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 577 records · Page 32

All sources NASA NTRS DOE OSTI

DOE OSTI2025

A structured framework for predicting sustainable aviation fuel properties using liquid-phase FTIR and machine learning

Sustainable aviation fuels have the potential to improve efficiency, reduce emissions, and enhance energy security. To help identify viable sustainable aviation fuels and accelerate research, machine learning models have been developed to predict relevant physicochemical properties. However, many models have limited applicability, leverage data from complex analytical techniques with confined spectral ranges, or use feature decomposition methods that offer limited interpretability. Using liquid-phase Fourier Transform Infrared (FTIR) spectra, this study presents a structured method for creating accurate and interpretable property prediction models for neat molecules, aviation fuels, and blends. Liquid FTIR spectra can be collected quickly and consistently, offering high reliability, sensitivity, and component specificity using less than 2 ml of sample. The method first decomposes FTIR spectra into fundamental building blocks using non-negative matrix factorization (NMF) to enable scientific analysis of FTIR spectra attributes and fuel properties. The NMF features are then used to create five ensemble models for predicting final boiling point, flash point, freezing point, density at 15°C, and kinematic viscosity at -20°C. All models were trained using experimental property data from neat molecules, aviation fuels, and blends. The models accurately predict key properties across a broad range of neat molecules and representative fuels and blends, while enabling interpretation of relationships between compositional elements, such as functional groups or chemical classes, and their resulting properties. This demonstrates strong potential to support sustainable aviation fuel research and development. The models and data are available on an interactive web tool.

Fourier transform infrared spectroscopy↗

DOE OSTI2024

Machine Learning Discrimination and Ultrasensitive Detection of Fentanyl Using Gold Nanoparticle-Decorated Carbon Nanotube-Based Field-Effect Transistor Sensors

The opioid overdose crisis is a global health challenge. Fentanyl, an exceedingly potent synthetic opioid, has emerged as a leading contributor to the surge in opioid-related overdose deaths. The surge in overdose fatalities, particularly due to illicitly manufactured fentanyl and its contamination of street drugs, emphasizes the urgency for drug-testing technologies that can quickly and accurately identify fentanyl from other drugs and quantify trace amounts of fentanyl. In this paper, gold nanoparticle (AuNP)-decorated single-walled carbon nanotube (SWCNT)-based field-effect transistors (FETs) are utilized for machine learning-assisted identification of fentanyl from codeine, hydrocodone, and morphine. The unique sensing performance of fentanyl led to use machine learning approaches for accurate identification of fentanyl. Employing linear discriminant analysis (LDA) with a leave-one-out cross-validation approach, a validation accuracy of 91.2% is achieved. Meanwhile, density functional theory (DFT) calculations reveal the factors that contributed to the enhanced sensitivity of the Au-SWCNT FET sensor toward fentanyl as well as the underlying sensing mechanism. Finally, fentanyl antibodies are introduced to the Au-SWCNT FET sensor as specific receptors, expanding the linear range of the sensor in the lower concentration range, and enabling ultrasensitive detection of fentanyl with a limit of detection at 10.8 fg mL –1 .

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2020

Prediction of the development of islet autoantibodies through integration of environmental, genetic, and metabolic markers

The Environmental Determinants of the Diabetes in the Young (TEDDY) study has prospectively followed, from birth, children at increased genetic risk of type 1 diabetes. We evaluated the potential of machine learning to identify new biomarkers that predict imminent (within 6 months) development of persistent islet autoantibodies to insulin, GAD or IA-2 in TEDDY participants through integration of time-invariant risk factors with time-varying metabolomics. The predictive modeling was initiated with over 220 potential biomarkers; through ensemble-based feature evaluation, the optimal model included 42 biomarkers, returning a cross-validated receiver operating characteristic area under the curve of 0.74. The model identified a principal set of 20 time-invariant markers, including 16 single nucleotide polymorphisms and two HLA-DR genotypes, gestational age, and exposure to a prebiotic formula. Integration of the metabolome identified 22 high-priority metabolites and lipids, including adipic acid and ceramide d42:0, that predicted development of islet autoantibodies, dependent upon the time horizon. The majority (86%) of metabolites that predicted development of islet autoantibodies belonged to 3 pathways: lipid oxidation, phospholipase A2 signaling, and pentose phosphate pathway. TEDDY data suggest that these metabolic processes may play a role in triggering islet autoimmunity.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2017

