ML25350C335

Task 3 - Evaluation of post-LOCA Criticality in PWR Core Following FFRD

ML25350C335
Person / Time
Issue date:	05/02/2025
From:	Michel Call, Muller S, Andrew Bielen, James Corson NRC/RES/DSA/FSCB
To:
References
Download: ML25350C335 (0)

Text

EVALUATION OF POST-LOCA CRITICALITY IN PWR CORE FOLLOWING FUEL FRAGMENTATION, RELOCATION AND DISPERSAL Mike Call, Steven Muller, Andrew Bielen and James Corson RES/DSA/FSCB May 2, 2025 1

INTRODUCTION The Office of Nuclear Regulatory Research (RES) is performing a series of analyses intended to address certain technical issues associated with fuel fragmentation, relocation and dispersal (FFRD) of high burnup/intermediate enrichment (i.e., Uranium-235 enrichment between 5 and 10 wt/o) (HBU/IE) nuclear fuel due to large-break loss of coolant accidents (LOCAs) per a request from the Office of Nuclear Reactor Regulation (NRR). Task requests that RES evaluate the criticality impacts of fuel dispersal and redistribution within the reactor core region following a LOCA. This work is intended to complement criticality evaluations previously performed by RES [1], which demonstrated that, even under excessively conservative assumptions, there is substantial margin to re-criticality in a debris bed of HBU/EE fuel in the lower plenum of a pressurized water reactor (PWR) reactor vessel (RV) [1].

Whereas in the prior work dispersed fuel is assumed to fall to the bottom of the RV following a LOCA, the focus of this analysis is to quantify reactivity effects of fuel redistribution within the core itself. A nuclear fuel assembly consists of an array of zirconium-alloy tubes (i.e., cladding) containing enriched uranium oxide fuel pellets. The array spacing is maintained by zirconium alloy or Inconel spacer grids. The posited scenario is that small fragments of fuel released from fuel rods through bursts formed in ballooned cladding during the core heat-up phases of a LOCA collects within the spacer grids. Any change in distribution of fissile material within a nuclear system carries with it an associated change in reactivity. Although it is expected that the intact lattice would be the most optimized for reactivity (otherwise, by this point in the industrys history, PWR nuclear designers would likely have developed alternative approaches to lattices),

it is helpful to have an explicit confirmation of this judgement to obliviate the need for additional related inquiry. The purpose of this work is to analytically determine the reactivity state of a reactor core following a LOCA over a range of potential FFRD-induced reconfigurations. If the computational evidence suggests a negligible to negative net reactivity effect, then core re-criticality following PWR LOCA is likely not of concern, like re-criticality in the lower plenum. On the other hand, if the computational models demonstrate significant reactivity increase under any conditions, additional work would be warranted to fully understand and disposition the potential consequences.

This report is organized in the following sections:

In Section 2, the computational model and approach are described In Section 3, criticality and flux distributions of various re-criticality scenarios are quantified In Section 4, conclusions are drawn and recommendations made based on the analytical evidence

2 DESCRIPTION OF MODELING APPROACH To address reactivity effects in reactor cores due to FFRD, a computational model with an adequate degree of fidelity is necessary to represent the physical processes which must be considered. For criticality calculations, these include neutron transport within the system of interest and associated interactions with the host materials as represented with cross section data. The model must explicitly include the effects of burnup on the fuel nuclide inventory to properly account for the reactivity effects of fissile and burnable absorber material depletion and build-in of actinides and fission products. Appropriate assumptions and approximations may be introduced to render the model computationally tractable but still representative of the system of interest.

To address these requirements, the following analytical tools from the SCALE neutronics code package [2] are employed:

ORIGAMI, to generate fuel isotopics within HBU/EE PWR lattices at specific burnups of interest based on pre-generated ORIGEN nuclide and reaction rate cross section libraries.

