Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “data set”

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 865 records · Page 48

Streaming Compression of Scientific Data via Weak-SINDy

Here, in this paper, a streaming weak-SINDy algorithm is developed specifically for compressing streaming scientific data. The production of scientific data, either via simulation or experiments, is undergoing a stage of exponential growth, which makes data compression important and often necessary for storing and utilizing large scientific data sets. As opposed to classical “offline” compression algorithms that perform compression on a readily available data set, streaming compression algorithms compress data “online” while the data generated from simulation or experiments is still flowing through the system. This feature makes streaming compression algorithms well suited for scientific data compression, where storing the full data set offline is often infeasible. This work proposes a new streaming compression algorithm, streaming weak-SINDy, which takes advantage of the underlying data characteristics during compression. The streaming weak-SINDy algorithm constructs feature matrices and target vectors in the online stage via a streaming integration method in a memory efficient manner. The feature matrices and target vectors are then used in the offline stage to build a model through a regression process that aims to recover equations that govern the evolution of the data. For compressing high-dimensional streaming data, we adopt a streaming proper orthogonal decomposition (POD) process to reduce the data dimension and then use the streaming weak-SINDy algorithm to compress the temporal data of the POD expansion. We propose modifications to the streaming weak-SINDy algorithm to accommodate the dynamically updated POD basis. By combining the built model from the streaming weak-SINDy algorithm and a small amount of data samples, the full data flow could be reconstructed accurately at a low memory cost, as shown in the numerical tests.

97 MATHEMATICS AND COMPUTING↗

HPC Analytics Support. Requirements for Uncertainty Quantification Benchmarks

This report outlines techniques for extending benchmark generation products so they support uncertainty quantification by benchmarked systems. We describe how uncertainty quantification requirements can be presented to candidate analytical tools supporting SPARQL. We describe benchmark data sets for evaluating uncertainty quantification, as well as an approach for using our benchmark generator to produce data sets for generating benchmark data sets.

benchmark↗

All-sky Search for Transient Astrophysical Neutrino Emission with 10 Years of IceCube Cascade Events

Abstract Neutrino flares in the sky are searched for in data collected by IceCube between 2011 and 2021 May. This data set contains cascade-like events originating from charged-current electron neutrino and tau neutrino interactions and all-flavor neutral-current interactions. IceCube’s previous all-sky searches for neutrino flares used data sets consisting of track-like events originating from charged-current muon neutrino interactions. The cascade data set is statistically independent of the track data sets, and while inferior in angular resolution, the low-background nature makes it competitive and complementary to previous searches. No statistically significant flare of neutrino emission was observed in an all-sky scan. Upper limits are calculated on neutrino flares of varying duration from 1 hr to 100 days. Furthermore, constraints on the contribution of these flares to the diffuse astrophysical neutrino flux are presented, showing that multiple unresolved transient sources may contribute to the diffuse astrophysical neutrino flux.

79 ASTRONOMY AND ASTROPHYSICS↗

Measurements of the 27 Al(α,n) Thick Target Yield near Threshold

In this work, we present results from direct measurements of the 27 Al(α,n) thick-target yield from laboratory incident energies E α ≈ 3 to 5 MeV, performed with the 3 HeBF 3 Giant Barrel (HeBGB) neutron detector at the Edwards Accelerator Laboratory. Our measurements have a small energy cadence in order to address discrepancies and sparseness of thick-target-yield data sets existing for this energy region. We find general agreement with existing data sets, including yields derived from cross-section data, while resolving a discrepancy between existing thick-target-yield data sets for E α ≈ 4–5 MeV. However, for E α < 3.5 MeV, our results are substantially lower than previous thick-target-yield data and somewhat larger than yields calculated from existing cross-section data. Our data complete the energy range needed for estimates of the 27 Al(α,n) contribution to neutrino and dark matter detector backgrounds and result in increased viability of 27 Al(α,n) as a plasma diagnostic tool at fusion facilities such as the National Ignition Facility.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Evaluating interaction energies of weakly bonded systems using the Buckingham-Hirshfeld method