A Disruptive Approach to Electric Vehicle Power Electronics Final Report

This project developed and demonstrated a new power electronics technology for electric vehicle drivetrains, that achieved significant improvements in power density and efficiency over standard drive cycles, relative to the commercial state of the art. In particular, the prototype measured loss and efficiency curves demonstrated a reduction in total loss over the US EPA US06 drive cycle by a factor of four. Similar reductions are observed for urban (UDDS) and highway (HWFET) drive cycles. Further, the project demonstrated a significant reduction in the total film capacitor requirements, by over an order of magnitude; this was achieved through fundamental improvements in converter circuit topologies. These nonincremental technology advances can lead to significant improvements in temperature rise, reliability, cost, and cooling system size and weight. These results were achieved through a new proposed composite converter architecture, and were demonstrated in a boost-type system similar to that employed by Toyota, Ford, and Honda. A 30 kW silicon prototype achieved a US06 average efficiency of 97.5%, which represents a reduction in average loss by a factor of approximately four relative to the 92.5% average efficiency of the commercial state of the art. A silicon carbide prototype achieved similar efficiency, while also achieving power densities of 23 kW/L (volumetric) and 20 kW/kg (gravimetric) in a boost dc-dc system including magnetics and film capacitors. This was achieved with a SiC switching frequency of 240 kHz, which allowed employment of planar magnetics. These power densities are roughly four times better than the commercial state of the art, and significantly exceed the DOE APEEM 2020 goals. In addition to the composite boost system, a SiC inverter was demonstrated at the 800 V 30 kW level. This inverter was demonstrated driving a permanent magnet machine over the US 06 drive cycle. The replacement of 1200 V Si IGBTs with 1200 V SiC MOSFETs led to improvement of CAFE average efficiency from 98.7% to 99.5%, which corresponds to a reduction of average loss by a factor of approximately two. The modular nature of the composite converter architecture opens the possibility of reuse of one or more converter modules during charging operations, providing a path for onboard charging capability at significant power levels while significantly reducing the volume and weight of the added charger module. This project demonstrated an integrated Level II charger whose add-on specific weight was approximately four times better than the DOE PHEV charger 2022 targets.

33 ADVANCED PROPULSION SYSTEMS↗

DOE OSTI2015

Stabilization of high-compression, indirect-drive inertial confinement fusion implosions using a 4-shock adiabat-shaped drive

Hydrodynamic instabilities and poor fuel compression are major factors for capsule performance degradation in ignition experiments on the National Ignition Facility. Using a recently developed laser drive profile with a decaying first shock to tune the ablative Richtmyer-Meshkov (ARM) instability and subsequent in-flight Rayleigh-Taylor growth, we have demonstrated reduced growth compared to the standard ignition pulse whilst maintaining conditions for a low fuel adiabat needed for increased compression. Using in-flight x-ray radiography of pre-machined modulations, the first growth measurements using this new ARM-tuned drive have demonstrated instability growth reduction of ∼4× compared to the original design at a convergence ratio of ∼2. Corresponding simulations give a fuel adiabat of ∼1.6, similar to the original goal and consistent with ignition requirements.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DOE OSTI2021

