ML19296C789

Supplemental Info to 790423 Application for Amend to License SNM-1067 Recriticality Safety Calculations of Touching Clad Rods in Horizontal Storage Packed in Hexagonal Lattice to Max Slab Thickness of 15-inches

ML19296C789
Person / Time
Site:	07001100
Issue date:	10/17/1979
From:	Lichtenberger H ABB COMBUSTION ENGINEERING NUCLEAR FUEL (FORMERLY
To:
References
14437, NUDOCS 8002280480
Download: ML19296C789 (31)

Text

{{#Wiki_filter:PDR. 70-yoo ~ C-E Power Systems Tet 2c3/63319 1 Cornbushen Engineenrg. Inc. Telex 99297 1000 Prospect Hill RCad W:ndsor, Connecticut C6c95 7 -' 3!"Td POWER Jf, 35u=Ei SYSTEMS P V RECEjygg & ET //0p C~i License Smi-1067 h.,,Qit'~~l}Q; a g cket 70-1100 October 17, 1979 ^ U. S. fluclear Regulatory Commission N/ Washington, D. C. 20555 d l g,,. Attention: Mr. L. C. Rouse, Chief Fuel Processing & Fabrication 3 ranch Division of Fuel Cycle & Material Safety

Reference:

Amendment Application dated April 23, 1979 and Supplements dated June 27, 1979 and August 7,1979 Gentlemen: In our original amendment application dated April 23, 1979 it was stated that an independent review of all criticality safety calculations would be performed by the fluclear Safety Commit ee. That review, including in-dependent analysis, has been completed and is hereby submitted as supple-mentary information to our original application. One section of our present license allows storage of touching clad rods in horizontal storage packed in a hexagonal lattice to a maximum slab thickness of 15 inches. This section was inadverter.tly cmitted in our anendment appli-cation and is included in this transmittal. It is requested that Page C-17 Revision 1 dated 7/16/79 be replaced by the corresponding attached page (Revi-sion 2) dated 10/8/79. If you have any questions regarding tins application, please contact ilr. G. J. Bakevich of my staff on extension 3150. Very truly yours, i /- [_,, 'bb/L.Lbs~ dy Q [' g H. V. Lichtenberger / Vice President-fluclear Fuel DOCKETED M tluclear Power Systems-itanufacturing USNRC Yf. 7 L., HVL/GJB/ssb .M I UI979 > Y, Enclosures 'S udygc S

