ML17320A414

Summary of New & Spent Fuel Storage Array Criticality Safety Analyses.

ML17320A414
Person / Time
Site:	Cook
Issue date:	02/28/1983
From:	INDIANA MICHIGAN POWER CO. (FORMERLY INDIANA & MICHIG
To:
Shared Package
ML17320A413	List: ML17320A412 ML17320A414 ML17320A415
References
AEP:NRC:0745B, AEP:NRC:745B, NUDOCS 8303080145
Download: ML17320A414 (23)

Text

ATTACHMENT NO.1 TO AEP:NRC:0745B DONALD C.COOK NUCLEAR PLANT UNIT NOS.1 AND 2

SUMMARY

OF NEW AND SPENT FUEL STORAGE ARRAY CRITICALITY SAFETY ANALYSES 8303080145 830888 PDR*DOCK 080003g8 P PDR

~~

~~1.0 SU!%MR OF CRI TY ANALYSIS K)R D C.QXK ENT HJEL RACK Criticality of fuel assemblies in the spent fuel storage rack is prevented by the design of the rack which limits fuel assembly interacticn.

This is done by fixing the zunian separaticn between assemblies and inserting neutron poison between assanblies.

The design basis for preventinp criticality outside the reactor is that, including uncertainties, there is a 95 percent probability at a 95 percent confidence level that the effective multiplication factor (K ff)of the fuel assarhly array will be less than 0.95 as reccamended in ANSI N210-1976 and in"NK'.Positicn for Revi~and Acceptance of Spent Bzel Storage and Handling Application." Xn meeting this design basis, scme of the conditions assumed are: fresh 15 x 15 Nestinghouse cptinuzed fuel assemblies (OFA)of 4.05 w/o U-235 are stored, the pool water has a density of 1.0 gm/cm, the storage array is infinite in lateral and axial extent which is narc reactive than the actual finite array, mechanical and method biases and uncert-~ties are included, the minimum poison loading is used, ard for scme accident conditions credit for the dissolved boron in the pool water is taken.The design rrathod which insures the criticality safety of fuel assemblies in the spent fuel storage rack uses the AMPX system of codes fbr cross-section generation and KENO XV for reactivity determinaticn.

A set of 27 critical experiments has been analyzed using the above m thcd to denanstrate its applicability to criti-cality analysis axxl to establish the method bias and variability which are then included'n the reactivity analysis of the rack.Th result of the above considerations is that the nuclear design of the rack will greet the requirements of NRC guidelines and criteria.

2-0 CRITICALITY A IS EOR D.C.CXXK SPRG'UEL CK 2-1 NEVZIKN NJLTIPLICATIOH FACIOR Criticality of fuel assemblies in the spent fuel storage rack is prevented by the design of the rack which limits fuel assembly interaction.

'Ihis is Gone by fixing the minimum separation between assemblies anR inserting neutrcn poiscn between assemblies.

The design basis for preventing criticality outside the reactor is that, including uncertainties, there is a 95 percent probability at a 95 percent confidence level that the effective multiplication factor (K ff)of the fuel assembly array will be less than 0.95 as eff reccmnended in ANSI 5210-1976 and in"NRC Position for Review and Acceptance of Spent Fuel Storage and Handling Applications".

The following are the conditions that are assumed in meeting this design basis.2.2 NORMAL SZORAGE a.'Ihe fuel assembly contains the highest enrichment authorized without any control rods ar any noncantained burnable poiscn and is at.its nost reactive point in life.Criticality analyses were done far Westinghouse 15 x 15 optimized fuel assanbly (OFA)with an enrichment of 4.05 w/o.'Ihe following'ssembly parameters were nadeled: Number of Fuel Rods per assembly Rod Zirc-4 Clad O.D.Clad Thickness Fuel Pellet O.D.Bml Pellet Density Fuel Pellet Dishy~Rd Pitch Nurrber Zirc-4 Guide Tubes Guide Tube O.D.Guide Tube Thickness 204 0-422" 0-0243" 0.3659" 955 Theoretical 1.190%0.5630" Square 21 0.546" 0-017" The ass lies are conservatively reeled with water replacing the assembly grid volum an2 no U-234 or U-236 in the fuel pellet..No U-235 burnup is assumed.b.The storage cell nanin-Q.gecmetry is shcam cn Figure l.c.'Ihe moderator is pure water at the temperature within the design limits of the pool which yields the largest reactivity.

