock case exceed the design spectra defined by the Applicant. Therefore, it is concluded that the design did not satisfy the requirements specified in the SSAR. Recently, the Applicant has performed additional i computations and developed modified design spectra. A comparison of the new spectra for the same nodes with the output fmm the soft rock case are shown in Figures E4 through E6 and indicate that the new design spectra suitably envelope the computed responses. E.2
E3. CONCLUSIONS On the basis of these results, the modified design response spectra appear appropriate for design of intemal equipment and components. The issue of design of the exterior walls and. basemat of the nuclear island remains to be evaluated. l l E.3
l l l 1 ZDR$C110N(VER11 CAL) NODE 3004 Y DIRECI1ON(EW) ) X DIRECTION (NS) i 1 l 2.0 . . DESIGN ENVELOPE t.s - 5% DAMP!NG SSAR,Rev.2, Fig.3.7.2 .
.... ...... Y-DR .
g,, _ 3D SASSI(H1+H2+V) _ UPPER BOUND ROCK CASE
> d
@ i Z { .
1.2 - . - , I* f.
- i
*g . !
1A = 9c-
. ' ' . =1 I *Efh -
! .1. . . . . _ _ _ . . . . . 1 1 : ! f,,1 . i
< :l .N<
3 0A -
/ !
1F d
/
) i
\ .
0.6 - i f/ 4
% t.
v> s 'i l .
. . 1. .
0.4 - l4 l %. .
, l *.....
*p.*
0.2 - - '
' A ^ ^ ^ ^' A a a a a a a g a m , ,
i 10 100 FREQUENCY (Hz) RESPONSE SPECTRA @ NODE 3004 AP6003D MOTION ALGEBRAICSUM tC.\
i - l ZDIRECHON(VERTICAL) i NODE 3016 Y DIRECTION (EW) X DIRECHON(NS) l 10.0 . I DESIGN ENVELOPE ' l 5% DAMPING SSAR Rev.2, Fig.3.7.2 9.0 - ' g,9y .
- I 3D SASSI (H1+H2+V) ,
l s.O - UPPER BOUND ROCK CASE. y
? '
8 1.0 - l Z ! O - P 6.0 - $ w,_ 1 l
,k 5.0 -
O I < kx 4.0 - ; g 3.0 - m 2.0 = i' !.0
- w .
~
l 0.0
^ '-- -' - - ' -
- l. --
1 10 200 FREQUENCY (Hz) RESPONSE SPECTRA @ NODE 3016 AP6003D MODON AISEBRAICSUM
(. i l l l l i ZDIRECDON(VERTICAL) ! l
, NODE 3110 Y DIRECDON(EW) I X DIRECDON (NS) ;
54 . DESIGN ENVELDPE ' i SSAR.Rev.2, Fig.3.7.2 l l 5% DAMP!NO ' i
- X DIR i
~!
48 - 3D SASSI(H1+H2+V) l l UPPER BOUND ROCK CASE- i i l '7 P l b "] z O \ 3.0 -
. i i-l l
L < 2.0 = l = l ) - if
< l m
th ~
~
i l CD ^ ^ ^ * ^ ^ -* -- l 1 10 100 FREQUENCY (Hz) RESPONSE SPECTRA @ NODE 3110 AP6003D MODON ALGEBRAICSUM L- _ - . - - _ - - -
1 l i l, 1 i 1 l l 2.0 - l WS 5% DAMPING , l
- spectra SRSS i
l
- 1 0 Spectra 100MOMO ! .g 1.5 ., l w
- g ---*--- Sum of H1, H2, V Time Histories '
O ' P ac l su ~
\ '
2 4 uJ 1 .0 - t - h J i i I f y 2 E -
= o.5 -
n 4 0.0 ' ' l 0.1 1 10 100 l FREQUENCY (Mz) RESPONSE SPECTRA @ NODE 3004, X-DIRECTION I 1 I i l l l td
i j l l I I 1 I i 1 1 o %- s l Io.0 ..a .,. 1 WS 5% DAMPING $
---a--- i' Specus SRSS .
, l 8.0 - - J m
0 Speers 100#0#0 -
- - * - -- San of Hl H2, V ~
Z Time Hinones O - b ,o g l ' w ' Il
.a .
6 u
\ ,
i i .
,,J 4.0 -
a>
< ii1, 2 . A 4
m
- c. <
l M ' l' - i - ! 0.0 - ' .# , , 0.1 3 10 300 , FREQUENCY (Hz) r RESPONSE SPECTRA @ NODE 3016, X-DIRECTION
/
b
10.0 s 5% DAMPING 3.0 - WS - n 3 . --.s.-- Spectra SRSS Z - 9 la 0 Spectra 100#0/40
< 6.0 -
oc - * - - - - gn Sum of H1, H2, V Time Histories - Gl .
) 4.0 -
4 - llc. l i til l l c ' M > l 2.0 - k 0.0 8 ' ' ' ' ' ' ' ' ' O.1
- 1. 10 100, FREQUENCY (Hz)
RESPONSE SPECIRA @ NODE 3110, X-DIRECHON t i l l}}