Engineering Papers⌕ Search

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.

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

At least 613 records · Page 34
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Uncovering acoustic signatures of pore formation in laser powder bed fusion

Abstract We present a machine learning workflow to discover signatures in acoustic measurements that can be utilized to create a low-dimensional model to accurately predict the location of keyhole pores formed during additive manufacturing processes. Acoustic measurements were sampled at 100 kHz during single-layer laser powder bed fusion (LPBF) experiments, and spatio-temporal registration of pore locations was obtained from post-build radiography. Power spectral density (PSD) estimates of the acoustic data were then decomposed using non-negative matrix factorization with custom $$\varvec{k}$$ k -means clustering (NMF $$\varvec{k}$$ k ) to learn the underlying spectral patterns associated with pore formation. NMF $$\varvec{k}$$ k returned a library of basis signals and matching coefficients to blindly construct a feature space based on the PSD estimates in an optimized fashion. Moreover, the NMF $$\varvec{k}$$ k decomposition led to the development of computationally inexpensive machine learning models which are capable of quickly and accurately identifying pore formation with classification accuracy of supervised and unsupervised label learning greater than 95% and 90%, respectively. The intrinsic data compression of NMF k , the relatively light computational cost of the machine learning workflow, and the high classification accuracy makes the proposed workflow an attractive candidate for edge computing toward in-situ keyhole pore prediction in LPBF.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Machine Learning for Joint Quality Control

The use of lightweight material combinations has been highly demanded in manufacturing automotive structures. However, making robust dissimilar material joints of such lightweight materials is still challenging. A significant barrier to achieving high-quality and repeatable joint performance is a deficient understanding of the relationship between the welding process, joint attributes, and joint performance. In this context, welding factors refer to material, equipment, environment, and process parameters, while joint features comprise specific microstructural attributes of the weld such as nugget size, heat affected zone (HAZ) topology, intermetallic layer thickness, and sheet thickness reduction. Joint performance is quantified in terms of strength (e.g., tensile shear, coach peel, cross-tension), weld size, and hardness, among other factors. While there have been many attempts to establish this process-structure-property relationship by developing a model derived from the associated physics and first principles, the complexity of the joining processes compounded by the complex interactions with different materials in an automotive assembly line environment, has hindered the usefulness of such attempts. The complexity is further exacerbated using different stacking materials, especially comprising dissimilar material combinations. In practice, the common approach has been the laborious process of creating welds, characterizing them, and then physically testing them through experimentation. With the emergence of artificial intelligence (AI) methods, an alternative pathway to eliciting the desired process-structure-property relationship at an accelerated pace is to use a data-driven approach by employing machine-learning (ML) techniques. This approach is benefitted by the availability of large streams of data, generated through years of research and testing by original equipment manufacturers, in the form of material, process, environmental, equipment, microstructural, and bulk-scale performance information from multimodal, multiscale sensors making measurements from laboratory-scale to production-scale processes. During Phase I efforts, which ended in fiscal year (FY) 2021, the Oak Ridge National Laboratory and Pacific Northwest National Laboratory (ORNL/PNNL) team demonstrated the effectiveness of different ML/AI frameworks in modeling complex relationships between resistance spot welding (RSW) process parameters, weld attributes, and joint properties using a subset of data from General Motors (GM). In FY 2022, the project team further refined and expanded their respective ML models to analyze additional welds with new weld stack-ups and materials to enhance the ML model predictive capability. ORNL extended its unified deep neural networks (DNN) ML training and prediction framework with new data streams of process parameters, and PNNL extended its model describing RSW process parameters’ associations with weld attributes. In FY 2023, the project team completed the development of the AI/ML architecture for analyzing aluminum/steel joints manufactured by GM via RSW and transitioned into the inline welding quality monitoring task for steel/steel RSW joints provided by GM.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Machine Learning for Geothermal Resource Exploration in the Tularosa Basin, New Mexico