OBIWAN, to convert ORIGAMI nuclide concentrations into a format readable by subsequent SCALE analysis sequences. OBIWAN is also used to blend HBU/EE fuel material with light water reactor coolant to create composition definitions that represent dispersed fuel particle beds at specific volume fractions trapped above spacer grids. In some instances, the ORIGAMI output in the (*_compBlock) file, which provides 29 nuclides important in criticality analyses of irradiated fuel in a format that can be directly added to criticality analysis sequence, was also used.

CSAS and CSAS (Criticality Safety Analysis Sequence) / Shift, which are employed to calculate the k-eigenvalue and neutron flux distribution within the system of interest. The CSAS model places materials within a combinatorial geometry that represents the physical arrangement of the system. Shift is a modern, massively parallelizable Monte Carlo neutron transport code employed within SCALE criticality sequences.

SCALEs flexibility in modeling nuclear systems gives users considerable power in defining problem domains that fit their needs. In this case, the most straightforward approach might be to explicitly define a three-dimensional model of a HBU/EE reactor core according to a known fuel loading pattern (e.g., the pattern employed within [3]), including the effect of spacer grids and associated debris layers. Even with modern computational tools, however, such models are extremely resource intensive to both develop and execute. An additional consideration is making this demonstration too problem specific. For example, it could potentially be difficult to draw general conclusions regarding FFRD-induced reactivity effects based on the performance of one single loading pattern.

For these reasons, two simpler, more flexible physical models were selected as representative of the post-LOCA core configuration upon which reactivity evaluations can be based. One model includes more of the assembly and core hardware, whereas the second model neglects these items and any parasitic neutron absorption resulting from their presence. Other model differences are identified as appropriate below. These model differences allow for capturing of any effects on reactivity trends that the assembly and core hardware might introduce. Use of the two models also provides further confirmation of reactivity impacts of the burst rods and resulting fuel slurry.

The approach is a three-dimensional adaptation of the colorset concept used to generate few-group nuclear data for light water reactor fuel lattices in which there are significant spectral differences between adjacent fuel assemblies (e.g., MOX assemblies adjacent to UO2). In thes models, fuel assemblies with initial enrichments of 8 wt % that are at or near the threshold burnup limit for FFRD (55 GWd/MTU) are placed in a radially infinite checkerboard array with lower burnup assemblies representing fuel near the end of its first cycle of irradiation (26~28 GWd/MTU). This arrangement represents the interior of a PWR core and conservatively neglects the effect of radial leakage on system reactivity. Illustrations of the CSAS/Shift and CSAS models are presented in Figures 1 and 2.

Figure 11: SCALE/CSAS model image: 2x2 array of 55 GWd/MTU (bluish) and 28 GWd/MTU (reddish) burned assemblies

(a) (b)

Figure 2: SCALE/CSAS Model Image: (a) axial view of 55 GWd/MTU assembly, no assembly and core hardware model; (b) axial view of color set assembly and core hardware model.

The following assumptions and approximations are made:

The fuel materials and dimensions, including spacer grids, are based on Westinghouse 17x17 Optimized Fuel Assembly (OFA) [4],[5], which are explicitly defined in Table 1 below. The fuel rods are explicitly represented within the model. The spacer grid materials are represented in one model, but they are neglected in the second model.

The models also consider the pellet-clad gap being filled with helium or water.

The model axial extent is from the top of the lower plenum to the upper core plate and includes the lower core plate, bottom nozzle, active fuel region, fuel rod plenum region, and upper nozzle. In the second model, the axial extent is limited to the active fuel region and has a 30 cm water reflector on each end.

Uniform axial and radial burnup profiles are assigned to each assembly within the models. This was chosen primarily for simplicity and can be revisited if necessary.

The models are executed with continuous energy (CE) cross sections to explicitly capture the nuclear datas detailed energy dependence, rather than applying a multigroup approximation.

