SAND2026-22860 SANDIA REPORT SAND2026-22860 Printed June 2026 Integrated Security-Safety Consequence Study of a Heat Pipe Reactor Christopher Faucett, Alan Evans, Ben Cipiti Prepared by Sandia National Laboratories Albuquerque, New Mexico 87185 and Livermore, California 94550 CONTENTS Abstract.................................................................................................................................................................3 Acknowledgements.............................................................................................................................................4 Executive Summary ............................................................................................................................................7 Acronyms and Terms .........................................................................................................................................8 1. Introduction ..................................................................................................................................................9 2. Modeling Approach ...................................................................................................................................10 2.1. Five Factor Formula........................................................................................................................10 2.2. MELCOR Modeling........................................................................................................................11 2.3. MACCS Modeling............................................................................................................................14 3. Modeling Results and Discussion............................................................................................................17 3.1. MELCOR Scenarios........................................................................................................................17 3.2. MACCS Results................................................................................................................................19 4. Conclusions and Path Forward................................................................................................................21 References ..........................................................................................................................................................22 Distribution........................................................................................................................................................23 LIST OF FIGURES Figure 1. MELCOR HPR reactor building nodalization............................................................................11 Figure 2. 95th quantile doses over distance for all MELCOR scenarios relative to 25 rem. .................19 LIST OF TABLES Table 1. MELCOR heat pipe reactor radionuclide class inventory. .........................................................12 Table 2. Recommended MACCS nuclides for LWR applications from NUREG/CR-7270 [3]. ........16 Table 3. MELCOR HPR scenarios................................................................................................................18 5 This page left blank 6 EXECUTIVE SUMMARY The U.S. Nuclear Regulatory Commission’s new 10 CFR 53 rulemaking enables certain physical security exemptions for advanced reactors if a licensee can demonstrate that consequences from a sabotage-induced radiological release would not exceed 25 rem at the site boundary two hours after release, without relying on mitigative actions. Because formal guidance for demonstrating compliance with this dose criterion is not yet available, integrated security/safety modeling methods are needed to support new reactor licensing. This report provides a preliminary demonstration of an integrated security/safety modeling approach for a hypothetical heat pipe microreactor subjected to direct sabotage using a 150 lb TNT- equivalent high-explosive device. A simplified radionuclide release estimate was developed using DOE-HDBK-3010-94 methods and implemented in the MELCOR severe accident analysis code to model transport and attenuation within reactor/building compartments. MELCOR outputs were then used as input to MACCS to compute offsite consequences, reported as D25rem, the distance from the release point at which dose falls below 25 rem. A set of six scenarios was evaluated to illustrate sensitivity to release pathway and building configuration and radionuclide physical form (vapor versus aerosols with different size distributions). Key