Geothermal energy is considered an essential renewable resource to generate flexible electricity. Geothermal resource assessments conducted by the U.S. Geological Survey showed that the southwestern basins in the U.S. have a significant geothermal potential for meeting domestic electricity demand. Within these southwestern basins, play fairway analysis (PFA), funded by the U.S. Department of Energy’s (DOE) Geothermal Technologies Office, identified that the Tularosa Basin in New Mexico has significant geothermal potential. This short communication paper presents a machine learning (ML) methodology for curating and analyzing the PFA data from the DOE’s geothermal data repository. The proposed approach to identify potential geothermal sites in the Tularosa Basin is based on an unsupervised ML method called non-negative matrix factorization with custom k-means clustering. This methodology is available in our open-source ML framework, GeoThermalCloud (GTC). Using this GTC framework, we discover prospective geothermal locations and find key parameters defining these prospects. Our ML analysis found that these prospects are consistent with the existing Tularosa Basin’s PFA studies. This instills confidence in our GTC framework to accelerate geothermal exploration and resource development, which is generally time-consuming.

15 GEOTHERMAL ENERGY↗
DOE OSTI2016

True beam commissioning experience at Nordland Hospital Trust, Norway

To evaluate the measured of all photon beam data of first Varian True Beam version 2.0 slim model, recently commissioned at Nordland Hospital Trust, Bodø. To compare and evaluate the possibility of beam matching with the Clinac2300, for the energies of 6MV and 15 MV. Materials/Methods: Measurements of PDD, OAR, and Output factors were realized with the IBA Blue-phantom with different detectors and evaluated between them for all photon energies: 6MV, 15MV, 6MV FFF and 10MV FFF. The ionization chambers used were Pin Point CC01, CC04, Semiflex CC13 and photon diode by Iba dosimetry. The data were processed using Beizer algorithm with a resolution of 1 mm. The measured depth dose curves, diagonals, OAR, and output factors were imported into Eclipse in order to calculate beam data for the anisotropic analytical algorithm (AAA version 10.0.28) for both the dataset measured with CC04 and CC13 and compared. The model head of 23EX was selected as the most near model to True Beam as a restriction of our version of Aria. It was seen that better results were achieved with the CC04 measured data as a result of better resolution. For the biggest field after 10 cm depth a larger difference is seen between measured and calculated for both dataset, but it is within the criteria for acceptance. Results: The Beam analysis criteria of 2 mm at 50% dose is achieved for all the fields accept for 40x40 that is within 3%. Depth difference at maximum dose is within 1 mm for all the fields and dose difference at 100 mm and 200 mm is lower than 1% for or all the fields. The PDD between two machines for all the fields differ after Dmax with less than 1%. For profiles in the field zone and outside field the difference is within 1% for all the fields. In the penumbra region the difference is from 2% up to 12% for big fields. As for diagonals they differ as a result of the head construction at the edge of the field and the penumbra region. The output factors differ for big fields within 5% and for the small fields within 3%. MU and dose distribution does not change for plans recalculated with the new modeled machine.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2020

Machine learning with bond information for local structure optimizations in surface science

Local optimization of adsorption systems inherently involves different scales: within the substrate, within the molecule, and between the molecule and the substrate. In this work, we show how the explicit modeling of different characteristics of the bonds in these systems improves the performance of machine learning methods for optimization. Furthermore, we introduce an anisotropic kernel in the Gaussian process regression framework that guides the search for the local minimum, and we show its overall good performance across different types of atomic systems. The method shows a speed-up of up to a factor of two compared with the fastest standard optimization methods on adsorption systems. Additionally, we show that a limited memory approach is not only beneficial in terms of overall computational resources but can also result in a further reduction of energy and force calculations.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2015

Identifying Important Ions and Positions in Mass Spectrometry Imaging Data Using CUR Matrix Decompositions

