SEARCH · Engineering Papers

Results for “factorization machine”

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.

SourceResearch connectionFrom yearThrough year

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

At least 937 records · Page 52

All sources NASA NTRS DOE OSTI

DOE OSTI2018

The plastic flow machining: A new SPD process for producing metal sheets with gradient structures

Highlights: • A new process named Plastic Flow Machining is presented to produce thin sheets by an extrusion operation from bulk material. • The produced sheets present ultra-fine grain structures, good mechanical properties and high formability. • A deformation gradient was found across the thickness of the sheet regarding plastic strain, microstructure and texture. - Abstract: A new severe plastic deformation process is presented – the plastic flow machining – an extrusion process during which a sheet is formed through a lateral opening of the die in a single operation. The sheet is heavily deformed presenting an ultra-fine grain structure. Experimental results are presented for commercially pure aluminum concerning microstructures, textures and mechanical properties. A gradient structure was found across the thickness of the sheet concerning plastic strain, grain size and texture. Only simple shear textures were seen. The material strength increased by a factor of three with significant ductility and with relatively high Lankford parameter.

36 MATERIALS SCIENCE↗

DOE OSTI2026

Resolving root causes of experiment discrepancies guided by machine learning

Abstract Scientists rely on accurate experimental data to explain nature and then harness this knowledge for applications addressing human needs. However, discrepancies between experiments of the same observable can impede scientific progress if one does not understand the underlying causes. Here, we developed a process that unravels data discrepancies by first using Bayesian machine learning to relate discrepancies to few of many, potentially biasing metadata features that encode experiment procedures. This machine learning output guides human experts to study discrepancy causes by simulating suspicious aspects of historical experiments or designing modern ones to address open questions. The study findings then lead to rejecting or correcting historical data on firm scientific bases. This process is demonstrated for the energy spectrum of neutrons emitted promptly (<1 ns) after fission of 252 Cf, a trusted nuclear physics Standard. It reduces the spread in experimental 252 Cf spectra by up to a factor of 6.

Neudecker, D. (ORCID:0000000339200627)↗

DOE OSTI2016

SU-F-T-230: A Simple Method to Assess Accuracy of Dynamic Wave Arc Irradiation Using An Electronic Portal Imaging Device and Log Files

Purpose: The Dynamic Wave Arc (DWA) technique, where the multi-leaf collimator (MLC) and gantry/ring move simultaneously in a predefined non-coplanar trajectory, has been developed on the Vero4DRT. The aim of this study is to develop a simple method for quality assurance of DWA delivery using an electronic portal imaging device (EPID) measurements and log files analysis. Methods: The Vero4DRT has an EPID on the beam axis, the resolution of which is 0.18 mm/pixel at the isocenter plane. EPID images were acquired automatically. To verify the detection accuracy of the MLC position by EPID images, the MLC position with intentional errors was assessed. Tests were designed considering three factors: (1) accuracy of the MLC position (2) dose output consistency with variable dose rate (160–400 MU/min), gantry speed (2.4–6°/s), ring speed (0.5–2.5°/s), and (3) MLC speed (1.6–4.2 cm/s). All the patterns were delivered to the EPID and compared with those obtained with a stationary radiation beam with a 0° gantry angle. The irradiation log, including the MLC position and gantry/ring angle, were recorded simultaneously. To perform independent checks of the machine accuracy, the MLC position and gantry/ring angle position were assessed using log files. Results: 0.1 mm intentional error can be detected by the EPID, which is smaller than the EPID pixel size. The dose outputs with different conditions of the dose rate and gantry/ring speed and MLC speed showed good agreement, with a root mean square (RMS) error of 0.76%. The RMS error between the detected and recorded data were 0.1 mm for the MLC position, 0.12° for the gantry angle, and 0.07° for the ring angle. Conclusion: The MLC position and dose outputs in variable conditions during DWA irradiation can be easily detected using EPID measurements and log file analysis. The proposed method is useful for routine verification. This research is (partially) supported by the Practical Research for Innovative Cancer Control (15Ack0106151h0001) from Japan Agency for Medical Research and development, AMED. Authors Takashi Mizowaki and Masahiro Hiraoka have consultancy agreement with Mitsubishi Heavy Industries, Ltd., Japan.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2025

