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 73 records · Page 4
All sourcesNASA NTRSDOE OSTI
DOE OSTI2015

Spreading of lithium on a stainless steel surface at room temperature

Lithium conditioned plasma facing surfaces have lowered recycling and enhanced plasma performance on many fusion devices and liquid lithium plasma facing components are under consideration for future machines. A key factor in the performance of liquid lithium components is the wetting by lithium of its container. We have observed the surface spreading of lithium from a mm-scale particle to adjacent stainless steel surfaces using a scanning Auger microprobe that has elemental discrimination. Here, the spreading of lithium occurred at room temperature (when lithium is a solid) from one location at a speed of 0.62 μm/day under ultrahigh vacuum conditions. Separate experiments using temperature programmed desorption (TPD) investigated bonding energetics between monolayer-scale films of lithium and stainless steel. While multilayer lithium desorption from stainless steel begins to occur just above 500 K (E des = 1.54 eV), sub-monolayer Li desorption occurred in a TPD peak at 942 K (E des = 2.52 eV) indicating more energetically favorable lithium-stainless steel bonding (in the absence of an oxidation layer) than lithium lithium bonding.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2021

Synthetic accessibility and stability rules of NASICONs

In this paper we develop the stability rules for NASICON-structured materials, as an example of compounds with complex bond topology and composition. By first-principles high-throughput computation of 3881 potential NASICON phases, we have developed guiding stability rules of NASICON and validated the ab initio predictive capability through the synthesis of six attempted materials, five of which were successful. A simple two-dimensional descriptor for predicting NASICON stability was extracted with sure independence screening and machine learned ranking, which classifies NASICON phases in terms of their synthetic accessibility. This machine-learned tolerance factor is based on the Na content, elemental radii and electronegativities, and the Madelung energy and can offer reasonable accuracy for separating stable and unstable NASICONs. This work will not only provide tools to understand the synthetic accessibility of NASICON-type materials, but also demonstrates an efficient paradigm for discovering new materials with complicated composition and atomic structure.

25 ENERGY STORAGE↗
DOE OSTI2020

Comparison of Linear and Rotary Electric Machine Topologies for a Hybrid Hydraulic Electric Architecture of Off-Highway Vehicles

Electrification of off-highway vehicles offers the benefits of improved energy efficiency, enhanced controllability, and reduction in greenhouse gas emissions. However, progress towards electrification has been limited by the low power density of electric components compared to hydraulic components. A recently proposed hybrid hydraulic electric vehicle architecture provides a potential solution to these power density challenges by minimizing the required power rating of the electric components. This paper investigates the optimal design of linear and rotary electric machines for this new system architecture. Key performance metrics are identified for the electric machine, several variants of electric machines are analytically sized and compared for two different power levels, and an FEA-assisted design optimization is performed on promising candidates. Based on the optimization results, it is shown that an axial flux rotary machine is most suitable for implements requiring high power ratings, while the linear machine's form factor makes it a compelling candidate for lower power implements.

42 ENGINEERING↗
DOE OSTI2021

Household Transportation Energy Affordability by Region and Socioeconomic Factors

Transportation fuel is an important component of household budgets, as 3.3% of total household expenditures are for vehicle fuel nationwide and over 50% of annual household expenditures on energy are for transportation. These average values vary geographically, and higher energy cost burdens are faced by households with lower incomes. Defining transportation energy affordability burden as the percentage of annual household income spent on vehicle fuel, this study aims to quantify affordability as a function of household characteristics and geography. Here, through analysis at the census tract level, this study (i) projects annual household vehicle miles traveled (VMT) based on demographic factors using machine-learning techniques, (ii) estimates local differences in vehicle fuel economy and fuel price, and (iii) quantifies resulting transportation energy affordability by census tract. This study found that the average burden by tract varies from 0.15% to 8%. The variation in affordability can be largely explained by income level and vehicle fuel efficiency. Suburban and rural households spend more on transportation energy compared with urban households because of the usage of less fuel-efficient vehicle technologies and higher annual VMT. Lower-income groups have a wide distribution of the percentage of income spent on transportation energy, 1.2% to 8%, whereas the range for the highest income group ($125,000+) is from 0.15% to 3.9%. This detailed transportation energy affordability analysis provides a better understanding of regional variations in household travel behavior, helps to determine where fuel-efficient vehicle technologies are more likely to be used, and improves estimates of vehicle ownership costs.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2014