\\',4 ccc.: tensive analyses of this sort are needed to support the calculations presented in the current license amendment. The subcommittec concludes from this review ti.at: {l) the calculatians in the license amendment application have been carried out with adecuate under-stanctng of the methods involved; and (2) althou::h the calculational results are subject to some uncertair.ty, the calculated margins to criticality are adequate to cover the uncertainties. The review has developed information useful in the assessment of uncertainties due to limitations of computer codes. Beyond that it has found only minor discrepancies in the o riginal calculations: a drawing error and an error in taking information from a drawing. Thes<. have proved to be inconsequential and the subcommittee finds no fault in the criticality safety argument presented ' a the amendment application. /R D T 4 J. R. Dietrich Chairman, Nuclear Safety Committee JRD:jd att. cc w/att. B F. J. Pianki G. J. Bakevich cc. w /o att. -B J. M. West R. S. Ha rding R. J. Klo t r. L. C. Noderer W. E. Abbott V. C. Hall R. L. Hellens A. Stathoplas S. Visner ~ Report to fluclear Safety Ccemittee on Review of tioclear Fuel anufacturing Application For An Increase In Limiting UO2 Enrichment From 3.5 to 4.1 w/o U-235 R.S.Hardingf R. J. Klotz J7. 7(' L. C. floderer / ( Combustion Engineerina, Inc. Power Systens Grous Windsor, Connecticut October 5, 1979 - g, .t., Summary and Ccnclusiens The amendrent to SNM License 1067 justifying an increase in the maximum enrichment fec= 3.5 to 4.1 w/o U-235, which was submitted to the Nuclear Reculatory Commission under a cover letter dated Acril 23, 1979, and two subsequent transmittals of additional inforration dated June 27, 1979 and August 7,1979 were reviewed frcm the standooint of the technical acceo-tability of the criticality evaluations. The r.ethods of review included detailed checkina of input to selected analyses, checking of data scurces and application of data to assessments of criticality safety by SIU methodology, and evaluation of acceptability of results by ccmoarison with other analyses. The conclusions of this review are as follows: On the basis of examining each of the analyses employing SIUs, 50% 1. of the multigroup KEU0 calculations, and rationalization of the few group KENO calculations, it was concluded that the original criti-cality analyses shown in the amendment were carried out in a satis-factory way. No significant fault was found in any calculation. 2. Adequate subcritical margin exists in the various operations when handling 4.1 w/o enriched U02 providing all administrative controls and area limits are adhered to by Manufacturing personnel. Detailed checking of selected Monte Carlo input resulted in only 3. a few ambiguities which, when resolved, resulted in lower multipli-cation factors. The use of 4 neutron energy groups in criticality evaluations tends 4. to underestimate the multiplication factor for a specific Icw moderator density condition (0.05 gr/cc) by an amount which is dependent upon the geometry being analyzed. For the case of interest, in-plant storaqa of fuel assemblies, the calculated multiplication factors may be -3% ak ncnconservative but the margin to criticality is calculated to be -20'; ;k. I ,0- B. Introducticn The fluclear Manufacturing Facility in uindsor has prepared a license amendment to justify increasinc the U-235 enrichment linit from 3.5 to 4.1 w/o U-235 in uranium dioxide. All steps in the fuel handling process were reviewed by fluclear fianufacturing personnel, criticality evaluations for various operations were carried out under the direction of the sucer-visor of f!uclear Licensing and Safety, and a license amendment was ore-pared and transmitted to the U.S. uclear Regulatory Cc=issien (l:RC) under a cover letter dated Aoril 23, 1979. In the cover letter it was stated that an independent review of all criticality safety calculations was being carried out-under the direction of the C-E t!uclear Safety Comittee and the results of this review would be forwarded under seoarate The purpose of this document is to su=arize the method and results cover. of the review. The basic objective of this review was to determine whether the results of the criticality safety analysis were reasonable and acceptable. The apolicable standards, regulatory guides, or branch technical positions which are available for guidance in this review acpear to be limited; con-sequently precedent and guidance from the *:RC reviewer would aopear to be a dominant factor in the overall review crocess. The only standard which apoears to be pertinent is AtlSI ti 16.1-1975, ituclear Criticality Safety in Ooerations with Fissionable Materials Gutside Reactors. Section 4.2.5, Subcritical Limits, states that where acplicable data are available, sub-critical limits shall be established on bases derived from excerimants, with adequate al'owance for uncertainties in the data. Plots of limiting values of U-235 mass, cylinder diameter, slab thickness, volume, and areal density are given as a function of w/o U-235; these curves are based on the data ccmpiled by H. K. Clark in Reference 1. Safe masses and dimensions are defined therein as those values resulting in an effective rultiolication f actor which lies 0.02 below the averane curve defined by analysis and nnr alized to exnerimental data. Deviations between individual experimen al data coints and average values of keff' given by the curves fall within t0.015 and for the most part within 10.01, consequently one may deduce that an effective multipli-cation factor as high as 0.995 is accectable under this standard. For this license application added conservatism is introduced in the definition of safe masses and dimensions beycnd that defined by this standard. Criteria on safe masses and dimensions are used quite extensively in this license amendment for demonstrating safe conditions for relatively simnle geometries of fissile materials or for specific configurations which are either closely related to a configuratien examined experimentally or con-vertible to a simple ceometry through areal density techniques. Criticality evaluations of the more ccmolex qecmetries are medelled on the computer using the Monte Carlo computer code, KENO I'/. t. ~ I 4 Extensive comoilations of experirental data are available for model verifi-cation (see for e <amnie, References 2, 3, and 4). However, there are nnticeable nars, the rnst nreminent can beino for lattices moderated by Icv density hvdrocenenus material sn as tn sirulate sn-called low density mist conditions. Mist conditions are oostulated to cccur for open arrays of fissionable material which mav he accessible to, for examnle, fire fichtino equipment such as sprinklers. foam or foo tyne noz:les. Tyoically, the most reactive moderator distribution is costulated so as to define the upper bounds to the multiplication factor. In reality, most fire fighting techniques are not able to give a sufficiently high density of mist so as t'o meet the postulated conditions. For example, overhead sprinklers can lead to conditions approximating water densities of 0.1% whereas fire fighting foams yield a hydrogen distributinn ecuivalent to a water density of aporoximately 3% (special foam materials can go as hich as 61). Over-head sprinklers are emoloyed in the manufacturing facility and fire hoses may be used but no foam is allowed for fire fighting in the facility. Consequently shere is the likelihncd that icw density mist conditions could occur and, under certain cnnditicns high local cnncentrations of water could occur for short periods of time, but the design and location of the facility is such that ccmnleteim ersicn of manufacturing operations in an environment of water having an effective density of greater than 0.1 gr/cc is highly improbable. The KEtt0 analyses carried out in suoport of this license amendment are of two types: (1) 16 group calculations eeploying the Hansen-Roach cross-section libraries, and (2) 4 group analyses where the broad group cross-sections are derived by the CEPAK lattice code. Reference 5 has been cited as a basis for validating the use of the KE!!0 with the Hansen-Roach cross-sections; however, the benchmark analyses were primarily reflected and unreflected clastic moderated criticals with uranium enrichments in the U-235 isotooe of 5!', or less and H/U-235 ratios in the range of ~130 to -970. The absence of benchmark exceriments on large geometry fissile arrays moderated by low density mist accears tn be a persistent problem. The 4 neutron croup apnrnach employing CEDA< as the cross-secticn generator has been emoloyed for criticality evaluaticns of snent fuel storane racks for rnny years. Appendix A provides information on the benchmarking of this tech-nique. However, there are still valid cuestions as to the uncertainty of this 4 group technioue when acclied to fissile conficurations containino a mist environment. This uncertainty is covered by c 1owing a large mar-1 gin to calculated criticality. Administrative controls olay an important cart in assuring that safe conditions exist in the fuel manufacturinc facility. Since Un2 fuel of less than 55 w/o U-235 cannot achieve criticality in the absence of hydrogeneous moderator and many of the cperations in manufacturing are dry, a key nbjective nf the administrative controls is to assure that in the event of floodinc of the facility by water (so as to produce an environrent comparable to part or full censity water), fuel configurations are such that criticality safety is assured as demonstrated by the criti-7 cality safety analyses. Administrative controls and/or procedures include the folicwino. All operations and chances in operations must be analy:ed . to establish safety limits and centrols and these analyses must undergo independent review. Safety limits and centrols are docurented by written procedures. Signs defining the criticality safety limits are posted at each work station; in cases where the safe scacing area extends beycnd the ecuipment boundaries, the boundary of the scacing areas are indicated by colored lines in the floor. All mass limited containers are labeled to indicate the enrichment and uranium content; procedures require labeling so that the identity of all the f el enrichments is known through-out the facility. Line management, includino the iuclear Licensing and Safety supervisor, are responsible for enforcing all administrative con-trols. g p. a C. Review Aoproach 1. Cateaoritation of Criticality Evaluations For purnoses of reviewinq the various criticality analyses, they are divided into four catenories as listed in Table 1. The first cateqory contains those analyses which use as their basis for demonstrating subcriticality the definition of Safe Individual Units (SIU). As noted earlier, safe masses and diFensions are defined by using exoeri-mental data to determine the maanitude of these parameters such that a suberitical multiolication factor is assured. The second cateacry consists of these criticality evaluatiens which employed the KENO-I'/ comnuter code and the 16 croup Hansen-Roach cross-section librarv. The ma.iority of these analyses involve hcmo-geneous mixtures of fuel and moderator. The third category consists of those KENO evaluations which employed four neutron group libraries generated by the CEPAK code. These cases involve heterogeneous drrays of fuel, moderate and in some cases, structural materials which were not amenable to volume homcgenizatien. The fcurth category involves those operations which are deduced to be subcritical by comoarisen with other evaluations. 