Mass spectrometry imaging enables label-free, high-resolution spatial mapping of the chemical composition of complex, biological samples. Typical experiments require selecting ions and/or positions from the images: ions for fragmentation studies to identify keystone compounds and positions for follow up validation measurements using microdissection or other orthogonal techniques. Unfortunately, with modern imaging machines, these must be selected from an overwhelming amount of raw data. Existing techniques to reduce the volume of data, the most popular of which are principle component analysis and non-negative matrix factorization, have the disadvantage that they return difficult-to-interpret linear combinations of actual data elements. In this work, we show that CX and CUR matrix decompositions can be used directly to address this selection need. CX and CUR matrix decompositions use empirical statistical leverage scores of the input data to provide provably good low-rank approximations of the measured data that are expressed in terms of actual ions and actual positions, as opposed to difficult-to-interpret eigenions and eigenpositions. We show that this leads to effective prioritization of information for both ions and positions. In particular, important ions can be found either by using the leverage scores as a ranking function and using a deterministic greedy selection algorithm or by using the leverage scores as an importance sampling distribution and using a random sampling algorithm; however, selection of important positions from the original matrix performed significantly better when they were chosen with the random sampling algorithm. Also, we show that 20 ions or 40 locations can be used to reconstruct the original matrix to a tolerance of 17% error for a widely studied image of brain lipids; and we provide a scalable implementation of this method that is applicable for analysis of the raw data where there are often more than a million rows and/or columns, which is larger than SVD-based low-rank approximation methods can handle. Lastly, these results introduce the concept of CX/CUR matrix factorizations to mass spectrometry imaging, describing their utility and illustrating principled algorithmic approaches to deal with the overwhelming amount of data generated by modern mass spectrometry imaging.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2019

Biochar application as a tool to decrease soil nitrogen losses (NH3 volatilization, N2O emissions, and N leaching) from croplands: Options and mitigation strength in a global perspective

Biochar application to croplands has been proposed as a potential strategy to decrease losses of soil-reactive nitrogen (N) to the air and water. However, the extent and spatial variability of biochar function at the global level are still unclear. Using Random Forest regression modelling of machine learning based on data compiled from the literature, we mapped the impacts of different biochar types (derived from wood, straw, or manure), and their interactions with biochar application rates, soil properties, and environmental factors, on soil N losses (NH3 volatilization, N2O emissions, and N leaching) and crop productivity. The results show that a suitable distribution of biochar across global croplands (i.e., one application of <40 t ha-1 wood biochar for poorly buffered soils, such as those characterized by soil pH<5, organic carbon<1%, or clay>30%; and one application of <80 t ha-1 wood biochar, <40 t ha-1 straw biochar, or <10 t ha-1 manure biochar for other soils) could achieve an increase in global crop yields by 222–766 Tg yr-1 (4%–16% increase), a mitigation of cropland N2O emissions by 0.19–0.88 Tg N yr-1 (6%–30% decrease), a decline of cropland N leaching by 3.9–9.2 Tg N yr-1 (12%–29% decrease), but also a fluctuation of cropland NH3 volatilization by -1.9–4.7 Tg N yr-1 (-12%–31% change). The decreased sum of the three major reactive N losses amount to 1.7–9.4 Tg N yr-1, which corresponds to 3%–14% of the global cropland total N loss. Biochar generally has a larger potential for decreasing soil N losses but with less benefits to crop production in temperate regions than in tropical regions.

biochar, soil amendment, nitrogen cycle, climate c↗
DOE OSTI2014

SU-E-T-553: Characterization of Plastic Scintillator Detectors for Radiation Therapy

Purpose: To characterize basic performance of plastic scintillator detectors (PSD) designed for dosimetry of radiation therapy. Methods: The Exradin W1 Scintillator is a plastic scintillating fiber-based detector designed for highly accurate measurement of small radiotherapy fields used in patient plan verification and machine commissioning and QA procedures. The Cerenkov emissions were corrected using spectral separation. The optical signal was converted to electronic signal with a photodiode. We measured its dosimetry performance, including percentage depth dose, output factor, dose and dose rate linear response. We compared the dosimetry results with reference ion chamber measurements. Results: The dosimetry results of PSD agree well with reference ion chamber measurements. For percentage depth dose, the differences between PSD and ion chamber results are on average 1.7±1.1% and 0.8±0.8% with a maximum of 3.5% and 2.7% for 6MV and 15MV beams, respectively. For the output factors, PSD measurements are within 2% from ion chamber results. The dose linear response is within 1% when dose is larger than 20 MU for both 6 MV and 15 MV. The dose rate linear response is within 1% for the entire dose rate used (100 MU/min to 600MU/min). Conclusions: The current design of PSD is feasible for the dosimtry measurement in radiation therapy. This combination of PSD and photodiode system could be extended to multichannel array detection of dose distribution. It might as well be used as range verification in proton therapy. The work is partially supported by: DOD (W81XWH-09-2-0174) and American Cancer Society (IRG-78-002-28)

