ML20106D653
| ML20106D653 | |
| Person / Time | |
|---|---|
| Site: | Sequoyah  |
| Issue date: | 10/08/1992 |
| From: | Burzynski M TENNESSEE VALLEY AUTHORITY |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| NUDOCS 9210130163 | |
| Download: ML20106D653 (15) | |
Text
.
E
, ee va, A w. g, 3 m va w v.o : v r, m m we un OCT 0 81992 TVA-SON-TS-92-01 10 CFR 50.90 U.S. Iluclear aegulatory Commission ATTH:
Document Control Desk Washington, DC 20555 Gentlemen:
In the Matter of
)
Dockot Hon. 50-327 Tennessee Valley As.thority
)
50-328 SEQUOYAN NUCLEAR PLANT (SQN) - RESPONSE TO QUESTIONS ON REQUEST FOR LICENSE AMENDMENT TO TECHNICAL SPECIFICATION (TS) - SPENT TUEL POOL STORAGE CAPACITY INCREASE On March 27, 1992, TVA requested a license amendment to the S011 technical specifications to support increased spent fuel storage capacity.
On September 1, 1992, wo received questions from NRC concerning the structural intogrity analysis of the proposed spent fuel storage racks.
The enclosed pages provide TVA's response to those questions.
Calculations referred to in this and previous submittals related to this amendment request were performed and issued by TVA's contract or, Holtec International.
The oppropriato TVA technical organizations have reviewed and concurred with the calculations. These calculations will be appropriately incorporated lato the TVA calculation system prior to actual fuel rack installation.
Please direct questions concerning this issue to C. R. Davis at (615) 751-7509.
Sincerely, m
1 u.
7 Mark J. Durzynski Manager Nuclear Licensing and Regulatory Aff-irs l
hh{ O 1 -)
Enclosures cc:
See page 2
/
9210130163 921008 PDR ADOCK 05000327
((/
P PDR
/
\\
_..m...-_._..._.__
t r
U.S. Nuclear Regi.atory Commission Page'2 OCT 0 81992 cc (Enclosures):
Mr. D. E. LaBarge, Project Manager U.S. Nuclear Regulatory Commission
[
One White Flint North 11555 Rockville Pike Rockville, Maryland 20852 l
Mr. Michael II. Mobley, Director (w/o Enclosures) l Division of Radiological Health i
T.E.R.R.A.
Building 150 9th Avenur, H Hashv1110, Tennessee 37203 URC Resident-Inspector i
Sequoyah Nuclear Plant 2600 Igou Ferry Road i
Soddy Daisy, Tennessee 37379 Mr. B. A. Wilson, Project Chief U.S. Nuclear Regulatory Commission Region II 101 Marietta Street, NW, Suite 2900 Atlanta, Georgia 30323 5
L t.
i=
l-
..a...
P ENCLOSURE 1.
Page 6-21 Provide a technical basis for the expression "Pa" for the compressive stress given in page 6-21 by means of derivation or by reference to an established code or both.
Please note that the rack wall has a possibility of side sway laterally in a direction normal to the wall, i.e. the top support of the column may move laterally away from the original position with respect to the bottom support of the column when a compressive load is applied to a box type structure such as a rack. Demonstrate that the expression given in pago 6-21 considers this possibility.
HRSPONSE l
T'e lord due to the weight of the fuel assemblies bears directly on n
I the baseplate of the fuel rack. Therefore, the only structural members subject to significant compression loadings are the support pedestals. The cellular portion of the rack experiences insignificant compressive loedings.
The term Pa includes the factor which accounts for the reduction in strength due to the slenderness effect of the structural member.
Since the pedestals have a very low slenderness ratio, there is practically no reduction in the allowable compressive s;rength in contrast to the tensile strength.
The expression for Pa owes its origin to civil / structural engineering literature and first appeared in the struciural engineering Code (Manual of Steel Construction, American Institute of Steel Construction NY, NY). The ASME Code had this formula in Appendix XVII of Section III until the 1983 Rdition and subsequently l
in Subsection NP (NP3322.2).
l Because of the relatively small axial compressive stress in the rack l
cellular region, there is a large margin of safety e~sinat buckling in that region.