2. Review Procedures a. Safe Individual Units The most exceditious way of reviewing those analyses employing SIU limits was determined to be an exar.ination of the data sources and a check of each anplication uncer the criteria adopted for this license amendment. b. 16 Grouo Monte Carlo Analyses Four out of seven of the KENO analyses in this category were selected for detailed review of the analysis (items C3.3, C3.4, C6.1 and C5.2) on the basis that they costulated moisture in the fuel and. involved the more complex gecmetries. c. 4 Group Monte Carlo Analyses Of the four analyses in this category, three exhibited maximum effective multiplication facters under full density water enviren-ments and the fourth, In-Plant 5 crage of Fuel Assemblies (C-8.7), exhibited its most reactive condition under a mist environment. The review of this cateoory focused on the cuestion of calcula-tional uncertainties associated with the four group approach for mist conditions; in particular, what bias may result from the use of only 4 neutron groups rather than a larger number. 1 1 l.. e 7 TABLE I Cateqorizatic>n of Criticality Evaluations A. Safe Individual Units 19.1 Limits for Individual Uni s 19.2 Interaction Analysis C-3.6 Pressing C-3.7 Devaxing and Sintering C-3.8 Final Sizing C-6.2 Pellet Storace Shelves - Additional Storage Evaluation C-6.3 Transfer of Material C-7.0 Pretreatment of Lcw Level Liquid Uastes C-8.9 Fuel Salvage C-8.10 In-Process Storage of Fuel Fellets in Containers C-3.11 Rod Transfer B. 16 Group Monte Carlo Analyses C-3.2 Virgin Powder Storace Area C-3.3 Batch itake-Uo C-3.4 Powder Preparation and Blending C-6.1 Concrete Block Storage Area C-6.2 Pellet Storage Shelves C-8.4 Fuel Rod Storage Area C-8.5 Double Shelf Red Storage Racks C. 4 Grouc Monte Carlo Analyses C-8.1 Pellet Alignment and Drying C-8.2 Rod Loading and Fuel Rod Transcort Carts C-8.7 In-Plant Storage of Fuel Assemblies C-8.8 Shipping Container Storage D. Comoarative Analyses C-3.5 Final Mixing C-8.3 Autoclave Corrosion Test C-8.6 Fuel Assembly Fabrication 't. I (e 8 D. Results of Review 1. Safe Individual Linits A safe individual unit limit is that mass or dimeasion which charac-terizes either a heteroceneous or herocenecus array of fuel and moderator known to be subcritical (safe) by comparison with experimental data. The definitien of safe masses and dimensions employed in this license acclication have added conservatism beyond that defined in ANST-N-16.1. Section 19.1 quotes safety factors of 2.3,1.3,1.1, and 1.2 for mass, volume, cylinder diameter and slab thickness, resoectively, as existing in the definition of safe individual unit limits. Attemots to verify the magnitude of the safety factors indicated that they may be underestimated relative to the data of Reference 2 but overesti a:ed relative' to the curves in ANSI-N-16.1. In any event it is concluded that there is substantial conservatism still remaining relative to either source of data. Additional ccnservatism has also been introduced in certain cases to meet criteria on maximum fraccion critical values. A review of section 19.2 concluded that Table 19.2 did not include the safe volume and spacine area for two story operatien with 4.1 w/o UO2 fuel to support, for examole, the pressinc operation in Exhibit C. Although the correct criteria are included in the text of 19.2 and the safe volume limit is given in Table 19.1, an additional statement in Table 19.2 is required for completeness. Based on a review of the references of section 19.1, it was concluded that the single level scacing criteria of Table 19.2 meet the fraction critical criteria outlined on oage XIX-4 Comments on individual sections of Exhibit C emoloying the SIU acproach are summarized below. a. Section C-1.6 pressina The writeun of this section of the license apolication is very brief and does not address criteria and controls fer all ascects of the cressing oceration. For examole, interactions between hoppers and presses are not discussed. Actually the hoooers are designed to be sa'e cylinders (3.5 w/c) or safe volumes (4.1 w/c) and are ' located on a mez:anine above the presses. Criteria cut-lined in Sections 19.1 and 19.2 for application to two story operation indicate that the safe volume hopcer for 4.1 w/o powder in combination with the already designated spacing areas for 3.5 w/o powder lead to safe conditions. The statements concerning maximum height of beats and a limit of only one boat at a oress work station at any given time do not adequately describe the jus-tification for safety of this eneration. Actually the boats have a wall height of only 215/16 inches, the cellets are loaded to a point below the lip of the boat and a card is placed on top of the pellets giving certinent data on the fuel. This procedure oro-vides reasonable assurance that the safe slab limit of 3.7 inches is met. In addition, a boat loaded with pellets stacked to the height of the sidewalls of the boat contains approximately 12 kg UO2 which is well below the safe mass limit of 24 kg for 4.1 w/o fuel. 9 -g-b. Section C-3.7 - Dewaxino and Sintering The criterion for safety of this creration is a safe slab limit of 3.7 inches which is achieved thecugh controls discussed above on tha height of pellets loaded in the boats. c. Section C-3.8 - Final Sizing Here again the safety criterion is a safe s'.ab lic.it of 3.7 inches. In most steps of this operation the fuel is distributed in a layer having a thickness of -1 pellet daimeter. The only part of the Operation requiring investigaticn was the infeeder where the pel-lets are dumced into a flooded bowl, one boat at a time to meet the 3.7 inch safe slab limit. The volume of the infeeder bowl is 31.2 less than 1.5, the criti-litres and for a volume ratio of H 0/UO2 2 cal volume is greater than 35 litres. Even if this bowl was fully filled with pellets it would be necessary to attain the large volume fraction of H 0, consequently the bowl is a safe volume 2 for this operation. At the out:ut of the grinder, the cellets are placed into " grinder trays" having a height of 1.5 inches and a cover assembly. The paragraph on drying of grinder sludge does not state what control is employed; cc=ents to the effect that the sludge is collected in volume limited SIU containers and trays having a maximum depth of 3.7 inches so that the oven is limited to a safe slab configuration, as stated in section 7.0, should have been included in the text. The centrifuge and grinder coolant sump meet the safe volume and scacing area criteria of sections 19.1 and 19.2 although the spacing area of the grinder coolant sump (9 ft ) is not stated. For the storage rack ('.l.S. P-20), the 2 safe slab limit is achieved by stcring grinder trays no more than 2 high. d. Section C-6.2 - Pellet Storage Shelves Of.S. P-15,19, 21) An increased slab thickness limit was adooted for this storage area so as to accommodate 3 levels of grinder storage trays. The consequence of usino this increased slab thickness is reduced cen-servatism relative to section 19 criteria. The rationale for deducing the 5.5 inch slab limit was reviewed and found to be con-servative relative to exneriments by a factor of 1.20 en the slab thickness when the H 0/UO2 volu e ratio is < l.0. In reality, 2 three levels of grinder storage trays would result in a fuel height of less than 4.5 inches which is significantly less than 5.5 inches but nore than the 3.7 inch slab limit used in other operations. e. Section C-6.3 - Transfer of Material The SIU limit in ecmbination with the cart dimensicns meets the criteria in sections 19.1 and 19.2. Transfers by hand occur in-frequently and are limited to one S!U. Attention is focused cn cart transfers since carts may te left unattended and the dimen-3' sion of the cart assures a safe scacino area. ' hen material is ^~ transferred by hand it is generally eiiher for a very short dis-tance or consists of an amount of fuel less than contained in 1 SIU. f. S,ection C-7.0 - Pretreatment of Low Level liquid '.!astes The rationale employed to deduce the safety of this operation was reviewed and found to be accentable. Althcugh the diameter of the tank exceeds the safe cylinter limit (9.8" in Table 19.1) by 0.2", it is smaller than the critical diameter of the infinitely long, fully reflected cylinder having optimun moderation of the fuel water mixture by 0.8 inches. This marcin in combiriation with the finite height of the settling portion of the tank (18 inches) should offer sufficient conservatism, g. Section C-3.9 - Fuel Salvace In addition to being mass lie.ited, safe containers are employed to receive the recovered fuel, h. Section C-8.11 - P,od Transfer The rationale for the increase in the safe slab limit to 5.5 inches is discussed in paragraph id, acove, on section C-6.2. 2. 16 Grouo f1onte Carlo Analyses The detailed review of the following analyses included checking the fuel composition for the most reactive case in each series of calcu-lations for nuclide number densities, enrichment, potential scattering, dimensions, geometrical representati0ns, and composition of other mate rial s. C 3.3 Batch Make-Up C 3.4 Powder Preparation and Blending . C 6.1 Concrete Block Storage C 6.:2 Pellet Storage Shelves The powder prenaration and blendinc station involves a comolex three dimensional gecretrical representation in KENO which maxes verifica-tion difficult. Hcwever, spotcheckinc of the geometry was done by reconstructing from the generalized gecretry equations. The following comments are provided for the indicated sections. a. Section C-3.2 - Virnin Powder Starace Area The fact that interseersed water moderation and flooding were not addressed does not belong in the listing of conservative assump-tions. The design of the facilf tv itself provides the principal argument that external moderatien is of no concern in this case. Internal moderation is addressed by the checking nrocedures of Section C-3.1 and the assum: tion in the criticality analysis of a higher value of moisture content than the limit defined in C-3.1. ..N _11 b. Section C-3.3 - Batch Make-Ua The introduction of internal and external moderation as two seoarate variables raises questions as to the rate of convernence of the interative process and the validity of the orncess in general. Two iterations were carried cut and tSere is no indication that the oracess is convercing. Fcaever, it Nould anocar from the trends of the analyses that any realistic conditions of rnderation would result in much lower multinlication factors than ccrouted here since: (1) the presence of near full density water external to the fuel containers is not probable under any conditions inside the hood, and (2) the optim.um moderation conditiens postulated for the fuel containers simultaneous with the postulated external moderation conditions are not realistic for this area of the facility. c. Section C-3.4 - Powder Precaration and Blending In subsection 3.4.2 under criticality analysis, the definition of optimum moderation should be stated more clearly as to the external moderation conditions. Cnce again the internal and external moderation variables aporoach is pursued. However,in this case the calculations appear to be convergent. The mul ti-plication factors indicate a high cegree of subcriticality even