The temperature field is uniform at 293 K with a corresponding atmospheric water density of 0.9982 g/cc assigned to the reactor coolant. No reactivity effect of coolant voiding or boron concentration are considered.

The base case is the unperturbed (intact) rod array geometry. Perturbations to this geometry are applied by assigning an additional region adjacent to a selected spacer grid with a homogeneous mixture of irradiated fuel and coolant to represent dispersal.

The prior lower plenum study demonstrated that self shielding was not an important consideration for the debris bed given expected dispersed fuel particle size (~1 mm) and that simple volume homogenization does not introduce significant bias into the calculation. The axial depth of the debris region, volume fraction of fuel within it, axial location (i.e., which grid it is assigned to), and radial location (i.e., within high burnup assembly, low burnup assembly, or both) of the bed are all varied parametrically to ensure adequate coverage of likely FFRD end states within the core. Specifically, the parametric variations were:

o Slurry with volume fraction (Vf) of 0.68 in the assembly central axial zone (where a zone is defined as between tops of two successive spacer grids):

slurry in the 1-cycle or the 2-cycle burned assemblies, no fuel missing from rods, slurry compositions: unirradiated, 1-cycle burned, and 2-cycle burned fuel, slurry also in second axial zone just below central zone for burned fuel slurry cases o Most reactive irradiated fuel slurry case from the above cases with varied Vfs o Most reactive Vf case and the 0.68 Vf case with different depths of slurry in a single (the central) axial zone o Most reactive Vf and slurry depth case with an equivalent burst rod layer next to and at the far end of the axial zone from the slurry o Most reactive Vf and slurry depth case with slurry in axial zone near top of the active fuel (zone second from top)

The collection of cases executed for this study, along with their results, are detailed in Table 4 in Appendix 1. Key cases and their results are shown in Table 3.

For each model, calculation parameters were selected to ensure adequate convergence of the models. These parameters are shown in Table 2. The Monte Carlo uncertainties for the CSAS6-Shift model (i.e., the model with assembly and core hardware) are shown in Table 3 and in Table 4 in the Appendix. For the CSAS6 model (i.e., the model without assembly and core hardware),

the calculation uncertainties were typically higher but did not exceed 33 pcm in k. Checks were also made of the convergence curves for k to ensure the calculations showed good convergence for the calculated kinf. The Shannon entropy results were also confirmed to indicate good convergence. Figures 3 and 4 are examples of the kinf and Shannon entropy curves for the calculation results for the CSAS6-Shift model. The CSAS6 model curves also indicated good convergence.

Table 1: Fuel design specifications Parameter Value Parameter Value Array size 17x17 Assembly length (in) 159.765 (ref)

Fuel Rods/assembly 264 Bottom nozzle length (in) 2.738 (ref)

Active fuel length (in) 144.0 Top nozzle length (in) 3.67 (ref)

Fuel pellet dia. (in) 0.3088 Assembly pitch (in) 8.466 (typ)

Fuel pellet length (in) 0.507 No. of spacer grids 2-end, 6-intermediate Fuel-clad gap (in) 0.0031 Pellet material UO2 at 95% TD Rod diameter (in) 0.360 Rod cladding Zircaloy-4 Rod length (in) 151.56 (ref)

Top nozzle Stainless steel 304 Rod pitch (in) 0.496 Bottom nozzle Stainless steel 304 Plenum length (in) 6.3 Intermediate spacers Zircaloy-4 End plugs length (total, in)1 1.3 Guide tubes Zircaloy-4 Assembly width (in) 8.434 End spacers Inconel-718 Axial locations of top of spacer grids from based of assembly (in, ref) 6.19, 31.08, 51.63, 72.18, 92.73, 113.28, 133.83, 153.96 Plenum spring Stainless steel 302

1.

