Notice: This manuscript has been authored by UT-Battelle, LLC, under contract DE-AC05-00OR22725 with the US Department of Energy (DOE). The US government retains and the publisher, by accepting the article for publication, acknowledges that the US government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for US government purposes. DOE will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doe-public-access-plan).

Detailed SCALE Dose Rate Evaluations for a Consolidated Interim Spent Nuclear Fuel Storage Facility

Georgeta Radulescu, Thomas M. Miller, Kaushik Banerjee, Douglas E. Peplow

Oak Ridge National Laboratory 1 Bethel Valley Road, Oak Ridge, Tenn., 37831, United States

radulescug@ornl.gov; millertm@ornl.gov; banerjeek@ornl.gov; peplowde@ornl.gov

INTRODUCTION

This paper presents a confirmatory dose rate calculation for a proposed consolidated interim spent nuclear fuel (SNF) storage facility (CISF) [1] using SCALE 6.2.1 [2]. Calculating a dose to the public is a necessary part of the licensing process for the construction of an SNF storage facility. During normal operations and anticipated occurrences, the annual dose to the whole body received by any real individual who is located beyond the controlled area is restricted to 0.25 mSv (25 mrem) (10 CFR 72.104) [3]. This dose rate limit is typically used in shielding analyses to determine the physical boundary of the controlled area for a newly designed SNF storage facility.

The dose rate at a location external to a storage cask is produced by photon and neutron radiation escaping from the cask and reaching that location unobstructed or through multiple scattering interactions with nuclei/atoms in the environment (e.g., air, soil, concrete). The contribution of the scattered radiation to the total dose rate increases with increasing distance from the storage cask [4]. Far away from the radiation source, this component dominates the total dose rate and radiation interaction with the environment can only be properly simulated with radiation transport methods.

Ample volumes of air and soil extending beyond the location of interest for dose rate calculation are typically included in the calculation model to properly simulate important radiation scattering events that contribute to the dose rate at that location. Hence, adequate simulation of radiation transport for a large SNF storage facility comprising many hundreds of storage casks would require detailed models and significant computing resources.

A two-step method that requires modest computing resources is often used in license applications. The purpose of the first step of the method is to determine energy and angular distributions on the cask external surface of the neutrons and photons exiting a storage cask. The surface source is then used in a new radiation transport calculation to determine dose rate as a function of distance from the

storage facility. One drawback of this approach is that the space/energy/angle distribution of the particles coming off the cask needs to be binned, stored, and then resampled in the second step. The method presented in this paper uses a detailed Monte Carlo radiation transport simulation in one step (from source to dose rate), which is expected to yield more accurate results and be simpler to setup and run than the typical analysis previously used.

DESCRIPTION OF THE ACTUAL WORK

Phase 1 CISF Model

The Phase 1 CISF analyzed in this paper consists of 467 (319 vertical and 148 horizontal) SNF storage casks that can store up to 5,000 metric tons of SNF [5]. The dimensions of the concrete pad were approximately 244 m ´ 107 m ´ 30 cm. The design of the Phase 1 CISF incorporates five vertical and three horizontal storage cask systems currently used at various independent spent fuel storage installations to store pressurized water reactor (PWR) and boiling water reactor (BWR) commercial SNF assemblies. The storage pad is organized in 15 arrays of vertical casks and one array of horizontal modules.

Cask system drawings and design basis fuel characteristics documented in final safety analysis reports (FSARs) were used to develop a Phase 1 CISF model for SCALE calculations consistent with that analyzed in the license application [1]. A variety of fuel assembly designs (e.g., PWR array sizes from 14 ´ 14 to 17 ´ 17 and the BWR 7 ´ 7 array) and bounding radiation characteristics were defined in these FSARS. Totally, there were eight different design basis fuel assemblies with average burnup ranging from 25 GWd/MTU to 62 GWd/MTU and cooling time ranging from 3 years to 21.4 years. The PWR design basis SNF assembly typically includes irradiated non-fuel hardware such as control rods, burnable poisons, or thimble plug assemblies with significant photon sources, primarily from the activation of the cobalt impurity in steel alloys.

Radiation Protection and Shielding: General 765

Transactions of the American Nuclear Society, Vol. 118, Philadelphia, Pennsylvania, June 17–21, 2018

The design basis fuel assembly is typically modeled as four axial regions: the upper end fitting region, the upper gas plenum region, the active fuel region, and the lower end fitting region. Each axial region is represented as homogenous material within its geometry boundary with a mass density and elemental compositions derived from assembly hardware materials and weights. The fuel assembly burnup is typically represented with an 18-zone axial profile for the PWR SNF and a 10-zone axial profile for the BWR SNF.

Two-dimensional (2-D) views of a representative vertical cask and a representative horizontal module are presented in Fig. 1. This figure shows the degree of geometry detail represented in the cask models. A 3-D view of the Phase 1 CISF model is presented in Fig. 2. The thickness of the concrete pad was 30 cm and the soil was modeled to one meter in depth to account for radiation scattering and absorption.

a b Fig. 1. Vertical cross-section views of the models for a representative a) vertical cask and b) horizontal module.

