Search NASA⌕ Search

SEARCH · Search NASA

Results for “representative sampling”

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 253 records · Page 14
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Analysis of Waste Material Feedstocks Using Laser-Induced Breakdown Spectroscopy and Machine Learning

Predicting properties such as heating value, ash fusion temperature, and mineral ash composition from Laser-Induced Breakdown Spectroscopy (LIBS) data can make gasifiers more flexible to different feedstocks. Understanding these feedstock properties in-situ improves feedstock conversion modelling methods that allow for consistent operation, higher carbon conversion, and reduced fouling and erosion rates. The purpose of this study is to demonstrate methods for model creation that take LIBS data as predictor features and estimate higher order material properties as a function of feedstock material properties. Six samples were chosen to represent a mixture of abundant and carbon rich waste materials. LIBS measurements were performed on these samples for elemental wavelengths and intensity values. Laboratory analytical results were obtained for each sample’s heating value, proximate and ultimate analysis, mineral ash composition, ash fusion temperatures, and viscosity temperatures. Thermal conductivity was measured using a HotDisk TPS 2500S. LIBS measurements were processed and used as predictor features for machine learning (ML) models to predict the sample’s material properties. Predictor feature selection algorithms, particularly minimum redundancy maximum relevance (mRMR), reduced the dimensionality of ML models. Many modelling methods such as Gaussian process regression (GPR), regression tree, neural networks (NN), and support vector machines (SVM) were demonstrated to be effective at predicting higher order properties; however, mRMR with GPR stood out as a clear winning combination.

01 COAL, LIGNITE, AND PEAT↗
DOE OSTI2020

Sample problems for Section III, Division 5 design by inelastic analysis of Grade 91 components

This report works through two sample design analyses of representative Grade 91 components using the ASME Boiler & Pressure Vessel Code Section III, Division 5, Appendix HBB-T rules for the strain and deformation and creep-fatigue design criteria using a design by inelastic analysis approach. The component geometries and loadings were selected to span a wide variety of potential applications. The sample components are an endplate geometry, which is fairly representative of typical pressure vessel construction, and a core block, representing a critical component in a heat pipe microreactor, which has a complicated geometry and complex thermal stress history. These sample problems serve several functions. First, they are complete worked examples of how to apply the current design by inelastic analysis rules. Second, they serve as test cases to evaluate the current rules and provide feedback on how they could be optimized to deal with complicated component geometries or to be better automated in modern finite element analysis software. Finally, the analysis here uses an inelastic constitutive model for Grade 91 developed by Argonne National Laboratory and proposed for inclusion in the ASME Boiler & Pressure Vessel Code Section III, Division 5 rules as a reference constitutive model. Using this constitutive model in the sample problem analyses tests the model in the context of the ASME Code rules. The final conclusions of this report are that the material model and the current Code rules are satisfactory, but that the Code rules could be optimized to simplify automation and reduce over conservatism. Future work should address the specific optimization topics discussed in this report.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2020

The Hα Dots Survey. II. A Second List of Faint Emission-line Objects

We present the second catalog of serendipitously discovered compact extragalactic emission-line sources—Hα Dots. These objects have been discovered in searches of moderately deep narrow-band images acquired for the ALFALFA Hα project. In addition to cataloging 119 new Hα Dots, we also present follow-up spectral data for the full sample. These spectra allow us to confirm the nature of these objects as true extragalactic emission-line objects, to classify them in terms of activity type (star-forming or AGN), and to identify the emission line via which they were discovered. We tabulate photometric and spectroscopic data for the all objects, and we present an overview of the properties of the full Hα Dot sample. The Hα Dots represent a broad range of star-forming and active galaxies detected via several different emission lines over a wide range of redshifts. The sample includes Hα-detected blue compact dwarf galaxies at low redshift, [O iii]-detected Seyfert 2 and Green Pea-like galaxies at intermediate redshifts, and QSOs detected via one of several UV emission lines, including Lyα. Despite the heterogeneous appearance of the resulting catalog of objects, we show that our selection method leads to well-defined samples of specific classes of emission-line objects with properties that allow for statistical studies of each class.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

