Search NASA⌕ Search

SEARCH · Search NASA

Results for “Distributed training”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,297 records · Page 72

Bayesian Cross-matching of High Proper-motion Stars in Gaia DR2 and Photometric Metallicities for ∼1.7 million K and M Dwarfs

We present a Bayesian method to cross-match 5,827,988 high proper-motion Gaia sources (μ > 40 mas yr{sup −1}) to various photometric surveys: Two Micron All Sky Survey, AllWISE data release from the Wide-field Infrared Explorer (WISE) mission, Galaxy Evolution Explorer, Radial Velocity Experiment, Sloan Digital Sky Survey, and Panoramic Survey Telescope and Rapid Response System (Pan-STARRS). To efficiently associate these objects across catalogs, we develop a technique that compares the multidimensional distribution of all sources in the vicinity of each Gaia star to a reference distribution of random field stars obtained by extracting all sources in a region on the sky displaced 2′. This offset preserves the local field stellar density and magnitude distribution, allowing us to characterize the frequency of chance alignments. The resulting catalog with Bayesian probabilities >95% has a marginally higher match rate than current internal Gaia data release 2 (DR2) matches for most catalogs. However, a significant improvement is found with Pan-STARRS, where ∼99.8% of the sample within the Pan-STARRS footprint is recovered, as compared to a low ∼20.8% in Gaia DR2. Using these results, we train a Gaussian process regressor to calibrate two photometric metallicity relationships. For dwarfs of 3500 < T{sub eff} < 5280 K, we use metallicity values of 4378 stars from the Apache Point Observatory Galactic Evolution Experiment and Hejazi et al. to calibrate the relationship, producing results with a 1σ precision of 0.12 dex and few systematic errors. We then indirectly infer the metallicity of 4018 stars with 2850 < T{sub eff} < 3500 K, which are wide companions of primaries whose metallicities are estimated with our first regressor, to produce a relationship with a 1σ precision of 0.21 dex and significant systematic errors. Additional work is needed to better remove unresolved binaries from this second sample to reduce these systematic errors.

47 OTHER INSTRUMENTATION↗

EXERGETIC: De-Risking Next-Generation Resilient Geothermal Hybrids via At-Scale Evaluation Using Virtual Emulation Digital Twin Environment for Efficient Operation

