Publications

The Integrated Tiger Series (ITS) generates a database containing energy deposition data. This data, when stored on an Exodus file, is not typically suitable for analysis within Sierra Mechanics for finite element analysis. The its2sierra tool maps data from the ITS database to the Sierra database.

NasGen provides a path for migration of structural models from NASTRAN bulk data format (BDF) into both an Exodus mesh file and an ASCII input file for Sierra Structural Dynamics (Salinas) and Solid Mechanics (Presto). Many tools at Sandia National Labs (SNL) use the Exodus format. NasGen was written specifically for Salinas and Presto but should be usable with a number of these packages.

We present a high-level architecture for how artificial intelligences might advance and accumulate scientific and technological knowledge, inspired by emerging perspectives on how human intelligences advance and accumulate such knowledge. Agents advance knowledge by exercising a technoscientific method—an interacting combination of scientific and engineering methods. The technoscientific method maximizes a quantity we call “useful learning” via more-creative implausible utility (including the “aha!” moments of discovery), as well as via less-creative plausible utility. Society accumulates the knowledge advanced by agents so that other agents can incorporate and build on to make further advances. The proposed architecture is challenging but potentially complete: its execution might in principle enable artificial intelligences to advance and accumulate an equivalent of the full range of human scientific and technological knowledge.

Efficient prediction of sampling-intensive thermodynamic properties is needed to evaluate material performance and permit high-throughput materials modeling for a diverse array of technology applications. To alleviate the prohibitive computational expense of high-throughput configurational sampling with density functional theory (DFT), surrogate modeling strategies like cluster expansion are many orders of magnitude more efficient but can be difficult to construct in systems with high compositional complexity. We therefore employ minimal-complexity graph neural network models that accurately predict and can even extrapolate to out-of-train distribution formation energies of DFT-relaxed structures from an ideal (unrelaxed) crystallographic representation. This enables the large-scale sampling necessary for various thermodynamic property predictions that may otherwise be intractable and can be achieved with small training data sets. Two exemplars, optimizing the thermodynamic stability of low-density high-entropy alloys and modulating the plateau pressure of hydrogen in metal alloys, demonstrate the power of this approach, which can be extended to a variety of materials discovery and modeling problems.

This work presents a 3D quantum mechanics based model to address the physics at band structure crossing/anti-crossing points in full band Monte Carlo (FBMC) simulations. The model solves the Krieger and Iafrate (KI) equations in real time using pre-computed coefficients at k-points spatially sampled within the first Brillouin zone. Solving the KI equations in real time makes this model applicable for all electric fields, which enables its use in FBMC device simulations. In this work, a two-level refinement scheme is used to aggressively sample regions in proximity to band crossings for accurate solutions to the KI equations and coarsely sample everywhere else to limit the number of k-points used. The presented sampling method is demonstrated on the band structure of silicon but is effective for the band structure of any semiconductor material. Next, the adaptation of the fully quantum KI model into the semi-classical FBMC method is discussed. Finally, FBMC simulations of hole transport in 4H silicon carbide with and without the KI model are performed. Results along different crystallographic directions for a wide range of electric fields are compared to previously published simulation and experimental values.

Both shock and shockless compression experiments were performed on laser powder bed fusion (LPBF) Ti-5Al-5V-5Mo-3Cr (Ti-5553) to peak compressive stresses near 15 GPa. Experiments were performed on the as-built material, containing a purely β (body centered cubic) microstructure, and two differing heat treatments resulting in a dual phase α (hexagonal close packed) and β microstructure. The Hugoniot, Hugoniot elastic limit (HEL), and spallation strength were measured and compared to wrought Ti-6Al-4V (Ti-64). The results indicate the LPBF Ti-5553 Hugoniot response is similar between heat treatments and to Ti-64. The HEL stress observed in the LPBF Ti-5553 was considerably higher than Ti-64, with the as-built, fully β alloy exhibiting the largest values. The spallation strength of the LPBF Ti-5553 was also similar to Ti-64. Clear evidence of initial porosity serving as initiation sites for spallation damage was observed when comparing computed tomography measurements before and after loading. Post-mortem scanning electron microscopy images of the recovered spallation samples showed no evidence of retained phase changes near the spall plane. The spall plane was found to have kinks aligned with the loading direction near areas with large concentrations of twin-like, crystallographic defects in the as-built condition. For the heat-treated samples, the concentrations of twin-like, crystallographic defects were absent, and no preference for failure at the interface between the α and β phases was observed.