observations include: • Building source term attenuation correlates with larger D25rem , as expected. • Reactor building compartmentalization strongly reduces consequences. Reducing the building from two stories to a single story lowered D25rem by roughly an order of magnitude compared to a cavity-only configuration, and the cavity-only case approached the consequences of a direct-to-environment release. • For the baseline two-story reactor building configuration, releasing radionuclides as vapor versus aerosols did not materially change LPF or D25rem on the spatial resolution used, though the smallest-bin aerosol case slightly increased LPF. Overall, the preliminary results suggest that the number of compartments/barriers between the reactor and the environment may be a dominant driver of sabotage consequence metrics such as D25rem . If supported by higher-fidelity shock physics and radionuclide transport analyses, this has implications for security design and adversary pathway evaluations, including tradeoffs between barriers that reduce releases and adversary resource allocation between creating leak paths and direct sabotage. However, the analysis relies on simplifying assumptions and data-limited representations of advanced-reactor sabotage response; results are provided as illustrative examples only and should not be used to draw conclusions about real advanced reactor designs’ ability to meet 10 CFR 53 criteria. Improved technical bases, better sabotage-response data, and more rigorous integrated modeling are needed to support credible exemption demonstrations. 7 ACRONYMS AND TERMS Acronym/Term Definition 10 CFR 53 Title 10 of the Code of Federal Regulations Parts 53 AR Advanced reactor ARF Airborne release fraction CF Control function CV Control volume DOE United States Department of Energy DR Damage ratio FP Flow path HED High-explosive device HPR Heat pipe-cooled microreactor LPF Leak path factor MAR Material-at-risk NRC United States Nuclear Regulatory Commission RB Reactor building RN Radionuclide (as used in the MELCOR code Radionuclide Package) TNT Trinitrotoluene 8 1. INTRODUCTION In anticipation of advanced reactor (AR) licensing, the United States Nuclear Regulatory Commission (NRC) has recently published rulemaking in Title 10 of the Code of Federal Regulations Part 53 (10 CFR 53) that allows for exemptions to physical security requirements if AR licensees can demonstrate that consequences from a sabotage-induced radiological source term do not exceed 25 rem at the site boundary 2 hours after the initial release without mitigative actions [1]. Currently, no guidance exists for demonstrating how the 10 CFR 53 dose criteria can be met. To facilitate AR licensing, new integrated security/safety modeling approaches are therefore needed. This report presents a preliminary demonstration of integrated security/safety modeling for direct sabotage of a hypothetical heat pipe-cooled microreactor (HPR) from a 150 lb. trinitrotoluene (TNT) equivalent high-explosive device (HED). Using United States Department of Energy (DOE) guidance from DOE-HDBK-3010-94, “Airborne Release Fractions/Rates and Respirable Fractions for Nonreactor Nuclear Facilities”, a simplified release scenario was modeled using the MELCOR severe accident analysis code. Results from MELCOR were used as input to MACCS, a consequence assessment code, to calculate D25rem i.e., the distance at which the dose from the radioactive plume released by adversary sabotage decreases below 25 rem. NOTE: The purpose of this work was to demonstrate how safety analysis tools could support justifying a security exemption under 10 CFR 53 rulemaking. However, because of data gaps regarding AR responses to direct sabotage, this work also made assumptions that may not reflect realistic reactor conditions or responses following a sabotage event. Therefore, all results should be treated as examples only, and the authors strongly advise against drawing conclusions about the feasibility of real-world advanced reactor designs meeting 10 CFR 53 requirements from this work. This report is organized as follows. Section 2 describes the MELCOR and MACCS modeling approach including the DOE-HDBK-3010-94 Five Factor Formula used to estimate the initial radionuclide (RN) release from an HED attack on the reactor vessel. Section 3 discusses preliminary results from the MELCOR and MACCS calculations. Finally, Section 4 summarizes high-level conclusions and highlights key areas for future work to improve the technical rigor of this preliminary analysis. 