This can be confirmed-by perusing the maximum stress factor (above the baseplate) provided in Tables 6.7.3 through 6.7.20 for various loading scenarios.
l l
f
2.
Page 6-23 Stress factors are discussed in page 6-23 Provide the most highly stressed examples of R1 and R6.
Identify what part of the rack these stresses correspond to and discuss the significance of the compressive stresses by providing the percentage of the compressive stress contribution to the R6.
RESPONSR Stress factors R1 and R6 have limiting values for the 12x14 spent fuel rack. The limiting values come from Table 6.7.30 (for a case where adjacent racks are assumed to move out-of-phase).
R1 R6 Gross Cross-Section
.042
.301 Just above baseplate on a section through the entire cellular region.
.287
.484 Just above baseplate on a section through one pedestal.
Por a case where the adjacent racks are assumed to move in-phase with the 12x14 rack (Table 6.7.21) the corresponding values are:
R1 R6
.039
.333
.282
.453 For the gross cross-section just above the baseplate (i.e., a cut-through the cellular region) the highest combined' stress will be at a corner cell. Only.042/.301 =.1395 (145) will be due to direct compression acting on the gross cross-section of 12x14 cells. In reality, the actual primary stress acting on the corner cell just above the baseplate will all be in compression since it is at "the extreme" fiber of the cross-section. The actual value of this stress on the outermost corner cell will be R6 x (.6S ) where.6S y
y approximates the allowable stress. Thus,.the maximum compressive primary stress.at the base of the outermost corner cell in the cell metal is.333 x.6Sy = 4995 pai.
We note that on the gross cross-section of the cellular region of the rack (just above the baseplate) direct compression plays only a small role.
Column buckling of the cellular structure as a beam is not a governing condition because there is only a small component of direct compression imposed during a seismic event (i.e., the heavy vertical' fuel load is imposed directly on the baseplate and is not uniformly distributed along the cells).
For the pedesta?. of course, the compressive load factors are a larger percentage (59-62%) of the total R. Buckling of the pedestal is-not a concern since the section is extremely compact.
3.
Page 6 24 The governing equation on page 6-24 does not have a damping term.
Please explain when a structural damping is used.
Discuss how the damping term is incorporated in the governing differential equation of motion. Also, justify the damping values used, referring to the Regulatory Guide 1.61.
RESPONSE
The matrix [Q) of the govert.ig equation of motion includes the damping term.
Structural damping follows established practice and is incorporated into the elastic portion of the model by introducing a structural damp 1'g natrix formed by associating linear structural damping coefficients of the form c = Ok with every linear spring in the model. Therefore, the Q matrix contains damping terms linearly proportional to velocity in addition to spring terms. 8 is a constant proportional to the specified damping percentage imposed on equipment subjected to the seismic event.
As required by the Updated Final Safety Analysis Report (UPSAR), 2 percent structural damping is used for the design basis event.
Four percent structural damping was used i
for the site specific event. The desian basis event is the plant commitment in the UFSAR where maximum 2 percent damping curves are developed (page 3 7-29 of UPSAR),
Por the proposed rerack, TVA also imposed an additional spectra, corresponding to an SSE event, to ba considered. Since this additional spectra is not part of the UPSAR, the damping value of 4 percent was obtained from Regulatory Guide 1.61 for welded structures.
It turned out that even with higher damping, limiting rack behavior was controlled by the additional site specific seismic event.
4.
Page 6-25 Provide a discussion regarding DYNAr.ACK verification. The discussion should emphasize the nonlinear portion of the analysis together with some linear response aspects.
Verification should include analytical calculation as well as experimental results, including full size tests.
RESPONSR The validation manual for DYh ARACK has been previously submitted on two acekets in the past year "TMI Unit one End D. C. Cook). A brief outline of the validation is provided in "
- following.
The validation of DYNARACK is in conformance with the provisions of the Holtec Quality Procedure HQP 5.2, Computer Programs, and demonstrates that 3YNARACK meets all validation requireuents of USNRC-SRP 3.8.1.
