Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “distributed inference”

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 37 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

Basic factors for acquiring, correcting, and interpreting probe current-voltage characteristic in moderate-collisional plasma for determining electron energy distribution

Methods for inferring the electron distribution function (EDF) from a probe current–voltage trace are reviewed for moderate-collisional plasma. The explicit accounting of particle collisions, modeling sheath depletion, and electron refilling processes are rarely performed when analyzing probe measurements, yet they are critical for regime validity, EDF-profile accuracy, and space potential quantification precision. Furthermore, the advantages, limitations, and utilization of large-area wall probes in moderate-collisional plasma, capable of highly resolved indications of energetic electrons in plasma, are outlined.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2022

Ion energy distribution in an electron beam ion trap inferred from simulations of the trapped ion cloud

For this work, we have inferred the energy distribution of trapped ions in an electron beam ion trap (EBIT) from simulations of the spatial distribution of Fe 13+ ions and a comparison with measured visible light images of the ion cloud. We simulated the cloud of Fe 13+ ions by computing ion trajectories in the EBIT for different ion energy distributions used to initialize the trajectories. We then performed a least-squares fit to infer the ion energy distribution that best reproduced the measured ion cloud. These best-fit distributions were typically non-Maxwellian. For electron beam energies of 395–475 eV and electron beam currents of 1–9 mA, we find that the average ion energy is in the range of 10–300 eV. We also find that the average ion energy increases with increasing beam current approximately as $\langle$E$\rangle$ ≈ 25 I e eV, where I e is the electron beam current in mA. We have also compared our results to Maxwell-Boltzmann-distribution ion clouds. We find that our best-fit non-thermal distributions have an $\langle$E$\rangle$ that is less than half that of the T from the best-fit Maxwell-Boltzmann distributions ($\langle$E$\rangle$/q)/T=0.41 ± 0.05.

74 ATOMIC AND MOLECULAR PHYSICS↗
DOE OSTI2023

Towards Realistic and High Fidelity Models for Nuclear Reactor Power Synthesis Simulation with Self-Powered Neutron Detectors

As presented in this report, a weighting function–based inferencing method is being applied to synthesize the power distribution in next-generation and university research reactors based on simulated self power neutron detector (SPND) responses. The overall goal is to assess the impacts of sensor uncertainty and true power distribution perturbations on the error in the synthesized power distribution. Regarding sensor uncertainty, the NuScale Small Modular Reactor (SMR) and the Westinghouse AP1000 serve as testbeds for analyzing the impact of varying the sensor uncertainty, as well as varying the number of sensors per sensor string in the reactor core. The reactor models are informed by Monte Carlo N-Particle (MCNP) neutron flux tallies. For the NuScale SMR and Westinghouse AP1000, the SPND response functions (i.e., the response of the SPNDs to individual segments of fuel) were determined homogeneously. Regarding an analysis of power distribution perturbation detection, the Texas A&M Testing, Research, Isotopes, General Atomics Reactor (TAMU TRIGA) reactor was used as a demonstration case with one particular arrangement of SPNDs; the response functions for this reactor model were determined heterogeneously, making this a uniquely high-fidelity demonstration of perturbation detection. Finally, SPND models generated in the Geometry and Tracking 4 (Geant4) code have been generated and tested for comparison with traditionally implemented analytical SPND models, with the intent for Geant4 integration with the full methodological framework. SPND current outputs as a function of distance from some fuel assembly segment in the NuScale SMR are compared with the analytically determined currents. Results from the sensor uncertainty simulations for the NuScale SMR and AP1000 indicate that the average error in the inferred power distribution on the fuel assembly segment level is reasonably low, being slightly less than the random uncertainty applied to all respective SPNDs in both cores. For example, if all SPNDs in the core have a random uncertainty of 5%, then the corresponding fuel assembly segment level error (i.e. difference between the true and inferred local power) is ~2–3%. However, the maximum error in the inferred power distribution on the fuel assembly segment level can be considerably high (>15%) when SPND random uncertainties start to exceed ~3%. In general, the average and maximum errors in the inferred power distribution were slightly higher in the AP1000 as opposed to the NuScale SMR for the sensor string configurations considered herein. Another result determined from analysis of the sensor uncertainty simulations was that increasing the number of SPNDs per string does not clearly reduce inferred power distribution error and can in fact make the error large in some cases; however, this assessment may skewed due to imposed iteration limits. Results from the perturbation detection demonstration using the high-fidelity TAMU TRIGA model indicate that, given the arrangement of 17 SPND strings and 4 SPNDs per string considered herein, there is a clear, provable ability to infer a localized Gaussian-type peak perturbation in the 3D power distribution. Such a perturbation was detected with an average fuel assembly segment level error of 0.19%, and the general visualization of the detected perturbation clearly indicates that the magnitude and shape were appropriately resolved. Finally, the electrical current output generated by the Geant4 modeled SPND indicates significant magnitude differences than the analytically modeled SPND, demonstrating the need for accurate SPND models which account for finite sensor geometry effects to inform the power synthesis work described herein.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2026