A conservative value of 1.0 gm/cm is used for the density of water-No dissolved borcn is included in the water.d.The nnunal case calculation is infinite in lateral and axial extent.e.Credit, is taken for the neutron absorption in full length structural materials arxl in solid materials added specifically for neutron absorption.

'Ihe minimum poison loading (0.02 gm-B10/an)is assumed in the poisoned cell walls.A bias is included in the reactivity calculation to account for the B4C particle self shielding.

g.A bias, with an uncert-~ty is included to account for the fact that the D.C.Cook racks have randem cells closer together than for the rarninal'design.

The minimum gap between adjacent cells may be as small as 0.953", canpared to the naninal gap of 1.139".I The.calculation m thod uncertainty and bias is discussed in'ection 2.4.2.3 POSHJLATED ACCXDEHIS I Rx;t accident conditions will mt result in an increase in K ff of eff the rack.Examples are the loss of cooling systans (reactivity decreases with decreasing water density)and dropping a fuel 0 assembly cn top of the rack.(the rack structure pertinent for criticality is not deformed and the assembly has mre than eight inches of water separatirg it, fran the active fuel in the rack which precludes interaction)

.Hmmver, accidents can be postulated which auld increase reactivity such as inadvertent drcp of an assembly between the outside periph-ery of the rack and the peal wall.'Iherefore, for accident condi-tions, the double contingency principle of AHS N16.1-1975 is a~lied.'Ihis states that it shall require two unlikely, inde-pendent, concurrent events to produce a criticality accident.Thus , for accident conditions, the-presence of soluble boron in the storage gxil water can be assumed as a realistic initial cccxiitim.

The presence of the approxinately 2000 pgn boron in the pool water will decrease reactivity by narc than 3(Sb,k.In perspective, this is narc negative reactivity than is present in the poisoned cell walls, (i.e., 24%b,k).Therefore, K ff far the rack would be less eff than 0.95 even if the cell walls were unpoisoned-

'Ihus Kef f~0-95 can be easily met for postulated accidents, since any reactivity increase will be much less than the negative worth of the dissolved Por fuel storage applications, water is usually present.~ever, accidental criticality when fuel assemblies are stored in the dry ccndition is also accounted for.Par this case, possible sources of naderaticn, such as those that'ould arise during fire fighting cperations, are included in the analysis.This"optinaxn naderation" accident is not a problem in poisoned fuel storage racks.'Xhe presence of poison plates raraves the conditions necessary far"option nxderation" so that K ff continually de-"3 3 creases as rraderator density decreases frcm 1.0 gm/cm to 0.0 gm/an in poiscn rack designs.

Figure 2 shows the behavior of K ff as a function of moderator'ff density far a typical PNR poisoned spent fuel storage rack.1 2.4 MEZHOD H)R CRITICALI'IY ANALYSIS'Ihe calculation aathod and cross-section va1ues are verified by canpariscn with critical experiment data for assemblies similar to those for~ch the racks'are designed.'Ihis benchmarking data is sufficient1y diverse to establish that the method bias and uncer-tainty will apply to rack cna9itions which include strong neutron absorbers, large water gaps ani lear rxderator densities.

The design mthod which ensures the criticality safety of fuel assemblies in the spent fuel storage rack uses the AMPX systan of codesC'for cross-section generation and KEHO IV for reactiv-C33 ity determinaticn.

The 218 energy group cross-section, library that is the ccrmen start~point far all cross-sections used far the benchnarks and the storage rack is generat,ed frun ENDF/8-XV data.'Ihe NITAWL program 3 includes, in this library, the shelf-shielded resonance cross-sections that are appropriate for each particular geanetxy.The Nordheim Integral Treatment is used.Energy and spatial weighting of cross-sections is performed by the XSDRNPM exp;am C23 which is a ane-Lunensional S transport theory code.These multi-group cross-section sets are then used as input to KENO IV which E33 is a three-dimensional Monte Carlo theory program designed for reactivity calculations.