The DOE-GTO-funded project, award number 5.1.2.12, entitled "EXERGETIC - De-risking Next Generation Resilient Geothermal Hybrids via at-Scale Evaluation Using a Virtual Emulation Digital Twin Environment for Efficient Operation," advances the solution to these challenges by developing and validating a geothermal co-emulation environment implemented at the National Laboratory of the Rockies (NLR)'s Advanced Research on Integrated Energy Systems (ARIES) platform. This framework enables the de-risking of next-generation geothermal and geothermal hybrid systems through high-fidelity modeling, real-time digital emulation, advanced control strategies, and techno-economic assessment. The project focused on geothermal hybrid configurations that integrate geothermal power plants with concentrated solar power and underground thermal energy storage, enabling enhanced efficiency, flexibility, and grid support capabilities. The main goal of this project was the development of a geothermal digital co-emulation environment to demonstrate the technical and economic value of geothermal hybrid systems and their contribution to grid stability and flexibility. The EXERGETIC framework combined physics-based models, controls, and real assets at ARIES, including digital real-time simulators (DRTS), a 20-MW-scale controllable grid interface (CGI), and a 2-MW conventional generator. Detailed transient models were developed for the key subsystems of a hybrid geothermal plant, including parabolic trough solar collectors, reservoir thermal energy storage (RTES), and a binary Organic Rankine Cycle (ORC) power plant. The ORC model explicitly captured thermal inertia and off-design operation and integrated control strategies to dynamically respond to electric load profiles. The models were validated against published experimental and numerical studies, demonstrating strong agreement and confirming the accuracy and robustness of the modeling approach. The resulting digital twin represents geothermal-solar-storage systems at multiple scales (1 MW to 100 MW) and enables realistic emulation of grid-connected operation. The control architecture allows the geothermal resource to provide stable baseload generation, while solar and stored thermal energy supply flexible, dispatchable support during periods of high demand or variable grid conditions. A key contribution of the EXERGETIC project is the demonstration that geothermal hybrid systems can be designed to be active grid assets rather than passive baseload generators. Using the ARIES platform, the digital twin was evaluated under multiple grid scenarios, including load following, voltage support at the distribution level, and frequency response at the transmission level. Results show that hybrid geothermal systems can respond effectively to dynamic grid conditions, providing inertia-like behavior, primary frequency support, and voltage regulation through coordinated control. In addition to the performance and grid services capability analysis of geothermal and hybrid geothermal systems, the EXERGETIC project also focused on scalability and techno-economic analysis of geothermal hybrid plants. In particular, for the scalability analysis, machine-learning (ML)-based surrogate models were trained using data generated from the geothermal digital twin under different grid-connected scenarios and plant capacities. These ML models demonstrated strong interpolation and extrapolation capabilities across plant sizes, accurately reproducing both steady-state and transient responses with very low errors. Regarding the techno-economic analysis, plant performance results were integrated with cost models for hybrid geothermal systems, and the levelized cost of electricity (LCOE) was used as the main economic metric to evaluate system performance across a range of system capacities, solar shares, solar multiples, and storage durations. Results indicate that economies of scale significantly reduce geothermal LCOE as plant capacity increases, with large-scale systems (25-100 MW) achieving substantially lower costs than small plants. Hybridization with solar thermal energy and storage further improves economic performance by increasing capacity utilization and enabling flexible dispatch. In addition, thermal storage plays a critical role in reducing LCOE by maximizing geothermal, solar, and stored energy resources. In summary, the results from this project demonstrate that geothermal hybrid systems represent a promising alternative for increasing the energy conversion efficiency of geothermal technologies, contributing to the preservation of geothermal resources, and supporting the transition of geothermal plants from traditional baseload resources into flexible, resilient, and cost-competitive energy conversion technologies.

15 GEOTHERMAL ENERGY↗

Universal Polarization Transformations: Spatial Programming of Polarization Scattering Matrices Using a Deep Learning‐Designed Diffractive Polarization Transformer

Abstract Controlled synthesis of optical fields having nonuniform polarization distributions presents a challenging task. Here, a universal polarization transformer is demonstrated that can synthesize a large set of arbitrarily‐selected, complex‐valued polarization scattering matrices between the polarization states at different positions within its input and output field‐of‐views (FOVs). This framework comprises 2D arrays of linear polarizers positioned between isotropic diffractive layers, each containing tens of thousands of diffractive features with optimizable transmission coefficients. After its deep learning‐based training, this diffractive polarization transformer can successfully implement N i N o = 10 000 different spatially‐encoded polarization scattering matrices with negligible error, where N i and N o represent the number of pixels in the input and output FOVs, respectively. This universal polarization transformation framework is experimentally validated in the terahertz spectrum by fabricating wire‐grid polarizers and integrating them with 3D‐printed diffractive layers to form a physical polarization transformer. Through this set‐up, an all‐optical polarization permutation operation of spatially‐varying polarization fields is demonstrated, and distinct spatially‐encoded polarization scattering matrices are simultaneously implemented between the input and output FOVs of a compact diffractive processor. This framework opens up new avenues for developing novel devices for universal polarization control and may find applications in, e.g., remote sensing, medical imaging, security, material inspection, and machine vision.

Optical neural networks↗

Hydrogeological Model Selection Among Complex Spatial Priors

Hydrogeological field studies rely often on a single conceptual representation of the subsurface. This is problematic since the impact of a poorly chosen conceptual model on predictions might be significantly larger than the one caused by parameter uncertainty. Furthermore, conceptual models often need to incorporate geological concepts and patterns in order to provide meaningful uncertainty quantification and predictions. Consequently, several geologically realistic conceptual models should ideally be considered and evaluated in terms of their relative merits. Here, we propose a full Bayesian methodology based on Markov chain Monte Carlo to enable model selection among 2-D conceptual models that are sampled using training images and concepts from multiple-point statistics. More precisely, power posteriors for the different conceptual subsurface models are sampled using sequential geostatistical resampling and Graph Cuts. To demonstrate the methodology, we compare and rank five alternative conceptual geological models that have been proposed in the literature to describe aquifer heterogeneity at the MAcroDispersion Experiment site in Mississippi, USA. We consider a small-scale tracer test for which the spatial distribution of hydraulic conductivity impacts multilevel solute concentration data observed along a 2-D transect. The thermodynamic integration and the stepping-stone sampling methods were used to compute the evidence and associated Bayes factors using the computed power posteriors. We find that both methods are compatible with multiple-point statistics-based inversions and provide a consistent ranking of the competing conceptual models considered.