Section II.4(e) of SRP 3.8.1 states that computer programs used in design and analysis should be described and validdted by any of the following procedures or criteria:
+
4
.n.
(1)
The somputer program is a recognized arogram in_the pub 10 l
~
73 domain, and has hn6 auffiaient history of uce to justify its j
4 applicability and validity without further demonstration.
?
The computer prograr colution to a series of test problems has been demonstrated to be substantially identical to those.
obtained by a similar and independently written and recognized program in the public domain. The test problems should be demonstrated to be similar to or within the range of applicabi)ily of the problems analyzed by the public
$k domain corguter program.
(iii)
Ths computer program solution to a series of test problems I
has been demonstrated to be e bstantially identical to those obtained from cic
's1 solution or from accepted experimental test e to analytical r36 b.m published in technical literato The test g40bleme veauld be h
demonstrated to be similar to or w. chin 'b range of
[
applicability of the classical probleas tr,z(d to justify L
c'
'% te of the program. A summary vmparison should be I
pn i e!. sor the results obtained in the validation c^ each cos,...or program.
Since DYNARACK t u a private domain program, the validation problems used for DYNARACK comply with criteria (ii) and (iii) above.
-" the DYNARACK Validation deport, it ir shown that DYNARACK meets the Teilowing criteria:
1.
All desired capabliities of the code perform ne expected.
2.
Results from DYNARACK are in excellent agreement with solutions obtained from other sources.
4 3.
The fluid coupling methodology in DYNARACK ia demonstrated to be in agreement with experimental results.
4.
The code exhibits excellent convergence when applied to both linear and nonlinear problems.
The experinental verification of DYNARACK had to be performed on a scaled model sinco a full scale testing would involve very large inertia, fluid, and friction forces which would outstrip the capability of calibrated testing in any U.S. laboratory. To our knowledge, the only effort at full scale testing was in Japan, which, too, falle short of the objective because some key loadir.gs such as the fluid coupling forces, were eliminated from the ext tment, presumably to Faep the testing effort managentle.
Alth?q,h attempts have been made t2 obtain it, the Japanese data has not i +e 7ade L7e
., and it vuuld be of limited value because of
abs entioned lie'+stions.
i 2
,,m
-. - - -.-i---
. Holtec's scaled model testing focused on the two key contributors to the dynamics of the racks--the fluid coupling and inertia forces.
The results from almost 100 experimente demonstrated remarkable agreement between the predictions of the Code and the experimental data.
Recognizing that empirical principles are used in constructing the DYNARACK equations of motion and that the Code has been ' -nchmarked against a wide array of linear and nonlinear problems in dynamics, the experimental validations have further reinforced +,he veracity of DYNARACK. To our knowledge, DYNARACK is the only Code with such a complete underlay of validations. This Code has been used in over 1000 dynamic simulations in over two dozen nuclear plant dockets since 1980.
- q) 5.
Page 6-28 Discuss the stress analysis of the various welds described in Page a..
4 M
6-28 and 6-29.
Provide a definition of the limit force and the g'
moment together with a numerical example of weld stress analysis of I
baseplate to rack and cell to cell.
In particular, expand the term function (P/Py, M/My).
RRSPONSR A copy of annotated back-up calculations for the welds is attached.
These computations were performed using MathCad (commercial calculation program) and show how final results reported in the licensing document were achieved. The limit analysis interuction formula is P
M
<1
=
+
Y M
Y y
where P. M are applied compressive force and net uending moment applied to a J-weld section, Fy = limit force = Wy A, and My=
limit moment calculated on basis of ideal plasticity with wy =
yield stress and Aw = weld effective stress eiea.
Use of a straight line interaction formula implied by the forecoing equation is conservative, as i tieglect of the gussets in the calculation of pir o;ic section moduli.
. 6.
Page 8-5 Provide the total weight of th0 s*cocture for the spent fuel pool (concrete, racks, fuel assemblies, ws'.er and oths-What is the increase in various loadv in going from the origin; design to the proposed high density design? Provide the amcunt of increase in strecmes due to the new loads.