Understanding the Quadrupole Mass Filter and Testing a High-Resolution QMS RGA for ITER

A common type of residual gas analyzer is the quadrupole mass spectrometer. One of the main components within this instrument is a mass filter known as the quadrupole. It is responsible for the selective throughput of the ionized gas particles - by ascending mass number - prior to ion impacts on the analyzer (or detector) surface from which the ion current signal is generated for processing. However, the quadrupole is not fully described in relation to the electric field characteristics and the function as an ion mass separator. This paper describes the basic origins of the electrical design, the intricate assembly criteria, and performance of the quadrupole within the spectrometer. A specialized quadrupole mass spectrometer is part of a configuration for a diagnostic gas analyzer system planned for ITER, a fusion research machine. It has a verified capability, essential as a diagnostic criterion for this reactor project, to successfully deconvolute the mass signals of Helium-4 and deuterium (reactor fuel exhaust gases, separated by only 0.026 atomic mass units), down to a relative three-percent concentration of the former gas. The associated preliminary testing, performed at the Oak Ridge National Laboratory, is also addressed. Finally, one of the key parameters used to express gas concentration, the relative sensitivity factor, will be explained, including an evaluation of dependency on other variables.

Marcus, Chris [ORNL] (ORCID:0000000190139636)↗

DOE OSTI2024

Assessment of Machine Learning for Ultrasonic Nondestructive Evaluation of Alkali–Silica Reaction in Concrete

Alkali–silica reaction (ASR) is a type of material degradation in concrete structures that leads to concrete cracking and rebar corrosion, thereby reducing the material’s structural integrity and the overall structure’s lifetime and raising safety concerns. Ultrasonic nondestructive evaluation (NDE) has been proven to be a valuable technique for assessing concrete properties and monitoring ASR progression in concrete. However, the deployment and analysis of ultrasonic NDE and its data requires specialized expertise, often relying on the engineer’s subjective interpretation. With the surge in computational power, artificial intelligence (AI) and machine learning (ML) algorithms have become popular in automating NDE data analysis. Various industrial sectors are increasingly adopting ML algorithms for NDE data analysis with a growing emphasis on AI–assisted automation. Regulatory agencies are also preparing for this technological shift, anticipating corresponding revisions in standards. Thus, there is an urgent need to identify the capabilities and limitations of current ML technologies for the evaluation of concrete material properties and damage status. Furthermore, the effects of various factors on ML model performance must be thoroughly investigated. The study summarized herein evaluated the effectiveness of two ML models (i.e., support vector regression (SVR) and deep neural network (DNN)) in predicting concrete material damage induced by ASR based on the long-term ultrasonic monitoring data. Four distinct concrete specimens were cast with artificially induced ASR, and over a period exceeding 500 days, ultrasonic signals and expansion data were continuously collected. For the SVR model, wave velocity and 12 other wave features were extracted from the ultrasonic signals, with 6 out of 13 features selected as input for the model. Different combinations of training and testing datasets were designed to explore factors influencing prediction performance, including the range of data within training and testing sets, in addition to various signal preprocessing methodologies. These findings suggest the importance of using a training dataset with a broader data range compared with testing datasets for improved model performance alongside consistent signal preprocessing across datasets.