with extremely high degrees of moderation in both regions. Drawing NFM-C-4065 shows 6 inches as the separation of the two one-half inch thick fuel layers on the conveyor belt whereas the calculation used 9 inches. The correce dimension is 9 inches; the drawing should be modified. In the discussion of the front end of the station, clarification of the term " optimum moderation" as employed in the conservative assumptions would be beneficial. The iteration on internal and external moderation conditions was truncated after the first itera-tion cn external moderator conditic9s and the resulting maximum multiplicaticn factor is in the rance of 0.94 to 0.95 for both enrichment cases. While there is less marcin to criticality than in the cases discussed above, it is difficult to see how the flooding of these areas with full censity water could occur. In the review of the KD0 geometry, two small wedges of fuel at the intersection of the powder scread funnel and the cowder tubes appeared to be improperly described in the original definition of the material regions. This should have a negligible effect upon reactivity but the case has been rerun to assess the inoact on the calculated multiplication fac:cr. The revised multiplication factor was 0.8484 0.C086 versus 0.5684 t 0.0101 computed earlier. d. Section C-6.1 - Concrete Block Storace Area The review of the calculations uncovered no problems or questions Other than those raised above pertaining to the treatment of 1-internal and external moderation as separate variables. - ~ - e. Section C-6.2 - Pellet Storage Shelves The discussion of the pellet storano shelves states that the shelves are limited to a slab thickness of'3.7 inches. Actually this is assured by limiting the number of the covered grinder trays stacked at any position to two wnich implies that the maximum fuel height is <3 inches. In the analysis, the iteration on internal and external moderation was truncated at the first iteration on external moderation and showed a relatively low maximum multiplication factor (-0.82) for an external modera-tion condition which is higher than one could attain from fire fighting equipment or a sprinkler systen. In the detennination of cross-sections for the pellet-water mixture, a volume homogenization crocedure was employed which is non-conservative for low enrichment fuel. However, it can be shown using data from Reference 2 that the non-conservatism of this approximation is more than offset by the assumption that the pellet water mixture is such that the fuel concentration is 2.7 gr U/cc. For a randem loading of tne trays, one would expect a higher fuel density, i.e. of the order of 5.9 gr U/cc. For 5 w/o U0 the folicwing critical masses (kg) are deduced frcm Reference 2:2 gr U/cc 2.7 5.9 Homogeneous -97 ~570 0.4" dia. rods ~86 -230 From these data one can see that the conservatism associated with the assumption of a fuel density of 2.7 gr U/cc in the hemoceneous approximation is equivalent to a reduction in the critical mass s of 463 kg whereas the non-conservatism of using the homogeneous approximation evaluated for a fuel density of 5.9 gr U/cc is equivalent to only 340 kg. The review of the KEfi0 analyses indicated that the eight inch thick hollow concrete block wall was represented as full density concrete. The effect nn reactivity is expected to be small but the case was rerun to evaluate the effect of reducing the wall thickness to five inches; the peak reactivity decreased from 0.8188 0.0088 to 0.7791 t 0.0081. f. Section C-8.4 - Fuel Rod Storace Area For UO2 of 4.1 w/o enrichment and no hydrogenecus moderation, this area is clearly subcritical. g. Section C-8.5 - Double Shelf Rod Storage Racks This analysis is similar to that of secticn C-8.4 with the excep-tion that the spacing between storage boxes is greater and no physical barriers have been used to exclude mist or personnel frcm between storage boxes. Previcus analyses for the 3.5 w/o t. enriched fuel employed 4 neutron groups and the DOT code for the .w spatial calculation. For the case of flooded boxes the analyses yielded a maximun multiplication factor of -0.87. In the case of 4.1 w/o fuel but with no water inside the boxes, a multiplica-tion factor of '0.89 was obtained. These two sets of results are in reasonable agreement if one assumes that the non-conservatism inherent to the use of only 4 neutron groups with uist between boxes is nearly offset by the increase in enrichment from 3.5 to 4.1 w/o and the elimination of the relati.ely small amount of room temperature water inside the boxes. 3. 4-Grouc "onte Carlo Analyses, a. Validity of 4 tieutron Group fiodel The analytical model employed in the criticality evaluation of fuel manufacturing ocerations considered in this section used CEPAK to generate 4 neutron arous cross-secticns and KEf;0 to solve for the spatial solution of the multiplication factor. Validation of this model is di:cusscd in Section 3.0 of Exhibit 0 of the license apolication. Results discussed in Exhibit 0 indi-cated very good agreement for lattices employing full density water. Analyses discussed in Apcendix A of this report provide added assurance of the validity of basic cross-section libraries and general methods of analyzino the reactivity of a broad variety of lattices. Thus, there is a high degree of confidence in the methodology nnt only at the full density water conditions but also at the reduced density conditions existing in a hot, full power reactor environment, i.e. water densities down to ~0.70. A question arises as to tne accuracy of the 4 group approach for mist conditions, viz. how strong is the dependence of the multi-plication on the number of broad neutron grouos emoloyed when two closely interacting regions of differing neutron spectrum are involved. To examine this coint, analyses of a 14x14 fuel assembly in a mist environment (0.05 gr/cc of H O) were carried out using 2 the DOT code for the spatial solution and alternate cross-section generators to prepare broad group cross-sections in differing num-bers of neutron groups. It shmid be noted that the CEPAK lattice code is not emoloyed to generate broad grouc cross-sactions in more than 3 non-thermal groups since it emoloys the MUFT tyoe solution to solve the multigrouc ecuations. The GA31 code was used as an alternate method of deriving the non-themal broad grouo constants and, in one lattice gec etry, the DIT code was employed to generate grouc constants. Both the GA'l-THERMOS and DIT calcula-tional models employed the same basic multigroup neutron cross section libraries (ENDF/S-IV); he..ever, differences in resonance self-shielding did exist between the two models. The DIT code is a C-E proprietary code which solves for the multi-group neutron spectrum in 85 neutron groups throuchout an entire fuel assembly and not simply the fuel pin cell as with CEPAK or GAtl. It is cresently employed for reactor lattice calculations where it solves for the soatially dependent multigroup soectrun in varicus subregicns of the hetercqeneous fuel assembly and provides few grouci, cross-sections for soecified subrecions. The DIT code is not cre sently programmed to deal explicitly with the larne geometrical arrays encountered in the present calculations, but it can be used to generate few group cross-sections more representative of spectral variations than either CEPAK or Gel-THER>iOS. This capability was . exploited in the generation of 4 and 9 group cross-sections for a DOT calculation of an infinite array of 14x14 fuel assemblies having an edge-to-edge spacing of 12 inches and a 0.05 gr/cc mist both internal and external to the fuel assembly. Figure 1 shows a ccmparison of the group dependence of the multi-plication factor computed with DIT and GAM-THERM 05 derived cross-sections in the 00T code. To excedite the GAM-THERMOS calculations, a modified 14x14 cell definition was used to represent the fuel assembly which preserved the volume of structure, fuel and moderator but redistributed the moderator and structure from the CEA guide tube regions to the moderator region of the unit cells. CEPAK and GAM were written to use slightly different resonance self-shielding algorithms; the former employs a fit to a Hellstrand experimental correlaticn whereas the latter used the Mordheim Integral Treatment based en resolved and unresolved resonance para-meters with a CEPAK derived Dancoff factor. The resulting difference in multiolication factors can be seen by comparing the 4 group "CEPAX(HETEROG)" point with the curve labeled GAM-THERROS(HCMOG). To show the few group trend, the latter curve was displaced downward to pass throuch the CE?AK coint and is plotted as a broken line. Th9 points of interest in Figure 1 are as folicws. First, the DIT calculation shows very little "few group" depen-dence because the multigroup solutien for the spatially decendent neutron soectra includes the influence of the mist environment in the generation of the few grouc constants. Second, the clas-sical Figner-Seitz cell approximation for the fuel pin and a separate calculation of' reflector constants employed in the GAM-THERMOS approach does not properly account for the fuel-reflector interaction in the generation of few group constants under the assumed mist conditions. Increasing the number of neutron groups gives a better approximatien to the energy dependent scatial flux solution but may still underestimate the multiplication factor because of a failure to adequately represent the effect of the interassembly mist environment into the calculation of primarily the. resonance escape calculation. Figure 2 shows the results of the DOT calculations for few group cross-sections derived by both 9A"-T'ER"05 and CE?AK for the larger i,terassembly scacings eM:yed in the storage of fuel assemblies in the manufacturinc plant: no DIT based calculation could be carried out for this conficuration since the ornblem size exceeds present cacabilities. The #irst noint of interest is that the qroup denendence is much less than in the closer spaced geometry of Figure 1. The second coint is that the bias in the calculation due to using only 4 neutron croupsappears to be in the range of 3" Ak. Analyses at full density water conditions exhibit significantly less dependence on the number of "few groups" employed. i, b. Comments on Applicable Sections (1) Section C-8.1 - Pellet Alice ent and Drying The first paragraph states that the pellet configuratien is limited to a 3.7 inch slab thickness; the text could state that this is an administrative centrol and the layer of pel-lets on the table is generally cne pellet high. Item 5 of the assumptions stated that the aluminum pellet troughs were not included in the analysis; no reference was made about the remaining aluminum structure inside the h inch thick stainless steel cylincer. The mean spacing of the pellet columns is sufficiently large that the assumed representation of the fuel / mist lattice is over-moderated in the fully flooded case. Therefere, whether or not substitu-tion of moderator for aluminum structure is a conservative assumption may depend upon the actual configuration of fuel, structure and moderator. In view of the latter question and the absence of data points bet..een water densities of 0.5 and 1.0 in Figure 0-1.7.1, a review was made of the analyses carried out to develoo Figure C-1.9.3 (Rev. 2, 9/16/74) which is labelled " Critical and Safe Cylinder Diameters as a Function of Rod Spacing