In lithium-metal batteries, grains of lithium can become electrically isolated from the anode, lowering battery performance. We state that experiments reveal that rest periods after battery discharge might help to solve this problem.

Replacement of conventional petroleum fuels with renewable fuels reduces net emissions of carbon and greenhouse gases, and affords opportunities for increased domestic energy security. Here, we present alkyl dialkoxyalkanoates (or DAOAs) as a family of synthetic diesel and marine fuel candidates that feature ester and ether functionality. These compounds employ pyruvic acid and fusel alcohols as precursors, which are widely available as metabolic intermediates at high titer and yield. DAOA synthesis proceeds in high yield using a simple, mild chemical transformation performed under air that employs bioderived and/or easily recovered reagents and solvent. The scalability of the synthetic protocol was proven in continuous flow with in situ azeotropic water removal, yielding 375 g of isolated product. Chemical stability of DAOAs against aqueous 0.01 M H2SO4 and accelerated oxidative conditions is demonstrated. The isolated DAOAs were shown to meet or exceed widely accepted technical criteria for sustainable diesel fuels. In particular, butyl 2,2-dibutoxypropanoate (DAOA-2) has indicated cetane number 64, yield soot index 256 YSI per kg, lower heating value 30.9 MJ kg−1 and cloud point < −60 °C and compares favorably to corresponding values for renewable diesel, biodiesel and petroleum diesel.

As ferroelectric hafnium zirconium oxide (HZO) becomes more widely utilized in ferroelectric microelectronics, integration impacts of intentional and nonintentional dielectric interfaces and their effects upon the ferroelectric film wake-up (WU) and circuit parameters become important to understand. In this work, the effect of the addition of a linear dielectric aluminum oxide, Al2O3, below a ferroelectric Hf0.58Zr0.42O2 film in a capacitor structure for FeRAM applications with niobium nitride (NbN) electrodes was measured. Depolarization fields resulting from the linear dielectric is observed to induce a reduction of the remanent polarization of the ferroelectric. Addition of the aluminum oxide also impacts the WU of the HZO with respect to the cycling voltage applied. Intricately linked to the design of a FeRAM 1C/1T cell, the metal-ferroelectric-insulator-metal (MFIM) devices are observed to significantly shift charge related to the read states based on aluminum oxide thickness and WU cycling voltage. A 33% reduction in the separation of read states are measured, which complicates how a memory cell is designed and illustrates the importance of clean interfaces in devices.

Recent findings suggest that ions are strongly correlated in atmospheric pressure plasmas if the ionization fraction is sufficiently high ( ≳ 10 − 5 ). A consequence is that ionization causes disorder-induced heating (DIH), which triggers a significant rise in ion temperature on a picosecond timescale. This is followed by a rise in the neutral gas temperature on a longer timescale of up to nanoseconds due to ion-neutral temperature relaxation. The sequence of DIH and ion-neutral temperature relaxation suggests a new mechanism for ultrafast neutral gas heating. Previous work considered only the case of an instantaneous ionization pulse, whereas the ionization pulse extends over nanoseconds in many experiments. Here, molecular dynamics simulations are used to analyze the evolution of ion and neutral gas temperatures for a gradual ionization over several nanoseconds. The results are compared with published experimental results from a nanosecond pulsed discharge, showing good agreement with a measurement of fast neutral gas heating.