36 MATERIALS SCIENCE↗

DOE OSTI2026

Predicting river turbidity in Pine Island Bayou using machine learning techniques coupled with variational mode decomposition

Elevated turbidity levels pose significant public health risks by facilitating the transport of harmful pollutants, including metals, organic compounds, and pathogenic microorganisms into the surface water. These conditions create serious challenges for public recreational water use and drinking water treatment, leading to economic losses and health risks. This study utilizes water monitoring data in Pine Island Bayou, Texas, and develops a Sequence-to-Sequence (S2S) model to predict turbidity using Attention-based Gated Recurrent Units with Encoder-Decoder (AT-GRU-ED) and Long Short-Term Memory (LSTM), coupled with Variational Mode Decomposition (VMD). Compared to the model without VMD, the model demonstrates satisfactory 72-hour turbidity prediction performance, achieving MAEs of 2.60 and 3.29 NTU (reductions of 53% and 58%), RMSEs of 21.08 and 31.49 NTU (reductions of 82% and 80%), and R² values of 0.96 and 0.84 on the validation and test sets, respectively. Feature importance analysis reveals that water temperature is the dominant factor influencing seasonal turbidity patterns, while real-time hourly rainfall significantly contributes to short-term variability. Turbidity typically peaks within 48 hours after rainfall events due to lagged effects from surface runoff and upstream flow. Findings suggest suspending recreational water use and water supply pumping for three days after heavy rainfall can benefit public health and improve water treatment processes. Discharges above 100 m3/s are found to accelerate sediment dilution and transport, reducing turbidity levels more quickly after the peak. In conclusion, the proposed model demonstrates reliable 72-hour turbidity prediction, supporting decision-making for water treatment plant operations and providing early warning for public recreational water use.

Deep learning↗

DOE OSTI2021

Electrical Systems of Pumped Storage Hydropower Plants: Electrical Generation, Machines, Power Electronics, and Power Systems

Adjustable-speed (AS) pumped storage hydropower (PSH) technology has the potential to become a large, consistent contributor to grid stability enabling higher penetrations of wind and solar energy on the future U.S. power system. AS-PSH has high-value characteristics, such as fast response to provide ancillary services to the grid, because it is a power converter interface with the grid (like battery storage), but at the same time it has the energy content large enough to supply both short-term (seconds-to-minutes) and long-term (minutes-to-hours) of energy needs, like more conventional power plants. However, designs must be optimized to lower the capital expenditure (CapEx) and to provide a high-quality grid interface capability (e.g. power quality, ancillary service provider, fault-tolerant or fault ride-through capability), which is a primary factor in the acceptance of AS-PSH into a utility’s generation mix. This CapEx will be greatly affected by the cost savings associated with the civil structure, turbine design, power electronics, control systems, or unique generator designs. A holistic design must be considered to get a full picture of the benefits of the technology proposed. The AS-PSH can be controlled to reduce the impact of transient disturbance on a power system and at the same time can be controlled to minimized subsequent component fatigue and potential oscillation modes within the plant, with the overall impact in reducing the operational expenditures (OpEx).Generating clean power to meet standards such as Institute of Electrical and Electronics Engineers (IEEE) 519 and International Electrotechnical Commission (IEC) 1000-3-2 will be a continuing challenge. For many technology developers, however, improved AS-PSH technologies will become a key component of generator-storage systems of the future given the prospects of increased performance and decreasing costs, and the ever-increasing penetration of renewables (e.g wind power and solar power).