A Nonparametric Method for the Inference of Halo Occupation Distributions

The galaxy–halo connection traces processes by which galaxies form and evolve. The halo occupation distribution (HOD) describes the relationship between galaxies and their host dark matter haloes. Measurements of the galaxy two-point correlation function (2PCF) allow us to extract information about the HODs of observed galaxy samples. Several parametric HOD models have been proposed in the literature, but the choice of parameterization restricts the space of possible HODs. To resolve this issue, we introduce a nonparametric HOD fitting method in which we train an emulator to learn the mappings among the galaxy 2PCF, physical properties used to select galaxy samples, and the HOD, all obtained from simulated past light cones constructed with the Santa Cruz semianalytic model. Implementing this emulator within a likelihood analysis framework, we derive constraints on the HOD of a galaxy sample when provided with a measurement of its 2PCF. Using the emulator to accelerate likelihood evaluations, we test the nonparametric HOD approach on a set of 2PCFs for mock galaxy samples drawn from the TNG100-1 simulation and selected above threshold values of stellar mass and star formation rate. Our framework is able to recover TNG100-1 HODs within 0.2 dex. We use the TNG100-1 mocks to tune the reported uncertainties to estimate those expected in the analysis of observations. Comparing to parametric HOD modelling routines applied to the same mock galaxy samples, our approach consistently infers the HOD with comparable or greater precision and accuracy.

Kennedy, Jacob [Rutgers Univ., Piscataway, NJ (Uni↗
DOE OSTI2020

Solving Stochastic Inverse Problems for Property–Structure Linkages Using Data-Consistent Inversion and Machine Learning

Determining process–structure–property linkages is one of the key objectives in material science, and uncertainty quantification plays a critical role in understanding both process–structure and structure–property linkages. In this work, we seek to learn a distribution of microstructure parameters that are consistent in the sense that the forward propagation of this distribution through a crystal plasticity finite element model matches a target distribution on materials properties. This stochastic inversion formulation infers a distribution of acceptable/consistent microstructures, as opposed to a deterministic solution, which expands the range of feasible designs in a probabilistic manner. Furthermore, to solve this stochastic inverse problem, we employ a recently developed uncertainty quantification framework based on push-forward probability measures, which combines techniques from measure theory and Bayes’ rule to define a unique and numerically stable solution. This approach requires making an initial prediction using an initial guess for the distribution on model inputs and solving a stochastic forward problem. To reduce the computational burden in solving both stochastic forward and stochastic inverse problems, we combine this approach with a machine learning Bayesian regression model based on Gaussian processes and demonstrate the proposed methodology on two representative case studies in structure–property linkages.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Simulating self-powered neutron detector responses to infer burnup-induced power distribution perturbations in next-generation light water reactors