The merger fraction of post-starburst galaxies in UNIONS

ABSTRACT Post-starburst galaxies (PSBs) are defined as having experienced a recent burst of star formation, followed by a prompt truncation in further activity. Identifying the mechanism(s) causing a galaxy to experience a post-starburst phase therefore provides integral insight into the causes of rapid quenching. Galaxy mergers have long been proposed as a possible post-starburst trigger. Effectively testing this hypothesis requires a large spectroscopic galaxy survey to identify the rare PSBs as well as high-quality imaging and robust morphology metrics to identify mergers. We bring together these critical elements by selecting PSBs from the overlap of the Sloan Digital Sky Survey and the Canada–France Imaging Survey and applying a suite of classification methods: non-parametric morphology metrics such as asymmetry and Gini-M20, a convolutional neural network trained to identify post-merger galaxies, and visual classification. This work is therefore the largest and most comprehensive assessment of the merger fraction of PSBs to date. We find that the merger fraction of PSBs ranges from 19 per cent to 42 per cent depending on the merger identification method and details of the PSB sample selection. These merger fractions represent an excess of 3–46× relative to non-PSB control samples. Our results demonstrate that mergers play a significant role in generating PSBs, but that other mechanisms are also required. However, applying our merger identification metrics to known post-mergers in the IllustrisTNG simulation shows that 70 per cent of recent post-mergers (≲200 Myr) would not be detected. Thus, we cannot exclude the possibility that nearly all PSBs have undergone a merger in their recent past.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

A comparison of compact, presumably young with extended, evolved radio active galactic nuclei

Context.The triggering and evolution of active galactic nuclei (AGNs) and the interaction of the AGN with its host galaxy is an important topic in extragalactic astrophysics. Radio sources with peaked spectra (peaked spectrum sources, PSS) and compact symmetric objects (CSO) are powerful, compact, and presumably young AGNs and therefore particularly suitable to study aspects of the AGN-host connection. Aims.We use a statistical approach to investigate properties of a PSS-CSO sample that are related to host galaxies and could potentially shed light on the link between host galaxies and AGNs. The main goal is to compare the PSS-CSO sample with a matching comparison sample of extended sources (ECS) to see if the two have significant differences. Methods.We analysed composite spectra, diagnostic line diagrams, multi-band spectral energy distributions (MBSEDs), star formation (SF) indicators, morphologies, and cluster environments for a sample of 121 PSSs and CSOs for which spectra are available from the Sloan Digital Sky Survey (SDSS). The statistical results were compared with those of the ECS sample, where we generally considered the two subsamples of quasi-stellar objects (QSOs) and radio galaxies separately. The analysis is based on a large set of archival data in the spectral range from the ultraviolet to mid-infrared. Results.We find significant differences between the PSS-CSO and the ECS sample. In particular, we find that the ECS sample has a higher proportion of passive galaxies with a lower star formation activity. This applies to both sub-samples (QSOs or radio galaxies) as well as to the entire sample. The star formation rates of the PSS-CSO host galaxies with available data are typically in the range ∼0 to 5 ℳ ⊙ yr −1 , and the stellar masses are in the range 3 × 10 11 to 10 12 ℳ ⊙ . Secondly, in agreement with previous results, we find a remarkably high proportion of PSS-CSO host galaxies with merger signatures. The merger fraction of the PSS-CSO sample is 0.61 ± 0.07, which is significantly higher than that of the comparison sample (0.15 ± 0.06). We suggest that this difference can be explained by assuming that the majority of the PSSs and CSOs cannot evolve to extended radio sources and are therefore not represented in our comparison sample.

Astronomy & Astrophysics↗
DOE OSTI2025

Prospects for silvicultural enhancement of fire resistance in mesic westside forests of the Pacific Northwest