Fig. 2. 3-D view of the Phase 1 concrete pad model.

Computer Codes

The source term and shielding calculation capabilities of the SCALE computer code system were used in the confirmatory dose rate calculations.

ORIGAMI [6], a SCALE code that performs fast ORIGEN depletion calculations using pre-generated ORIGEN cross-section libraries, was used to calculate the photon and neutron sources in the fuel region and the 60Co

activation source in fuel assembly hardware regions and non-fuel hardware.

Dose rate calculations were performed with the MAVRIC shielding analysis sequence in SCALE, which employs the state of the art hybrid variance reduction capabilities [7, 8] developed at Oak Ridge National Laboratory (ORNL) to generate high-fidelity dose results. MAVRIC features (1) multigroup and continuous-energy Monte Carlo transport capabilities, (2) automated variance reduction, (3) mesh, point detector, and region tallies, (4) simplified source term descriptions (e.g., direct use of radiation source terms from ORIGEN binary files), and (5) utilities for post-processing mesh tally files. A variance reduction method referred to as forward-weighted consistent adjoint driven importance sampling (FW-CADIS) enables estimation of dose rate with good statistical accuracy everywhere outside a heavily shielded system. This method (respWeighting) performs both forward and adjoint discrete ordinates calculations with the Denovo discrete ordinates code [9] to determine energy- and space-dependent source biasing and particle importance parameters.

Instrumental to the development of the fuel assembly and cask models was the TemplateEngine utility developed for the Used Nuclear Fuel-Storage, Transportation & Disposal Analysis Resource and Data System (UNF-ST&DARDS) [10], an integrated data and analysis tool for cask-specific SNF criticality, thermal, dose rate, and containment analyses. TemplateEngine is a string substitution program designed to conduct template imports and attribute replacement, thereby generating a complete input file.

MAVRIC Input Parameters

For the FW-CADIS method in MAVRIC, the user supplies a mesh for the coarse-mesh forward and adjoint deterministic calculations that are used to create the importance map and consistent biased source distribution. The mesh, while coarse compared to a production-level discrete ordinates calculation, still needs enough refinement to capture the major material changes and divide the source region for biasing. A mesh that captured all the details of the vast number of sources present in a CISF model would require more memory than current computers have. Since the contribution to the site boundary dose rate from one source is independent of the other sources (superposition principle), the strategy employed in this work was to parallelize on source.

A series of simulations was made using the complete site geometry (all casks present) but with only one cask containing source. For the importance map and source biasing calculations, a cask-specific Denovo mesh grid was used with finer mesh for the volume of the analyzed storage cask and its adjacent casks, but a coarse mesh elsewhere. A fine geometry mesh enables adequate representations of materials and radiation sources in most important regions of

Radiation Protection and Shielding: General766

Transactions of the American Nuclear Society, Vol. 118, Philadelphia, Pennsylvania, June 17–21, 2018

the problem (i.e., the analyzed cask). Equally important is the use a coarse Denovo mesh in the geometry regions where radiation interactions with matter are rare events, so that the computer time dedicated to those events is optimized.

Sensitivity of dose rate to important modeling parameters was also evaluated. The analyzed parameters included the volume of the surrounding air, Denovo mesh definition, parameters describing the angular scattering in Denovo SN calculations, tally mesh size, and computer time. The values for these parameters were selected to minimize computer time requirements per particle without compromising the quality of the calculated dose rate tally.

The volume of dry air beyond the expected location of the controlled area boundary was gradually increased until minimal effects were noticed on the total dose rate at that location. This volume of dry air around the concrete pad was approximately 2700 m ´ 2600 m ´ 960 m. The adjoint source was specified within a volume of air outside the storage pad and extending from the soil top to 2 m above the soil. This is the geometry region of the model where the dose rate is expected to be calculated with significant accuracy because variance reduction parameters were optimized for dose rate in this region.

The ANSI/ANS 6.1.1-1977 neutron and photon flux-to-dose-rate conversion factors [11] were applied to the flux estimates to obtain dose rates. Dose rates were calculated within a geometry mesh with a voxel size of 15.24 m ´ 15.24 m ´ 2 m. The mesh tally only had one cell in the vertical direction, which extended from the top of the soil to 2 m above the soil.

A total of 467 independent calculations were performed to obtain the dose rate maps produced by each storage cask. All cases were run on the Panacea cluster at ORNL. A single processor was used for each case. The specified total number of particles was 2 ´ 1010 and the selected maximum computer time was 1440 minutes.

Preliminary calculations indicated that all these parameters would ensure a maximum statistical relative error < 20% for the highest dose produced by each storage cask along the estimated controlled area boundary. Hence, a relative error of a few percent was expected for the total dose rate produced by the 467 storage casks.