07 ISOTOPE AND RADIATION SOURCES↗
DOE OSTI2016

Strengthening LLNL Missions through Laboratory Directed Research and Development in High Performance Computing

High performance computing (HPC) has been a defining strength of Lawrence Livermore National Laboratory (LLNL) since its founding. Livermore scientists have designed and used some of the world’s most powerful computers to drive breakthroughs in nearly every mission area. Today, the Laboratory is recognized as a world leader in the application of HPC to complex science, technology, and engineering challenges. Most importantly, HPC has been integral to the National Nuclear Security Administration’s (NNSA’s) Stockpile Stewardship Program—designed to ensure the safety, security, and reliability of our nuclear deterrent without nuclear testing. A critical factor behind Lawrence Livermore’s preeminence in HPC is the ongoing investments made by the Laboratory Directed Research and Development (LDRD) Program in cutting-edge concepts to enable efficient utilization of these powerful machines. Congress established the LDRD Program in 1991 to maintain the technical vitality of the Department of Energy (DOE) national laboratories. Since then, LDRD has been, and continues to be, an essential tool for exploring anticipated needs that lie beyond the planning horizon of our programs and for attracting the next generation of talented visionaries. Through LDRD, Livermore researchers can examine future challenges, propose and explore innovative solutions, and deliver creative approaches to support our missions. The present scientific and technical strengths of the Laboratory are, in large part, a product of past LDRD investments in HPC. Here, we provide seven examples of LDRD projects from the past decade that have played a critical role in building LLNL’s HPC, computer science, mathematics, and data science research capabilities, and describe how they have impacted LLNL’s mission.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Design Choices in Anomaly Detection for Industrial Control Systems: Insights from Gas Pipeline Data

Industrial control systems (ICS) remain vulnerable to increasingly sophisticated cyberattacks, yet evaluating anomaly detection models in these environments is challenging due to temporal dependencies, missing-not-at-random patterns, and extremely imbalanced datasets. These factors make common practices—especially random data splits and naïve imputation—prone to severe temporal leakage, which can inflate reported performance and obscure real-world limitations. In this work, we systematically examine classical machine learning models, temporal deep learning architecture, and tensor-decomposition–based methods on a gas-pipeline dataset using a fully temporally separated evaluation pipeline designed to mimic realistic deployment conditions. Our findings show that proper temporal handling and MNAR-aware preprocessing significantly alter the relative performance of popular anomaly-detection methods, providing practical guidance for designing reliable, leakage-resistant ICS intrusion-detection systems.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Utah FORGE 2-2439: A Multi-Component Approach to Characterizing In-Situ Stress: Laboratory, Modeling and Field Measurement - Workshop Presentation

This is a presentation on A Multi-Component Approach to Characterizing In-Situ Stress at the U.S DOE FORGE EGS Site: Laboratory, Modeling and Field Measurement project by Battelle [Columbus, OH], presented by Mark Kelley. The project's objective was to characterize stress in the Utah FORGE EGS reservoir using three methods: a laboratory rock-core stress estimation combined with a Machine Learning approach for estimation of in-situ stress from field sonic-log data, a field based in-situ measurement (min-frac) approach, and a modeling approach. This presentation was featured in the Utah FORGE R&D Annual Workshop on September 7, 2023. The workshop provided a valuable opportunity to explore the progress made in each of the 17 Research and Development projects funded under Solicitation 2020-1 which aim to enhance our understanding of the crucial factors influencing the development of Enhanced Geothermal Systems (EGS) reservoirs and resources.