We present the finalized Buckingham-Hirshfeld method (BHD-DFT) for the evaluation of interaction energies of non-bonded dimers with Density Functional Theory (DFT). In the method, dispersion energies are evaluated from static multipole polarizabilities, obtained on-the-fly from Coupled Perturbed Kohn-Sham calculations and partitioned into diatomic contributions using the iterative Hirshfeld partitioning method. The dispersion energy expression is distributed over four atoms and has therefore a higher delocalized character compared to the standard pairwise expressions. Additionally, full multipolar polarizability tensors are used as opposed to effective polarizabilities, allowing to retain the anisotropic character at no additional computational cost. A density dependent damping function for the BLYP, PBE, BP86, B3LYP, and PBE0 functionals has been implemented, containing two global parameters which were fitted to interaction energies and geometries of a selected number of dimers using a bi-variate RMS fit. The method is benchmarked against the S22 and S66 data sets for equilibrium geometries and the S22x5 and S66x8 data sets for interaction energies around the equilibrium geometry. Best results are achieved using the B3LYP functional with mean average deviation values of 0.30 and 0.24 kcal/mol for the S22 and S66 data sets, respectively. This situates the BHD-DFT method among the best performing dispersion inclusive DFT methods. Effect of counterpoise correction on DFT energies is discussed.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Generating and executing programs for a floating point single instruction multiple data instruction set architecture

Mechanisms for generating and executing programs for a floating point (FP) only single instruction multiple data (SIMD) instruction set architecture (ISA) are provided. A computer program product comprising a computer recordable medium having a computer readable program recorded thereon is provided. The computer readable program, when executed on a computing device, causes the computing device to receive one or more instructions and execute the one or more instructions using logic in an execution unit of the computing device. The logic implements a floating point (FP) only single instruction multiple data (SIMD) instruction set architecture (ISA), based on data stored in a vector register file of the computing device. The vector register file is configured to store both scalar and floating point values as vectors having a plurality of vector elements.

97 MATHEMATICS AND COMPUTING↗

The Power of Environmental Observatories for Advancing Multidisciplinary Research, Outreach, and Decision Support: The Case of the Minnesota River Basin

Observatory-scale data collection efforts allow unprecedented opportunities for integrative, multidisciplinary investigations in large, complex watersheds, which can affect management decisions and policy. Through the National Science Foundation-funded REACH (REsilience under Accelerated CHange) project, in collaboration with the Intensively Managed Landscapes-Critical Zone Observatory, we have collected a series of multidisciplinary data sets throughout the Minnesota River Basin in south-central Minnesota, USA, a 43,400-km 2 tributary to the Upper Mississippi River. Postglacial incision within the Minnesota River valley created an erosional landscape highly responsive to hydrologic change, allowing for transdisciplinary research into the complex cascade of environmental changes that occur due to hydrology and land use alterations from intensive agricultural management and climate change. Data sets collected include water chemistry and biogeochemical data, geochemical fingerprinting of major sediment sources, high-resolution monitoring of river bluff erosion, and repeat channel cross-sectional and bathymetry data following major floods. The data collection efforts led to development of a series of integrative reduced complexity models that provide deeper insight into how water, sediment, and nutrients route and transform through a large channel network and respond to change. These models represent the culmination of efforts to integrate interdisciplinary data sets and science to gain new insights into watershed-scale processes in order to advance management and decision making. Finally, the purpose of this paper is to present a synthesis of the data sets and models, disseminate them to the community for further research, and identify mechanisms used to expand the temporal and spatial extent of short-term observatory-scale data collection efforts.