A reduction in wake effects in large wind farms through wake-aware control has considerable potential to improve farm efficiency. This work examines the success of several emerging, empirically derived control methods that modify wind turbine wakes (i.e., the pulse method, helix method, and related methods) based on Strouhal numbers on the (Formula presented.). Drawing on previous work in the literature for jet and bluff-body flows, the analyses leverage the normal-mode representation of wake instabilities to characterize the large-scale wake meandering observed in actuated wakes. Idealized large-eddy simulations (LES) using an actuator-line representation of the turbine blades indicate that the (Formula presented.) and (Formula presented.) modes, which correspond to the pulse and helix forcing strategies, respectively, have faster initial growth rates than higher-order modes, suggesting these lower-order modes are more appropriate for wake control. Exciting these lower-order modes with periodic pitching of the blades produces increased modal growth, higher entrainment into the wake, and faster wake recovery. Modal energy gain and the entrainment rate both increase with streamwise distance from the rotor until the intermediate wake. This suggests that the wake meandering dynamics, which share close ties with the relatively well-characterized meandering dynamics in jet and bluff-body flows, are an essential component of the success of wind turbine wake control methods. A spatial linear stability analysis is also performed on the wake flows and yields insights on the modal evolution. In the context of the normal-mode representation of wake instabilities, these findings represent the first literature examining the characteristics of the wake meandering stemming from intentional Strouhal-timed wake actuation, and they help guide the ongoing work to understand the fluid-dynamic origins of the success of the pulse, helix, and related methods.

Pyrimidine has two in-plane CH(δ+)/N̈(δ–)/CH(δ+) binding sites that are complementary to the (δ–/2δ+/δ–) quadrupole moment of CO₂. For this study, we recorded broadband microwave spectra over the 7.5–17.5 GHz range for pyrimidine-(CO₂)_n with n = 1 and 2 formed in a supersonic expansion. Based on fits of the rotational transitions, including nuclear hyperfine splitting due to the two ¹⁴N nuclei, we have assigned 313 hyperfine components across 105 rotational transitions for the n = 1 complex and 208 hyperfine components across 105 rotational transitions for the n = 2 complex. The pyrimidine-CO₂ complex is planar, with CO₂ occupying one of the quadrupolar binding sites, forming a structure in which the CO₂ is stabilized in the plane by interactions with the C–H hydrogens adjacent to the nitrogen atom. This structure is closely analogous to that of the pyridine-CO₂ complex studied previously by (Doran, J. L. J. Mol. Struct. 2012, 1019, 191–195). The fit to the n = 2 cluster gives rotational constants consistent with a planar cluster of C_2v symmetry in which the second CO₂ molecule binds in the second quadrupolar binding pocket on the opposite side of the ring. The calculated total binding energy in pyrimidine-CO₂ is –13.7 kJ mol^–1, including corrections for basis set superposition error and zero-point energy, at the CCSD(T)/ 6-311++G(3df,2p) level, while that in pyrimidine-(CO₂)₂ is almost exactly double that size, indicating little interaction between the two CO₂ molecules in the two binding sites. The enthalpy, entropy, and free energy of binding are also calculated at 300 K within the harmonic oscillator/rigid-rotor model. This model is shown to lack quantitative accuracy when it is applied to the formation of weakly bound complexes.

Detecting changepoints in functional data has become an important problem as interest in monitoring of climate phenomenon has increased, where the data is functional in nature. The observed data often contains both amplitude ((Formula presented.) -axis) and phase ((Formula presented.) -axis) variability. If not accounted for properly, true changepoints may be undetected, and the estimated underlying mean change functions will be incorrect. In this article, an elastic functional changepoint method is developed which properly accounts for these types of variability. The method can detect amplitude and phase changepoints which current methods in the literature do not, as they focus solely on the amplitude changepoint. This method can easily be implemented using the functions directly or can be computed via functional principal component analysis to ease the computational burden. We apply the method and its nonelastic competitors to both simulated data and observed data to show its efficiency in handling data with phase variation with both amplitude and phase changepoints. We use the method to evaluate potential changes in stratospheric temperature due to the eruption of Mt. Pinatubo in the Philippines in June 1991. Using an epidemic changepoint model, we find evidence of a increase in stratospheric temperature during a period that contains the immediate aftermath of Mt. Pinatubo, with most detected changepoints occurring in the tropics as expected.