Understanding how 3D power distribution will be monitored throughout reactor core volumetric space in next-generation nuclear power reactors is crucial to the design, deployment, and licensing of these reactors. Although numerous techniques exist for 3D power distribution monitoring based on the response of both in situ and ex situ sensors currently implemented or proposed for use in the US reactor fleet, crucial details about these techniques are often unclear. The publicly available documentation does not include information such as how well these techniques are characterized and optimized in their implementations and the levels of uncertainty in the inferred 3D power distribution. The work described herein investigated a recently developed 3D power distribution inferencing method as applied to two next-generation reactor simulations: (1) the NuScale small modular reactor design and (2) the Westinghouse AP1000 design, both of which contain in-core strings of vanadium self-powered neutron detectors (SPNDs). This investigation considered a range of SPND string sensor densities, as well as a range of 3D power distribution axial segment sizes. In this work, SPND response simulation is informed by neutron flux calculations in representative homogenized cores. For the different sensor densities and power distribution axial segment sizes in these simulations, the average solution error, solver iterations, and run time were tracked to parameterize the sensor-core configuration.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2024

Machine learning assisted bayesian inference of mix and hot-spot conditions in NIF implosions

Experiments on the National Ignition Facility (NIF) have provided clear evidence of ablator material mixing into the Hot-Spot, leading to degraded performance. However, inferring the amount of mix and Hot-Spot conditions from typical experimental observations (e.g. x-ray spectra and images) is highly challenging. Here, we have developed an analysis method that utilizes machine learning assisted Bayesian inference to find the probability distributions of the Hot-Spot and mix conditions. This approach uses a neural network, trained on an idealized 2-dimensional representation of the Hot-Spot and mix distribution, and Bayesian inference to find the statistical distributions of Hot-Spot conditions that provide a match with observations. We have tested this method with synthetic data from simulations.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2021

Stochastic Gradient-Based Distributed Bayesian Estimation in Cooperative Sensor Networks

Distributed Bayesian inference provides a full quantification of uncertainty offering numerous advantages over point estimates that autonomous sensor networks are able to exploit. However, fully-decentralized Bayesian inference often requires large communication overheads and low network latency, resources that are not typically available in practical applications. In this paper, we propose a decentralized Bayesian inference approach based on stochastic gradient Langevin dynamics, which produces full posterior distributions at each of the nodes with significantly lower communication overhead. We provide analytical results on convergence of the proposed distributed algorithm to the centralized posterior, under typical network constraints. Finally, we also provide extensive simulation results to demonstrate the validity of the proposed approach.

42 ENGINEERING↗
DOE OSTI2023

Inferring the Energy and Distance Distributions of Fast Radio Bursts Using the First CHIME/FRB Catalog

Abstract Fast radio bursts (FRBs) are brief, energetic, typically extragalactic flashes of radio emission whose progenitors are largely unknown. Although studying the FRB population is essential for understanding how these astrophysical phenomena occur, such studies have been difficult to conduct without large numbers of FRBs and characterizable observational biases. Using the recently released catalog of 536 FRBs published by the Canadian Hydrogen Intensity Mapping Experiment/Fast Radio Burst (CHIME/FRB) collaboration, we present a study of the FRB population that also calibrates for selection effects. Assuming a Schechter function, we infer a characteristic energy cut-off of E char = 2.38 − 1.64 + 5.35 × 10 41 erg and a differential power-law index of γ = − 1.3 − 0.4 + 0.7 . Simultaneously, we infer a volumetric rate of [ 7.3 − 3.8 + 8.8 (stat.) − 1.8 + 2.0 ( sys . ) ] × 10 4 Gpc −3 yr −1 above a pivot energy of 10 39 erg and below a scattering timescale of 10 ms at 600 MHz, and find we cannot significantly constrain the cosmic evolution of the FRB population with star-formation rate. Modeling the host’s dispersion measure (DM) contribution as a log-normal distribution and assuming a total Galactic contribution of 80 pc cm −3 , we find a median value of DM host = 84 − 49 + 69 pc cm −3 , comparable with values typically used in the literature. Proposed models for FRB progenitors should be consistent with the energetics and abundances of the full FRB population predicted by our results. Finally, we infer the redshift distribution of FRBs detected with CHIME, which will be tested with the localizations and redshifts enabled by the upcoming CHIME/FRB Outriggers project.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