58 GEOSCIENCES↗

Real-time confinement regime detection in fusion plasmas with convolutional neural networks and high-bandwidth edge fluctuation measurements

Abstract A real-time detection of the plasma confinement regime can enable new advanced plasma control capabilities for both the access to and sustainment of enhanced confinement regimes in fusion devices. For example, a real-time indication of the confinement regime can facilitate transition to the high-performing wide-pedestal (WP) quiescent H-mode, or avoid unwanted transitions to lower confinement regimes that may induce plasma termination. To demonstrate real-time confinement regime detection, we use the 2D beam emission spectroscopy (BES) diagnostic system to capture localized density fluctuations of long wavelength turbulent modes in the edge region at a 1 MHz sampling rate. BES data from 330 discharges in either L-mode, H-mode, quiescent H (QH)-mode, or WP QH-mode were collected from the DIII-D tokamak and curated to develop a high-quality database to train a deep-learning classification model for real-time confinement detection. We utilize the 6×8 spatial configuration with a time window of 1024 µ s and recast the input to obtain spectral-like features via fast Fourier transform preprocessing. We employ a shallow 3D convolutional neural network for the multivariate time-series classification task and utilize a softmax in the final dense layer to retrieve a probability distribution over the different confinement regimes. Our model classifies the global confinement state on 44 unseen test discharges with an average F 1 score of 0.94, using only ∼1 ms snippets of BES data at a time. This activity demonstrates the feasibility for real-time data analysis of fluctuation diagnostics in future devices such as ITER, where the need for reliable and advanced plasma control is urgent.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Stringent σ8 constraints from small-scale galaxy clustering using a hybrid MCMC + emulator framework

ABSTRACT We present a novel simulation-based hybrid emulator approach that maximally derives cosmological and Halo Occupation Distribution (HOD) information from non-linear galaxy clustering, with sufficient precision for DESI Year 1 (Y1) analysis. Our hybrid approach first samples the HOD space on a fixed cosmological simulation grid to constrain the high-likelihood region of cosmology + HOD parameter space, and then constructs the emulator within this constrained region. This approach significantly reduces the parameter volume emulated over, thus achieving much smaller emulator errors with fixed number of training points. We demonstrate that this combined with state-of-the-art simulations result in tight emulator errors comparable to expected DESI Y1 LRG sample variance. We leverage the new abacussummit simulations and apply our hybrid approach to CMASS non-linear galaxy clustering data. We infer constraints on σ8 = 0.762 ± 0.024 and fσ8(zeff = 0.52) = 0.444 ± 0.016, the tightest among contemporary galaxy clustering studies. We also demonstrate that our fσ8 constraint is robust against secondary biases and other HOD model choices, a critical first step towards showcasing the robust cosmology information accessible in non-linear scales. We speculate that the additional statistical power of DESI Y1 should tighten the growth rate constraints by at least another 50–60 ${{\ \rm per\ cent}}$, significantly elucidating any potential tension with Planck. We also address the ‘lensing is low’ tension, which we find to be in the same direction as a potential tension in fσ8. We show that the combined effect of a lower fσ8 and environment-based bias accounts for approximately $50{{\ \rm per\ cent}}$ of the discrepancy.