In cases where the stresses are deceeased in spite of the increase in the load, state the reasons for such an outcome (such as difference in analysis methods).
RRSPONSE The attached table shows that there is a 6.3 percent increase in total bearing weight caused by the proposed rerack.
If we do not count the concrete in the table, then the percent increase becomes 7939.2 - 6264.9
% Increase = 100 X
= 25.2%
6264.9 The previous analysis of the pool st'.acture used a mixture of analytical cale:uie".19ns on reduced mouels plus sexe finite element computatione on selecta4 portions of the structure.
The ner cc lysis used a total finite elenent based analysis. The criteria need for acceptance is comparicori of moments and shears with Amer'.can Concrete Institute allowab'e va'aes.
The limiting section of the structure, both in the current and prc. posed configuration, is the 18" intermediate wall between the cask pit and the main pool.
It is difficult to quantify tl.
actual increase $n moment at critical sections because of a 1sek of 1 to 1 correspondence in the models.
The new analysis includes amplification of the response due to low re6onant frequency modes on the intermediate wall (the wall separating the cask pit from the spent fuel pool).
The increase la moment is acceptable with the established requirement that the cask pit remain flooded.
'3
l 7.
Page 8-9 l
Provide a detailed discussion, in terms of numerical values, as to L
how the maximum stress of c2992 psi is obtained on the liner.
Discuss the desiga criterion that is based un an ultimate strength.
The discussion should include the data basis for the ultimate strength and how the ultimate strength addresses-bearing, tearing I:
(fracture), denting or any other type of failure mode of the lircr.
RESPONSR The maximum liner stress of 22992 pai is obtained from a finite element analysis of a portion of the lit 3r subject to imposed loads in the vertical and tangential direction. The purpose of this analysis is to assess whether reracking imposes the potential for liner damage due to the increased loads. The estimate of liner stress is obtained by considering the highest peak load from any pedestal in the pool during the governing seismic event to be-applied uniformly over a load patch equal to the nominal size of a bearing pad.
For conservatism, it is also assumed that friction forces are applied equa?. to.8 x the peak normal load in each of two direction.s.
(It is recalled that the bounding value of the interface coefficient of friction for stainless sheet in water.is 0.8) The liner is simulated as a 1/4" thick plate in contact with an ela-::1c foundation (the concrete). A representative section of l
the liner is considered and it is assumed that the load patch is l
applied near one corner of the liner section considered (roughly 5" away from a weld senm).
l The corresponding elastostatic solution encompassing the three components of load is obtained and the maximum bending stress in the liner determined from the finita element analysis. The result for maximum elastic plate bending atress is 22992 psi.
As expected,
-thin-maximum stress is near the edge of the sezm weld.
l The rimary inten+ of tl.c analysis is to calculate maximum stress i
leve. in the liner and at e welds to ossess potential overstress I
and possible rupture of thu liner. Theregis no criteria established for assessment oi liner stress level ~in the NRC OT Postion Paper;-
the margin with respect to the liner ultimate stress provides a-I measure of the safety against in-plane rupture.
In this case, since the stresses remain low, in the elastic range, rupture of the liner is not possible, i
l.
i 8.
Page 8-9.
The concept of cumulative damage factor (CDP) is used in addressing the adequacy of the pool liner.
Provide a basis for the use of CDF-by reference, noting that the nature of paismic loading represents a' i
1 w cycle fatigue with relatively high stresses.
j
RESPONSE
It is recognized that the vibratory motion of the re.ck due to the seismic event inducce cyclic stressea in the pool liner. _ If the amplitude of the cyclic stress ir above the endurance limit, then the most likely actuating mechanism for failure is low cycle fatigue. -The governing design code for high density racks, Subsection NF of Section 3, Clsse 3 does not contain techniques for fatigue analysis. We refer' to ASME,Section III, Subuection NB-3222.4 for thy appropriate methodology.
The uce of a fatigue criter.on for liner assessment is another measure that is useful for considering implications of the rerack.