54 ENVIRONMENTAL SCIENCES↗

Putting Our Industry's Data to Work: A Case Study of Large-Scale Data Aggregation: Preprint

With increasing deployment of Advanced Metering Infrastructure (AMI), Building Automation System (BAS) controls, Internet of Things (IoT) network devices, and data-driven evaluation, measurement, and verification studies, the building sector is currently generating a staggering amount of energy-related data. In the right hands, these data sets can contribute to increased comfort and energy savings for building occupants and a more reliable electrical grid; however, due to a combination of factors, including significant privacy concerns, much of the data that are presently generated and stored are not used outside of basic operational applications. In the past year, our team has dedicated over 2,000 person-hours to accessing building energy data for a project funded by the U.S. Department of Energy (DOE) Building Technologies Office. We sought whole-building or end-use (e.g., lighting) timeseries data at the individual-building or equipment level where possible or aggregated information, such as timeseries averages and quartiles by building type (e.g., office, retail, hospital), where sharing individual building information was not an option. We are additionally working with IoT and BAS data sets to derive information important to the project. We have assembled an extensive data set that will enable the development of publicly available end-use load profiles to benefit the U.S. building and electricity industries. Here we present an overview of the data set that we have assembled to date, the motivators and approaches that got us here, and the lessons we learned through our efforts. We also discuss work underway that presents additional options for future data access.

building energy data↗

Uncertainty-driven dynamics for active learning of interatomic potentials

Machine learning (ML) models, if trained to data sets of high-fidelity quantum simulations, produce accurate and efficient interatomic potentials. Active learning (AL) is a powerful tool to iteratively generate diverse data sets. In this approach, the ML model provides an uncertainty estimate along with its prediction for each new atomic configuration. If the uncertainty estimate passes a certain threshold, then the configuration is included in the data set. Here we develop a strategy to more rapidly discover configurations that meaningfully augment the training data set. The approach, uncertainty-driven dynamics for active learning (UDD-AL), modifies the potential energy surface used in molecular dynamics simulations to favor regions of configuration space for which there is large model uncertainty. The performance of UDD-AL is demonstrated for two AL tasks: sampling the conformational space of glycine and sampling the promotion of proton transfer in acetylacetone. The method is shown to efficiently explore the chemically relevant configuration space, which may be inaccessible using regular dynamical sampling at target temperature conditions.

97 MATHEMATICS AND COMPUTING↗

Bright high z SnIa: A challenge for {lambda}CDM

It has recently been pointed out by Kowalski et. al.[Astrophys. J. 686, 749 (2008).] that there is 'an unexpected brightness of the SnIa data at z>1'. We quantify this statement by constructing a new statistic which is applicable directly on the type Ia supernova (SnIa) distance moduli. This statistic is designed to pick up systematic brightness trends of SnIa data points with respect to a best fit cosmological model at high redshifts. It is based on binning the normalized differences between the SnIa distance moduli and the corresponding best fit values in the context of a specific cosmological model (e.g. {lambda}CDM). These differences are normalized by the standard errors of the observed distance moduli. We then focus on the highest redshift bin and extend its size toward lower redshifts until the binned normalized difference (BND) changes sign (crosses 0) at a redshift z{sub c} (bin size N{sub c}). The bin size N{sub c} of this crossing (the statistical variable) is then compared with the corresponding crossing bin size N{sub mc} for Monte Carlo data realizations based on the best fit model. We find that the crossing bin size N{sub c} obtained from the Union08 and Gold06 data with respect to the best fit {lambda}CDM model is anomalously large compared to N{sub mc} of the corresponding Monte Carlo data sets obtained from the best fit {lambda}CDM in each case. In particular, only 2.2% of the Monte Carlo {lambda}CDM data sets are consistent with the Gold06 value of N{sub c} while the corresponding probability for the Union08 value of N{sub c} is 5.3%. Thus, according to this statistic, the probability that the high redshift brightness bias of the Union08 and Gold06 data sets is realized in the context of a (w{sub 0},w{sub 1})=(-1,0) model ({lambda}CDM cosmology) is less than 6%. The corresponding realization probability in the context of a (w{sub 0},w{sub 1})=(-1.4,2) model is more than 30% for both the Union08 and the Gold06 data sets indicating a much better consistency for this model with respect to the BND statistic.