SU-E-T-585: Commissioning of Electron Monte Carlo in Eclipse Treatment Planning System for TrueBeam

Purpose: To commission electron Monte Carlo (eMC) algorithm in Eclipse Treatment Planning System (TPS) for TrueBeam Linacs, including the evaluation of dose calculation accuracy for small fields and oblique beams and comparison with the existing eMC model for Clinacs. Methods: Electron beam percent-depth-dose (PDDs) and profiles with and without applicators, as well as output factors, were measured from two Varian TrueBeam machines. Measured data were compared against the Varian TrueBeam Representative Beam Data (VTBRBD). The selected data set was transferred into Eclipse for beam configuration. Dose calculation accuracy from eMC was evaluated for open fields, small cut-out fields, and oblique beams at different incident angles. The TrueBeam data was compared to the existing Clinac data and eMC model to evaluate the differences among Linac types. Results: Our measured data indicated that electron beam PDDs from our TrueBeam machines are well matched to those from our Varian Clinac machines, but in-air profiles, cone factors and open-filed output factors are significantly different. The data from our two TrueBeam machines were well represented by the VTBRBD. Variations of TrueBeam PDDs and profiles were within the 2% /2mm criteria for all energies, and the output factors for fields with and without applicators all agree within 2%. Obliquity factor for two clinically relevant applicator sizes (10×10 and 15×15 cm{sup 2}) and three oblique angles (15, 30, and 45 degree) were measured for nominal R100, R90, and R80 of each electron beam energy. Comparisons of calculations using eMC of obliquity factors and cut-out factors versus measurements will be presented. Conclusion: eMC algorithm in Eclipse TPS can be configured using the VTBRBD. Significant differences between TrueBeam and Clinacs were found in in-air profiles and open field output factors. The accuracy of the eMC algorithm was evaluated for a wide range of cut-out factors and oblique incidence.

07 ISOTOPE AND RADIATION SOURCES↗
DOE OSTI2022

Empirical relationships between environmental factors and soil organic carbon produce comparable prediction accuracy as the Machine Learning

Accurate representation of environmental controllers of soil organic carbon (SOC) stocks in Earth System Model (ESM) land models could reduce uncertainties in future carbon-climate feedback projections. Using empirical relationships between environmental factors and SOC stocks to evaluate land models can help modelers understand prediction biases beyond what can be achieved with the observed SOC stocks alone. In this study, we used 31 observed environmental factors, field SOC observations (n = 6,213) from the continental US, and two Machine Learning approaches [Random Forest (RF) and Generalized Additive Modeling (GAM)] to (1) select important environmental predictors of SOC stocks, (2) derive empirical relationships between environmental factors and SOC stocks, and (3) use the derived relationships to predict SOC stocks and compare the prediction accuracy of simpler model developed with the machine learning predictions. Out of the 31 environmental factors we investigated, 12 were identified as important predictors of SOC stocks by the RF approach. In contrast, the GAM approach identified six (of those 12) environmental factors as important controllers of SOC stocks: potential evapotranspiration, normalized difference vegetation index, soil drainage condition, precipitation, elevation, and net primary productivity. The GAM approach showed minimal SOC predictive importance of the remaining six environmental factors identified by the RF approach. Our derived empirical relations produced comparable prediction accuracy as the GAM and RF approach using only a subset of environmental factors. The empirical relationships we derived using the GAM approach can serve as important benchmarks to evaluate environmental control representations of SOC stocks in ESMs, which could reduce uncertainty in predicting future carbon-climate feedbacks.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Nonnegative canonical tensor decomposition with linear constraints: nnCANDELINC

Abstract There is an emerging interest for tensor factorization applications in big‐data analytics and machine learning. To speed up the factorization of extra‐large datasets, organized in multidimensional arrays (also known as tensors), easy to compute compression‐based tensor representations, such as, Tucker and tensor train formats, are used to approximate the initial large‐tensor. Further, tensor factorization is used to extract latent features that can facilitate discoveries of new mechanisms and signatures hidden in the data, where the explainability of the latent features is of principal importance. Nonnegative tensor factorization extracts latent features that are naturally sparse and parts of the data, which makes them easily interpretable. However, to take into account available domain knowledge and subject matter expertise, often additional constraints need to be imposed, which lead us to canonical decomposition with linear constraints (CANDELINC), a canonical polyadic decomposition with rank deficient factors. In CANDELINC, Tucker compression is used as a preprocessing step, which lead to a larger residual error but to more explainable latent features. Here, we propose a nonnegative CANDELINC (nnCANDELINC) accomplished via a specific nonnegative Tucker decomposition; we refer to as minimal or canonical nonnegative Tucker. We derive several results required to understand the specificity of nnCANDELINC, focusing on the difficulties of preserving the nonnegative rank of a tensor to its Tucker core and comparing the real valued to nonnegative case. Finally, we demonstrate nnCANDELINC performance on synthetic and real‐world examples.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2016