9 2. MODELING APPROACH The MELCOR and MACCS codes were used to calculate radionuclide transport to the environment and the resulting dose over distance curves, respectively [2]. The baseline sabotage scenario examined for this work was direct sabotage of a 5 MW HPR using a 150 lbs charge. The 150 lbs TNT equivalent charge was chosen as a conservative reference explosive based on subject matter expert opinion. In alignment with 10 CFR 53, no mitigating actions were modeled. 2.1. Five Factor Formula To simplify modeling, the Five Factor Formula from DOE-HDBK-3010-94 was used to approximate the radiological source term resulting from direct sabotage to reactor fuel [1]. The Five Factor Formula is: 𝑆 = 𝑀𝐴𝑅 ∗ 𝐷𝑅 ∗ 𝐴𝑅𝐹 ∗ 𝐿𝑃𝐹 ∗ 𝑅𝐹 Equation 1 Where S is the radiological source term, material-at-risk (MAR) is the radiological inventory, damage ratio (DR) is the fraction of the MAR impacted by accident conditions the airborne release fraction (ARF) is the fraction of radioactive material suspended as an aerosol that is available to transport to the environment, the leak path factor (LPF) is the fraction of radionuclides that transports out of containment/confinement to the environment, and respirable fraction (RF) is the fraction of radionuclides that can be inhaled into the human respiratory system (i.e., whose aerodynamic equivalent diameter is less than 10 μm). Ideally, shock physics simulations of reactor sabotage would be used to calculate the radiological source term resulting from direct sabotage. However, shock physics simulations are time consuming, computationally expensive, and require material equations of state (whose derivation for new materials like advanced reactor fuels is also labor intensive). Therefore, the following bounding approximation from DOE-HDBK-3010-94 was used to estimate MAR*DR [1]: Accordingly, for detonations in or contiguous to aggregate materials, a respirable release of the mass of inert material equal to the calculated TNT equivalent is assessed to be bounding. Using this approximation, DR was calculated as: 𝑀𝑇𝑁𝑇 𝐷𝑅 = Equation 2 𝑅𝑁𝑐𝑜𝑟𝑒 Where MTNT is the TNT equivalent mass converted from lbs to kg and RNcore is the whole-core radionuclide mass in kg. The Five Factor Formula parameters for this work are: • For a 150 lbs. TNT equivalent HED and a 4.57 MTU core, DR is equal to 1.487x10-2. • In the absence of a estimated particle size distribution estimate, RF was assumed to be 1. 10 • ARF was assumed to be 1. • LPF was calculated using MELCOR. The assumption that ARF and RF are both equal to 1 is inherently conservative as it minimizes radionuclide retention in the reactor vessel and maximizes dose from radionuclides released to the environment. 2.2. MELCOR Modeling The MELCOR HPR model developed for the MELCOR Advanced Reactor Source Term Demonstration project was used as the representative microreactor design [1]. The MELCOR HPR is a 5 MWth heat pipe-cooled reactor. The HPR core mass is 4.57 MTU. At end of life, the HPR burnup is 1.998 GWd/MTU. The HPR is housed in a below-grade cavity underneath a two-story reactor building (RB). The reactor cavity and RB floors are connected by a series of flow paths (FPs) to enable natural circulation. The reactor vessel and RB both were modeled with FPs representing leak paths to the RB and environment, respectively. Figure 1. MELCOR HPR reactor building nodalization. To simplify radionuclide transport, MELCOR groups radionuclides into classes of chemically similar elements1 [2]. The HPR inventory at 1.998 GWd/MTU inventory is listed inTable 1 where the element in parenthesis is the representative element for the class. With respect to the Five Factor Formula, MAR is the summed mass of all radionuclide classes (i.e., the sum of all rows inTable 1). 1 The MELCOR RN class naming convention uses the uppercase representative element to refer to all elements within the class. For example, CS refers to all elements in the MELCOR CS class, not Cs specifically. 