79 ASTRONOMY AND ASTROPHYSICS↗

Impact of Domain Knowledge on the Property Prediction of Specialized Machine Learning Models

Developing transferable machine learning models is trending in data-driven materials research. However, how to apply such models to a specific research domain remains unclear. Here, in this work, we choose high-entropy materials as a platform with a specialized data set containing 145,323 DFT-relaxed materials. This data set is used to explore the role of domain-specific knowledge in training effective models. Our tests with three representative graph neural network architectures indicate the model complexity has much smaller influence on performance than the data itself. Specifically, the consideration of low-energy atomic ordering, structures with diverse elemental coverage, and high-order interactions significantly influences the model performance. We also find that domain knowledge-driven sampling can greatly enhance unsupervised learning techniques. This research highlights that developing specialized data sets is more beneficial than further complicating deep learning architectures. Additionally, physics-inspired sampling algorithms are crucially needed for better machine learning models for a specific materials research domain.

36 MATERIALS SCIENCE↗

A TEMPERATURE AND ABUNDANCE RETRIEVAL METHOD FOR EXOPLANET ATMOSPHERES

We present a new method to retrieve molecular abundances and temperature profiles from exoplanet atmosphere photometry and spectroscopy. We run millions of one-dimensional (1D) atmosphere models in order to cover the large range of allowed parameter space. In order to run such a large number of models, we have developed a parametric pressure-temperature (P-T) profile coupled with line-by-line radiative transfer, hydrostatic equilibrium, and energy balance, along with prescriptions for non-equilibrium molecular composition and energy redistribution. The major difference from traditional 1D radiative transfer models is the parametric P-T profile, which essentially means adopting energy balance only at the top of the atmosphere and not in each layer. We see the parametric P-T model as a parallel approach to the traditional exoplanet atmosphere models that rely on several free parameters to encompass unknown absorbers and energy redistribution. The parametric P-T profile captures the basic physical features of temperature structures in planetary atmospheres (including temperature inversions), and fits a wide range of published P-T profiles, including those of solar system planets. We apply our temperature and abundance retrieval method to the atmospheres of two transiting exoplanets, HD 189733b and HD 209458b, which have the best Spitzer and Hubble Space Telescope data available. For HD 189733b, we find efficient day-night redistribution of energy in the atmosphere, and molecular abundance constraints confirming the presence of H{sub 2}O, CO, CH{sub 4}, and CO{sub 2}. For HD 209458b, we confirm and constrain the dayside thermal inversion in an average 1D temperature profile. We also report independent detections of H{sub 2}O, CO, CH{sub 4}, and CO{sub 2} on the dayside of HD 209458b, based on six-channel Spitzer photometry. We report constraints for HD 189733b due to individual data sets separately; a few key observations are variable in different data sets at similar wavelengths. Moreover, a noticeably strong CO{sub 2} absorption in one data set is significantly weaker in another. We must, therefore, acknowledge the strong possibility that the atmosphere is variable, both in its energy redistribution state and in the chemical abundances.

79 ASTRONOMY AND ASTROPHYSICS↗

Model-independent plotting of the cosmological scale factor as a function of lookback time