A set of 27 critical experiments has been analyzed using the above method to demonstrate its applicability to criticality analysis and to establish the method bias and variability.

'Ihe experiments range frcn water moderated, oxide fuel arrays separated by various materials (Boral, steel and mter)that simulate GR fuel shipping and storage conditions

'to dry, harder spe~~uraru.un metal C4,53.cylinder arrays with various interspersed materials (Plexiglass,~C63 steel ard air that deaenstrate t¹wide e of applicability of the m thod.The results and scna descriptive facts about each of the 27 bench-V mark critical experiments are given in Table 1.The average K ff of eff the benchmarks is 0.9998 which denanstrates that there is no bias associated with the methcd.The standard deviation of the K ff eff values is 0.0057 dk.The 95/95 one sided tolerance limit factor for 27 values is 2.26.Thus, there is a 95 percent.probability with a 95 r percent confidence level that the uncertainty in reactivity, due to the method, is not greater than 0.0136,k.'The total uncertainty (TU)is to be added to a criticality calcula-txcn xs.~=E(ks)method

+(~)~~al+(ks)~h 3 where (ks)~~is 0.013 as discussed above, (ks),~is the statistical uncertainty associated with the particular KENO calculation heirs.used, (ks)~is the statistical uncert-unty associated with randan gap reduction between adjacent storage cells.For a single can it is fourri that reactivity does not increase significantly because the increase in reactivity due to the water gap reduction on one side of the can is offset by the decrease in reactivity due to the increased water gap on the cpposite side of this can.The analysis, for the ef fect of m chanical tolerances, 1xxmer, assum s a"worst" case of a rack ccmgosed of an array of groups of four cans where the water gap between the four cans is reduced to 0.953 inch.KEG calculations using this minimum gap result in a bias of 0-002lldk ard.a 95K/95K uncert-sty of 0.00454.Scne nechanical tolerances are rat included in the analysis because worst case assumptions are used in the ncminal case analysis.An example of this is eccentric assenibly position.Calculations were performed whi s~that the mast rea ve cxnxU.tion is the assembly centered in the can w'hich is assumed in the rxminal case.The final result of the uncm~inty analysis is that the criticality design criteria are mt when the calculated effective multiplication factor, plus the total uncm~ty (1U)anR any biases, is less than 0.95.These methods conform with'ANSI N18.2-1973,"Nuclear Safety Criteria fcr the Design of Stationary Pressurized Water Reactor Plants", MSX 5210-1976,"Design Cbjectives for LNR Spent&el Storage Facilities at Nuclear Poorer Stations", ANSI N16.9-1975,'Validation of Calculational Methods for Nuclear Criticality Safety";HRC 1 Standard Review Plan, ard the NEC Guidance,"NRC Positicn for'Revim: and Acceptance of Spent Keel Storage and Handling Applications".

2.5 CRITICALI'IY

RESULTS The spent fuel storage cell is shown in Fi,gure l.'Ihe minismm B loadie in the poisoned cell walls is 0.02 gm-B/cm.The sensi-10 2 ,t'vity of storage 1 tt'm K ff to U-235~ie nt of the f el assembly, the storage lattice pitch, and B loading in the poison plates as requested by the NRC for prison racks is given in Figures 3~For rmrmal operation and using the rrathod described in the above sections, the K ff far the rack is detexmined in the follcwing eff manner.:~+B~+B~+B~+2 2 1/2 where: K.al=amninal case KEHO K ff K ff bias to account for the fact that mchanical eff tolerances can result in water gaps between poison plates less than ncminal B~=m~bias determined frcm benchmark crit.ical ccnpari-sons B=bias to account for poison particle self-shielding ks~~=95/95 uncertainty in the ncaunal cae KENT K ff ks~.=95/95 uncertainty in the calculation due to KENO analysis of mech-mical tolerances ks~~=95/95 uncertainty in the method bias Substituting calculated values, the results are the folly'Lng:

K ff 0 92837++00211+0 0+0025+t (006494)+(004539)2 2 eff+(.013)3=.9482 Since K ff is less than 0.95 including uncertainties at a 95/95 eff probability/confidence level, the acceptance criteria far critical-ity is m t.2 6 ACCEPI'ANCE CRITERIA FOR CRITICALITY The neutrcn multiplicaticn factor in spent fuel pools shall be less than or ecyal to 0.95, including a11 uncertainties, under all conditions.