Predictive Models of Genetic Redundancy in Arabidopsis thaliana

Abstract Genetic redundancy refers to a situation where an individual with a loss-of-function mutation in one gene (single mutant) does not show an apparent phenotype until one or more paralogs are also knocked out (double/higher-order mutant). Previous studies have identified some characteristics common among redundant gene pairs, but a predictive model of genetic redundancy incorporating a wide variety of features derived from accumulating omics and mutant phenotype data is yet to be established. In addition, the relative importance of these features for genetic redundancy remains largely unclear. Here, we establish machine learning models for predicting whether a gene pair is likely redundant or not in the model plant Arabidopsis thaliana based on six feature categories: functional annotations, evolutionary conservation including duplication patterns and mechanisms, epigenetic marks, protein properties including posttranslational modifications, gene expression, and gene network properties. The definition of redundancy, data transformations, feature subsets, and machine learning algorithms used significantly affected model performance based on holdout, testing phenotype data. Among the most important features in predicting gene pairs as redundant were having a paralog(s) from recent duplication events, annotation as a transcription factor, downregulation during stress conditions, and having similar expression patterns under stress conditions. We also explored the potential reasons underlying mispredictions and limitations of our studies. This genetic redundancy model sheds light on characteristics that may contribute to long-term maintenance of paralogs, and will ultimately allow for more targeted generation of functionally informative double mutants, advancing functional genomic studies.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2015

Output calculation of electron therapy at extended SSD using an improved LBR method

Purpose: To calculate the output factor (OPF) of any irregularly shaped electron beam at extended SSD. Methods: Circular cutouts were prepared from 2.0 cm diameter to the maximum possible size for 15 × 15 applicator cone. In addition, two irregular cutouts were prepared. For each cutout, percentage depth dose (PDD) at the standard SSD and doses at different SSD values were measured using 6, 9, 12, and 16 MeV electron beam energies on a Varian 2100C LINAC and the distance at which the central axis electron fluence becomes independent of cutout size was determined. The measurements were repeated with an ELEKTA Synergy LINAC using 14 × 14 applicator cone and electron beam energies of 6, 9, 12, and 15 MeV. The PDD measurements were performed using a scanning system and two diodes—one for the signal and the other a stationary reference outside the tank. The doses of the circular cutouts at different SSDs were measured using PTW 0.125 cm{sup 3} Semiflex ion-chamber and EDR2 films. The electron fluence was measured using EDR2 films. Results: For each circular cutout, the lateral buildup ratio (LBR) was calculated from the measured PDD curve using the open applicator cone as the reference field. The effective SSD (SSD{sub eff}) of each circular cutout was calculated from the measured doses at different SSD values. Using the LBR value and the radius of the circular cutout, the corresponding lateral spread parameter [σ{sub R}(z)] was calculated. Taking the cutout size dependence of σ{sub R}(z) into account, the PDD curves of the irregularly shaped cutouts at the standard SSD were calculated. Using the calculated PDD curve of the irregularly shaped cutout along with the LBR and SSD{sub eff} values of the circular cutouts, the output factor of the irregularly shaped cutout at extended SSD was calculated. Finally, both the calculated PDD curves and output factor values were compared with the measured values. Conclusions: The improved LBR method has been generalized to calculate the output factor of electron therapy at extended SSD. The percentage difference between the calculated and the measured output factors of irregularly shaped cutouts in a clinical useful SSD region was within 2%. Similar results were obtained for all available electron energies of both Varian 2100C and ELEKTA Synergy machines.

07 ISOTOPE AND RADIATION SOURCES↗

DOE OSTI2023

High-throughput functional trait testing for bacterial pathogens

Functional traits are characteristics that affect the fitness and metabolic function of a microorganism. There is growing interest in using high-throughput methods to characterize bacterial pathogens based on functional virulence traits. Traditional methods that phenotype a single organism for a single virulence trait can be time consuming and labor intensive. Alternatively, machine learning of whole-genome sequences (WGS) has shown some success in predicting virulence. However, relying solely on WGS can miss functional traits, particularly for organisms lacking classical virulence factors. We propose that high-throughput assays for functional virulence trait identification should become a prominent method of characterizing bacterial pathogens on a population scale. This work is critical as we move from compiling lists of bacterial species associated with disease to pathogen-agnostic approaches capable of detecting novel microbes. We discuss six key areas of functional trait testing and how advancing high-throughput methods could provide a greater understanding of pathogens.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2015

Dislocation model of nucleation and development of slip bands and their effect on service life of structural materials subject to cyclic loading