15 GEOTHERMAL ENERGY↗
DOE OSTI2025

Mapping tree height in complex terrain of northern China using ultra-high-resolution images

Tree height is a key parameter for estimating forest biomass and carbon sequestration. In recent years, notable progress has been made in mapping tree height using satellite imagery. However, existing tree height products show low accuracy in mountainous and complex terrains, and few studies typically addressed tree height estimations in mountain areas. This study examines the Mentougou district of Beijing, China, characterized by complex terrain and mountainous landscapes. We analyzed two methods for estimating tree height: one using only spectral features and another combining spectral features with topographic factors (elevation, slope, aspect). We used 3-m resolution PlanetScope 8-band multispectral imagery, with 710 field-measured individual tree heights averaged to obtain 471 pixel-level tree height values as ground-truth, to develop tree height prediction models using eXtreme Gradient Boosting (XGBoost), Random Forest (RF), and Gradient Boosting Machine (GBM) models. The results show that the XGBoost model consistently presented the highest accuracy for both methods evaluated. Specifically, the XGBoost model that combined spectral data with elevation and slope variables with an R² of 0.75 and an RMSE of 2.69 m. Using the XGBoost model, we generated the tree height map for the Mentougou area at 3 m resolution, showing tree heights ranging from 0.5 to 30.4 m, and the model’s prediction error standard deviations ranged from 2.50 to 4.71 m, indicating reliable performance across varied terrain. Additionally, we compared and evaluated the global tree height products, identifying limitations in the accuracy within complex terrains. This study demonstrates the potential for accurately predicting tree heights by combining high-resolution multispectral satellites with a terrain factor modeling approach.

Complex terrain↗
DOE OSTI2024

Macroscopic Traffic Modeling Using Probe Vehicle Data: A Machine Learning Approach

Abstract The macroscopic fundamental diagram (MFD) captures an orderly relationship among traffic flow, density, and speed at the network level. It is a simple yet powerful tool for modeling traffic dynamics in large urban networks with broad application in traffic control and management. However, empirically derived MFDs in urban regions require high-resolution traffic data from the network. Having the network flow and vehicular density estimated at the (granular) census tract level using vehicle probe data, we apply machine learning methods to predict the MFDs across U.S. urban areas and capture the impacts of location-specific input features on the network flow–density relationships at a large scale. The results show that, among the four tested machine learning approaches (Random Forest, XGBoost, Support Vector Machine, and Neural Network), XGBoost delivers the best performance in predicting network traffic flow based on vehicular density and location attributes. Using interaction Shapley Additive explanation (SHAP) values and partial correlation analysis, we examine the factors influencing MFD shapes across different locations. Our empirical findings reveal that across U.S. urban areas, network topology, transportation infrastructure, and land use are primary factors shaping MFD curves, while demand and trip-related factors play a lesser role. Specifically, higher ranking roads, centrality, and development levels correlate positively with network capacity and critical density, whereas negative associations are observed for network connectivity, mixed-use development, and road roughness levels.

Jin, Ling↗
DOE OSTI2020

Recent advances in constraint and machine learning-based metabolic modeling by leveraging stoichiometric balances, thermodynamic feasibility and kinetic law formalisms

Understanding the governing principles behind organisms’ metabolism and growth underpins their effective deployment as bioproduction chassis. A central objective of metabolic modeling is predicting how metabolism and growth are affected by both external environmental factors and internal genotypic perturbations. The fundamental concepts of reaction stoichiometry, thermodynamics, and mass action kinetics have emerged as the foundational principles of many modeling frameworks designed to describe how and why organisms allocate resources towards both growth and bioproduction. Furthermore, this review focuses on the latest algorithmic advancements that have integrated these foundational principles into increasingly sophisticated quantitative frameworks.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2015