• and "cderator Censity".

Although these analyses are for 3.5 w/o fuel rods, they are informa-tive. They indicate that indeed the oven, if fully loaded and ficoded with full density water, is overmoderated and more than 500 rods at the average scacing employed in the oven would be required for a critical configuration. The dif-ference in reactivity due to the uranium enrichment from 3.5 to 4.1 w/o should be more than offset by the introduction of the poison rods. Consequently, if the oven could be fully flooded, it should be subcritical. Item 6 of the conservative ass. motions states that four group cross sections were cenerated by the CEPAK code for fuel and poison regions; actually a -1"v?R-0TF sequence of calculations was used fnr these lumoed ccisens. The discussion at the tcp of page 19b is unclear as to z.'rether the internal and external moderator conditions were varied simultaneously or indepen-dently; actually they were trea:ed as a single variable. (2) Section C-8.2 - Rod Loadinc and Fuel Rcd Transcort Carts The description of the fuel configuraticn is not too clear; actually there are 250 fuel rods arranged in 5 concentric rings formed by 4 spacer rings within an annulus of ~14.6" 10 and 25.7" 00. The analytical results are judged to be reasonable on the basis of the relative larce mean scacing of fuel rods, the maximum nur.ber of rods (250) and decoupled (annular) array of fuel rods in the fixture. i~ ~