Perovskite solar cells (PSCs) are emerging photovoltaic (PV) technologies capable of matching power conversion efficiencies (PCEs) of current PV technologies in the market at lower manufacturing costs, making perovskite solar modules (PSMs) cost competitive if manufactured at scale and perform with minimal degradation. PSCs with the highest PCEs, to date, are lead halide perovskites. Lead presents potential environmental and human health risks if PSMs are to be commercialized, as the lead in PSMs are more soluble in water compared to other PV technologies. Therefore, prior to commercialization of PSMs, it is important to highlight, identify, and establish the potential environmental and human health risks of PSMs as well as develop methods for assessing the potential risks. Here, we identify and discuss a variety of international standards, U.S. regulations, and permits applicable to PSM deployment that relate to the potential environmental and human health risks associated with PSMs. The potential risks for lead and other hazardous material exposures to humans and the environment are outlined which include water quality, air quality, human health, wildlife, land use, and soil contamination, followed by examples of how developers of other PV technologies have navigated human health and environmental risks previously. Potential experimentation, methodology, and research efforts are proposed to elucidate and characterize potential lead leaching risks and concerns pertaining to fires, in-field module damage, and sampling and leach testing of PSMs at end of life. Lastly, lower technology readiness level solutions to mitigate lead leaching, currently being explored for PSMs, are discussed. PSMs have the potential to become a cost competitive PV technology for the solar industry and taking steps toward understanding, identifying, and creating solutions to mitigate potential environmental and human health risks will aid in improving their commercial viability.

Cast Monel alloys are used in many industrial applications that require a combination of good mechanical properties and excellent resistance to corrosion. Despite relative widespread use, there has been limited prior research investigating the fundamental composition–structure–property relationships. In this work, microstructural characterization, thermal analysis, electron probe microanalysis, tensile testing, and Varestraint testing were used to assess the effects of variations in nominal composition on the solidification path, microstructure, mechanical properties, and solidification cracking susceptibility of cast Monel alloys. It was found that Si segregation caused the formation of silicides at the end of solidification in grades containing at least 3 wt pct Si. While increases to Si content led to significant improvements in strengthening due to the precipitation of β1-Ni3Si, the silicide eutectics acted as crack nucleation sites during tensile loading which severely reduced ductility. The solidification cracking susceptibility of low-Si Monel alloys was found to be relatively low. However, increases to Si concentration and the onset of associated eutectic reactions increased the solidification temperature range and drastically reduced cracking resistance. Increases in the Cu and Mn concentrations were found to reduce the solubility limit of Si in austenite which promoted additional eutectic formation and exacerbated the reductions in ductility and/or weldability.

Ignitions of solid materials from very high heat fluxes (>200 kW/m2) are differentiated from more common lower flux ignition because the required total energy input can be lower, and the process is much faster. Prior work has characterized ignition thresholds via thermal properties of the solids, flux, and fluence. The historical data, however, neglect to provide similar focus on the initiation of pyrolysis. The initiation of pyrolysis is of key relevancy because it represents an absolute threshold below which ignition is of zero probability. It is also a metric of potentially higher reliability for assessing material response because surface material properties such as absorptivity, conductivity, and density tend to change upon initial pyrolysis due to charring or other transformations. Recent data from concentrated solar flux for a variety of materials and exposures are analyzed here to explore the nature of trends and thresholds for onset of pyrolysis at high heat flux. This work evaluates initiation threshold data and provides a theoretical technique for further model development. The technique appears to be functionally appropriate to evaluate trends to aid in predicting material response to high flux exposures.