Generally, the acceptance criteria for postulated accident condi-tions can be K ff<0.98 because of the accuracy of the methods used eff-I coupled with the lear probability of occurrence.

Bar instance, in ANSI H210-1976 the acceptance criteria for the"option mcderation" condition is K ff 0.98.~ever, for storage pools, which contain dissolved bore+, the use of realistic, initial conditions ensures that K ff<<0.95 for postulated accidents as discussed in Section eff 2.3.Thus, for simplicity, the acceptance criteria far all condi-tions will be K ff<0.95.eff 3.0 CRI1LITY ANALYSIS FOR O.C.COOKI FUEL RACK 3.1 NEUTRON MULTIPLICATION FACTOR Criticality of fuel assemblies in the new fuel storage rack is prevented by the design of the rack which limits fuel assembly interaction.

This is done by fixing the minimum separation between assemblies to take advantage of neutron absorption in water and stainless steel.The design basis for preventing criticality outside the reactor is that, including uncertainties, there is a 95 percent probability at a 95 per-cent confidence level that the effective multiplication factor (K ff)of the fuel assembly array will be less than 0.98 as recommended in ANSI N18.2-1973.

'he following are the conditions that are assumed in meeting this design basis for the D.C.Cook new fuel storage racks.3.2 NORMAL STORAGE a.The fuel assembly contains the highest enrichment authorized without any control rods or any noncontained burnable poison and is at its most reactive point in life.Because the Westinghouse 17xl7 and 15xl5 are very similar neutronically

, only the 17x17 will be examined.Sufficient margin will be maintained to,cover any reac-tivity differences.

The enrichment of the 17x17 Westinghouse stan-dard fuel assembly is 4.5 w/o U-235 with no depletion or fission product buildup.The assembly is conservatively modeled with the assembly grid volume removed and no U-234 and U-236 in the fuel pellet.b.The array is either infinite in lateral extent or is surrounded by a conservatively chosen reflector, whichever is appropriate for the design.The nominal case calculation is infinite in lateral and axial extent.Calculations show that the'inite rack is less reac-tive than the nominal case infinite rack.Therefore, the nominal case of an infinite array of cells is a conservative assumption.

2407F: 6 c.Nechanica1 uncpieties and biases due to mac!ica1 to1erances I during construction are treated by either using"worst case" condi-tions or by performing sensitivity studies to obtain the appropriate values.The~tems included in the analysis are:-stainless steel thickness-cell ID-center-to-center spacing-asymmetric assembly position The calculation method uncertainty and bias is discussed in Sec-tion 4.d.Credit is taken for the neutron absorption in full length stainless steel structural material.3.3 POSTULATED ACCIDENTS Most accident conditions will not result in an increase in Keff of the rack.An example is the dropping of a fuel assembly on top of the rack (the rack structure pertinent for criticality is not deformed and the assembly has more than eight inches separating it from the.active fuel in the rest of the rack which precludes interaction).