Increasing wildfire activity in mesic, temperate Pacific Northwest forests west of the Cascade Range crest has stimulated interest in understanding whether alternative forest management practices could reduce risk of stand-replacing fire. To explore how management can enhance fire resistance in these forests and assess tradeoffs among resistance enhancement, carbon sequestration and storage, and economic returns, we conducted 40-year simulations of stand development with BioSum, a framework for conducting landscape analysis with the Forest Vegetation Simulator (FVS), utilizing a statistically representative and spatially balanced sample of Forest Inventory and Analysis (FIA) plots. Simulation outcomes under business-as-usual silviculture were contrasted with fire-aware silviculture, and treatment optimization logic was developed and applied to represent landscape-scale outcomes under business-as-usual and fire-focused management scenarios. Simulation results indicate that fire-aware prescriptions and fire-focused management can meaningfully enhance stand- and landscape-scale fire resistance of westside forests under less than extreme fire weather, but at the cost of lower economic returns and reduced net carbon storage and sequestration over the 40-year analysis window. Shifting from business-as-usual regeneration harvests with short rotations to fire-aware, episodic selection harvest improved fire resistance the most, especially in young privately-owned forests, and with only modest tradeoffs in carbon and economic outcomes. While fire-aware treatments generally reduced net present value from forest operations over business-as-usual, most treatments still generated positive net present value and could be implemented without subsidy. Fire-aware prescriptions that removed and utilized non-merchantable harvest residues instead of burning them, via either pile or broadcast burning, partially mitigated carbon emissions associated with fire-aware treatments, with about the same improvement in fire resistance. Given the currently limited institutional and financial capacity to implement fire resistance enhancing treatments at scale, the insights from this analysis may aid managers seeking to elevate fire resistance to prioritize where and how to manage.

Science & Technology - Other Topics↗
DOE OSTI2023

Impact-Driven Sampling Strategies for Hybrid Attack Graphs

Cyber-Physical Systems (CPSs) have a large input space, with discrete and continuous elements across multiple layers. Hybrid Attack Graph (HAG) provide a flexible and efficient approach to generate attack sequences for a CPS. Analysis and testing of large-scale HAGs are prohibitively costly. We propose a dimension reduction via property-preserving multi-layer graph sampling algorithms. Existing property-preserving graph sampling approaches generate a representative subgraph of an original large-sized graph while preserving the key properties, such as node and edge distribution, clustering coefficients, and betweenness. On the other hand, we propose impact-driven sampling strategies to transform the input data to a lower-dimensional representation while retaining key properties of the data.

Subasi, Omer↗
DOE OSTI2021

Long-Term Monitoring the Damage Performance of Multilayer Dielectric Grating Samples Residing Inside the Compressor Chamber of the OMEGA EP Laser

Understanding of the laser-induced damage threshold and impact of air–vacuum cycling of the optical components in short-pulse laser systems is of fundamental importance. We report the results of a damage-testing campaign that monitored representative pulse compression grating samples that were positioned inside the OMEGA EP grating compressor vacuum chamber during normal operation and routinely damage tested on every quarterly vent for a period of about 10 years. Here, the evolution of their damage resistance under 10-ps and 100-ps pulse lengths are associated with a significant degradation of the laser-induced damage thresholds, which is comparatively larger at 100 ps, is described.

42 ENGINEERING↗
DOE OSTI2020

High-Resolution Comonomer Sequencing of Blocky Brominated Syndiotactic Polystyrene Copolymers Using 13C NMR Spectroscopy and Computer Simulations

This work demonstrates the first high-resolution comonomer sequencing of Blocky brominated syndiotactic polystyrene (sPS-co-sPS-Br) copolymers based on pentad assign- ments of the quaternary carbon region of the nuclear magnetic resonance spectrum. Copolymers containing p-bromostyrene (Br-Sty) units were prepared in matched sets using postpolymerization bromination methods carried out in the heterogeneous gel state (Blocky) and homogeneous solution state (Random). Quantitative information from the quaternary carbon spectra, heteronuclear multiple bond correlation spectroscopy, electronic structure calculations, and simulated statistically random copolymers was correlated to confirm the carbon resonance assignments for all 20 possible pentad comonomer sequences. Using the experimental pentad sequence prevalences, a computer code was developed to simulate chains with microstructures typical of each sample as a means to visually represent the copolymer blockiness with quantitative precision. Based on the microstructure and distribution of run lengths in these chains, the simulations revealed that the Blocky copolymers contain a high degree of blockiness. By comparing the run lengths in the simulated chains to the average number of styrene units in a crystalline segment of sPS (found by small-angle X-ray scattering), copolymer crystallizability was predicted. For the simulated Blocky B-21% (21 mol % Br-Sty) chain, the probability of randomly selecting a styrene unit in a crystallizable block was 25.8%, while that in the simulated Random R-18% was zero, in excellent agreement with the experimental crystallization behavior measured by differential scanning calorimetry. Additionally, these predictions confirmed that the simulated chains accurately represent the ensemble of chains in their respective copolymer samples. Furthermore, each simulated Blocky chain contained one or more long sPS blocks that paralleled the measured 38-40 styrene units spanning a crystalline segment within the sPS/CCl4 gel. This finding affirmed that the long sPS segments originated from the precise lamellar structure within the heterogeneous gel morphology (i.e., block length is correlated with lamellar thickness). Overall, the ability to tailor the copolymer microstructure through control of the semicrystalline gel morphology opens the door to synthesizing ordered copolymers by postpolymerization functionalization processes with unprecedented levels of compositional control.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