11 Table 1. MELCOR heat pipe reactor radionuclide class inventory. MELCOR RN Class RN Class Mass (kg) Noble Gases (Xe) 1.323E+00 Alkali Metals (Cs) 1.184E+00 Alkaline Earths (Ba) 7.211E-01 Halogens (I) 5.265E-02 Chalcogens (Te) 1.383E-01 Platinoids (Ru) 7.598E-01 Early Transition Elements (Mo) 1.226E+00 Tetravalent (Ce) 6.755E+00 Trivalents (La) 2.287E+00 Uranium (U) 4.561E+03 More Volatile Main Group (Cd) 1.083E-02 Less Volatile Main Group (Ag) 1.697E-02 As a lumped parameter control volume (CV) code, MELCOR cannot capture direct sabotage phenomena like shockwave propagation. Rather than attempting to simulate the physical response of the HPR vessel to an explosive attack, this work focused on simulating the post-detonation blowdown following the explosion-induced vessel overpressure. A simplified conservation of mass and energy approach was used to approximate the vessel overpressure transient: • Complete combustion of the 150 lbs. TNT equivalent HED was assumed (i.e., no soot was formed). The duration of the HED detonation2 transient was 1 s. • A time-independent source CV was connected to the reactor vessel CV by velocity-specified flow path (FP). The time-independent CV was entirely filled with dry air (80% N2, 20% O2) as a surrogate for combustion products3. The gas temperature of the source CV was set by a control function (CF) to match the ambient temperature of the reactor vessel CV atmosphere. • The energy of detonation of the 150 lb. TNT equivalent HED was assumed to deposit entirely into reactor fuel4. To model this, the HED was converted to a “heat source” by 2 The actual reaction rate of a HED is on the order of μs. However, MELCOR is not designed to resolve μs time scales, a 1 s transient was modeled for code stability. Given that the timescale of interest for simulations was 2 hours, the duration of the explosive transient is not expected to have impacted results. 3 The actual detonation of an HED will produce a mixture of combustion products (H O, CO, CO , N and carbon 2 2 2,, soot). Because MELCOR treats non-condensable gas thermal-hydraulics identically, dry air was used to simplify modeling. 4 Realistically, the energy produced by detonation of the HED will deposit not only in fuel material, but also structural material inside the reactor vessel, gases filling the free volume of the reactor core between fuel elements, and the reactor vessel wall itself. Without running shock physics simulations, it is impossible to predict the amount of energy deposited into each of the different reactor materials. 12 divided the energy released by HED combustion by the duration of the detonation transient. The HED “heat source” was added to the HPR’s decay heat curve in MELCOR over the 1 s detonation transient. This approach neglects spatial effects of the HED as the HED energy is deposited over the entire core volume. This approach also neglects the internal energy added to the HPR model from surrogate combustion gas source5. • The total energy released by 150 lb. TNT equivalent was converted to power by dividing the total energy by the 1 s (i.e., the duration of the detonation period) and adding the HED “heat source” to the reactor’s decay heat curve in MELCOR. The HED “heat source” is only active during over the 1 s detonation transient • The velocity of the velocity-specified FP was set so the time-integrated mass flow rate over the initial overpressure equaled 150 lbs. The HPR model was run for a 20,000 s initialization period prior to the transient to allow materials in the reactor vessel to reach steady-state temperature conditions. The transient begins at 0 s. Beginning at 0 s, the following events were modeled: • Fission products from all RN classes were sourced into the reactor core CV at a linear rate over the 1 s transient period. The per-class mass of RN sourced into the MELCOR model was equal to the class mass scaled by DR calculated using Equation 2. The original RN inventory in the MELCOR model was not reduced to account for the RN source, meaning a small amount of additional decay heat was added when fission products were sourced in. • Because the aerosol size distribution resulting from sabotage of HPR fuel elements is unknown, fission products were sourced into the HPR vessel as vapor (i.e., an underlying aerosol size distribution was not assumed). Normally, the release of fission products as vapor would be a conservative assumption as fission product vapors in MELCOR are more mobile than aerosols. Aerosol agglomeration was not disabled, so vapors were able to form aerosols after they were sourced in. • A 6” diameter hole representing the vessel breach from the explosion opens when the transient initiates. This implicitly assumes that the HED penetrates through the entire core and creates an exit hole. The presence of an exit hole induces flow that carries material directly out of the vessel. Compared to a partial, single penetration, full vessel penetration can increase the aerosol release by a factor of 10. To promote natural circulation following the overpressure transient, the 6” vessel penetration was modeled using 2 FPs representing the upper and low halves of the flow area. Countercurrent flow effects were not modeled. • During modeling, it was observed that MELCOR’s