SU-F-T-267: A Clarkson-Based Independent Dose Verification for the Helical Tomotherapy

Purpose: There have been few reports for independent dose verification for Tomotherapy. We evaluated the accuracy and the effectiveness of an independent dose verification system for the Tomotherapy. Methods: Simple MU Analysis (SMU, Triangle Product, Ishikawa, Japan) was used as the independent verification system and the system implemented a Clarkson-based dose calculation algorithm using CT image dataset. For dose calculation in the SMU, the Tomotherapy machine-specific dosimetric parameters (TMR, Scp, OAR and MLC transmission factor) were registered as the machine beam data. Dose calculation was performed after Tomotherapy sinogram from DICOM-RT plan information was converted to the information for MU and MLC location at more segmented control points. The performance of the SMU was assessed by a point dose measurement in non-IMRT and IMRT plans (simple target and mock prostate plans). Subsequently, 30 patients’ treatment plans for prostate were compared. Results: From the comparison, dose differences between the SMU and the measurement were within 3% for all cases in non-IMRT plans. In the IMRT plan for the simple target, the differences (Average±1SD) were −0.70±1.10% (SMU vs. TPS), −0.40±0.10% (measurement vs. TPS) and −1.20±1.00% (measurement vs. SMU), respectively. For the mock prostate, the differences were −0.40±0.60% (SMU vs. TPS), −0.50±0.90% (measurement vs. TPS) and −0.90±0.60% (measurement vs. SMU), respectively. For patients’ plans, the difference was −0.50±2.10% (SMU vs. TPS). Conclusion: A Clarkson-based independent dose verification for the Tomotherapy can be clinically available as a secondary check with the similar tolerance level of AAPM Task group 114. This research is partially supported by Japan Agency for Medical Research and Development (AMED)

60 APPLIED LIFE SCIENCES↗
DOE OSTI2025

Flood Susceptibility Mapping Using Machine Learning and Geospatial-Sentinel-1 SAR Integration for Enhanced Early Warning Systems

This study presents a comprehensive framework for flood susceptibility mapping by integrating geospatial factors with both statistical and machine learning models. Thirteen Flood-related factors, including DEM, slope, TWI, NDVI, etc., are extracted as features of models, and historical flood data derived from Sentinel-1 SAR from 2018 to 2023 are used as the target variables of the models. These datasets are analyzed using a frequency-based statistical model and three machine learning models, including Random Forest, XGBoost, and CNN, to generate flood susceptibility maps. The performance of each model is evaluated through AUC; and SHAP scores are separately generated for Machine learning (ML) models to explain each feature contribution in the ML model. The generated susceptibility maps are validated by high-flood-risk locations monitored by flood sensors, BLE inundation models, and flood-prone areas suggested by the Local Community Task Force. The results indicate that the XGBoost model outperforms all other models, with an AUC of 0.92 and demonstrates the highest alignment with recommended high-flood-risk locations, while the frequency-based statistical model showed the weakest performance with an AUC of 0.65. SHAP value graphs highlight the elevation, slope, and TWI as the most influential features across all models. The susceptibility maps generated by the machine learning model show strong agreement with the BLE map and high-flood-risk areas identified by the local Community Task Force.

Google Engine↗
DOE OSTI2023

Large-scale genomic analyses with machine learning uncover predictive patterns associated with fungal phytopathogenic lifestyles and traits