Demonstration of superconducting micromachined cavities

Superconducting enclosures will be key components of scalable quantum computing devices based on circuit quantum electrodynamics. Within a densely integrated device, they can protect qubits from noise and serve as quantum memory units. Whether constructed by machining bulk pieces of metal or microfabricating wafers, 3D enclosures are typically assembled from two or more parts. The resulting seams potentially dissipate crossing currents and limit performance. In this letter, we present measured quality factors of superconducting cavity resonators of several materials, dimensions, and seam locations. We observe that superconducting indium can be a low-loss RF conductor and form low-loss seams. Leveraging this, we create a superconducting micromachined resonator with indium that has a quality factor of two million, despite a greatly reduced mode volume. Inter-layer coupling to this type of resonator is achieved by an aperture located under a planar transmission line. The described techniques demonstrate a proof-of-principle for multilayer microwave integrated quantum circuits for scalable quantum computing.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2026

Entire four-graviton EFT from the duality between color and kinematics

The Bern-Carrasco-Johansson (BCJ) double-copy construction reveals a fundamental structural connection between gauge and gravity theories. At its core, the BCJ double copy is directly due to a duality between the algebraic relations of a color root and those of a kinematic root. We generalize this principle beyond the conventional Lie algebra structure of tree-level Yang-Mills theory. By demanding color-kinematics duality for the complete basis of four-point color structures—including those involving the symmetric 𝑑 𝑎⁢𝑏⁢𝑐 constants—we define the universal double copy. We systematically classify the bases of all such parity-even generalized gauge-theory numerators and, independently, the space of all parity-even four-graviton higher-derivative operators. We demonstrate that our universal double-copy construction precisely spans the entire tower of parity-even four-graviton amplitudes in any dimension, except for the Lovelock 𝑅 3 contribution in 𝐷 > 6 which we can express in terms of a particularly simple universal triple-copy involving gauge theories coupled to scalars. Explicit machine-readable expressions for the complete basis of gauge-theory numerators and fundamental gravitational building blocks are provided in the Supplemental Material. This establishes that all possible four-point gravitational interactions can be factorized into products of gauge-theory building blocks governed by this universal notion of color-kinematics duality.

Carrasco, John Joseph M. [Northwestern Univ., Evan↗
DOE OSTI2026

The Impact of Time-Aware Design Choices in ICS Anomaly Detection

Industrial control systems (ICS) remain vulnerable to increasingly sophisticated cyberattacks, yet evaluating anomaly detection models in these environments is challenging due to temporal dependencies, missing-not-at-random patterns, and extremely imbalanced datasets. These factors make common practices—especially random data splits and na¨ıve imputation— prone to severe temporal leakage, which can inflate reported performance and obscure real-world limitations. In this work, we systematically examine classical machine learning models, temporal deep learning architecture, and tensordecomposition– based methods on a gas-pipeline dataset using a fully temporally separated evaluation pipeline designed to mimic realistic deployment conditions. Our findings show that proper temporal handling and MNAR-aware preprocessing significantly alter the relative performance of popular anomaly-detection methods, providing practical guidance for designing reliable, leakage-resistant ICS intrusion-detection systems.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Building Stock Models for Embodied Carbon Emissions—A Review of a Nascent Field

Building stock modeling emerges as a critical tool in the strategic reduction of embodied carbon emissions, which is pivotal in reshaping the evolving construction sector. This review provides an overall view of modern methodologies in building stock modeling, homing in on the nuances of embodied carbon analysis in construction. Examining 23 seminal papers, our study delineates two primary modeling paradigms—top-down and bottom-up—each further compartmentalized into five innovative methods. This study points out the challenges of data scarcity and computational demands, advocating for methodological advancements that promise to refine the precision of building stock models. A groundbreaking trend in recent research is the incorporation of machine learning algorithms, which have demonstrated remarkable capacity, improving stock classification accuracy by 25% and urban material quantification by 40%. Furthermore, the application of remote sensing has revolutionized data acquisition, enhancing data richness by a factor of five. This review offers a critical examination of current practices and charts a course toward an environmentally prudent future. It underscores the transformative impact of building stock modeling in driving ecological stewardship in the construction industry, positioning it as a cornerstone in the quest for sustainability and its significant contribution toward the grand vision of an eco-efficient built environment.

Hu, Ming (ORCID:0000000325831161)↗