Most of the destructions of machine parts are of fatigue character. Under cyclic loading, the surface layer, in which hardening–softening processes rapidly occur, is formed almost at once after its beginning. The interaction of plastic-deformation traces with each other and with other structural elements, such as grains, results in the formation of a characteristic microstructure of the machine-part surface subject to cyclic loadings. The character of accumulation of slip bands and their shape (narrow, wide, twisting, and broken) depends on the conditions under which (under what factors) the cyclic loading occurs. The fatigue-resistance index expressed in terms of the slope of left portion of the fatigue curve linearized in logarithmic coordinates also depends on the set of relevant factors. The dependence of the surface damageability on the fatigue resistance index makes it possible to implement the method of predicting the fatigue curve by the description of the factors acting on a detail or construction. The position of the inflection point on the curve in the highcycle fatigue region (the endurance limit and the number of loading cycles, the ordinate and abscissa of the inflection point on the fatigue curve, respectively) also depends on the set of relevant factors. In combination with the previously obtained value of the slope of the left portion of the curve in the high-cycle fatigue region, this makes it possible to construct an a priori fatigue curve, thus reducing the scope of required fatigue tests and, hence, high expenses because of their long duration and high cost. The scope of tests upon using the developed method of prediction may be reduced to a minimum of one or two samples at the predicted level of the endurance limit.

36 MATERIALS SCIENCE↗

DOE OSTI2019

Report for 2.2.1 Task 3: Develop and Document an Advanced Human System Interface for the Generic Pressurized Water Reactor Simulator

The report explains the development of four new Operator Work Displays (OWDs). The work is performed for Idaho National Laboratory (INL) by the Institute for Energy Technology (IFE). The OWDs are implemented, and initially tested at the Generic Pressurized Water Reactor (gPWR) simulator in HAlden Man Machine LABoratory (HAMMLAB). The OWDs are designed to bridge the gap between large overview displays and detailed operator workstation displays. The report presents their design rationale, which stem from: human factors findings, industrial standards and guidelines and research into visual perception for computer displays. The report presents initial experiences using the OWDs, and topics for further work.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

DOE OSTI2019

Passive method to measure strength of turbulence

Disclosed is a method to passively measure and calculate the strength of turbulence via the index of refraction structure constant C.sub.n.sup.2 from video imagery gathered by an imaging device, such as a video camera. Processing may occur with any type computing device utilizing a processor executing machine executable code stored on memory. This method significantly simplifies instrumentation requirements, reduces cost, and provides rapid data output. This method combines an angle of arrival methodology, which provides scale factors, with a new spatial/temporal frequency domain method. As part of the development process, video imagery from high speed cameras was collected and analyzed. The data was decimated to video rates such that statistics could be computed and used to confirm that this passive method accurately characterizes the atmospheric turbulence. C.sub.n.sup.2 accuracy from this method compared well with scintillometer data through two full orders of magnitude and more capability is expected beyond this verification.

O'Neill, Mary Morabito↗

DOE OSTI2021

Passive method to measure strength of turbulence

Disclosed is a method to passively measure and calculate the strength of turbulence via the index of refraction structure constant Cn2 from video imagery gathered by an imaging device, such as a video camera. Processing may occur with any type computing device utilizing a processor executing machine executable code stored on memory. This method significantly simplifies instrumentation requirements, reduces cost, and provides rapid data output. This method combines an angle of arrival methodology, which provides scale factors, with a new spatial/temporal frequency domain method. As part of the development process, video imagery from high speed cameras was collected and analyzed. The data was decimated to video rates such that statistics could be computed and used to confirm that this passive method accurately characterizes the atmospheric turbulence. Cn2 accuracy from this method compared well with scintillometer data through two full orders of magnitude and more capability is expected beyond this verification.