79 ASTRONOMY AND ASTROPHYSICS↗

Adaptive X-ray imaging with reinforcement learning

X-ray imaging is a powerful technique to scan samples in a variety of contexts including biological, environmental and materials science, but commonly requires a synchrotron light source to produce X-rays at sufficient intensity. As these facilities are expensive to operate, the available beam time is limited and always in high demand. Particularly if the illuminated samples are sparse, standard raster scanning methods can be time-consuming, with a majority of that time being spent on areas of the image that carry little information. To increase the efficiency and maximize the information gain for a given time budget, we split the scanning process into a series of steps where previous measurements are used to inform the decision making and adapt the exposure distribution at later stages of the sequence. We formulate this task as a reinforcement learning problem where the goal is to produce a sequence of exposure maps that maximize a predefined scalar metric. We demonstrate the potential of this approach in simulations where the adaptive illumination can accelerate the measurement process by up to an order of magnitude compared with standard raster scanning. Finally, we present the first results from deploying the trained agents on an X-ray fluorescence beamline at the Stanford Synchrotron Radiation Lightsource.

Reinforcement Learning↗

Demand or Request: Will Load Behave?

Power planning engineers are trained to design an electric system that satisfies predicted electrical demand under stringent conditions of availability and power quality. Like responsible custodians, we plan for the provision of electrical sustenance and shelter to those in whose care regulators have given us the responsibility to serve. Though most customers accept this nurturing gladly, a growing number are concerned with the economic costs and environmental impacts of service at a time when technology (particularly distributed generation, storage, automation, and information networks) offers alternatives for localized control and competitive service. As customers’ and their systems mature, a new relationship with the electricity provider is emerging. Demand response is perhaps the first unsteady step where the customer participates as a partner in system operations. This paper explores issues system planners need to consider as demand response matures to significant levels beyond direct load control and toward a situation where service is requested and bargains are reached with the electricity provider based on desired load behavior. On one hand, predicting load growth and behavior appears more daunting than ever. On the other, for the first time load becomes a new resource whose behavior can be influenced during system operations to balance system conditions.

24 POWER TRANSMISSION AND DISTRIBUTION↗

HydraGNN_OPF_GFM_2026 - Ensemble of predictive graph foundation models for power grid applications

This dataset supports research on graph foundation models for optimal power flow (OPF) on electric grids using HydraGNN. It contains heterogeneous graph representations of PGLib-OPF cases spanning systems from 14 to 13,659 buses, together with packed HDF5 datasets for pretraining, feasibility classification, and N-1 contingency analysis. The release includes OPF solution data, downstream fine-tuning datasets, pretrained HeteroSAGE and HeteroHEAT model checkpoints, hyperparameter-optimization summaries across multiple heterogeneous GNN architectures, and aggregated fine-tuning results for sample-efficiency studies. The dataset is designed to enable scalable training, evaluation, and transfer-learning studies for OPF surrogate modeling, including node-level AC-OPF solution prediction, graph-level prediction, feasibility classification, operating-condition generalization, and contingency-response tasks.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Bayesian computational approaches for gene regulation studies of bioethanol and biohydrogen production

It has recently become clear that regulatory RNAs play a major role in regulation of gene expression in bacteria. RNA secondary structures play a major role in the function of many regulatory RNAs, and structural features are often key to their interaction with other cellular components. Thus, there has been considerable interest in the prediction of the secondary structures for RNA families. A paper describing our new algorithm, RNAG, to predict consensus secondary structures for unaligned sequences using the blocked Gibbs sampler has been published[1]. This sampling algorithm iteratively samples from the conditional probability distributions: P(Structure | Alignment) and P(Alignment | Structure). Subsequent to publication of the RNAG paper we have employed the technology from RNAG in the development of an RNA motif finding algorithm. To develop and RNA motif finding algorithm, RGibbs, we capitalized on our long experience in DNA motif finding and RNA secondary structure prediction. We applied RGibbs to three data sets from the literature and compared it to existing methods: one for training and two others for tests sets. In both test sets we found RGibbs out performed existing procedures.