The role to which a realistic inflow turbulent boundary layer (TBL) influences transient and mean large-scale pool fire quantities of interest (QoIs) is numerically investigated. High-fidelity, low-Mach large-eddy simulations that activate low-dissipation, unstructured numerics are conducted using an unsteady flamelet combustion modeling approach with mutiphysics coupling to soot and participating media radiation transport. Three inlet profile configurations are exercised for a large-scale, high-aspect rectangular pool that is oriented perpendicular to the flow direction: a time-varying, TBL inflow profile obtained from a periodic precursor simulation, the time-mean of the transient TBL, and a steady power-law inflow profile that replicates the mean TBL crosswind velocity of 10.0 m/s at a vertical height of 10 m. Results include both qualitative transient flame evolution and quantitative flame shape with ground-level temperature and convective/radiative heat flux profiles. While transient fire events, which are driven by burst-sweep TBL coupling, such as blow-off and reattachment are vastly different in the TBL case (contributing to increased root mean square QoI fluctuation prediction and disparate flame lengths), mean surface QoI magnitudes are similar. Quadrant analysis demonstrates that the TBL configuration modifies burst-sweep phenomena at windward pool locations, while leeward recovery is found. Positive fluctuations of convective heat flux correlate with fast moving fluid away from the pool surface due to intermittent combustion events.

Ducted fuel injection (DFI) is a strategy to improve fuel/charge-gas mixing in direct-injection compression-ignition engines. DFI involves injecting fuel along the axis of a small tube in the combustion chamber, which promotes the formation of locally leaner mixtures in the autoignition zone relative to conventional diesel combustion. Previous work has demonstrated that DFI is effective at curtailing engine-out soot emissions across a wide range of operating conditions. This study extends previous investigations, presenting engine-out emissions and efficiency trends between ducted two-orifice and ducted four-orifice injector tip configurations. For each configuration, parameters investigated include injection pressure, injection duration, intake manifold pressure, intake manifold temperature, start of combustion timing, and intake-oxygen mole fraction. For both configurations and across all parameters, DFI reduced engine-out soot emissions compared to conventional diesel combustion, with little effect on other emissions and engine efficiency. Emissions trends for both configurations were qualitatively the same across the parameters investigated. The four-duct configuration had higher thermal efficiency and indicated-specific engine-out nitrogen oxide emissions but lower indicated-specific engine-out hydrocarbon and carbon monoxide emissions than the two-duct assembly. Both configurations achieved indicated-specific engine-out emissions for both soot and nitrogen oxides that comply with current on- and off-road heavy-duty regulations in the United States without exhaust-gas aftertreatment at an intake-oxygen mole fraction of 12%. High-speed in-cylinder imaging of natural soot luminosity shows that some conditions include a second soot-production phase late in the cycle. The probability of these late-cycle events is sensitive to both the number of ducted sprays and the operating conditions.

Chromium self-diffusion through stainless steel (SS) matrix and along grain boundaries is an important mechanism controlling SS structural materials corrosion. Cr diffusion in austenitic SS was simulated using canonical ab initio molecular dynamics with realistic models of type-316 SS bulk, with and without Cr vacancies, and a low-energy Σ3 twin boundary typically observed at active corrosion sites. Cr self-diffusion coefficients at 750 and 850 °C calculated using Einstein's diffusion equation are 4.2 × 10−6 and 8.1 × 10−6 Å2 ps−1 in pristine bulk, 3.8 × 10−3 and 5.5 × 10−3 Å2 ps−1 in bulk including Cr vacancies, and 9.5 × 10−2 and 1.0 × 10−1 Å2 ps−1 at a Σ3[1 1 1]60° twin boundary.

Accurately modeling large biomolecules such as DNA from first principles is fundamentally challenging due to the steep computational scaling of ab initio quantum chemistry methods. This limitation becomes even more prominent when modeling biomolecules in solution due to the need to include large numbers of solvent molecules. We present a machine-learned electron density model based on a Euclidean neural network framework that includes a built-in understanding of equivariance to model explicitly solvated double-stranded DNA. By training the machine learning model using molecular fragments that sample the key DNA and solvent interactions, we show that the model predicts electron densities of arbitrary systems of solvated DNA accurately, resolves polarization effects that are neglected by classical force fields, and captures the physics of the DNA-solvent interaction at the ab initio level.