lower (reactor vessel) head failure (LHF) model was predicting lower head failure produced by strain in the lower head from the vessel overpressure. To promote vessel leakage, 2 FPs representing the lower head breach were also added to the model. The FP flow areas are dynamically set so that the sum of the flow areas equals the LHF breach size predicted by MELCOR. The MELCOR HPR model includes a user-defined pressure gradient across RB to environment leak paths that represents Bernoulli pressure generated from external wind effects. Wind effects are often 5 The velocity of surrogate combustion gases was calculated to be approximately 387 m/s by MELCOR. At this velocity, the kinetic energy of 150 lbs. of combustion products (~68 kg) is about 4.98 MJ. For comparison, complete detonation of 1 kg of TNT produces 4.184 MJ. Internal energy of the surrogate combustion gases should be considered in future work. 13 included in safety analyses when internal building driving forces like reactor decay heat are weak. Details of how the wind pressure gradient is calculated are available in the HPR report. The wind speed was assumed to be 0 mph for this work for MELCOR simulations, so no external pressure gradient was applied. Potentially important phenomena that were not modeled include: • Sodium combustion from failed heat pipes: It is likely that direct sabotage to the reactor core will fail heat pipes, causing sodium to leak and oxidize. Energy released by sodium combustion could drive the release out of the vessel or the reactor building. Given the low decay heat of the HPR, additional energy sources from a sodium fire could be impactful. • Dose emitting radionuclides in activated cladding, structural materials, or sodium: The activity of activated non-fuel materials was assumed to be inconsequential compared to irradiated fuel. Not modeling releases of activated structural materials is a normal severe accident modeling practice for conventional reactors, although the authors note that severe accident modeling practices for advanced reactors are still being developed. • Radionuclide release and blowdown timing: The radionuclides and combustion gases were concurrently sourced into the reactor vessel. Realistically, there will be a time delay between the initial overpressure and the release of fission products where fission products are entrained and swept out of the reactor vessel as opposed to being pushed out while they are being generated. • Heat pipe failure: Compression of heat pipes from the HED shockwave could result in heat pipe failure along the heat pipe including in the condenser region (assuming the heat pipe working fluid behaves like an ideal gas6). Heat pipe failure in the condenser region would create a confinement bypass leak path where radionuclides could transport directly to the environment. LPF was calculated for each radionuclide class using Equation 3: 𝑀𝑒𝑛𝑣, 𝑖 𝐿𝑃𝐹𝑖 = Equation 3 𝑀𝐴𝑅𝑖 ∗ 𝐷𝑅 Where Menv,i and MARi are the mass of radionuclide class i in the environment after 2 hours and the radionuclide class inventory (i.e., MAR) from Table 1. 2.3. MACCS Modeling The MACCS consequence analysis code was used to predict D25rem using the release fractions predicted by MELCOR. Developed by Sandia National Laboratories for the NRC, MACCS is a consequence modeling tool used to simulate the atmospheric transport and dispersion of radionuclides released to the environment and their associated consequences [5]. MACCS has 6 The MELCOR HPR model includes a heat pipe creep failure model where creep accumulation is calculated using the heat pipe’s internal pressure and wall temperature. If the heat pipe working fluid behaves like an ideal gas and the heat pipe does not rupture, the reduction in free volume from shock compression will increase the internal pressure, resulting in elevated creep and potential creep failure. 14 detailed modeling capabilities including simulation of meteorology, protective actions, health effects, and economic factors. For this work, the meteorology and health effects models were used. The MACCS HPR consequence model published by Nosek in “MACCS Consequence Analysis Demonstration Calculations for an Example Heat Pipe Reactor Source Term” was used for this work [6]. Nosek’s input was based developed to demonstrate HPR consequence modeling using best practice values from NUREG/CR-7270, “Technical Bases for Consequence Analysis Using MACCS (MELCOR Accident Consequence Code System)”. A complete description of Nosek’s input is available in its companion report. Details of Nosek’s input include: • Releases were modeled as