Abstract Invasive plant pathogenic fungi have a global impact, with devastating economic and environmental effects on crops and forests. Biosurveillance, a critical component of threat mitigation, requires risk prediction based on fungal lifestyles and traits. Recent studies have revealed distinct genomic patterns associated with specific groups of plant pathogenic fungi. We sought to establish whether these phytopathogenic genomic patterns hold across diverse taxonomic and ecological groups from the Ascomycota and Basidiomycota, and furthermore, if those patterns can be used in a predictive capacity for biosurveillance. Using a supervised machine learning approach that integrates phylogenetic and genomic data, we analyzed 387 fungal genomes to test a proof-of-concept for the use of genomic signatures in predicting fungal phytopathogenic lifestyles and traits during biosurveillance activities. Our machine learning feature sets were derived from genome annotation data of carbohydrate-active enzymes (CAZymes), peptidases, secondary metabolite clusters (SMCs), transporters, and transcription factors. We found that machine learning could successfully predict fungal lifestyles and traits across taxonomic groups, with the best predictive performance coming from feature sets comprising CAZyme, peptidase, and SMC data. While phylogeny was an important component in most predictions, the inclusion of genomic data improved prediction performance for every lifestyle and trait tested. Plant pathogenicity was one of the best-predicted traits, showing the promise of predictive genomics for biosurveillance applications. Furthermore, our machine learning approach revealed expansions in the number of genes from specific CAZyme and peptidase families in the genomes of plant pathogens compared to non-phytopathogenic genomes (saprotrophs, endo- and ectomycorrhizal fungi). Such genomic feature profiles give insight into the evolution of fungal phytopathogenicity and could be useful to predict the risks of unknown fungi in future biosurveillance activities.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

Low-cost sensors and analytics for a digital factory

ORNL Manufacturing Demonstration Facility and Perisense worked on a low-cost solution for machining process monitoring to enable a digital factory. Enabling machine connectivity for digital factors is expensive for many small to medium manufacturers. The hypothesis for the research project is that a lowcost sensor suite, which can be retrofitted to a legacy machine tool or a new machine tool, can effectively provide real-time, in-situ sensing for manufacturing processes, which can enable improvements in process performance and efficiency. The Perisense sensor node was installed on the Haas TM1 machine with a current sensor and an accelerometer. Cutting tests were performed at different process parameters and sensor data was recorded. Results show a correlation of the current sensor to the spindle load. The accelerometer, mounted on the machine base, showed a correlation to the spindle speed as well as the spindle load. The sensor data can be used to monitor the machining process and provide insights on the machine spindle load utilization.

42 ENGINEERING↗
DOE OSTI2015

SU-E-T-517: Investigation of Factors Contributing to Extracranial Radiation Doses From Leksell Gamma Knife

Purpose The purpose of this study is to investigate dominant factors for doses to extracranial sites in treatment with Leksell Gamma Knife (LGK). Methods Monte Carlo simulation was implemented using EGS5 version 1.4.401. The simulation was divided into two major steps for the purpose of efficiency. As the first step, phase-space files were obtained at a scoring plane located just below patient-side surface of the collimator helmet of LGK. Scored particles were classified into three groups, primary, leakage and scatter, using their history information until their arrival to the scoring plane. Then classification was used at the following second step simulation to investigate which type of particle is dominant in the deposited energy at extra-cranial sites. In the second stage, a cylindrical phantom with a semisphere shaped head was modeled such that the geometrical center of the phantom’s head corresponds to the unit center point (UCP) of LGK. Scoring regions were arranged at 10 cm intervals from the UCP to 70 cm away on the central axis of the phantom. Energy deposition from each type of particles and location of interaction were recorded. Results The dominant factor of deposited energy depended on the collimator size. In the case of smaller collimator size, leakage was dominant. However, contribution of leakage was relatively small in the case of larger collimator size. The contribution of internal scatter varied with the distance from the UCP. In the proximal areas, internal scatter was dominant, whereas in the distal areas, particles interacting with machine components became dominant factor. Conclusion The Result of this study indicates that the dominant factor to dose to an extracranial site can vary with the distance from UCP and with collimator size. This means that the variation of this contribution must be considered for modeling of the extracranial dose especially in the distal area. This work was partly supported by the JSPS Core-to-Core Program (No. 23003)

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

A Tool for Assessing the Text Legibility of Digital Human Machine Interfaces

A tool intended to aid qualified professionals in the assessment of the legibility of text presented on a digital display is described. The assessment of legibility is primarily for the purposes of designing and analyzing human machine interfaces in accordance with NUREG-0700 and MIL-STD 1472G. The tool addresses shortcomings of existing guidelines by providing more accurate metrics of text legibility with greater sensitivity to design alternatives.