The rate of electric vehicle (EV) adoption, powered by the Li-ion battery, has grown exponentially; largely driven by technological advancements, consumer demand, and global initiatives to reduce carbon emissions. As a result, it is imperative to understand the state of stability (SoS) of the cells inside an EV battery pack. That understanding will enable the warning of or prevention against catastrophic failures that can lead to serious injury or even, loss of life. The present work explores rapid electrochemical impedance spectroscopy (EIS) coupled with gas sensing technology as diagnostics to monitor cells and packs for failure markers. These failure markers can then be used for onboard assessment of SoS. Experimental results explore key changes in single cells and packs undergoing thermal or electrical abuse. Rapid EIS showed longer warning times, followed by VOC sensors, and then H2 sensors. While rapid EIS gives the longest warning time, with the failure marker often appearing before the cell vents, the reliability of identifying impedance changes in single cells within a pack decreases as the pack complexity increases. This provides empirical evidence to support the significant role that cell packaging and battery engineering intricacies play in monitoring the SoS.

In this report we discuss training a deep learning seismic signal detection model on 3-component stations from the International Monitoring System (IMS) using the PhaseNet architecture. Using 14 years of associated signals from the International Data Centre’s (IDC) Late Event Bulletin (LEB), we auto-curated training data consisting of signal windows containing associated arrivals, and noise windows that contain no LEB-associated signals. We trained several models using different waveform window durations (30 seconds and 100 seconds), with and without bandpass filtering. We evaluated the effectiveness of our models using associated signals from the Unconstrained Global Event Bulletin (UGEB) and found that several of our models outperformed the signal detections from the IDC’s Selected Event List 3 (SEL3) arrival table. The SEL3 bulletin evaluated on the UGEB dataset with 100-second waveform windows registered a precision and recall of .15 and .48, respectively, versus .19 and .59 for our filtered-data model. For the 30-second waveform window dataset, the SEL3 bulletin achieved a precision and recall of .31 and .47, respectively, versus .32 and .60 for our filtered-data model. Finally, our models detected signals from all source-to-receiver distances, suggesting it is feasible to use a single PhaseNet model for the IMS network.

Abstract not provided.

Abstract not provided.

A dry etching process to transfer the pattern of a photonic integrated circuit design for high-speed laser communications is described. The laser stack under consideration is a 3.2-µm-thick InGaAs/InAlAs/InAlGaAs epitaxial structure grown by molecular beam epitaxy. The etching was performed using Cl2-based inductively-coupled-plasma and reactive-ion-etching (ICP-RIE) reactors. Four different recipes are presented in two similar ICP-RIE reactors, with special attention paid to the etched features formed with various hard mask compositions, in-situ passivations, and process temperatures. The results indicate that it is possible to produce high-aspect-ratio features with sub-micron separation on this multilayer structure. Additionally, the results of the etching highlight the tradeoffs involved with the corresponding recipes.

Abstract not provided.

Abstract not provided.

This report documents the results of a long-term (5.79 year) exposure of 4-point bend corrosion test samples in the inlet and outlet vents of four spent nuclear fuel dry storage systems at the Maine Yankee Independent Spent Fuel Storage Installation. The goal of the test was to evaluate the corrosiveness of salt aerosols in a realistic near-marine environment, providing a data set for improved understanding of stress corrosion cracking of spent nuclear fuel dry storage canisters. Examination of the samples after extraction showed minor corrosion was present, mostly on rough-ground surfaces. However, dye penetrant testing showed that no SCC cracks were present. Dust collected on coupons co-located with the corrosion specimens was analyzed by scanning electron microscopy and leached to determine the soluble salts present. The dust was mostly organic material (pollen and stellate trichomes), with lesser detrital mineral grains. Salts present were a mix of sea-salts and continental salts, with chloride dominating the anions, but significant amounts of nitrate were also present. Both corrosion samples and dust samples showed evidence of wetting, indicating entry of water into the vents. The results of this field test suggest that the environment at Maine Yankee is not highly aggressive, although extrapolation from the periodically wetted vent samples to the hot, dry, canister surface may be difficult. No stress corrosion cracks were observed, but minor corrosion was present despite high nitrate concentrations in the salts. These observations may help address the ongoing question of the importance of nitrate in suppressing corrosion and SCC.