59 BASIC BIOLOGICAL SCIENCES↗

AGR-3/4 Irradiation Test Train Disassembly and Component Metrology First Look Report

The AGR-3/4 experiment was designed to study fission product transport within graphitic matrix material and nuclear-grade graphite. To this end, this experiment consisted of 12 capsules, each fueled with 4 compacts containing UCO TRISO particles as driver fuel and 20 UCO designed-to-fail (DTF) fuel particles in each compact. The DTF fuel was fabricated with a thin pyrocarbon layer which was intended to fail during irradiation and provide a source of fission products. These fission products could then migrate through the compact and into the surrounding concentric rings of graphitic matrix material and/or nuclear graphite. Through post-irradiation examination (PIE) of the rings (including physical sampling and gamma scanning) fission product concentration profiles within the rings can be determined. These data can be used to elucidate fission product transport parameters (e.g. diffusion coefficients within the test materials) which will be used to inform and refine models of fission product transport. After irradiation in the Advanced Test Reactor (ATR) had been completed in April 2014, the AGR-3/4 experiment was shipped to the Hot Fuel Examination Facility (HFEF) at the Materials and Fuels Complex (MFC) for inspection, disassembly, and metrology. The AGR-3/4 test train was received at MFC in two separate shipments between February and April 2015. Visual examinations of the test train exterior did not indicate dimensional distortion, and only two small discolored areas were observed at the bottom of Capsules 8 and 9. No corresponding discoloration was found on the inside of these capsules, however. Prior to disassembly, the two test train sections were subject to analysis via the Precision Gamma Scanner (PGS), which did not indicate that any gross fuel relocation had occurred. A series of specialized tools (including clamps, cutters, and drills) had been designed and fabricated in order to carry out test train disassembly and recovery of capsule components (graphite rings and fuel compacts). This equipment performed well for separating each capsule in the test train and extracting the capsule components. Only a few problems were encountered. In one case, the outermost ring (the sink ring) was cracked during removal of the capsule through tubes. Although the sink ring will be analyzed in order to obtain a mass balance of fission products in the experiment, these cracks do not pose a major concern because the sink ring will not be analyzed in detail to obtain the spatial distribution of fission products. In Capsules 4 and 5, the compacts could not be removed from the inner rings. Strategies for removing the compacts are being evaluated. Sampling the inner rings with the compacts in-place is also an option. Dimensional measurements were made on the compacts, inner rings, outer rings, and sink rings. The diameters of all compacts decreased by 0.5 to 2.0 %. Generally, the extent of diametric shrinkage increased linearly with increasing neutron fluence. Most compact lengths also decreased. Compact lengths decreased with increasing fluence, reaching maximum shrinkage of about 0.9 % at a fast fluence of 4.0x10 25 n/m 2 E > 0.18 MeV. Above this fluence, the extent of length shrinkage appeared to decrease with fluence, and two compacts from Capsule 7 were found to have slightly increased in length (< 0.1 %) after a fluence of 5.2x10 25 n/m 2 .

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

AGR-3/4 Irradiation Test Train Disassembly and Component Metrology First Look Report

The AGR-3/4 experiment was designed to study fission product transport within graphitic matrix material and nuclear-grade graphite. To this end, this experiment consisted of 12 capsules, each fueled with 4 compacts containing UCO TRISO particles as driver fuel and 20 UCO designed-to-fail (DTF) fuel particles in each compact. The DTF fuel was fabricated with a thin pyrocarbon layer which was intended to fail during irradiation and provide a source of fission products. These fission products could then migrate through the compact and into the surrounding concentric rings of graphitic matrix material and/or nuclear graphite. Through post-irradiation examination (PIE) of the rings (including physical sampling and gamma scanning) fission product concentration profiles within the rings can be determined. These data can be used to elucidate fission product transport parameters (e.g. diffusion coefficients within the test materials) which will be used to inform and refine models of fission product transport. After irradiation in the Advanced Test Reactor (ATR) had been completed in April 2014, the AGR-3/4 experiment was shipped to the Hot Fuel Examination Facility (HFEF) at the Materials and Fuels Complex (MFC) for inspection, disassembly, and metrology. The AGR-3/4 test train was received at MFC in two separate shipments between February and April 2015. Visual examinations of the test train exterior did not indicate dimensional distortion, and only two small discolored areas were observed at the bottom of Capsules 8 and 9. No corresponding discoloration was found on the inside of these capsules, however. Prior to disassembly, the two test train sections were subject to analysis via the Precision Gamma Scanner (PGS), which did not indicate that any gross fuel relocation had occurred. A series of specialized tools (including clamps, cutters, and drills) had been designed and fabricated in order to carry out test train disassembly and recovery of capsule components (graphite rings and fuel compacts). This equipment performed well for separating each capsule in the test train and extracting the capsule components. Only a few problems were encountered. In one case, the outermost ring (the sink ring) was cracked during removal of the capsule through tubes. Although the sink ring will be analyzed in order to obtain a mass balance of fission products in the experiment, these cracks do not pose a major concern because the sink ring will not be analyzed in detail to obtain the spatial distribution of fission products. In Capsules 4 and 5, the compacts could not be removed from the inner rings. Strategies for removing the compacts are being evaluated. Sampling the inner rings with the compacts in-place is also an option. Dimensional measurements were made on the compacts, inner rings, outer rings, and sink rings. The diameters of all compacts decreased by 0.5 to 2.0 %. Generally, the extent of diametric shrinkage increased linearly with increasing neutron fluence. Most compact lengths also decreased. Compact lengths decreased with increasing fluence, reaching maximum shrinkage of about 0.9 % at a fast fluence of 4.0x1025 n/m2 E > 0.18 MeV. Above this fluence, the extent of length shrinkage appeared to decrease with fluence, and two compacts from Capsule 7 were found to have slightly increased in length (< 0.1 %) after a fluence of 5.2x1025 n/m2.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