• caption on figure says rod si::e other than rod spacing.

2 (3) Secticn C-8.7 - In-Plant Sterace of Fuel Assemblies The postulated representation of the fuel array for this cate-gory of analyses resulted in a raximum multiplication factor for reduced water density ccnditiens (? < 0.5 gr/cc), of -0.80. Consequently a margin of aoproximately 0.20 ak exists to the critical condition. This margin should be more than adecuate to cover the bias a;sociated with the use of 4 neu-tron groups as well as other biases and calculational uncer-tainties. (4) Section C-8.8 - Shinoino Container Storage If the shinoing containers are olaced in contad there will be a minimum of approxirately la inches separating fuel assemblies in adjacent containers. Under fully flooded con-ditions this is adequate to orevent interaction. Based en analyses of fuel transfer tubes and ucenders, the calculated ' multiplication factor is reascnable. 4. Comoarative Analyses a. Section C-3.5 - Final Mixing The writeup neglecgs to state that the spacing area for this operation is 27 ft which meets the requirement of Section 19 of the License. b. Section C-8.3 - Autoclave Corrcsicn Test The raticnale for criticality safety contained in the license acplication is in error on one coint, viz. that the fuel rods in a fuel assembly are spaced at the mest reactive citch. Actually ennineering judgement..'culd say that the autoclave is highly subcritical with 32 fuel reds present. The previous defense of this aceration for 3.5.-:/o fuel showed that the system was suberitical with 120 f.el reds. For d.1 w/o fuel the number of rodshas been reduced by 73-' and the amount of U-235 has been increased by only 17~: in addition, the rod scacing has been increased so as to further decrease the infinite multi-plication factor below that which e.ould exist if the rods were spaced so as to accommodate 120 reds in the autoclave. Clearly this system will be subcritical. Section C-8.6 - Fuel Assembly Fabrication c. The logic emoloyed for deducing safety of this operation is acceptable. I References 1. H. K. Clark, "Cri tical and Sa6. :2sses and Dimensions of Lattices Rods in Water," CP-10l *, Savannah River Laboratory (1966). of U and U02 2. UKAEA Handbcok AHSB(1)- Handbook of Criticality Data, Compiled by J. H. Chairers, Pa. Walker, and J. Pugn, United Kingdom Atcmic Enerav Authority (1965) 3. G. A. Price, " Uranium-Water Lattice C:noilation Part I, Cf!L Exponential Assemblies," Ri!L-50035, Brookhaven National Laboratory (1966). 4 E. Bernocchi and R. Martinelli, "Lichc '..'ater Lattice Data," NEACRP-U-190, C ?!Eil-Quaderno KIT / !S (77)l. 5. G. R. Handley and C. li. Hooper, "'.'alidation of the "KEM0" Code for fluclear Criticality Safety Calculati ns of Moderated, Low-Enriched Uranium Systens," Y-1948, Oak Ridq? Y-12 Plant (1974). g I ~ . ~. - l P,- 'h.65Q.3 [) .J MAL ~fd' . z. i 1"$3W .1. i'y j .i,l i 7% i 6 I-l ( i i i i l l <.} t l i s. O 71 ,T L4 . i.. j' .o y;t 1 ~a e i 6-i - a .l l l w N -3 i-i i ~ l

• I.

h.v9 V.- i i I. i sa i t 9 fd 'l 'N t . i 8 ,.,,, p. .y k Q g i-i i I s i .J ..g s 1; i 1 i

i. - *

.I g% y

- \\

..P a b l j I I h\\ b: l + I i I \\. i i y .l. h., 9, b s 3 .Yt i i N ,u t .7, %\\ 1 w 1 j i M- % I D I: .\\ . /. i .i.... .l - 3' 2l i i y s1 1.: i i. .l .i-. L i. I D } _ ..g.. .O N.. .i. !,.1,..