Sinco fatigue analyntu methods are not spelled out in the NF section of the Code, we refer to ASME,Section III, subsection NB-3222.4 for Class 1 components.
The procedure outlined in NB-3222.4 alto refers to and requires use of Sections NB-3E22.2, NB-32228.5, NS-3215, and NB-3216.
Appropriate fatigue curves for obtaining cyclic life versus alternating streas range are givea la Section III for austenitic steel.
Another refercnce wl.ere the concept of the Cumulative Damage Factor (CDP) is comprehensively explained is the text by David Burgreen,
" Design Metheds for Power Plant Structures," Arcturus Publishers (1975).
We use the tinte-history results of pedestal loads from the whole pool multi-rack analysis to determine the peak impact vertical load and make a conservative estimate of friction loads at the same instant. Per the requirements of the fatigue method, stress intensities are computed frcm the finite eiement analy is, and cycles are estimated from the time-history pedestal laad files.
In this case, since the stresses in the liner are low (see response number 7), the' cumulative damage factor is less than 0.1 (allowable = 1.0).
,i:u
'JJ n
N W s3
'W ATTACilMENT FOR RESPONSE TO NRC QUESTION IS VELD STRESSES BETWEN r..L AND RASEPLATE tot c be cett width, t be cell thickness, tw be weld length per side of cell, tw be weld size. Asswae wielding on 4 sides,
c :: 8.75
- .060 8, :: 7
- , ::.060 A.g :: c't*4 A.a = 2.1 sq.in.
c c
A.a :: 1, * *, *.7071
- 4 A,,w
= 1.158 sq.in.
R ::
R = 1.768 A,,a For TVA Sequoyah, the critical value for R6 just above the baseplate, is obtained for nai do12x14a.rfS. From the sunnary tables 6.7.2 in the licensing report, we obtain:
q Ce {l p.6, g
%N a mLAA h.
ca si. e A
s./
%W *U 6 tog amqA-
&,p 4
w A
u/
~
e vMc.
k ll
Rg::.333 3
s h p :: R g
- 15000
- R
$ bp = 8.83 *10
~ This is less than the attowable weld shear stress.
For f aulted conditions, we use.42 x ultima'e stress as the shear limit per ASME NF which refers you to Appendix D for dealing witn f aulted conditions. Note that this weld stress is less than the weld stress allowable from the NF table for normal operation.
(21000 24000 psi)
Su :: 71000 psia so that the weld attowable. ear for SSE (f aulted conditions) is t a
- 2 982'l t a * *'2'S u psi tIELD BETWEEli St.PPORT FOOT PAD AND 8ASEPLATE
"*g Q
Tha weld between baseplate and swport pedeytet is checked by using y
a timit analysis. This weld is a groov~ # 1. Additional wetd area is 1
provided by gussets applied at nf ety degree locations.
b d +_e b I I' The fornuta used for the limit analysis is (basic pedestat is circular) *-
yb'a ym F/f y + (Mx""2+My**2)**1/2*1/MY <= 1 where F,Mx,My are the calculated soments, FY,MY are the yield force and moment for the weld section. Ve now calculate the appropriate quantitles.
The allowable weld Limit stress is taken as
.t a :: 42
- S u WE liECLh,f THE EFFECT OF THE GUSSEITSilli. We will include the gussets in this calculation only if we need them for the timit qualification c :: 4.5 in.
t, ::.625 t a = 2.982 *10' A ::.7071
- p
- t w *[ 2
- c ]
A = 12.495 sq in. of weld area per leg 5
FY : A*t n = 3.726*10 a
lbs.
6 MY ::
A* 2**
- ta MY = 1.067 *10 in.lb.
D We check the welds for critical cases. The case to c5eck is run di12x144.rfd (table 6.7.2) whict has the critical value of R4 for the upper support locatIcos.
Frera table 6.7.30 of Licensing Report referenced above R6 :: 484 the values for the individual load factocs are R, ::.287 on the some foot Then we est ute the total bending losd f actor as (Rb Is bending effect en circular section.)