FY17 ISCR Scholar End-of-Assignment Report - Robbie Sadre

Throughout this internship assignment, I did various tasks that contributed towards the starting of the SASEDS (Safe Active Scanning for Energy Delivery Systems) and CES-21 (California Energy Systems for the 21st Century) projects in the SKYFALL laboratory. The goal of the SKYFALL laboratory is to perform modeling and simulation verification of transmission power system devices, while integrating with high-performance computing. The first thing I needed to do was acquire official Online LabVIEW training from National Instruments. Through these online tutorial modules, I learned the basics of LabVIEW, gaining experience in connecting to NI devices through the DAQmx API as well as LabVIEW basic programming techniques (structures, loops, state machines, front panel GUI design etc).

24 POWER TRANSMISSION AND DISTRIBUTION↗

Neutron multiplicity analysis tool

I describe the capabilities of the EXCOM (EXcel based COincidence and Multiplicity) calculation tool which is used to analyze experimental data or simulated neutron multiplicity data. The input to the program is the count-rate data (including the multiplicity distribution) for a measurement, the isotopic composition of the sample and relevant dates. The program carries out deadtime correction and background subtraction and then performs a number of analyses. These are: passive calibration curve, known alpha and multiplicity analysis. The latter is done with both the point model and with the weighted point model. In the current application EXCOM carries out the rapid analysis of Monte Carlo calculated quantities and allows the user to determine the magnitude of sample perturbations that lead to systematic errors. Neutron multiplicity counting is an assay method used in the analysis of plutonium for safeguards applications. It is widely used in nuclear material accountancy by international (IAEA) and national inspectors. The method uses the measurement of the correlations in a pulse train to extract information on the spontaneous fission rate in the presence of neutrons from ({alpha},n) reactions and induced fission. The measurement is relatively simple to perform and gives results very quickly ({le} 1 hour). By contrast, destructive analysis techniques are extremely costly and time consuming (several days). By improving the achievable accuracy of neutron multiplicity counting, a nondestructive analysis technique, it could be possible to reduce the use of destructive analysis measurements required in safeguards applications. The accuracy of a neutron multiplicity measurement can be affected by a number of variables such as density, isotopic composition, chemical composition and moisture in the material. In order to determine the magnitude of these effects on the measured plutonium mass a calculational tool, EXCOM, has been produced using VBA within Excel. This program was developed to help speed the analysis of Monte Carlo neutron transport simulation (MCNP) data, and only requires the count-rate data to calculate the mass of material using INCC's analysis methods instead of the full neutron multiplicity distribution required to run analysis in INCC. This paper describes what is implemented within EXCOM, including the methods used, how the program corrects for deadtime, and how uncertainty is calculated. This paper also describes how to use EXCOM within Excel.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Quantitative Insight to Fission Gas Bubble Distribution in Irradiated Annular U-10Zr Metallic Fuel Using Deep Learning