a point source. • All aerosols were assumed to be in the smallest particle size bin. • The Regulatory Guide 1.145 Full Model was used for nearfield modeling. • A phantom, non-evacuating cohort was used to evaluate ambient dose. The phantom cohort takes no protective actions and receives no shielding. After 2 h, the phantom cohort receives perfect shielding so no further dose is accumulated7. For this analysis, the standard radionuclides recommended for light water reactor applications in NUREG/CR-7270 shown in Table 1 were used. It should be noted that nuclides shown in Table 1 for LWR applications may not necessarily account for unique isotopic inventories of advanced reactors that could significantly impact health consequences. The following changes were made to Nosek’s input: • Weather conditions were sampled using stratified sampling with 1-hour intervals (i.e., 24 samples per day accounting for all weather conditions at a site). Weather conditions were sampled from 2025 site data for the Surry Power Station. • The Power Model for plume buoyancy was used. The decision to use the Power Model was a holdover from earlier versions of this work that did not use MELCOR where values for mass release rates and plume densities needed for the Density and Flow Model were not available. • The source term was modeled as a single plume segment. A linear release over 3600 s was assumed based on MELCOR results. • The release height was assumed to be 1 m from the ground as it is expected that adversaries would breach the reactor building close to the ground (i.e., as opposed to the roof). • The spatial grid was redefined to use logarithmic intervals with 0.1 km increments between 0 to 1 km, 1 km increments from 1 to 10 km, and 10 km increments from 10 to 100 km. 7 During model review, errors were found in the MACCS input where shielding was incorrectly applied over the first hour of the scenario, not after the first two hours as intended. Additionally, the phantom cohort began evacuating after the first hour had passed. These errors were corrected during modeling so shielding was only applied after two hours as intended. Correcting the errors increased D25rem by a factor of approximately 10x. 15 Table 2. Recommended MACCS nuclides for LWR applications from NUREG/CR-7270 [3]. MACCS Nuclides Am-241 Ba-137m Ba-139 Ba-140 Ce-141 Ce-143 Ce-144 Cm-242 Cm-244 Co-58 Co-60 Cs-134 Cs-136 Cs-137 I-131 I-132 I-133 I-134 I-135 Kr-85 Kr-85m Kr-87 Kr-88 La-140 La-141 La-142 Mo-99 Nb-95 Nb-97 Nb-97m Nd-147 Np-239 Pr-143 Pr-144 Pr-144m Pu-238 Pu-239 Pu-240 Pu-241 Rb-86 Rb-88 Rh-103m Rh-105 Rh-106 Ru-103 Ru-105 Ru-106 Sb-127 Sb-129 Sr-89 Sr-90 Sr-91 Sr-92 Tc-99m Te-127 Te-127m Te-129 Te-129m Te-131 Te-131m Te-132 Xe-133 Xe-135 Xe-135m Y-90 Y-91 Y-91m Y-92 Y-93 Zr-95 Zr-97 16 3. MODELING RESULTS AND DISCUSSION The MELCOR and MACCS codes were used to estimate D25rem for estimated direct sabotage source terms from the HPR. The MACCS results are presented as 95th quantile of peak doses calculated over each interval of the spatial grid over all weather samples8. NOTE: The results shown in this section were produced by simulations with conservative, unvalidated assumptions and inputs that may not reflect realistic reactor conditions or responses following a sabotage event. The authors advise against drawing conclusions about the feasibility of real-world advanced reactor designs meeting 10 CFR 53 security requirements from these results. 3.1. MELCOR Scenarios To demonstrate consequence modeling capabilities, a set of scenarios were run with initial conditions designed to promote RN transport to the environment. The scenarios are listed in Table 3 with a brief of the modeled initial conditions, LPFs for XE and non-XE RN classes9, and D25rem. The scenarios modeled were: • Scenario 1 (Full release): The RN source term was released directly to the environment with no attenuation from the reactor building i.e., LPF equals 1.0. Scenario 1 was modeled for comparison purposes only. • Scenario 2 (Base scenario): The MELCOR HPR model was run as described in Section 2.2 (RN sourced in as vapor, 2 story RB above reactor cavity). The pathway for RN to transport to the environment includes the reactor cavity and both floors of the RB. • Scenario 3 (Cavity-only RB) and Scenario 4 (Single story RB): To investigate how different RB configurations could affect LPF, two scenarios were run with a modified RB – Scenario 3 where the reactor cavity leaks directly to the environment and Scenario 4 where the RB was reduced to a single floor above the reactor cavity. • Scenario 5 (Smallest-bin aerosol release) and Scenario 6 (Uniform aerosol release): RN released from direct sabotage of reactor fuel are expected to be in the form of aerosols. Aerosols are less mobile than vapors, meaning an aerosol source term is expected to have a smaller LPF. To investigate the effects of different aerosol size distributions, two scenarios were run where RN were sourced as aerosols. Scenario 5 assumed direct sabotage produced aerosols only in the smallest MELCOR aerosol size bin10, and Scenario 6 assumed aerosols were uniformly distributed across all size bins. XE was sourced as an aerosol for both scenarios. 