Chemistry imaging and distribution analysis of rare earth elements in coal using LIBS and LA-ICP-MS instruments

Currently, demand for rare earth elements (REEs) increased significantly. Coal is actively evaluated as potential economic sources for extraction of REEs. Here, in this work, laser-induced breakdown spectroscopy (LIBS) was evaluated for rapid estimation of REEs content and their distribution in the natural coal samples. The results were compared with similar laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) measurements. Thirteen coal samples (nine standard samples and five natural samples) were used in this study. Powder samples were pressed into pellets while coal chunks were directly ablated for data recording. Pellets of the powder standard samples were used to optimize the data acquisition system and then data recorded with this optimized system was used to identify the proper data acquisition and analysis models. After establishing the proper data acquisition system and analysis model using the standard samples, natural coal samples in powder form and their chunks were utilized to record LIBS and LA-ICP-MS spectra. Multivariate calibration models were developed using four of the natural samples, which were evaluated by predicting the REE content in the fifth sample. Principal component analysis was performed on the LIBS data obtained from the natural samples and it classified all the samples with high accuracy. Two-dimensional (2D) elemental mapping on coal chunk samples was also performed using both LIBS and LA-ICP-MS to study the distribution of REEs in the samples. The resulting elemental images and their correlations can be used to infer mineral distributions.

01 COAL, LIGNITE, AND PEAT↗
DOE OSTI2025

Accelerating Hamiltonian Monte Carlo for Bayesian inference in neural networks and neural operators

Hamiltonian Monte Carlo (HMC) is a powerful and accurate method to sample from the posterior distribution in Bayesian inference. However, HMC techniques are computationally demanding for Bayesian neural networks due to the high dimensionality of the network’s parameter space and the non-convexity of their posterior distributions. Therefore, various approximation techniques, such as variational inference (VI) or stochastic gradient MCMC, are often employed to infer the posterior distribution of the network parameters. Such approximations introduce inaccuracies in the inferred distributions, resulting in unreliable uncertainty estimates. In this work, we propose a hybrid approach that combines inexpensive VI and accurate HMC methods to efficiently and accurately quantify uncertainties in neural networks and neural operators. The proposed approach leverages an initial VI training on the full network. We examine the influence of individual parameters on the prediction uncertainty, which shows that a large proportion of the parameters do not contribute substantially to uncertainty in the network predictions. This information is then used to significantly reduce the dimension of the parameter space, and HMC is performed only for the subset of network parameters that strongly influence prediction uncertainties. This yields a framework for accelerating the full batch HMC for posterior inference in neural networks. We demonstrate the efficiency and accuracy of the proposed framework on deep neural networks and operator networks, showing that inference can be performed for large networks with tens to hundreds of thousands of parameters. Finally, we show that this method can effectively learn surrogates for complex physical systems by modeling the operator that maps from upstream conditions to wall-pressure data on a cone in hypersonic flow.

Bayesian inference↗
DOE OSTI2021

Resource selection functions based on hierarchical generalized additive models provide new insights into individual animal variation and species distributions