Abstract not provided.

Abstract not provided.

Sandia National Laboratories (SNL) performed a high-altitude nuclear electromagnetic pulse (HEMP) critical generation station component vulnerability test campaign with a focus on high-frequency, conducted early-time (E1) HEMP for the Department of Energy (DOE) Office of Cybersecurity, Energy Security, and Emergency Response (CESER). This report provides vulnerability test results to investigate component response and/or damage thresholds to reasonable HEMP threat levels that will help to inform site vulnerability assessments, mitigation planning, and modeling calibrations. This work details testing of North American Electric (NAE) magnetic motor starters to determine the effects of conducted HEMP environments. Motor starters are the control elements that provide power to motors throughout a generating plant; a starter going offline would cause loss of power to critical pumps and compressors, which could lead to component damage or unplanned plant outages. Additionally, failed starters would be unable to support plant startup. Six industrial motor starters were tested: two 2 horsepower (HP) starters with breaker disconnects and typical protection equipment, two 20 HP starters with breaker disconnects, and two 20 HP starters with fused disconnects. Each starter was placed in a circuit with a generator and inductive motor matching the starter rating. The conducted EMP insult was injected on the power cables passing through the motor starter, with separate tests for the generator and motor sides of the starter.

The On-Line Waste Library is a website that contains information regarding United States Department of Energy-managed high-level waste, spent nuclear fuel, and other wastes that are likely candidates for deep geologic disposal, with links to supporting documents for the data. This report provides supporting information for the data for which an already published source was not available.

The final quality of any AI/ML system is directly related to the quality of the input data used to train the system. In this case, we are trying to build a reliable image classifier that can correctly identify electrical components in x-ray images. The classification confidence is directly related to the quality of the labels in the training data, which are used in developing the AI/ML classifier. Incorrect or incomplete labels can substantially hinder the performance of the system during the training process, as it tries to compensate for variations that should not exist. Image labels are entered by subject matter experts, and in general can be assumed to be correct. However, this is not a guarantee, so developing ways to measure label quality and help identify or reject bad labels is important, especially as the database continues to grow. Given the current size of the database, a full manual review of each component is not feasible. This report will highlight the current state of the “RECON” x-ray image database and summarize several recent developments to try to help ensure high quality labeling both now and in the future. Questions that we hope to answer with this development include: 1) Are there any components with incorrect labels? 2) Can we suggest labels for components that are marked “Unknown”? 3) What kind of overall confidence do we have in the quality of the existing labels? 4) What systems or procedures can we put in place to maximize label quality?

Abstract not provided.

Abstract not provided.

Abstract not provided.

Quantifying the radioactive sources present in gamma spectra is an ever-present and growing national security mission and a time-consuming process for human analysts. While machine learning models exist that are trained to estimate radioisotope proportions in gamma spectra, few address the eventual need to provide explanatory outputs beyond the estimation task. In this work, we develop two machine learning models for a NaI detector measurements: one to perform the estimation task, and the other to characterize the first model’s ability to provide reasonable estimates. To ensure the first model exhibits a behavior that can be characterized by the second model, the first model is trained using a custom, semi-supervised loss function which constrains proportion estimates to be explainable in terms of a spectral reconstruction. The second auxiliary model is an out-of-distribution detection function (a type of meta-model) leveraging the proportion estimates of the first model to identify when a spectrum is sufficiently unique from the training domain and thus is out-of-scope for the model. In demonstrating the efficacy of this approach, we encourage the use of meta-models to better explain ML outputs used in radiation detection and increase trust.

Abstract not provided.

Abstract not provided.