human factors↗
DOE OSTI2018

Nonnegative tensor factorization for contaminant source identification

Unsupervised Machine Learning (ML) is becoming increasingly popular for solving various types of data analytics problems including feature extraction, blind source separation, exploratory analyses, model diagnostics, etc. In this work, we have developed a new unsupervised ML method based on Nonnegative Tensor Factorization (NTF) for identification of the original groundwater types (including contaminant sources) present in geochemical mixtures observed in an aquifer. Frequently, groundwater types with different geochemical signatures are related to different background and/or contamination sources. The characterization of groundwater mixing processes is a challenging but very important task critical for any environmental management project aiming to characterize the fate and transport of contaminants in the subsurface and perform contaminant remediation. This task typically requires solving complex inverse models representing groundwater flow and geochemical transport in the aquifer, where the inverse analysis accounts for available site data. Usually, the model is calibrated against the available data characterizing the spatial and temporal distribution of the observed geochemical types. Numerous different geochemical constituents and processes may need to be simulated in these models which further complicates the analyses. Additionally, the application of inverse methods may introduce biases in the analyses through the assumptions made in the model development process. Here, we substitute the model inversion with unsupervised ML analysis. The ML analysis does not make any assumptions about underlying physical and geochemical processes occurring in the aquifer. Our ML methodology, called NTFk, is capable of identifying (1) the unknown number of groundwater types (contaminant sources) present in the aquifer, (2) the original geochemical concentrations (signatures) of these groundwater types and (3) spatial and temporal dynamics in the mixing of these groundwater types. These results are obtained only from the measured geochemical data without any additional site information. In general, the NTFk methodology allows for interpretation of large high-dimensional datasets representing diverse spatial and temporal components such as state variables and velocities. NTFk has been tested on synthetic and real-world site three-dimensional datasets. Finally, the NTFk algorithm is designed to work with geochemical data represented in the form of concentrations, ratios (of two constituents; for example, isotope ratios), and delta notations (standard normalized stable isotope ratios).

58 GEOSCIENCES↗
DOE OSTI2019

Machine Learning

Presentation overview: Why Machine Learning?; Why ML-Informed Physics?; Why Physics-Informed ML?; LANL ML strategy related to National & Energy Security; ML-Informed Physics for climate modeling and induced seismicity (LDRD-DR); Physics-Informed ML for blind source separation (LDRD-DR); ML forecasting production from tight rock reservoirs (Chevron)

58 GEOSCIENCES↗
DOE OSTI2022

Land-based wind turbines with flexible rail-transportable blades – Part 2: 3D finite element design optimization of the rotor blades

Abstract. Increasing growth in land-based wind turbine blades to enable higher machine capacities and capacity factors is creating challenges in design, manufacturing, logistics, and operation. Enabling further blade growth will require technology innovation. An emerging solution to overcome logistics constraints is to segment the blades spanwise and chordwise, which is effective, but the additional field-assembled joints result in added mass and loads, as well as increased reliability concerns in operation. An alternative to this methodology is to design slender flexible blades that can be shipped on rail lines by flexing during transport. However, the increased flexibility is challenging to accommodate with a typical glass-fiber, upwind design. In a two-part paper series, several design options are evaluated to enable slender flexible blades: downwind machines, optimized carbon fiber, and active aerodynamic controls. Part 1 presents the system-level optimization of the rotor variants as compared to conventional and segmented baselines, with a low-fidelity representation of the blades. The present work, Part 2, supplements the system-level optimization in Part 1 with high-fidelity blade structural optimization to ensure that the designs are at feasible optima with respect to material strength and fatigue limits, as well as global stability and structural dynamics constraints. To accommodate the requirements of the design process, a new version of the Numerical Manufacturing And Design (NuMAD) code has been developed and released. The code now supports laminate-level blade optimization and an interface to the International Energy Agency Wind Task 37 blade ontology. Transporting long, flexible blades via controlled flapwise bending is found to be a viable approach for blades of up to 100 m. The results confirm that blade mass can be substantially reduced by going either to a downwind design or to a highly coned and tilted upwind design. A discussion of active and inactive constraints consisting of material rupture, fatigue damage, buckling, deflection, and resonant frequencies is presented. An analysis of driving load cases revealed that the downwind designs are dominated by loads from sudden, abrupt events like gusts rather than fatigue. Finally, an analysis of carbon fiber spar caps for downwind machines finds that, compared to typical carbon fibers, the use of a new heavy-tow carbon fiber in the spar caps is found to yield between 9 % and 13 % cost savings.

17 WIND ENERGY↗