Habitat selection studies are designed to generate predictions of species distributions or inference regarding general habitat associations and individual variation in habitat use. Such studies frequently involve either individually indexed locations gathered across limited spatial extents and analyzed using resource selection functions (RSFs) or spatially extensive locational data without individual resolution typically analyzed using species distribution models. Both analytical methodologies have certain desirable features, but analyses that combine individual- and population-level inference with flexible non-linear functions may provide improved predictions while accounting for individual variation. Here, we describe how RSFs can be fit using hierarchical generalized additive models (HGAMs) using widely available software, providing a means to explore individual variation in habitat associations and to generate species distribution maps. We used GPS tracking data from golden eagles Aquila chrysaetos from across eastern North America with four environmental predictors to generate monthly distribution models. We considered three model structures that assumed different amounts of individual variation in the functional relationship between predictors and habitat use and used k-fold cross-validation to compare model performance. Models accounting for individual variability in shape and smoothness of functional responses performed best. Eagles exhibited the least amount of individual variation in response to land cover variables during winter months, with most individuals more closely adhering to the population-level trend. During the summer months, eagles exhibited more substantial individual variation in shape and smoothness of the functional relationships, suggesting some need to account for individual variation in eagle habitat use for both inferential and predictive purposes, during this time of year. Because they allow users to blend flexible functions with random effects structures and are well-supported by a variety of software platforms, we believe that HGAMs provide a useful addition to the suite of analyses used for modeling habitat associations or predicting species distributions.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Quantifying electron temperature distributions from time-integrated x-ray emission spectra

K-shell x-ray emission spectroscopy is a standard tool used to diagnose the plasma conditions created in high-energy-density physics experiments. In the simplest approach, the emissivity-weighted average temperature of the plasma can be extracted by fitting an emission spectrum to a single temperature condition. It is known, however, that a range of plasma conditions can contribute to the measured spectra due to a combination of the evolution of the sample and spatial gradients. In this work, we define a parameterized model of the temperature distribution and use Markov Chain Monte Carlo sampling of the input parameters, yielding uncertainties in the fit parameters to assess the uniqueness of the inferred temperature distribution. Here we present the analysis of time-integrated S and Fe x-ray spectroscopic data from the Orion laser facility and demonstrate that while fitting each spectral region to a single temperature yields two different temperatures, both spectra can be fit simultaneously with a single temperature distribution. We find that fitting both spectral regions together requires a maximum temperature of $1310^{+90}_{-70}$ eV with significant contributions from temperatures down to 200 eV.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2022

High Resolution Ion Energy Distribution Measurements for Energization Studies on Madison Symmetric Torus RFP Plasmas

We propose to continue a comprehensive and aggressive collaborative study of the critical issues of RFP and fusion physics at MST by enhancing the inferred ion distribution function and the high (spatial and temporal) resolution of magnetic fluctuation measurements. For this, we propose to make novel use of a retarding potential before a neutral particle analyzer as well as a novel millimeter magnetic probe. The measurement of relationships between ion and electron temperatures and magnetic field fluctuations can provide otherwise unobtainable insights on the role of turbulence in fusion plasmas with special focus on the impact of magnetic field turbulence on plasma confinement and scaling. These insights can be particularly valuable when they are coupled with an increased understanding of the issues which drive the turbulence onset, enhance steady-state sustainment, and enable turbulent dissipation. With these new insights, MST will be particularly well positioned to provide important insights in a manner portable to a broad class of fusion devices. We propose to exploit this opportunity and, at the same time, take full advantage of the associated educational enhancements for plasma physics students as FAMU continues to expand the diversity of the fusion scientific workforce.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2024

Thinking Bayesian for plasma physicists

Bayesian statistics offers a powerful technique for plasma physicists to infer knowledge from the heterogeneous data types encountered. To explain this power, a simple example, Gaussian Process Regression, and the application of Bayesian statistics to inverse problems are explained. The likelihood is the key distribution because it contains the data model, or theoretic predictions, of the desired quantities. By using prior knowledge, the distribution of the inferred quantities of interest based on the data given can be inferred. Because it is a distribution of inferred quantities given the data and not a single prediction, uncertainty quantification is a natural consequence of Bayesian statistics. The benefits of machine learning in developing surrogate models for solving inverse problems are discussed, as well as progress in quantitatively understanding the errors that such a model introduces.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2023

Software-Hardware Co-design of Heterogeneous SmartNIC System for Recommendation Models Inference and Training