Model assumes top and bottom plugs of equal length (0.65 inches each)

Table 2: Calculation Parameters for Convergence Model generations neutrons/gen skipped generations total histories CSAS6-Shift 500 150000 50 6.75e7 CSAS6 450 21000 100 7.35e6 Figure 3. Example average kinf across generations

Figure 4. Example Shannon Entropy Curve 3

RESULTS Results presented in this report are for the model that includes the assembly and core hardware and was run using CSAS6-Shift unless specifically noted otherwise. The first set of calculations staff conducted considered variations of three parameters. These parameters are the slurry isotopic composition, slurry depth (on a large scale), and the irradiated assembly in which the slurry accumulates. The composition variations include unirradiated fuel and isotopics for fuel irradiated in one or two cycles. While a uniform axial burnup profile is applied to the assemblies, and so the fuel isotopics are axially uniform, the burnup and isotopics vary axially, though that variation gets pushed further toward the axial ends of the fuel with increasing assembly burnups. Using slurries of these three different isotopic compositions allows for conservative evaluation of reactivity impacts from pellet debris originating from burst rods in different axial zones of the assembly. With consideration of pellet debris from one assembly accumulating in a neighboring assembly, it allows for conservative evaluation of reactivity impacts of pellet debris from burst rods in different axial zones of the adjacent assembly. It should be noted, however, that there is no expectation that 1-cycle burned assemblies or that axial fuel zones of such low burnup that they are similar to 1-cycle burned fuel or unirradiated fuel will burst and that fuel form a water-fuel mixture like that considered in this analysis. Having a slurry accumulation that fills the space between consecutive spacer grids in one or two adjacent axial zones very conservatively bounds the amount of fuel expected to be released from bursting assembly rods (~20% or 40% of the pellets in the assembly for a slurry volume fraction of 0.68). Previous studies had indicated that a slurry volume fraction of 0.68 was one of the more reactive slurry compositions; so, it was used for these initial calculations. Since the fuel compositions in the fuel rods are axially uniform, the slurry is placed in the centermost axial zone. When two zones have slurry, the second zone is the zone just below the centermost zone. This was done based on the reasonable expectation that reactivity effects will be greatest for changes occurring in this part of the assembly. The fuel rods are assumed to be filled with fuel. Subsequent calculations included changes to these assumptions and parameters. The most notable reactive cases are described in Table 3 below.

As can be seen in the results in Appendix 1 for this initial set of calculations, none of the variations from the base case of intact assemblies resulted in a higher reactivity than the based case. Although a scenario where rods in all assemblies had burst and created pellet debris accumulations in all assemblies was not analyzed, the results in Appendix 1 indicate that such a

scenario would also be less reactive than the base case. Of the evaluated cases involving irradiated fuel slurry, a slurry of either 1-cycle or 2-cycle irradiated fuel accumulating in central axial zone of the 2-cycle burned assemblies was the most reactive variation. The staff used the 1-cycle irradiated fuel slurry in the 2-cycle burned assemblies case to perform further calculations.

The next set of calculations addressed the impacts of varying the volume fraction of the fuel in the slurry of the above identified most reactive variation case. This was done to confirm the staff was in fact using the most reactive volume fraction. The results in Appendix 1 indicate that a volume fraction of 0.68 is not the most reactive. Consistent with the staffs expectation and the initial calculation results showing the intact assemblies case to be most reactive, reactivity increased as the slurry volume fraction approached zero (i.e., the intact assemblies condition).

The behavior of the k-eigenvalue with slurry volume fraction can be clearly seen in Figure 5 below. The figure also shows a small increase in the k-eigenvalue for volume fractions greater than 0.5, peaking somewhere between 0.6 and 0.7 with the result at 0.68 being the most reactive in that range. Based on this outcome and that a volume fraction of 0.68 had been identified in previous investigations as one of the more reactive slurry compositions, the staff used both this volume fraction and the smallest analyzed volume fraction (0.05) in further analyses.