The U.S. Department of Energy (DOE) Office of Cybersecurity, Energy Security, and Emergency Response (CESER), and Office of Electricity (OE) commissioned the National Renewable Energy Laboratory (NREL) to develop a method and tool to enable electric utilities to understand and manage the risk of cybersecurity events that can lead to physical effects like blackouts. This tool, called Cyber100 Compass, uses cybersecurity data elicited from cybersecurity experts, then incorporates that data into a tool designed to be usable by cybersecurity non-experts who understand the system itself. The tool estimates dollar-valued risks for a current or postulated future electric power digital control configuration, in order to enable utility risk planners to prioritize among proposed cybersecurity risk mitigation options. With the development of the Cyber100 Compass tool for quantification of future cyber-physical security risks, NREL has taken an initial bold step in the direction of enabling and indeed encouraging electric utilities to address the potential for cybersecurity incidents to produce detrimental physical effects related to electric power delivery. As part of the Cyber100 Compass development process, DOE funded NREL to seek out an independent technical review of the risk methodology embodied in the tool. NREL requested this review from Sandia National Laboratories, and made available to Sandia a very late version of the project report, as well as NREL personnel to provide clarification and to respond to questions. This paper provides the result of the independent review activity.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Derivative computation is a key component of optimization, sensitivity analysis, uncertainty quantification, and the solving of nonlinear problems. Automatic differentiation (AD) is a powerful technique for evaluating such derivatives, and in recent years, has been integrated into programming environments such as Jax, PyTorch, and TensorFlow to support derivative computations needed for training of machine learning models, facilitating wide-spread use of these technologies. The C++ language has become the de facto standard for scientific computing due to numerous factors, yet language complexity has made the wide-spread adoption of AD technologies for C++ difficult, hampering the incorporation of powerful differentiable programming approaches into C++ scientific simulations. This is exacerbated by the increasing emergence of architectures, such as GPUs, with limited memory capabilities and requiring massive thread-level concurrency. C++ AD tools must effectively use these environments to bring novel scientific simulations to next-generation DOE experimental and observational facilities. In this project, we investigated source transformation-based automatic differentiation using LLVM compiler infrastructure to automatically generate portable and efficient gradient computations of Kokkos-based code. We have demonstrated that our proposed strategy is feasible by investigating the usage of a prototype LLVM-based source transformation tool to generate gradients of simple functions made of sequences of simple Kokkos parallel regions. Speedups of up to 500x compared to Sacado were observed on NVIDIA V100 GPU.

Abstract not provided.

We can improve network pruning by leveraging the loss-topography extraction techniques used by projective integral updates for variational inference. Low-variance Hessians facilitate more aggressive pruning by providing better loss approximations when a parameter is removed.

As machine learning models for radioisotope quantification become more powerful, likewise the need for high-quality synthetic training data grows as well. For problem spaces that involve estimating the relative isotopic proportions of various sources in gamma spectra it is necessary to generate training data that accurately represents the variance of proportions encountered. In this report, we aim to provide guidance on how to target a desired variance of proportions which are randomly when using the PyRIID Seed Mixer, which samples from a Dirichlet distribution. We provide a method for properly parameterizing the Dirichlet distribution in order to maintain a constant variance across an arbitrary number of dimensions, where each dimension represents a distinct source template being mixed. We demonstrate that our method successfully parameterizes the Dirichlet distribution to target a specific variance of proportions, provided that several conditions are met. This allows us to follow a principled technique for controlling how random mixture proportions are generated which are then used downstream in the synthesis process to produce the final, noisy gamma spectra.

Abstract not provided.

Abstract not provided.

Commercial nuclear power plants typically use nuclear fuel that is enriched to less than five weight percent in the isotope ²³⁵U. However, recently several vendors have proposed new nuclear power plant designs that would use fuel with ²³⁵U enrichments between five weight percent and 19.75 weight percent. Nuclear fuel with this level of ²³⁵U enrichment is known as “high assay low-enriched uranium.” Once it has been irradiated in a nuclear reactor and becomes used (or spent) nuclear fuel, it will be stored, transported, and disposed of. However, irradiated high assay low-enriched uranium differs from typical irradiated nuclear fuel in several ways, and these differences may have economic effects on its storage, transport, and disposal, compared to typical irradiated nuclear fuel. This report describes those differences and qualitatively discusses their potential economic effects on storage, transport, and disposal.