Rh Rg~Rg Rb = 0.197
f Therefore,.the sctual stresses (based on support eres ced Inertie,
,not weld area, inertin) are s b :: 15000 psi
=.6*sa 3
s o :: s b
- n 3 s o = 4.305 *10 s 1 := s b' Ab sj = 2.M5 '103 Knowing the 54mrt area AS and support inertit !$ input into the analysis runs, we can back figure the actual direct toad and bending sonents as follows:] (get data from ser 6 of this report dealing with input to predyna)
AS :: 45.859 sq.in, is :: 369.04 in**4 l
= 1.974 *1E F1 :: SO*AS F1 i
M1 :: b8* s j 5
ui = 2.423*10 l ::
+
I = 0.757 e 1
- 0. :
Note that this neglects the added inertia and area of the g'ssetsill u
ANALYS!$ OF SPOT WELOS Raf. Hottec drawings 852,853, we can tocate the spot welds, es:h weld is considered as having e.fective diameter.5 inch. There are two welds at any Levet. Therefore, the weld area avaltable for shear transfer is p :: 3.14159 dw ::,5 t a = 2.9M *10' psi 2
d a, :: 2*p*p a, = 0.393 sq. In.
The capacity of the weldc (2 at any level) is: Pe :: a
- t Pc
- 1.171*10 lbs.
w a We coropare the weld capacity at any levet by the load that need be transferred by any impact.
' For a weto analysis, assune tnat adjacent boxes rre net moving and that the impact tv.d is being transferred from the box being fepected by the fuel to an adjacent box Assuno that each weld set transfers ispects at two locations (simultaneously) in the box P. :: 2*P P = 599 lbs.
a Another shear check can be made at the bottom of the rack where we can take the shear leading at the worst location and see If the avaltable weld spots can transfer the load We make the worst case assuption that the adjacent boxes are fixed.. Frcrn Table 6.7.4 of the licensing report, the limit value of R2 is n2 :: 041 Therefore, the maximori elastic shear stress is Smar:: 1.5
- R2
- sy 3
S
= 1.538*10 pst P,,, :: s
- a P
= 603.774 lb.
mu mn w
max Pmax is less than ?c for both esses considered. Also, note that at the cottom of the rack, there are two closely spaced set of spot welds so that the actual capacity is doubled.
N.
'Aat. HIVE CALCULATIONS FOR IVA SEQUOYAN REGULAN FUEL HrAD fl(e \\mcad\\tvargrpt.med August. 28,1991 IMPACT LOAD BETWELN FUEL ASSEMBLY AND CELL VALL Design calculations are made using Section 1 of N!-89330.
WCs nuntaer of to-Med celta, LOAbstotal Load. From Table 6.7.2 of Licmsing repurt,H191670, the highest rack to fuel impact load la for run di13x14c.rf8. Also see, table 6.7.12 NC :: 182.
LOAD ::54509 therefore, the irpact load per loaded cett is
-l LOAD p.
NC P = 299.5 pounds
-)
We use equ.1.1 and 1.2 of Part !!! of the g meral seismic report to cocpute the cell capacity. We assune an inpact over a length L. Data below conws frm Hottec fuetrack data sheet attached to this calculation.
we cell width, asfuel width w :: 8.75 a: 8.426 L ;: 10
]
sy: 25000.
t ::.050 e ::
+,
c = 0.162 OL !" 8 y ' * : *[.5 ]
1 pse cett including a FURTHER safet) factor of 2.0 The shear Load timit is 4
- ( a +L )
O w = 1.382 *10 og :: S t
y There will be no damage to the fuel assembly due to this load.
The fuel assently manuf acturer can attest that this load la less than that required to fait the assembly.
'n ATTACHMENT FOR RPSPONSE TO NRC QUESTION NO. 6 WEIGHT OF SPENT FUEL POOL (in Kips)
PROPOSED CURRENT CONCRETE 18694.4 18694.4 WATER 3918.6 3911.6 RACKS 330.8 198.0 SPENT FUEL 3589.8 2148.3
.)
TOTAL 26533.6 24959.3 26533.6 - 24959.3-
%=
X 100-24959.3
= 6.3%
.m
-