However, accidents can be postulated (under flooded conditions) which would increase reactivity such as inadvertent drop of an assembly be-tween the outside periphery of the rack and pool wall.Therefore, for accident conditions;-=the double contigency principle of ANS N16.1-1975 is applied.This states that it is unnecessary to assume two unlikely, independent, concurrent events to ensure protection against a criti-cality accident.Thus, for accident conditions, the absence of water in-the storage pool can be assumed as a realistic initial condition since assuming its presence would be a second unlikely event.2407F:6 The absence of wat in the storage pool guarantee subcriticality for enrichments less than 5 w/o.Thus any postulated accidents other E13 than the introduction of water into the storage area will not preclude the pool from meeting the Keff<0.98 limit.Because the most limiting accident is the introduction of moderation into the storage pool, this accident will be considered in determining the maximum K ff for the storage pool.For this accident, possible eff sources of moderation, such as those that could arise during fire fight-ing operations, are included in the analysis.This"optimum moderation" accident is not a problem in new fuel storage racks because physically achievable water densities (caused, for instance, by sprinklers, foam generators or fog nozzles)are considerably too low (<<0.01 gm/cm)to yield K values higher than full density water.The optimum achievable moderation occurs with water at 1.0 gm/cm.Pre-ferential water density reduction between cells (i.e., boiling between cells)is prevented by the rack design.3.4 METHOD FOR CRITICALITY ANALYSIS'I The most important effect on reactivity of the mechanical tolerances is the possible reduction in the center-to-center spacing between adjacent assemblies.-The nominal gap between adjacent cells for D.C.Cook is 11.0 inches.The design also guarantees that the average center-to-center storage cell spacing for a module of cells will be 21.0 inches.(See Figure 4).Therefore, any reduction of cell-to-cell gap on one side of a can will produce a gap increase on the opposite side of the can-The KENO model for the gap reduction analysis consists of an infinite array of clusters of 4 cells with the gap between adjacent cells in each clus-ter reduced to 10.97 inches.Another center-to-center spacing reduction can be caused by the asym-metric assembly position within the storage cell.The inside dimensions of a nominal storage cell are such that if a fuel assembly is loaded into the corner of the cell, the assembly centerline will be displaced J 2407F: 6 only 0.284 inches fQ the cell centerline.

This ns that adjacent asymmetric fuel assemblies would have their center-to-center distance reduced by 0.568 inches from the nominal.Analysis shows that the combined effect of the worst mechanical toler-ances and the asymnetric assembly positioning may increase reactivity by 0.00lhk.This will be treated as a bias although the individual devi-ations will be random.The final result of the uncertainty analysis is that the criticality design criteria are met when the calculated effective multiplication factor, plus the total uncertainty (TU)and any biases, is less than 0.98.These methods conform with ANSI N18,2-1973,"Nuclear Safety Criteria for the Design of Stationary Pressurized Water Reactor Plants", Section 5.7, Fuel Handling System;ANSI N16.9-1975,"Yalidation of Calculational Methods for Nuclear Criticality Safety".3.5 CRITICALITY ANALYSIS FOR RACK DESIGN For normal operation and using the method in the above section, the K ff for the rack is determined in the following manner.eff I K=K'+B+B+ef f nominal mech method nominal method~Where: nominal nominal case KENO Keff mech K ff bias to account for the fact that mechanical eff tolerances can result in spacings between assemblies less than nominal 2407F:6 B method me~bias determined from benchm critica1 compari-I sons nominal 95/95 uncertainty in the nominal case KENO K ff ks=95/95 uncertainty in the method bias Substituting calculated values in the order listed above, the result is: K ff=0.9189+0.0010+0.0+f.(.0062)+(.013)]=.9343 eff Since Keff is less than 0.98 including uncertainties at a 95/95 pr o-bability/confidence level, the acceptance criteria'or criticality is met.2407F:6 REFERENCES 1.M.E.Ford III, et al,"A 218-Group Neutron Cross-Section Library in the AMPX Master Interface Format for Criticality Safety Studies," ORNL/CSD/TM-4 (July 1976).2.N.M.Green, et al,"AMPX: A Modular Code System for Generating Coupled Multigroup Neutron-Gamma Libraries from ENDF/B," ORNL/TM-3706 (March 1976).3.L.M.Petrie and N.F.Cross,"KENO IY-An Improved Monte Carlo Criti-cality Program," ORNL-4938 (November 1978).4.S.R.Bierman, et al,"Critical Separation Between Subcritical Clus-2 2 ters of 2.35 wt X UO Enriched UO Rods in Mater with Fixed Neutron Poisons," Battelle Pacific Northwest Laboratories PNL-2438 (October 1977)-5.S.R.Bierman, et al,"Critical Separation Between Subcritical Clus-ters of 4.29 wt'X UO Enriched VO Rods in Mater with Fixed 2 2 Neutron Poisons," Battelle Pacific Northwest Laboratories PNL-2614 (March 1978).6.J.T.Thomas,"Critical Three-Dimensional Arrays of U (93.2)-Metal Cylinders," Nuclear Science and Engineering, Volume 52, pages 350-359 (1973).7.Letter No.AEP:NRC:00105 dated November 22, 1978..2407F: 6