ORNL Analysis of Leach-Burn-Leach Round-Robin Test Samples

An international round-robin test to examine the consistency in leach-burn-leach (LBL) analysis of tristructural-isotropic- (TRISO-) coated particle fuel was conducted by three research organizations from the Generation IV International Forum member countries of the People’s Republic of China, the Republic of Korea, and the United States of America. Two sets of round-robin test samples were exchanged for analysis. One set of samples consisted of a series of nonuranium-bearing, TRISO-coated zirconium dioxide particles seeded with up to four depleted uranium-bearing, TRISO-coated uranium dioxide (UO 2 ) particles, which had intentionally damaged coating layers to simulate either particles with either exposed-kernel defects (i.e., particles with a cracked TRISO coating that should be detected during preburn leaching) or particles with silicon carbide (SiC) defects (i.e., particles with an intact pyrocarbon coating and a hole in the SiC layer that should be detected during postburn leaching). These simulated samples also contained added powder with known quantities of impurities from a coal standard. The other sample set consisted of representative sublots of UO 2 -TRISO particles fabricated in a production-scale coater, except they all contained depleted uranium instead of enriched uranium. In this report, the methodology used at Oak Ridge National Laboratory to conduct LBL analysis of the round-robin samples is presented, and the general results are summarized.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2023

Molybdenum in basalt-hosted seafloor hydrothermal systems: Experimental, theoretical, and field sampling approaches

Here, seafloor hydrothermal vents represent potential sources of Mo and other biologically relevant transition metals to the global ocean, complementing continental runoff. Here, we use a combination of experimental, theoretical, and field-sampling approaches to investigate the behavior of Mo in basalt-hosted seafloor hydrothermal systems to provide insight into the processes controlling Mo concentrations in hydrothermal fluids and to derive estimates of vent fluid Mo concentrations and fluxes. Results of this study demonstrate that reaction fluids generated from 350°C, 500 bar hydrothermal basalt alteration experiments contain 775–801 nmol/kg Mo and are thus comparable to a recently collected time series of natural seafloor vent fluids that contained 200–220 nmol/kg Mo at 302°C and 29–30 nmol/kg at 281–282°C (Evans et al., 2023). Synchrotron-based analyses of experimentally altered basalt produced in this study and additional natural samples of altered oceanic crust originally collected from Pito Deep Rift reveal the presence of Mo-rich particles consistent with trace molybdenite. Comparisons of Mo:Cu ratios in natural vent fluids and near-vent sediment trap samples from Main Endeavour Field indicate that vent fluid Mo is readily incorporated into buoyant plume particles and advected out of the near-vent field, analogous to previous mass balance studies of Cu in this region. Thermodynamic calculations of molybdenite solubility in the context of mineral-buffered hydrothermal fluids and comparisons with natural and experimental hydrothermal fluids suggest that high-temperature vent fluids contain 30–1500 nmol/kg Mo. While a minor component of the modern Mo budget, hydrothermal Mo fluxes are estimated to have constituted 0.3–200× the contemporaneous continental weathering fluxes prior to the ~2.4 Ga ago “Great Oxidation Event” and widespread oxidative continental weathering. Overall, identification of hydrothermal vents as a source of Mo-rich plume particles with potential for dispersal into the wider marine environment has significant implications for hypotheses regarding the co-evolution of Life and Earth’s environments, specifically the form and availability of Mo in anoxic Archean-Eon oceans, where Mo-dependent enzymatic pathways are thought to have emerged and subsequently evolved.