O'Neill, Mary Morabito↗

DOE OSTI2025

Accelerating multilevel Markov Chain Monte Carlo using machine learning models

Here, this work presents an efficient approach for accelerating multilevel Markov Chain Monte Carlo (MCMC) sampling for large-scale problems using low-fidelity machine learning models. While conventional techniques for large-scale Bayesian inference often substitute computationally expensive high-fidelity models with machine learning models, thereby introducing approximation errors, our approach offers a computationally efficient alternative by augmenting high-fidelity models with low-fidelity ones within a hierarchical framework. The multilevel approach utilizes the low-fidelity machine learning model (MLM) for inexpensive evaluation of proposed samples thereby improving the acceptance of samples by the high-fidelity model. The hierarchy in our multilevel algorithm is derived from geometric multigrid hierarchy. We utilize an MLM to accelerate the coarse level sampling. Training machine learning model for the coarsest level significantly reduces the computational cost associated with generating training data and training the model. We present an MCMC algorithm to accelerate the coarsest level sampling using MLM and account for the approximation error introduced. We provide theoretical proofs of detailed balance and demonstrate that our multilevel approach constitutes a consistent MCMC algorithm. Additionally, we derive the expression for cost reduction due to machine learning model to facilitate cost analysis of the hierarchical sampling algorithm. Our technique is demonstrated on a standard benchmark inference problem in groundwater flow, where we estimate the probability density of a quantity of interest using a four-level MCMC algorithm. Our proposed algorithm accelerates multilevel sampling by a factor of two while achieving similar accuracy compared to sampling using the standard multilevel algorithm.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2021

Impact of short-read sequencing on the misassembly of a plant genome

Abstract Background Availability of plant genome sequences has led to significant advances. However, with few exceptions, the great majority of existing genome assemblies are derived from short read sequencing technologies with highly uneven read coverages indicative of sequencing and assembly issues that could significantly impact any downstream analysis of plant genomes. In tomato for example, 0.6% (5.1 Mb) and 9.7% (79.6 Mb) of short-read based assembly had significantly higher and lower coverage compared to background, respectively. Results To understand what the causes may be for such uneven coverage, we first established machine learning models capable of predicting genomic regions with variable coverages and found that high coverage regions tend to have higher simple sequence repeat and tandem gene densities compared to background regions. To determine if the high coverage regions were misassembled, we examined a recently available tomato long-read based assembly and found that 27.8% (1.41 Mb) of high coverage regions were potentially misassembled of duplicate sequences, compared to 1.4% in background regions. In addition, using a predictive model that can distinguish correctly and incorrectly assembled high coverage regions, we found that misassembled, high coverage regions tend to be flanked by simple sequence repeats, pseudogenes, and transposon elements. Conclusions Our study provides insights on the causes of variable coverage regions and a quantitative assessment of factors contributing to plant genome misassembly when using short reads and the generality of these causes and factors should be tested further in other species.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2015

AUTOMATED TRANSIENT IDENTIFICATION IN THE DARK ENERGY SURVEY

We describe an algorithm for identifying point-source transients and moving objects on reference-subtracted optical images containing artifacts of processing and instrumentation. The algorithm makes use of the supervised machine learning technique known as Random Forest. We present results from its use in the Dark Energy Survey Supernova program (DES-SN), where it was trained using a sample of 898,963 signal and background events generated by the transient detection pipeline. After reprocessing the data collected during the first DES-SN observing season (2013 September through 2014 February) using the algorithm, the number of transient candidates eligible for human scanning decreased by a factor of 13.4, while only 1.0% of the artificial Type Ia supernovae (SNe) injected into search images to monitor survey efficiency were lost, most of which were very faint events. Here we characterize the algorithm’s performance in detail, and we discuss how it can inform pipeline design decisions for future time-domain imaging surveys, such as the Large Synoptic Survey Telescope and the Zwicky Transient Facility. An implementation of the algorithm and the training data used in this paper are available at at http://portal.nersc.gov/project/dessn/autoscan.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2023