• p....,.. _

.l 'I i O L i l \\ '... .4 O'

=

1 ). g .v (1,S l 'q, j' W \\i \\: iG I i Q g. 3 .1 : ... _.. j... ... _.. k Th g( i ' l" l' 4{ \\ -j I i i j l j j j R i y Q,' c. g N., ._. 3 >g l- 'i j. 'o u] o~ i, \\'l -O 'S i 9 i i I t i o w tu Q w(, i \\. i ci e.; l' i-O I.

l. '

l l i g i 'g p--" -

l

-'-,\\- j 3g l \\ h l-l -D 1, l

P.

fJ Q i N 1 i y-l k N1 l l l l l i.. i qy o 8 i N l N* r s 'O I l 6 N I j . tg \\ i '/1f) j.. I I i s ,4, i (a I l 'l t g i i f l h l '-' i l t [ i l I 'l i w ' c ~. i i s i y$ I i I l A i- .T i t I i i l l l A i i i { f f l j f* O o Q 4 I i O l i en. i i n t s-4 l 5'. l p ~ I i i i i 3 \\s 9 4 i i l l l i I T i i 8 i .j. l 'I f i I l t i I l 'I l i i l .., _ l ._...I..... I s I 6 l.. s h I i i i u i l. Y;- i .n i -{, { ]LW i i j i q I I g i i 6 i j j i l l t i i i bN 1 i l t I f [ f N. w f l I: i i.! . o y. l., u i s, i l 8-s~ t l l . l. r I N i l. , k it I i l i g.,. .j.. ..t... g, , ~. . j g I I l i _~ i j T ll~ i

• ^

j 'Y. % .I l__...i. a l i'l l ...i... .r._ i i I j i.. _ m %,,u i i '5 ! l .i l l i.. o: 1 l .i.' !~ I i.. 7. 4,* I s t i - (t t_., T-I i i I: I. i oi i n e. n u! g i 1 s (. i j n\\ I q. L. s- -- d...I,=== 7 . _..'l --.O-- x.i p. . N. i g 1 i

ti j i

i i 0 :. I i .T . 5. '},- 1 l ~! ~l s$ rg i 22 y....g .. l

n:

1. _

it,,. d l O= t i l j A- .l.. i, i w I . + 0.:. l I. . ; g,s.'i . t..'., ye t l , c ~ > i + t l 3 C{ !\\ l i A< ,\\. ?. l .......'t.. .\\. eu D. ' L. j t-y !.... l %j i. .g .. I, i i, .t1 iv i I ,a i i. 9 ... i.1. :..,. ;. i 9 l y! b,, q: i- 'l ....i [. .i. g p g. q l i i-I, -.\\..' Q,. t .1... 'D t \\: w i r r -o l i tJ.i - ..t.....t.. .j. ...l. . t. 4. t i .ti.. 3 {...

s..

i \\ l' f l i s,',. 5 y.3

M..,

.. l i t j "~.r .. j i y y g l. I i\\ l ~ < 0,U i l l l l i i j ( ~M b 5 s.o ...i. ; \\ ...i.. 4 w i \\i { .- 1,d c i I .t ( i

• Q'

7 q.

M - i \\t..i

(

1. - _...

..n '). d. 'tr 2 i .o -43 l \\., l [ l t g s F. sc '. .t s; Nuw C = yg ! ',,l l o A'S E l l '..k-m; i 4-i .s v> x w .I I i D l 1 l i j l t I i i i j . i, l i i w. i i t i i i. i i i i i N t Q j 9 r,.. 5 l i j Q i t i 3 Q '? g l g N t I I s !.....'..L.'.._!...lL.!..j._._.._I't' i l I f I l i i I I'I i i a. APPENDIX A Oualification of Analytical Method Employed in Criticality Evaluation of Fuel Handling Ocerations I. Purpose The purpose of this Appendix is to orovide qualification of the calcula-tional model and evaluation of calculational uncertainties and/or bias factors used in perforning criticality evaluations of fuel handling operations with structural and/or fixed oisons in the form of steel boxes and baron carbide plate. This qualificaticn is based on the analysis of a variety of reactor and laboratory experiments. The methods of cross section ceneration are essentially those of C-E's physics design proce-dures modified appropriately for use in fcur group transport, discrete ordinate method criticality calculations, and Monte Carlo codes. II. Calculational Uncertainty and Sias The results of the~ analysis of a series of UO2 critical experiments are summarized in Table I. These are calculated using the CEPAK 2.3 lattice code as a few grcup neutron cross-section generator. Table I includes the mean and standard deviation for this CEPAK model. These calculations support use of the differential cross-section data base and broad group cross-section generation codes. To assess the accuracy of the calculaticnal model in predicting the mul-tiplication factor of fuel assemblies havir.g a secaration distance suffi-ciently larae so as to be isolated, analyses were carried out for a group of subcritical exoonential experiments en clusters of 3.0 w/o UOp fuel pins clad with tyce 304 S.S. and moderated by Hi0 (cage 165 of Reference 7). The cluster sizes analyzed vary fecm 181 to 301 fuel rods so as to encem-pass the rance of sizes typical of current ?'..R fuel assemblies. In these analyses, the spatial flux solution was obtained di ectly with the trans-port code, ANISN. The multiplication fact:rs fcc the lattices analyzed using axial bucklings deduced frcm the reper:ed relaxation lengths are tabulated below. No. of Fuel Rods Keff 181 0.9966 211 1.0011 235 0.9966 265 0.9983 301 0.9984 . g_ These results indicate that the calculaticnal model predicts the multipli-cation factor for small clusters of fuel reds in a water environment to a high degree of accuracy, i.e. a bias of .0017. ~~