58 GEOSCIENCES↗
DOE OSTI2025

Strategy to safely enable X-ray computed tomography examination of highly radioactive tristructural isotropic nuclear fuel

Nondestructive post-irradiation examination of nuclear fuels and materials is useful in collecting data to inform commercial licensing of new nuclear concepts. This work details the planning, shielding design, and workflow of an X-ray computed tomography examination of an irradiated AGR-5/6/7 fuel compact with a dose rate of 1318 R/hr on contact from both β and ɤ-ray radiation with 120 R/hr of the dose rate coming exclusively from ɤ-rays. Post-irradiation examination of highly radioactive samples can help reduce the timeframe from conceptualization of novel nuclear concepts to their commercial implementation as not only is less time spent waiting for experiments to decay away to acceptable levels for examinations, but data from these experiments can more quickly inform model efforts and subsequent experiments. While this work represents the hottest radiological sample that has been openly transported and examined at Idaho National Laboratory’s Irradiated Materials Characterization Laboratory to date, it is anticipated that this work and its lessons learned will facilitate future examinations of similar or even more radioactive specimens.

42 - ENGINEERING↗
DOE OSTI2025

Temporal relationships among lunar crustal rocks

Temporal relationships among the three most common suites of lunar crustal rocks have been investigated by obtaining new high precision ages on Felsic/Alkali-suite Quartz monzodiorite Clast B from breccia 15405 and Magnesian-suite norite 78235/6/8/55/56 and comparing them to previously dated ferroan anorthosite sample 60025. The weighted average age of 4337.19 ± 0.49 Ma of 15405 Clast B is defined by zircon U-Pb and Pb-Pb ages as well as mineral isochron Sm-Nd and Nd-Nd ages. It is identical to the weighted average age for Apollo 17 norite 78235/6/8/55/56 of 4334.1 ± 3.5 Ma which is defined by Pb-Pb ages measured on baddeleyites in this investigation and less precise Pb-Pb and Sm-Nd ages reported in the literature. Both ages are ∼ 25 Ma younger than the weighted average of Sm-Nd and Pb-Pb ages reported in the literature on ferroan anorthosite 60025 of 4359.3 ± 2.3 Ma. The fact that ages of all three samples are defined by multiple U-Pb, Pb-Pb, Sm-Nd, and 142 Nd- 143 Nd chronometers provide confidence that they record the igneous crystallization history of the samples and do not represent disturbances or mixing lines with no temporal significance. Here, the extent to which these three ages represent broader scale magmatism is difficult to evaluate. Nevertheless, the age defined for 15405 Clast B, 78235/6/8/55/56, and 60025 are contemporaneous with the peak of ages observed in detrital zircons from the Apollo 12, 14, 15, and 17 landing sites (4340 ± 20 Ma), a Mg-suite Sm-Nd whole rock isochron defined by samples from Apollo 14, 15, 16, and 17 landing sites (4348 ± 25 Ma), and a Ferroan Anorthosite-suite Sm-Nd whole rock isochron defined by samples from the Apollo 15 and 16 landing sites (4354 ± 29 Ma). This implies that Ferroan Anorthosite-suite magmatism is temporally distinct and earlier than magmatism associated with the Mg-suite and the Felsic/Alkali-suite, as predicted by the lunar magma ocean model of lunar differentiation. The short 35 ± 10 Ma interval between primary ferroan anorthosite magmatism and secondary magmatism suggests that the lunar crust formed over a limited period of time. Although heat from decay of long-lived isotopes, large impacts, tidal heating associated with interactions between the Earth and Moon, and density driven overturn of the magma ocean have all been invoked to explain production of ancient secondary crustal magmatism, only tidal heating and cumulate overturn are consistent with the apparent short duration of secondary crustal magmatism and the great depth of crystallization implied for some Mg-suite samples. The initial ε 143 Nd values derived from the 15405 Clast B and 78238 Mg-suite norite isochrons, as well as a Mg-suite whole rock isochron are −0.23 ± 0.11, −0.27 ± 0.74, and −0.25 ± 0.09, respectively. They are identical within uncertainty indicating that Mg-suite and Felsic/Alkali-suite magmas were derived from materials that had the same time averaged Sm/Nd ratios since the formation of the solar system. This, combined with the contemporaneous nature of 15405 Clast B and 78235/6/8/55/56 Mg-suite norite, is consistent with evolution of both samples, and likely both magma suites, from a common source through closed system fractional crystallization or partial melting processes.