13 HYDRO ENERGY↗

DOE OSTI2020

Shear localization in metallic materials at high strain rates

Three factors govern adiabatic shear localization: strain hardening (or softening), strain-rate hardening, and thermal softening. It is typically associated with large shear strains (>1), high strain rates (10 3 –10 7 s –1 ), and high temperatures (40–100% of melting point), all of which happen within narrow regions with widths of about 1–200 μm. It is often an undesirable phenomenon, leading to failure, but there are situations where it is desirable, e. g., the generation of machining chips. Here, we review the development of both theoretical and experimental achievements, from the initiation of shear bands to their propagation with emphasis on three aspects: novel experimental techniques, novel materials, and nano/microstructural effects. The principal characteristics of adiabatic shear bands in metallic materials at the nano- and micro-scale are described. Bands that were formerly identified as transformed actually consist of nanocrystalline/ultrafine grains. These grains result from the breakup of the microstructure by a rotational recrystallization process. The evolution of the microstructure inside shear bands and their interactions for hcp, bcc, and fcc alloys, high-entropy alloys, nanocrystalline alloys, and metallic glasses are analyzed mechanistically. In conclusion, the gaps in the field and opportunities for future research are identified. Modern experimental characterization and computational techniques enable a more profound and predictive understanding of adiabatic shear localization and its avoidance in advanced materials.

36 MATERIALS SCIENCE↗

DOE OSTI2024

Deep-Learning Interatomic Potential Connects Molecular Structural Ordering to the Macroscale Properties of Polyacrylonitrile

Polyacrylonitrile (PAN) is an important commercial polymer, bearing atactic stereochemistry resulting from nonselective radical polymerization. As such, an accurate, fundamental understanding of governing interactions among PAN molecular units is indispensable for advancing the design principles of final products at reduced processability costs. While ab initio molecular dynamics (AIMD) simulations can provide the necessary accuracy for treating key interactions in polar polymers, such as dipole–dipole interactions and hydrogen bonding, and analyzing their influence on the molecular orientation, their implementation is limited to small molecules only. Herein, we show that the neural network interatomic potentials (NNIPs) that are trained on the small-scale AIMD data (acquired for oligomers) can be efficiently employed to examine the structures and properties at large scales (polymers). NNIP provides critical insight into intra- and interchain hydrogen-bonding and dipolar correlations and accurately predicts the amorphous bulk PAN structure validated by modeling the experimental X-ray structure factor. Furthermore, the NNIP-predicted PAN properties, such as density and elastic modulus, are in good agreement with their experimental values. Overall, the trend in the elastic modulus is found to correlate strongly with the PAN structural orientations encoded in the Hermans orientation factor. In conclusion, this study enables the ability to predict the structure–property relations for PAN and analogues with sustainable ab initio accuracy across scales.

36 MATERIALS SCIENCE↗

DOE OSTI2014

Geometry optimization for micro-pressure sensor considering dynamic interference

Presented is the geometry optimization for piezoresistive absolute micro-pressure sensor. A figure of merit called the performance factor (PF) is defined as a quantitative index to describe the comprehensive performances of a sensor including sensitivity, resonant frequency, and acceleration interference. Three geometries are proposed through introducing islands and sensitive beams into typical flat diaphragm. The stress distributions of sensitive elements are analyzed by finite element method. Multivariate fittings based on ANSYS simulation results are performed to establish the equations about surface stress, deflection, and resonant frequency. Optimization by MATLAB is carried out to determine the dimensions of the geometries. Convex corner undercutting is evaluated. Each PF of the three geometries with the determined dimensions is calculated and compared. Silicon bulk micromachining is utilized to fabricate the prototypes of the sensors. The outputs of the sensors under both static and dynamic conditions are tested. Experimental results demonstrate the rationality of the defined performance factor and reveal that the geometry with quad islands presents the highest PF of 210.947 Hz{sup 1/4}. The favorable overall performances enable the sensor more suitable for altimetry.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

DOE OSTI2024

A Data-Driven Approach for High-Impedance Fault Localization in Distribution Systems

Accurate and quick identification of high-impedance faults (HIFs) is critical for the reliable operation of distribution systems. Unlike other faults in power grids, HIFs are very difficult to detect by conventional overcurrent relays due to the low fault current. Although HIFs can be affected by various factors, the voltage-current characteristics can substantially imply how the system responds to the disturbance and thus provides opportunities to effectively localize HIFs. In this work, we propose a data-driven approach for the identification of HIF events. To tackle the nonlinearity of the voltage-current trajectory, first, we formulate optimization problems to approximate the trajectory with piecewise functions. Then we collect the function features of all segments as inputs and use the support vector machine approach to efficiently identify HIFs at different locations. Numerical studies on the IEEE 123-node test feeder demonstrate the validity and accuracy of the proposed approach for real-time HIF identification.