._7 A-2 To ascertain whether the calculational model can predict the reactivity characteristics of subcritical clusters of fuel separated by water channels of various thicknesses and, in some cases, with thick stain-less steel plates and boron poisoned plates inserted in the water channels, an analysis was made of the exoeriments on critical separations of 2.35 The results w/o U-235 UO2 subcritical clusters recorted in Reference S.

using the Monte Carlo code KEM0 IV are shown in Table II. The calculation methods for these experimental comcarisons, which are also used in criticality evaluations for fuel storace racks, fuel shipping containers, plus other fuel configurations found in fuel manufacturing areas, are based on CEPAK 2.3 cross-sections. Using an anarcoriate buck-ling value and taking procer account of resonance absorption, three fast groups are collaosed from 55 fine energy mesh grouos in FORM and the one thermal group is collapsed frc7 29 thermal energy grouos in THERMOS. In addition, each component such as water gan, or poison olate has its thermal cross-section determined by a slab TFERMOS calculation emnloying an approcriate fuel environment. FOP 11 and THERMOS are sub-programs of CEPAK. For one dimensional analyses such as the BNL exponential experiments the discrete ordir.:tes code ANISN (Reference 9') is used. For two dimensional analyses DOT-2W (Reference 10) is used. For three dimensional analyses (such as the critical separation experiments) KEN 0 IV (Reference 11) is used. The above analyses indicate a mean error between predicted and measured multiplication factors of +.00135 and a calculational uncertainty of 0.0071a at the 95/95 confidence level for the ccmolete series of U02 experiments.

References:

1. T. C. Engelder, et al, " Spectral Shift Control Reactor, Basic Physics Program," S&N-1273, November 1963. 2. R. H. Clark, et al, " Physics Verification Procram Final Report," B&W-3647-3, March 1967. 3. P. W. Davison, et al, " Yankee Critical Excerirents," YAEC-94, April,1959. 4. W. J. Eich and W. P. Rocacik, " Reactivity and Neutron Flux Studies in Multi-Region Loaded Cores," WCAP-1443, 1961. 5. F. J. Fayers, et al, "An Evaluation of Scme Uncertainties in the Com-parison Between Theory and Experiments for Regular Light Water Lattices, Brit. Nuc. En. Soc. J., 6,,;.11 1967. 6. J. R. Brown, et al, " Kinetic and Bucklino Measurements on lattices of ~ Reds in Light Water," WAPD-176, 1958. Slightly Enriened Uranium and UO2 7. G. A. Price, " Uranium - Water Lattice Ccmoilation Part I, BNL Exponential Assemblies," BNL-50035 (T-449), December 1955.

A-3 8. S. R. Biernan, E. D. Claytnn and R. M. Curst. " Critical Separation Between Subcritical Clusters of 2.35 w/o U-215 Enriched UO2 Reds in Mater !4ith Fixed Neutron Poisons," PNL-2433, October 1977. 9. Ward W. Engle, Jr., "A Users Manual for ;"!S.'!, A Cne Dimensional Discrete Ordinates Transoort Code With Anisotrocic Scattering K-1693, March 30,1967. 10. R. G. Sottesy, R. K. Disney, A Collier, " User's Manual for the DOT-IIU Discrete Ordinates Transoort Computer Code," '4ANL-TME-1982, December 1969. 11. L. M. Petrie and M. F. Cross, "KE'10 IV, An Imnroved Monte Carlo Criticality Program," OR.'il-4938, November 1975. N.

A-4 TABLE I Results of Analysis of Critical UO2 Systems 0 tio. Lattice ol "eff 1 B&W (IT I .88-2 1.00121 2 II .172-2 1.00534 3 X- .79-2 .99838 4 XIII .701-2 1.00419 5 XX .202-2 1.00550 6 B&W (2} l .861-2 1.00269 7 2 .420-2 -1.00443 8 Yankee (3) 1 .408-2 1.00088 9 2' .531-2 1.00115 10 3 .633-2 1.0013G 11 Yankee (4) 4 .688-2 1.00244 Win f ri th '(5) 12 Rl-20 .660-2 1.00214 13 RI-80 .626-2 .99942 14 23 .510-2 1.00422 15 Bettis (6) 1 .326-2 1.00053 '6 2 .355-2 1.00046 .7 3 .342-2 1.00106 Average 1.00208 +.00206

• Using calculated radial bucklings and measured axial bucklings.

't, p..

A-9 TriCLE !! Calculated kerf Values For Separation Ex:eri.ments Monte Carlo Expt d Type Poison Plate Keff 6(STD Deviation) 15 None 1.00227 .00534 04 Mone 0.99912 .00540 49 None 1.00221 .00473 18 !!ane 1.00813 .00489 21 None 0.99559 .G0461 28 304 5 Steel 0.0 w/o Baron 1.00393 .00308 OS* 304 5 Steel 0.0 w/o Coren 1.cn329 .00303 ?) 304 5 Steel 0.0 w/o Bor-a 1.00271 00302 27 304 5 Steel 0.0 w/o Coron 1.01413 .00273 26 304 5 Steel 0.0 w/o Boren 0.39811 .00279 34 304 S Steel 0.0 w/o Baron 0.03793 .00297 35 304 5 Steel 0.0 w/o Baron 1.00435 .00290 32 304 5 Steel 1.05 w/o Baron 0.99970 .00524 33 304 S Steel 1.05 w/o Baron 1.01173 .00491 38 304 S Steel 1.62 w/o Baron ' 00289 .00512 39 304 5 Steel 1.62 w/o Baron 1.00208 .00506 20 Boral C.99535 .00301 16 Boral 1.00020 .00288 17 Boral. 0.99519 .00285 ltean Kef f '/alue I.C0157 Std. deviation .00419 9 I i 1 W437

}}