Chronology↗
DOE OSTI2026

Accessing terapascal pressures on two-stage light-gas guns for high-energy-density research

We present pressure amplifiers that have been developed to enable the study of material behavior at terapascal (TPa) pressures using two-stage light-gas gun (2SLGG) facilities. When impacted by a hyper-velocity projectile, Mach reflections within the pressure amplifier eventually merge to form a Mach stem, generating a planar output shock that greatly exceeds the initial drive. Validation experiments demonstrated that peak pressures of 0.6 and 1.073 TPa were achieved in quartz samples for two different amplifier designs. These pressures represent an approximately fivefold increase over the sample pressures typically attainable at 2SLGG facilities. The temporal and spatial uniformity of the pressure drive is suitable for high-precision equation-of-state measurements of materials relative to a shock standard. This work greatly extends the capabilities of 2SLGG facilities and opens new avenues of study in high-energy-density physics using these drivers.

Gorman, Martin G. [First Light Fusion, Oxford (Uni↗
DOE OSTI2024

Differentiable Multiphysics Codes: A Breakthrough Technology for Simulation and Computing

This document summarizes the findings of a strategic planning exercise commissioned by the Weapons Simulation and Computing, Computational Physics (WSC/CP) program at the Lawrence Livermore National Laboratory (LLNL) in FY24. During the year, the committee met with multiple stakeholder communities to gather input, opinions, suggestions and concerns which have been incorporated throughout this document. The key findings from this exercise are summarized: • The development of multiphysics modelling and simulation (mod/sim) codes and software technologies, their deployment on exascale compute platforms, and their broad adoption across the NNSA is a major success of the Advanced Simulation and Computing (ASC) program and the Exascale Computing Project (ECP). Sustained investment in these core technologies is essential. • Today’s state of the art involves running ensembles of O(100K) simulations to perform uncertainty quantification (UQ) and design studies using multiple statistical methods such as Bayesian optimization to understand sensitivities of our models and explore parameterized design spaces. Even with exascale computing, we are practically limited to O(10) parameters in these studies since the number of simulations required to sample the space scales exponentially with the number of design parameters. • The data from these simulation ensembles is increasingly being used to train machine learned (ML) surrogates (or reduced order models, ROMs) which can then be used for optimization or real time design exploration. However, the trained surrogates are still limited in the number of parameters they can represent due to the sampling limitations previously noted. • Augmenting our suite of integrated multiphysics simulation codes, both current and emerging, with the ability to compute gradients (solution derivatives) of arbitrary simulation outputs with respect to (some or all) simulation inputs would be a breakthrough technology, opening the door to a new era of efficient and automated inverse design based on verified and validated mod/sim capabilities. • This capability, which we refer to as differentiable multiphysics codes (DMCs), would revolutionize both UQ and optimization studies by breaking the curse of dimensionality that presently limits our “gradient-free” ensemble based computing approach. A similar breakthrough occurred in the AI/ML community once the ability to compute gradients of arbitrary loss functions using back-propagation became commonplace. Gradient information from the multiphysics codes can also be used to dramatically improve the efficiency and scale of training of ML/ROM surrogates for rapid assessments. • Achieving this in our suite of codes will be a grand challenge, similar to the amount of effort that was required to transition from CPU to GPU computing. It will require buy-in from the entire WSC/CP program and beyond, including all integrated codes, physics and engineering models, third-party library dependencies and performance portability abstractions. It will also require investment in research and development of numerical methods for computing adjoints of coupled physics across multiple adaptively refined moving meshes and of stochastic (Monte Carlo) and mesh free (SPH) methods. • New software and numerical techniques, largely pioneered by the AI/ML community, make this feasible. Chief among these is automatic differentiation (AD), the ability to employ AD at point-wise locations in a physics calculation (instead of traditional black-box approaches) and the ability to perform “back-propagation in time” (or reverse mode AD) for non-linear partial differential equations (PDEs). Fundamentally, the conclusion of this strategic planning exercise is that the time is right to undertake a large scale effort in WSC, centered on the existing integrated codes, to continue the natural evolution of mod/sim in the age of AI/ML. Instead of attempting to replace mod/sim with purely data driven AI/ML models, we believe the key to success is to integrate AI/ML by building on top of the decades of hard-won knowledge and the verified/validated multiphysics modelling capability that is the hallmark of the ASC program.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Deep reinforcement learning for class imbalance fault diagnosis of equipment in nuclear power plants