RESULTS

The dose rate map for the entire facility, presented in Fig. 3, was obtained by summing the uncorrelated space-dependent dose rate values of the mesh tallies from the 467 separate calculations. The 25 mrem/yr dose rate contour corresponding to the controlled area boundary is delimited by a rectangle with the dimensions of 1.25 km ´ 1.25 km.

The contributions of the dose rate from individual storage casks to the total dose rate at a selected location on the controlled area boundary are shown in Fig. 4. As seen in the figure, contributions from individual casks at a specific

location vary considerably and the total dose rate may be dominated by several storage casks. A cask array reduces the effectiveness of cask external radiation source because a fraction of radiation escaping from individual casks is absorbed or attenuated in adjacent casks. Individual cask contribution depends on the cask shielding design, design basis radiation sources, and its relative location within the cask arrays. For the same cask system, peripheral casks have a larger contribution to the total dose rate at the control area boundary than casks placed at interior locations.

The typical dose rate error at the controlled area boundary was less than 10% (see Fig. 5). However, a large error bar (~40% uncertainty) was noticed for at least one cask, which is an indication that that particular calculation had not converged. The calculation for the impacted casks is planned to be repeated with a refined Denovo grid mesh/longer computer time.

Fig. 3. Total dose rate (mrem/h) surrounding the Phase 1 CISF. The black line encloses the 25 mrem/yr contour and denotes the smallest rectangle that could serve as the controlled area boundary. Total area shown is 2700 m ´ 2575 m.

Fig. 4. Contributions of the dose rate from individual casksto the total dose rate at a point along the south line of the controlled area boundary.

Radiation Protection and Shielding: General 767

Transactions of the American Nuclear Society, Vol. 118, Philadelphia, Pennsylvania, June 17–21, 2018

Fig. 5. Relative statistical error associated with dose rate values.

CONCLUSIONS

A proposed CISF was modeled in detail using SCALE 6.2.1 to determine the radiation dose map of the surrounding areas of the facility and determine the location of the controlled area boundary based on 10 CFR 72.104 requirements. The spatial dose rate distribution was determined as the total dose rate contributions by each of the 467 storage casks comprising the CISF. A rectangular boundary outside of which no dose rate exceeds 25 mrem/yr has the dimensions of 1.25 km ´ 1.25 km. It denotes the smallest rectangle that could serve as the controlled area boundary. The results (dose map) show the unique capability of the SCALE code system for a high-resolution radiation dose simulation of a large SNF storage facility.

ACKNOWLEDGMENTS

The work described in this paper was accomplished with funding provided by the US Nuclear Regulatory Commission.

REFERENCES

1. Application for a License for a Consolidated InterimSpent Fuel Storage Facility, US Nuclear RegulatoryCommission, Waste Control Specialists LLC, DocketNO. 72-1050, ADAMS ML16133A100 (2017).

2. B. T. REARDEN and M. A. JESSEE, Eds., “SCALECode System,” ORNL/TM-2005/39, Version 6.2, OakRidge National Laboratory (2016). Available fromRadiation Safety Information Computational Center asCCC-834.

3. 10 CFR 72, “License Requirements for the IndependentStorage of Spent Nuclear Fuel, High-Level RadioactiveWaste, and Reactor-Related Greater that Class CWaste,” US Government Publishing Office,Washington, DC (2006).

4. J. K. SHULTIS, “Hybrid Skyshine Calculations forComplex Neutron and Gamma Ray Sources,” Nucl. Sci.Eng. 136(2), 294–304 (2000).

5. “WCS Consolidated Interim Storage Facility SafetyAnalysis Report,” Waste Control Specialists LLC,ADAMS ML16133A134 (2017).

6. W. A. WIESELQUIST et al., “ORIGAMI AutomatorPrimer: Automated ORIGEN Source Terms and SpentFuel Storage Pool Analysis”, ORNL/TM-2015/409,Oak Ridge National Laboratory (2016).

7. D. E. PEPLOW, “Monte Carlo Shielding AnalysisCapabilities with MAVRIC,” Nucl. Technol. 174(2),289–313 (2011).

8. J. C. WAGNER, D. E. PEPLOW, and S. W. MOSHER,“FW-CADIS Method for Global and Regional VarianceReduction of Monte Carlo Radiation TransportCalculations,” Nucl. Sci. Eng., 176(1), 37-57 (2014).

9. T. M. EVANS et al., “Denovo: A New Three-Dimensional Parallel Discrete Ordinates Code InSCALE,” Nucl. Technol. 171(2), 171–200 (2010).

10. R. A. LEFEBVRE et al. “Development of StreamlinedNuclear Safety Analysis Tool for Spent Nuclear FuelApplications,” Nucl. Technol., 199(3), 227-244 (2017).

11. American National Standard Neutron and Gamma-RayFlux-to-Dose-Rate Factors, ANSI/ANS 6.1.1-1977,American Nuclear Society (1977).

Radiation Protection and Shielding: General768

Transactions of the American Nuclear Society, Vol. 118, Philadelphia, Pennsylvania, June 17–21, 2018