8 When performing weather sampling, MACCS calculates statistical values for peak doses including the mean, the absolute peak dose, and various quantiles. Normally, the 95th quantile is reported. 9 RN in the MELCOR XE class do not form aerosols and, as a result, have a higher LPF than the other RN classes. 10 MELCOR bins aerosols into user-defined size bins. The aerosol size bins are logarithmically distributed between a user-defined minimum and maximum aerosol geometric diameters. Aerosols larger than the maximum user-defined aerosol diameter are assumed to settle out. 17 Table 3. MELCOR HPR scenarios. Leak Path Factor (-) Scenario Scenario Name Scenario Description D25rem Number XE Other RN Direct release of source term to environment modeled in 1 Full release 1.0 1.0 10 to 30 km MACCS. 2 Base scenario MELCOR HPR model run as described in Section 2.2. 0.0058 0.0042 to 0.0045 0.6 to 0.8 km RB reduced to only the cavity. RB leak paths were redefined 3 Cavity-only RB 0.99 0.60 to 0.61 10 to 30 km to connect reactor cavity directly to environment. RB reduced to only the cavity and Floor 1. RB leak paths were redefined to connect Floor 1 of the RB to 4 Single story RB 0.99 0.079 to 0.083 6 to 8 km environment. FPs connecting the cavity to the first floor of the RB were unchanged. Smallest-bin RN sourced in as aerosol particles in the smallest aerosol 5 0.0052 0.0042 to 0.0044 0.6 to 0.8 km aerosol release. size bin. Uniform aerosol RN source in as aerosol with particles distributed evenly 6 0.004 0.0038 to 0.0041 0.6 to 0.8 km release across all bins (i.e., a log-uniform distribution). 18 3.2. MACCS Results Figure 2. 95th quantile doses over distance for all MELCOR scenarios relative to 25 rem. Figure 2 shows the 95th quantile peak dose over distance calculated by MACCS for the six MELCOR scenarios listed in Table 3. The following results were observed: • LPF and D25rem showed a positive correlation. This result was expected as larger RN releases should produce higher doses at the same distance. • RN attenuation in the RB created a noticeable reduction in D25rem. D25rem for the single-story RB model was a factor of 10 lower than the cavity-only model. D25rem for the cavity-only scenario was effectively the same as a direct release to the environment. The dose for the direct release scenario was slightly higher than the cavity-only scenario which indicates some RN attenuation from the reactor cavity. • Sourcing RN as vapor or aerosols did not influence LPF for the base 2 story RB configuration. Sourcing in RN in the smallest aerosol size bin slightly increased LPF but did not produce an observable difference D25rem on the spatial grid used for this work. Tentatively, the results of the MACCS simulations suggest that the number of compartments separating the reactor from the environment has the strongest influence on D25rem. If this conclusion holds, this has multiple implications for security design and adversary pathway analyses. With respect to security design, increasing the number of barriers or compartments between the reactor and the 19 environment could require adversaries to use explosives (bounded by the design basis threat) to create leak pathways which reduces the lbs. of explosives available for direct sabotage. This creates an optimization problem for pathway analyses where identifying the “worst case scenario” requires balancing LPF against DR. However, stronger technical bases are needed before definitive conclusions can be drawn from this type of analysis. 