The next set of calculations evaluated smaller amounts of pellet debris being released from the rods to create slurries at the volume fractions 0.68 and 0.05. The different depths of slurry are characterized by how many pellets from all the fuel rods in the assembly that amount of slurry equates to. For example, one pellet layer is a slurry of such a depth that, at the specified volume fraction, contains fuel equivalent to 264 pellets, one from each fuel rod in the assembly lattice. As Note 2 to Table 4 in Appendix 1 indicates, a slurry of 0.68 volume fraction and just under 0.9 cm depth contains this amount of fuel whereas it takes just over 12.2 cm of slurry with a 0.05 volume fraction to have the same amount of fuel. For each volume fraction, the single pellet layer depth of slurry is the most reactive, with the single pellet layer for a 0.68 volume fraction being the most reactive of all the cases investigated to this point and the only case that exceeds the reactivity of the base case, though only slightly, as also shown inTable 3.

Table 3: Most Reactive Cases from Computations Most reactive slurry depth and Vf case with pellets missing from rods (base kinf=

1.337189, MC uncertainty =0.000088)

Case ID Slurry depth Missing pellets per rod Missing pellet layer location kinf k v. base (pcm)

MC Uncertainty (pcm) 2B_68_1 1 pellet layer N/A N/A 1.33786 67 31 2B_68_1s 1 pellet layer 1x264 rods Next to slurry 1.33817 98 50 2B_68_1u 1 pellet layer 1x264 rods

~Top of zone 1.33781 62 30 Most reactive slurry depth and Vf case in different axial zone Case ID Slurry depth Zone v.

center kinf k v. base (pcm)

MC Uncertainty (pcm) 2B_68_1t 1 pellet layer 2 up 1.33759 40 01

Figure 5: Behavior of the k-eigenvalue vs. slurry volume fraction for the central axial zone of the 2-cycle irradiated assembly filled with 1-cycle irradiated slurry.

The last few calculations show the impact of removing an amount of fuel from the fuel rods equal to that in the layer of slurry and the impact of the location of missing fuel in the rods versus the slurry layer. They also show the impact of the slurry being in a different axial zone other than the central zone. Since the most reactive case is the 0.68 volume fraction slurry that is one pellet layer deep, this case was used for these last few cases. The model replaces fuel in the rods as well as the rods cladding with moderator for the length of one pellet. As can be seen in Table 3, having the fuel pellets missing from the rods just above the slurry layer is the most reactive configuration and is more reactive than the base case, but only by 98 pcm. Any significant distance between the burst rod location where pellets are missing from the rods and the slurry layer, whether in the same axial zone as or in a separate axial zone from the slurry layer (i.e., the accumulated fuel debris) does not produce any impact on reactivity. Figure 6 illustrates these two cases in Table 3 but is taken from the model without hardware.

As shown in Table 3, moving the slurry to a zone closer to the axial ends of the assembly reduces the reactivity impact of the slurrys presence. This is reasonable for an assembly modeled with an axially uniform fuel composition. It is expected that this may also be true even for models that consider a burnup profile that reflects the actual burnup distribution of the assembly. This expectation is based in part on the results of the initial set of configurations analyzed in this study as shown at the top of Table 2. It was more reactive to have slurry in the 2-cycle irradiated assemblies than to have slurry in the 1-cycle irradiated assemblies, the latter having burnup and isotopic compositions that would be approximately those of the ends of the fuel of the former.

Staff also examined the fission neutron generation distribution among the assemblies. Figure 7 below shows the production within assemblies in an axial slice just below the slurry of fragmented fuel, with higher fission rates within the 28 GWd/MTU assemblies. This visualization of distributions is reasonable given the modeling parameters.

(a)

(b)

Figure 6: Two-Cycle Irradiated Assembly with One-Cycle Irradiated Pellet Debris Slurry (No Assembly and Core Hardware Model) (a) Burst Rods at Top of Same Axial Zone as Slurry, (b)

Burst Fuel Rod Layer Next to the Slurry Figure 7: Neutron Production Distribution for 2x2 Assembly Array Sliced at the Axial Location Below Burst Fuel Rods.