U-10Zr Metal fuel is a promising nuclear fuel candidate for next-generation sodium-cooled fast spectrum reactors. Porosity is one of the key facts to impact the performance of metallic fuel. Additionally, a mechanical understanding of fission gas bubbles evolution behavior is a prerequisite for fuel development and qualification. Previous study of fission gas bubbles relied on a simple threshold method working on low resolution optical microscopy images, which has challenges in recognizing bubble boundaries, and caused inaccurate statistics of bubble properties. In this paper, a pre-trained deep learning model on Scanning Electron Microscopy (SEM) images from an annular U-10Zr fuel (AF1), was applied to another U-10Zr annular fuel (AF2). More accurate fission gas bubble segmentation results were generated, which leads to more precise qualitative analysis on the morphology, size, density, and orientation of bubbles. Furthermore, we investigated the lanthanide movement along the radial temperature gradient and obtained conclusive findings.

36 MATERIALS SCIENCE↗

Extracting structural motifs from pair distribution function data of nanostructures using explainable machine learning

Characterization of material structure with X-ray or neutron scattering using e.g. Pair Distribution Function (PDF) analysis most often rely on refining a structure model against an experimental dataset. However, identifying a suitable model is often a bottleneck. Recently, automated approaches have made it possible to test thousands of models for each dataset, but these methods are computationally expensive and analysing the output, i.e. extracting structural information from the resulting fits in a meaningful way, is challenging. Our Machine Learning based Motif Extractor (ML-MotEx) trains an ML algorithm on thousands of fits, and uses SHAP (SHapley Additive exPlanation) values to identify which model features are important for the fit quality. We use the method for 4 different chemical systems, including disordered nanomaterials and clusters. ML-MotEx opens for a type of modelling where each feature in a model is assigned an importance value for the fit quality based on explainable ML.

36 MATERIALS SCIENCE↗

Demonstration of Decentralized Physics-Driven Learning

In typical artificial neural networks, neurons adjust according to global calculations of a central processor, but in the brain, neurons and synapses self-adjust based on local information. Contrastive learning algorithms have recently been proposed to train physical systems, such as fluidic, mechanical, or electrical networks, to perform machine-learning tasks from local evolution rules. However, to date, such systems have only been implemented in silico due to the engineering challenge of creating elements that autonomously evolve based on their own response to two sets of global boundary conditions. Here, we introduce and implement a physics-driven contrastive learning scheme for a network of variable resistors, using circuitry to locally compare the response of two identical networks subjected to the two different sets of boundary conditions. Using this method, our system effectively trains itself, optimizing its resistance values without the use of a central processor or external information storage. Once the system is trained for a specified allostery, regression, or classification task, the task is subsequently performed rapidly and automatically by the physical imperative to minimize power dissipation in response to the given voltage inputs. Here, we demonstrate that, unlike typical computers, such learning systems are robust to extreme damage (and thus manufacturing defects) due to their decentralized learning. Our twin-network approach is therefore readily scalable to extremely large or nonlinear networks, where its distributed nature will be an enormous advantage; a laboratory network of only 500 edges will already outpace its in silico counterpart.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Using AI for Management of Field Emission in SRF Linacs

Field emission control, mitigation, and reduction is critical for reliable operation of high gradient superconducting radio-frequency (SRF) accelerators. With the SRF cavities at high gradients, the field emission of electrons from cavity walls can occur and will impact the operational gradient, radiological environment via activated components, and reliability of CEBAF’s two linacs. A new effort has started to minimize field emission in the CEBAF linacs by re-distributing cavity gradients. To measure radiation levels, newly designed neutron and gamma radiation dose rate monitors have been installed in both linacs. Artificial intelligence (AI) techniques will be used to identify cavities with high levels of field emission based on control system data such as radiation levels, cryogenic readbacks, and vacuum loads. The gradients on the most offending cavities will be reduced and compensated for by increasing the gradients on least offensive cavities. Training data will be collected during this year’s operational program and initial implementation of AI models will be deployed. Preliminary results and future plans are presented.

Carpenter, Adam↗