explainable artificial intelligence↗

DOE OSTI2022

A Catalog of Candidate Double and Lensed Quasars from Gaia and WISE Data

Making use of strong correlations between closely separated multiple or double sources and photometric and astrometric metadata in Gaia Early Data Release 3 (EDR3), we generate a catalog of candidate double- and multiply imaged lensed quasars and active galactic nuclei (AGNs), comprising 3140 systems. It includes two partially overlapping parts: a sample of distant (redshifts mostly greater than 1) sources with perturbed data; and systems that have been resolved into separate components by Gaia at separations less than 2''. For the first part, which is roughly one-third of the published catalog, we synthesized 0.617 million redshifts using multiple machine-learning prediction and classification methods, using independent photometric and astrometric data from Gaia EDR3 and the Wide-field Infrared Survey Explorer, with accurate spectroscopic redshifts from the Sloan Digital Sky Survey (SDSS) as a training set. Using these synthetic redshifts, we estimate a 4.9% rate of interlopers with spectroscopic redshifts below 1 in this part of the catalog. Unresolved candidate double and dual AGNs and quasars are selected as sources with a marginally high BP/RP excess factor (phot_bp_rp_excess_factor), which is sensitive to source extent, limiting our search to high-redshift quasars. For the second part of the catalog, additional filters on measured parallax and near-neighbor statistics are applied to diminish the propagation of the remaining stellar contaminants. The estimated rate of the positives (double or multiple sources) is 98%, and the estimated rate of dual (physically related) quasars is greater than 54%. A few dozen serendipitously found objects of interest are discussed in more detail, including known and new lensed images, planetary nebulae, young IR stars of peculiar morphology, and quasars with catastrophic redshift errors in SDSS.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2022

Need for advanced research reactors for the next-generation reactor physics, analysis tools, and technology

Full text of publication follows. There is an urgent need for design and deployment of advanced research and test reactors in support of design, licensing and operation of advanced power reactors and education of next generation nuclear workforce. Existing research reactors mostly were designed and constructed decades ago with the main objectives of training operators, performing reactor physics experiments, and educating nuclear engineers and scientists. There are already gaps and significant concern about future capabilities for the existing research reactor facilities to address modern instrumentation and/or flexible environments for: performing reactor physics studies for advanced designs which have significantly different core materials forms and compositions, reactor shapes and size; validation of advanced high-fidelity software; development of machine learning algorithms for enhancement of human-machine collaboration in support of reactor monitoring, operation and safeguards; and, effective education of the next-generation workforce. The authors will focus on the need for advanced research reactors to improve and validate fast and accurate simulation tools for high-fidelity modeling and analysis of nuclear reactors in support of their design, optimization, licensing, operation, and monitoring. In the past, simulation tools were limited to relatively coarse models using approximate methodologies that benefited from two main factors: i) allowance for large margins and tolerances; ii) ability to construct prototype (e.g., zero power) reactors for adjustment of approximate methodologies. The next generation reactors have to be designed mainly by using novel high-fidelity computational tools that are accurate and fast, and therefore can be used for parametric studies and uncertainty quantification. To sufficiently demonstrate the accuracy of these tools, advanced research reactors are needed. The authors argue the need for new computational paradigms such as the MRT (Multistage, Response- function Transport) methodology which has resulted in the development of the novel high-fidelity RAPID (Real-time Analysis for Particle-transport and In-situ Detection) code system. Such code systems have to be robust in modeling any complex system, and should be fast and accurate, henceforth their uncertainties can be quantified at reasonable costs. Again, advanced research reactors are needed for the validation of the fidelity and accuracy of new computational tools. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