The results for the model without assembly and core hardware shows similar trends and comparisons in kinf versus the base case as the results described in this section for the model that includes that hardware. There are some slight differences, but these differences do not affect the conclusions that can be drawn from the results presented here. Those differences can be summed up as follows. The perturbation cases are consistently less reactive than the base case with the exception of the two cases with slurry equivalent to a single layer of pellets with a zone of burst rods of a single pellet height. Those two cases are slightly more reactive than the base case by 1.6 and 2.2 pcm. Also, the trend in kinf as a function of slurry volume fraction did not indicate kinf at a volume fraction of 0.68 would be noticeably higher than at other volume fractions in the range of 0.6 to 0.8. However, the results also did not indicate that use of a 0.68 volume fraction would be inappropriate for the remaining calculations.

4 CONCLUSIONS The purpose of this study was to evaluate the criticality impacts of fuel dispersal and redistribution within the reactor core region following a LOCA. The scenario that this study considers is that small fuel fragments are released from fuel rods that burst during core heat-up phases of a LOCA and that these fuel fragments collect within or on the assembly spacer grids.

The study considered a several variations of parameters that would characterize such a scenario. These parameters include isotopic composition of the fuel fragments, the depth and concentration of accumulated fuel fragments (represented by a homogenous slurry of fragments and un-borated water moderator), axial location of the accumulated fuel fragments, and equivalent amounts of fuel missing from the fuel rods at different locations relative to the accumulated fuel fragments. Based on the results of the calculations done for this study, the staff concludes that fuel dispersal and redistribution with the reactor core region (accumulation of fuel fragments on assembly spacer grids), results in at most a negligible net positive reactivity effect. Thus, core re-criticality following PWR LOCA is likely not of concern.

5 APPENDIX & REFERENCES Appendix 1) Full Case Results Tables Table 24: Summary of parametric cases and their results1 Filled axial zone, central fuel zone (and 1 below), no missing fuel in rods, slurry Vf= 0.68 Case ID FA w/

slurry Slurry specs

1. of slurry zones kinf k v. base (pcm)

MC Uncertainty (pcm)

Base N/A N/A N/A 1.33719 9

1A 1x burn Fresh 1

1.32973 9

1B 1x burn 1x burn 1

1.32837

-882 12 1C 1x burn 2x burn 1

1.32990

-729 8 1D 1x burn 1x burn 2

1.32601

-1118 10 1E 1x burn 2x burn 2

1.32330

-1389 8 2A 2x burn Fresh 1

1.33363

-356 11 2B 2x burn 1x burn 1

1.33105

-614 41 2C 2x burn 2x burn 1

1.33105

-614 10 2D 2x burn 1x burn 2

1.32725

-994 11 2E 2x burn 2x burn 2

1.32511

-1208 9 Most reactive irradiated fuel slurry case above with variable slurry Vf Case ID Vol.

fraction kinf k v.

base (pcm)

MC Uncertainty (pcm) 2B_76 0.76 1.33042

-677 49 2B_72 0.72 1.33022

-697 36 2B_68 (2B) 0.68 1.331052

--614 41 2B_66 0.66 1.33067

-652 31 2B_64 0.64 1.33081

-638 28 2B_62 0.62 1.33050

-669 18 2B_60 0.60 1.33051

-668 39 2B_50 0.50 1.32987

-732 15 2B_40 0.40 1.33057

-662 35 2B_30 0.30 1.33079

-640 34 2B_20 0.20 1.33279

--440 24 2B_10 0.10 1.33370

-349 32

2B_05 0.05 1.33634

-85 29 Vf=0.68 & 0.05 cases with variable slurry depth in pellet layers (264 pellets/layer)2 Case ID Vol.

fraction Slurry depth kinf k v. base (pcm)