Human Factors for Advanced Reactors

Existing light water reactors in the U.S. are primarily large baseload electricity generating facilities. The concept of operations for these plants remains largely unchanged since the advent of commercial nuclear power—the main control room serves as the hub of plant activities and is staffed with multiple licensed operators who work in tandem under the shift supervisor, and staff such as field workers support the control room remotely. While newer plants have brought the advent of digital human-machine interfaces to replace earlier analog and mechanical instrumentation and controls, much of the control process remains unchanged and manual. It is simply a newer version of legacy concepts. Advanced reactors potentially bring considerable changes to the size, fuel type, automation, and staffing of nuclear power plants, necessitating a fundamental shift not just from analog to digital, but further from human to automation, from onsite to remote, from control to monitoring, and from many to few operators. Despite this multitude of parallel evolutions in reactor designs, many of the vendors developing the next generation of reactors represent smaller research and development enterprises. It is therefore not feasible to address all aspects of plant design at the same time. In particular, the competing design aspects of new reactors present a significant challenge to the development of robust and human factored systems at the plant. As vendors develop new reactor designs, much of the early focus is naturally on the fuel and reactor system technology. Looming behind these early advances is the daunting prospect of first-of-a-kind control concepts that have not yet been developed or validated. A failure to address the human element of reactor design early will lead to missed opportunities. The quickest development process is the replication of existing concepts of operations at legacy plants, even when such systems were long ago surpassed by better human-machine technologies outside the nuclear industry. Conversely, attempting to undertake novel concepts of operations late in the design life cycle of a plant could result in protracted development efforts and delays in licensing and deployment. This does not have to happen, and it is imperative that human factors be considered now, early in the design of new reactors.

99 GENERAL AND MISCELLANEOUS↗

DOE OSTI2023

Massively scalable Kerr comb-driven silicon photonic link

Abstract The growth of computing needs for artificial intelligence and machine learning is critically challenging data communications in today’s data-centre systems. Data movement, dominated by energy costs and limited ‘chip-escape’ bandwidth densities, is perhaps the singular factor determining the scalability of future systems. Using light to send information between compute nodes in such systems can dramatically increase the available bandwidth while simultaneously decreasing energy consumption. Through wavelength-division multiplexing with chip-based microresonator Kerr frequency combs, independent information channels can be encoded onto many distinct colours of light in the same optical fibre for massively parallel data transmission with low energy. Although previous high-bandwidth demonstrations have relied on benchtop equipment for filtering and modulating Kerr comb wavelength channels, data-centre interconnects require a compact on-chip form factor for these operations. Here we demonstrate a massively scalable chip-based silicon photonic data link using a Kerr comb source enabled by a new link architecture and experimentally show aggregate single-fibre data transmission of 512 Gb s −1 across 32 independent wavelength channels. The demonstrated architecture is fundamentally scalable to hundreds of wavelength channels, enabling massively parallel terabit-scale optical interconnects for future green hyperscale data centres.

Rizzo, Anthony (ORCID:000000034752797X)↗

DOE OSTI2026

Mass of 101 Sn and Bayesian extrapolations to the proton drip line

The favorable energy configurations of nuclei at magic numbers of 𝑁 neutrons and 𝑍 protons are fundamental for understanding the evolution of nuclear structure. The 𝑍 = 50 (tin) isotopic chain is a frontier for such studies, with particular interest at and around the doubly magic 100 Sn isotope, for which the mass is a topic of debate. Precise mass values for neutron-deficient isotopes provide necessary anchor points for mass models to test extrapolations near the proton drip line, where experimental studies remain out of reach. In this work, we report a Penning trap mass measurement of 101 Sn . The determined mass excess of −59889.89⁢(96) keV for 101 Sn represents a factor-of-300 improvement over the current precision and indicates that 101 Sn is less bound than previously thought. Mass predictions from a recently developed Bayesian model combination framework employing statistical machine learning and nuclear masses computed within seven global models based on nuclear density functional theory agree within 1⁢𝜎 with experimental masses from the 48 ≤ 𝑍 ≤ 52 isotopic chains. The framework's resilience to new mass data gave confidence in the extrapolation of tin masses down to 𝑁 = 46. Our calculations suggest that 96 Sn is a two-proton drip line nucleus and predict a mass excess of −58090⁢(800) keV for 100 Sn , showing a preference within 1⁢𝜎 for the mass of 100 Sn derived from the 𝛽-delayed 𝑄 value measured at GSI.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