20 4. CONCLUSIONS AND PATH FORWARD This report provides a preliminary demonstration of an integrated security/safety modeling approach for a hypothetical heat pipe microreactor subjected to direct sabotage using a 150 lb TNT- equivalent high-explosive device. A simplified radionuclide release estimate was developed using DOE-HDBK-3010-94 methods and implemented in the MELCOR severe accident analysis code to model transport and attenuation within reactor/building compartments. MELCOR outputs were then used as input to MACCS to compute offsite consequences, reported as D25rem, the distance from the release point at which dose falls below 25 rem. Overall, the preliminary results suggest that the number of compartments/barriers between the reactor and the environment may be a dominant driver of sabotage consequence metrics such as D25rem . If supported by higher-fidelity shock physics and radionuclide transport analyses, this has implications for security design and adversary pathway evaluations, including tradeoffs between barriers that reduce releases and adversary resource allocation between creating leak paths and direct sabotage. As a preliminary demonstration analysis, there are multiple areas where research is needed to build a defensible technical basis suitable for licensing activities. • Shock physics simulations of advanced reactor direct sabotage are needed to estimate DR, ARF, and RF. Coupling of shock physics codes with MELCOR may be required to accurate simulate the RB response to detonation of an HED. • Estimates for LPF could be refined by using to simulate the reactor vessel response following direct sabotage. Hot gases expelled from the reactor vessel following detonation of an HED are expected to create buoyant internal plumes inside the RB that would more readily transport to the environment. Capturing this phenomenon will require coupled explosion and computational fluid dynamics simulations. • The temperatures produced by HED interactions with fuel may volatilize fission products, further increasing the release from sabotage fuel. Integration of fuel particulate temperature histories derived from shock physics calculations and fuel performance codes (e.g., BISON) may be necessary to accurately predict the total RN release from direct sabotage. Ideally, experimental validation of direct sabotage simulations will be also performed. However, given the costs and timeframes associated with large experimental campaigns compared to near-term advanced reactor licensing timelines, heavy reliance on modeling/simulation to support advanced reactor licensing may be required. 21 REFERENCES [1] United States Nuclear Regulatory Commission, "Risk-Informed, Technology-Inclusive Regulatory Framework for Advanced Reactors," Washington, D.C., 2026. [2] L. L. Humphries, B. A. Beeny, F. Gelbard, D. L. Louie, J. Phillips, R. C. Schmidt and N. E. Bixler, "MELCOR Computer Code Manuals, Vol. 1: Primer and Users’ Guide, Version 2.2.18019," Sandia National Laboratories, Albuquerque, NM and Livermore, CA, 2021. [3] United States Department of Energy, "Airborne Release Fractions/Rates and Respirable Fractions For Nonreactor Nuclear Facilities Volume I - Analysis of Experimental Data," United States Department of Energy, Washington, D.C., 1994. [4] K. Wagner, C. Faucett, R. Schmidt and D. Luxat, "MELCOR Accident Progression and Source Term Demonstration Calculations for a Heat Pipe Reactor," Sandia National Laboratories, Albuquerque, NM, 2022. [5] Sandia National Laboratories, "About MACCS," 2026. [Online]. Available: https://maccs.sandia.gov/software/maccs. [Accessed 12 June 2026]. [6] A. J. Nosek, "MACCS Consequence Analysis Demonstration Calculations for an Example Heat Pipe Reactor Source Term," U.S. Nuclear Regulatory Commission, Washington, D.C., 2023. [7] N. Bixler, K. Compton, M. Dennis, L. Eubanks, R. Haaker, J. Jones, M. Kimura, K. McFadden, A. Nosek, A. Outkin and F. Walton, "Technical Bases for Consequence Analyses Using MACCS (MELCOR Accident Consequence Code System)," U.S. Nuclear Regulatory Commission Office of Nuclear Regulatory Research, Washington, D.C., 2022. 22 DISTRIBUTION Email—Internal Name Org. Sandia Email Address Technical Library 1911 sanddocs@sandia.gov Bland, Jesse John 6817 jjbland@sandia.gov Pickrell, Brent 8854 bmpickr@sandia.gov Email—External Name Company Email Address Company Name Jim Beardsley jdb5@nrc.gov Nuclear Regulatory Commission Kase Brock kase.brock@inl.gov Idaho National Laboratory Chris Bunting christopher.bunting@nuclear.energy.gov Department of Energy John McKirgan john.mckirgan@nrc.gov Nuclear Regulatory Commission Stacy Prasad stacy.prasad@nrc.gov Nuclear Regulatory Commission Joe A. Renevitz joseph.renevitz@inl.gov Idaho National Laboratory Anthony Ulses anthony.ulses@nrc.gov Nuclear Regulatory Commission Dan Warner daniel.warner@nuclear.energy.gov Department of Energy 23 This page left blank 24 Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia LLC, a wholly owned subsidiary of Honeywell International Inc. for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.