BENCHMhBK CRETXCAL EXPERIMIKIS 2~3~4~5~6.7~8.9~10.12.13.14.15.16.17.18.19.20.21.22.23~24.25'6.27.U metal cyliners General M2 rod lattice UO2 rcd lattice UO2 rod lattice UO2 rod lattice UO2 rod lattice UO2 rcd lattice U)2 rod lattice UO2 rcd lattice UO2 rcd lattice UO2 lcd lat tice UO2 rcd lattice UO2 rcd lattice UO2 rod lattice UO2 rod lattice UO2 rod lattice U metal cyliners U metal cyliners U metal cyliners U metal cyliners U metal cyliners U metal cyliners U metal cyliners U metal cyliners U metal cyliners U metal cyliners U metal cyliners Enrichment w/o V235 2.35 2.35 2.35 2.35 2.35 2.35 2.35 2.35 2.35 2.35 2.35 4.29 4.29 4.29 4.29 93.2 93.2 93.2 93.2 93.2 93.2 93.2 93.2 93.2 93.2 93.2 93.2 Reflector water water water water water water water water water water water water water water water bare paraffin bare paraffin bare paraffin bare paraffin bare paraf fin bare Separating Material water water water water stainless steel stainless steel stainless steel stainless steel boral boral boral water stainless steel stainless steel boral air air air air air air plexiglass plexiglass plexiglass plexiglass steel plexiglass, steel Characterizing Se ation (cm)11.92 8.39 6.39 4.46 10.44 11.47 7.76 7.42 6.34 9-03 5.05 10.64 9.76 8.08 6.72 15.43 23.84 19.97 36.47 13.74'.23.48 15.74 24.43 21.74 27.94 14-74 16.67 1.004+0.993+1.005+0.994+1.005+0.992+0.992+1.004+1.005+0.992+1.001+0.999+0.999+0.998+0.998+0.998+1.006+1.005+1.001+1.005+1.005+1.010+1.006+0.999+0.994+1.000+0.996+.004.004.004.004.004.004.004.004.005.005.006.005.003.00$.004-003.004.003.004~003.005.003.003 tl I r I>++g I~i I~$I Pw+IL~Ei I I I I FIGU 2 K~~VS.hVTER YiODERAQ eff FOR A TYPICAL"POISONED" SPENT FUEL STORAGE RACK'1.0 0.9 eif 0.8 0.7 0.6 TYPE OF RACK C-C, 10.25 INCH POISON LOADING, 0.02 gm-B/cm 10 2 FUEL, 3.5 M/0 M 17 x 17 0 0.2 0.4 0.6 0.8 1.0 MODERATOR DENSITY (gm/cm)~0-'

~I~I U.Kerf AS A FUt>CTIOH OF C-C SPACING, , POI LOAOIt(G CttD Et<RICHNEhT FO ltd STI tsut{OUSE 15 x 15 OFA FUEL FOR O.C.COOK SPEtlT.FUEL RACK 1.0 C-C SPACING.98 POISON LOADING.96 ENR I CHiMENT.94.92.90.88~v<t s i 3.55-P n>>ck~pnO (/o)-4.05 10.0-9P-"<<~~Q-C (-Q~z)-10.5 0.01-L,>~~y (~-8'/z-3-0.02 For enric'r,;-,.ant c~r.e, C-C=10.5", lo=ding=For spacing curve, w/o=4.05, loading=0.02 For loading curve, w/o=4.05, C-C=10.5" 2 0.,02 c.---B/cm ,".Ia/, 2 4.55 11.0 0.03 FIGURE 4 STRUCTURE BARS INTERi'MEDIATELY SPACED (NOT INCLUDED IN KENO i%0EL)REFLECTIVE f ANGLE IRONS (FULL LENGTH)0.25" 0.25"~I, l FUEL ASSEMBLY 17 x 17 M STD.8.432".I 9.0" 21.0"