Deep Learning Recommendation Models (DLRMs) are critical applications in various domains and have evolved as one of the single largest machine learning applications. Trillions of DLRM parameters exceed the on-chip memory capacity of GPUs. Large-scale multi-node systems are required for distributed DLRM inference and training, which suffer from the all-to-all communication bottleneck, mainly limiting the scalability of ever-growing DLRMs. In recent years, SmartNICs have evolved with coupled computation and communication capabilities providing opportunities for a powerful heterogeneous device in the system. However, there isn't such a distributed system that fully leverages the abundant smartNIC resources that resolve the scalability issue of DLRMs. In this work, we proposed a software-hardware co-design of a heterogeneous smartNIC system that resolves the communication bottleneck of distributed DLRMs, mitigates the memory bandwidth pressure, and improves computation efficiency. We provide a set of smartNIC designs of cache systems (including local cache and remote cache) and smartNIC computation kernels which reduce data movement, relieve memory lookup intensity, and improve the GPU's computation efficiency. In addition, we propose a graph algorithm that improves the data locality of queries within batches which optimizes the overall system performance with higher data reuse. Our evaluation shows that our system achieves 2.1x latency speedup for inference and 1.6x throughput speedup for training.

Guo, Anqi↗
DOE OSTI2020

The age distribution of global soil carbon inferred from radiocarbon measurements

Soils contain more carbon than the atmosphere and vegetation combined. An increased flow of carbon from the atmosphere into soil pools could help mitigate anthropogenic emissions of carbon dioxide and climate change. Yet we do not know how quickly soils might respond because the age distribution of soil carbon is uncertain. Here we used 789 radiocarbon (Δ 14 C) profiles, along with other geospatial information, to create globally gridded datasets of mineral soil Δ 14 C and mean age. We found that soil depth is a primary driver of Δ 14 C, whereas climate (for example, mean annual temperature) is a major control on the spatial pattern of Δ 14 C in surface soil. Integrated to a depth of 1m, global soil carbon has a mean age of 4,830 ± 1,730 yr, with older carbon in deeper layers and permafrost regions. In contrast, vertically resolved land models simulate Δ 14 C values that imply younger carbon ages and a more rapid carbon turnover. Overall, our data-derived estimates of older mean soil carbon age suggest that soils will accumulate less carbon than predicted by current Earth system models over the twenty-first century. Reconciling these models with the global distribution of soil radiocarbon will require a better representation of the mechanisms that control carbon persistence in soils.

58 GEOSCIENCES↗
DOE OSTI2023

Reliable extrapolation of deep neural operators informed by physics or sparse observations

Deep neural operators can learn nonlinear mappings between infinite-dimensional function spaces via deep neural networks. As promising surrogate solvers of partial differential equations (PDEs) for real-time prediction, deep neural operators such as deep operator networks (DeepONets) provide a new simulation paradigm in science and engineering. Pure data-driven neural operators and deep learning models, in general, are usually limited to interpolation scenarios, where new predictions utilize inputs within the support of the training set. However, in the inference stage of real-world applications, the input may lie outside the support, i.e., extrapolation is required, which may result to large errors and unavoidable failure of deep learning models. Here, we address this challenge of extrapolation for deep neural operators. First, we systematically investigate the extrapolation behavior of DeepONets by quantifying the extrapolation complexity, via the 2-Wasserstein distance between two function spaces and propose a new strategy of bias–variance trade-off for extrapolation with respect to model capacity. Subsequently, we develop a complete workflow, including extrapolation determination, and we propose five reliable learning methods that guarantee a safe prediction under extrapolation by requiring additional information—the governing PDEs of the system or sparse new observations. The proposed methods are based on either fine-tuning a pre-trained DeepONet or multifidelity learning. We demonstrate the effectiveness of the proposed framework for various types of parametric PDEs. Furthermore, our systematic comparisons provide practical guidelines for selecting a proper extrapolation method depending on the available information, desired accuracy, and required inference speed.

42 ENGINEERING↗