MC Uncertainty (pcm) 2B_68_1 0.68 1 pellet layer 1.33786 67 31 2B_68_2 0.68 2 pellet layers 1.33697

-22 6

2B_68_4 0.68 4 pellet layers 1.33686

-33 6

2B_68_8 0.68 8 pellet layers 1.33572

-147 9

2B_68 0.68 1 zone 1.331052

-614 41 2B_05_1 0.05 1 pellet layer 1.33689

-30 10 2B_05_2 0.05 2 pellet layers 1.33577

-142 9

2B_05_3 0.05 3 pellet layers 1.33536

-183 17 2B_05_4 0.05 4 pellet layers 1.33407

-312 8

2B_05 0.05 1 zone 1.33634

-85 29 Most reactive slurry depth and Vf case with pellets missing from rods Case ID Slurry depth Missing pellets per rod Missing pellet layer location kinf k v. base (pcm)

MC Uncertainty (pcm) 2B_68_1 1 pellet layer N/A N/A 1.33786 67 31 2B_68_1s 1 pellet layer 1x264 rods Next to slurry 1.33817 98 50 2B_68_1u 1 pellet layer 1x264 rods

~Top of zone 1.33781 62 30 Most reactive slurry depth and Vf case in different axial zone Case ID Slurry depth Zone v. center kinf k v. base (pcm)

MC Uncertainty (pcm) 2B_68_1t 1 pellet layer 2 up 1.33759 40 2

1. The results for the model without assembly and core hardware, run in CSAS6, were all less reactive than the base case with the exception of the equivalent cases for 2B_68_1s and 2B_68_1u in the table. These two cases were slightly more reactive than the base case, similar to the results shown in the table for the model including the assembly and core hardware and run in CSAS6-Shift.

2. Slurry height for per pellet layer is 0.89836 cm at Vf=0.68 and 12.21794 cm at Vf=0.05 Regarding the convergence discussion noted above, for the model without hardware, that model used the noted parameters, resulting in calculations for 7.35e6 particle histories. All calculations passed the Chi-square test at the 95% level and have MC uncertainties no greater than 33 pcm.

The k-inf convergence curves were confirmed to show good convergence of k-inf for the calculations. Also, the tests for Shannon Entropy were also checked. All calculations passed the test for final source convergence and adequate active generations after source convergence. The majority of the calculations, including the base case and the most reactive

slurry cases of Vf = 0.68 and 0.05, and all cases varying the slurry height and including burst rods passed the test that all active generations are within an epsilon of 0.1 of the average. For those cases that did not pass this latter test, the worst was an epsilon of 0.54 and no more than 190 active generations.

[1] Esmaili, H., 2024. Evaluation of Re-Criticality Within the Lower Plenum Due to Fuel Fragmentation, Relocation, and Dispersal, Memorandum from H. Esmaili (NRC) to S. Krepel (NRC), November 21, 2024 (ADAMS Accession No. ML24319A262).

[2] W.A. Wieselquist, R.A. Lefebvre, Eds., SCALE 6.3.2 User Manual, ORNL/TM-2024/3386, UT-Battelle, Oak Ridge National Laboratory (February 2024).

[3] Bielen, A., J. Corson, and J. Staudenmeier, 2023. NRCs Methodology to Estimate Fuel Dispersal during a Large Break Loss of Coolant Accident, Proceedings of the 20th International Topical Meeting on Nuclear Reactor Thermal Hydraulics (NURETH-20), Washington, DC, August 2023, pp. 5938-5951 (ADAMS Accession No. ML23116A214).

[4] WCAP-9500-A, Reference Core Report 17x17 Optimized Fuel Assembly (May 1982).

[5] DOE/RW-0184, Characteristics of Spent Fuel, High-Level Waste, and Other Radioactive Wastes Which May Require Long-Term Isolation," Appendix 2A Physical Descriptions of LWR Fuel Assemblies (December 1987).