In equipment fault diagnosis in nuclear power plants, there may be far more samples in one class (e.g., a health state) than in another class (e.g., a fault state). The distribution of data in each class is highly skewed. Most machine learning algorithms are suitable for balanced training datasets. When faced with imbalanced samples, these algorithms tend to provide good identification for the majority classes and bias for the minority classes. However, the misclassification of minority classes can lead to high costs. To address the above problem, this paper develops a deep reinforcement learning-based diagnosis method that models fault diagnosis as a sequential decision-making process. At each time step, the agent receives the state of the environment represented by the training samples and then takes a diagnosis action guided by a policy. If the action is correct/incorrect, the agent receives a positive/negative reward. The reward for minority classes is higher than that for majority classes. The agent’s goal is to obtain as many cumulative rewards as possible in the process, i.e., to identify the sample as correctly as possible. Six demonstration scenarios are constructed, depending on the selected fault datasets and the designed model structures. Experiments show that the proposed method achieves a higher weighted-averaged F1 score than the classical supervised learning method in most cases of class imbalance. Finally, the proposed method has potential applications in the field of class imbalance fault diagnosis of equipment in nuclear power plants.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2023

Characterizing Reactor Operations from Realistic Simulated Environmental Samples: Combining High-Performance Computing and Data Analytics

Environmental sampling is a common technique employed by inspectors and facility operators in nuclear safeguards, proliferation detection, and process monitoring contexts. Interpreting measurements performed on samples or collections of samples and ensuring the information extracted is accurate and precise is difficult. To date, these analyses have relied on simulated data to enable systematic studies; however, these models are inherently limited by the fidelity of the models and the implicit spatial averaging of isotopic composition or other signatures of interest. To advance this capability, we have refined the spatial discretization and expanded the range of physics in the simulation codes we use to perform reactor simulations and depletion calculations. This allows us to generate data that are more representative of real environmental samples, especially for the length scale of the isotopic composition and associated variation. Accordingly, these new data allow a more realistic assessment of traditional and new data analytic analysis methods. Here we present motivation for developing reactor simulations using high-performance computing methods and resources, impacts of these new simulations on our assessment of data analysis and interpretation methods, and initial results of developing and systematically testing data analytic methods designed to overcome the challenges expected of real-world samples. We also quantify the performance of these analyses using defensible statistical methods.

Dayman, Ken J.↗
DOE OSTI2023

Measured CO 2 sorption isotherms with 25 Bakken Petroleum System rock samples from the Lower and Upper Shales, Middle Bakken, and Three Forks formations

Isotherms were measured with 25 Bakken Petroleum System (BPS) rock samples using a magnetic suspension balance at reservoir conditions of 110 °C and pressures up to 345 bar. Samples from four wells represented the Middle Bakken (MB) and Three Forks (TF) production zones, and the Upper Bakken Shale (UBS) and Lower Bakken Shale (LBS) source shales. UBS and LBS CO 2 sorption capacities at 345 bar averaged 12–14 g/kg rock compared to 1.3 to 2.5 g/kg rock for MB and TF. Increasing pressure from 150 to 345 bar resulted in only small increases in CO 2 sorption for all lithofacies. For the UBS and LBS source shales, the higher average total organic carbon (TOC) values (14.7 and 11.4 wt.%) and clay values (25 and 28 wt.%) corresponded to their higher CO 2 sorption values while for the MB and TF samples, lower TOCs (0.3 and 0.5 wt.%) and clay values (8.2 and 18.2 wt.%) corresponded to lower CO 2 sorption values. Here, when all 25 samples (from all lithofacies) were evaluated as a group, sorption capacities were highly correlated with TOC (Spearman's rho values +0.84 to +0.85), with rock density (rho values -0.75 to -0.76), and clay content (rho values +0.66). However, correlations with porosity and thermal maturity (Tmax) were weak. Measured sorption capacities and rock porosities showed that sorption accounted for about 10-times as much CO 2 storage as the available pore space for the UBS and LBS lithofacies. In contrast, the pore space dominated the storage resource of the TF and MB lithofacies.

54 ENVIRONMENTAL SCIENCES↗