DOE OSTI2024

A Data-Driven Approach for High-Impedance Fault Localization in Distribution Systems: Preprint

explainable artificial intelligence↗

DOE OSTI2018

Evaluation of Factors that Influence Residential Solar Panel Installations

Though rooftop photovoltaic (PV) systems are the fastest growing source of distributed generation, detailed information about where they are located and who their owners are is often known only to installers and utility companies. This lack of detailed information is a barrier to policy and financial assessment of solar energy generation and use. To bridge the described data gap, Oak Ridge National Laboratory (ORNL) was sponsored by the Department of Energy (DOE) Office of Energy Policy and Systems Analysis (EPSA) to create an automated approach for detecting and characterizing buildings with installed solar panels using high-resolution overhead imagery. Additionally, ORNL was tasked with using machine learning techniques to classify parcels on which solar panels were automatically detected in the Washington, DC, and Boston areas as commercial or residential, and then providing a list of recommended variables and modeling techniques that could be combined with these results to identify attributes that motivate the installation of residential solar panels. This technical report describes the methodology, results, and recommendations in greater detail, including lessons learned and future work.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

DOE OSTI2022

Improved Architectures and Training Algorithms for Deep Operator Networks

Operator learning techniques have recently emerged as a powerful tool for learning maps between infinite-dimensional Banach spaces. Trained under appropriate constraints, they can also be effective in learning the solution operator of partial differential equations (PDEs) in an entirely self-supervised manner. In this work we analyze the training dynamics of deep operator networks (DeepONets) through the lens of Neural Tangent Kernel theory, and reveal a bias that favors the approximation of functions with larger magnitudes. To correct this bias we propose to adaptively re-weight the importance of each training example, and demonstrate how this procedure can effectively balance the magnitude of back-propagated gradients during training via gradient descent. We also propose a novel network architecture that is more resilient to vanishing gradient pathologies. Taken together, our developments provide new insights into the training of DeepONets and consistently improve their predictive accuracy by a factor of 10-50x, demonstrated in the challenging setting of learning PDE solution operators in the absence of paired input-output observations.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2023

Latent Dirichlet Allocation modeling of environmental microbiomes

Interactions between stressed organisms and their microbiome environments may provide new routes for understanding and controlling biological systems. However, microbiomes are a form of high-dimensional data, with thousands of taxa present in any given sample, which makes untangling the interaction between an organism and its microbial environment a challenge. Here we apply Latent Dirichlet Allocation (LDA), a technique for language modeling, which decomposes the microbial communities into a set of topics (non-mutually-exclusive sub-communities) that compactly represent the distribution of full communities. LDA provides a lens into the microbiome at broad and fine-grained taxonomic levels, which we show on two datasets. In the first dataset, from the literature, we show how LDA topics succinctly recapitulate many results from a previous study on diseased coral species. We then apply LDA to a new dataset of maize soil microbiomes under drought, and find a large number of significant associations between the microbiome topics and plant traits as well as associations between the microbiome and the experimental factors, e.g. watering level. This yields new information on the plant-microbial interactions in maize and shows that LDA technique is useful for studying the coupling between microbiomes and stressed organisms.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2021

Compact representations of structured BFGS matrices

For general large-scale optimization problems compact representations exist in which recursive quasi-Newton update formulas are represented as compact matrix factorizations. For problems in which the objective function contains additional structure, recent structured quasi-Newton methods exploit available second-derivative information and approximate unavailable second derivatives. Here, this article develops the compact representations of two structured Broyden-Fletcher-Goldfarb-Shanno update formulas. The compact representations enable efficient limited memory and initialization strategies. Two limited memory line search algorithms are described for which extensive numerical results demonstrate the efficacy of the algorithms, including comparisons to IPOPT on large machine learning problems, and to L-BFGS on a real world large scale ptychographic imaging application.

97 MATHEMATICS AND COMPUTING↗