In this work we describe a model-independent method of developing a plot of scale factor a(t) versus lookback time t{sub L} from the usual Hubble diagram of modulus data against redshift. This is the first plot of this type. We follow the model-independent methodology of Daly and Djorgovski used for their radio-galaxy data. Once the a(t)data plot is completed, any model can be applied and will display as described in the standard literature. We then compile an extensive data set to z = 1.8 by combining Type Ia supernovae (SNe Ia) data from SNLS3 of Conley et al., high-z SNe data of Riess et al., and radio-galaxy data of Daly and Djorgovski to validate the new plot. We first display these data on a standard Hubble diagram to confirm the best fit for ΛCDM cosmology, and thus validate the joined data set. The scale factor plot is then developed from the data and the ΛCDM model is again displayed from a least-squares fit. The fit parameters are in agreement with the Hubble diagram fit confirming the validity of the new plot. Of special interest is the transition time of the universe, which in the scale factor plot will appear as an inflection point in the data set. Noise is more visible in this presentation, which is particularly sensitive to inflection points of any model displayed in the plot, unlike on a modulus-z diagram, where there are no inflection points and the transition-z is not at all obvious by inspection. We obtain a lower limit of z ≥ 0.6. It is evident from this presentation that there is a dearth of SNe data in the range z = 1-2, exactly the range necessary to confirm a ΛCDM transition-z around z = 0.76. We then compare a 'toy model' wherein dark matter is represented as a perfect fluid with an equation of state p = –(1/3) ρ to demonstrate the plot sensitivity to model choice. Its density varies as 1/t {sup 2} and it enters the Friedmann equations as Ω{sub dark}/t {sup 2}, replacing only the Ω{sub dark}/a {sup 3} term. The toy model is a close match to ΛCDM, but separates from it on the scale factor plot for similar ΛCDM density parameters. It is described in the Appendix. A more complete transition time analysis will be presented in a future paper.

79 ASTRONOMY AND ASTROPHYSICS↗

A revised partiality model and post-refinement algorithm for X-ray free-electron laser data

An updated partiality model and post-refinement algorithm for XFEL snapshot diffraction data is presented and confirmed by observing anomalous density for S atoms at an X-ray wavelength of 1.3 Å. Research towards using X-ray free-electron laser (XFEL) data to solve structures using experimental phasing methods such as sulfur single-wavelength anomalous dispersion (SAD) has been hampered by shortcomings in the diffraction models for X-ray diffraction from FELs. Owing to errors in the orientation matrix and overly simple partiality models, researchers have required large numbers of images to converge to reliable estimates for the structure-factor amplitudes, which may not be feasible for all biological systems. Here, data for cytoplasmic polyhedrosis virus type 17 (CPV17) collected at 1.3 Å wavelength at the Linac Coherent Light Source (LCLS) are revisited. A previously published definition of a partiality model for reflections illuminated by self-amplified spontaneous emission (SASE) pulses is built upon, which defines a fraction between 0 and 1 based on the intersection of a reflection with a spread of Ewald spheres modelled by a super-Gaussian wavelength distribution in the X-ray beam. A method of post-refinement to refine the parameters of this model is suggested. This has generated a merged data set with an overall discrepancy (by calculating the R{sub split} value) of 3.15% to 1.46 Å resolution from a 7225-image data set. The atomic numbers of C, N and O atoms in the structure are distinguishable in the electron-density map. There are 13 S atoms within the 237 residues of CPV17, excluding the initial disordered methionine. These only possess 0.42 anomalous scattering electrons each at 1.3 Å wavelength, but the 12 that have single predominant positions are easily detectable in the anomalous difference Fourier map. It is hoped that these improvements will lead towards XFEL experimental phase determination and structure determination by sulfur SAD and will generally increase the utility of the method for difficult cases.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

ON THE RADIO AND OPTICAL LUMINOSITY EVOLUTION OF QUASARS

We calculate simultaneously the radio and optical luminosity evolutions of quasars, and the distribution in radio loudness R defined as the ratio of radio and optical luminosities, using a flux-limited data set containing 636 quasars with radio and optical fluxes from White et al. We first note that when dealing with multi-variate data it is imperative to first determine the true correlations among the variables, not those introduced by the observational selection effects, before obtaining the individual distributions of the variables. We use the methods developed by Efron and Petrosian which are designed to obtain unbiased correlations, distributions, and evolution with redshift from a data set truncated due to observational biases. It is found that the population of quasars exhibits strong positive correlation between the radio and optical luminosities. With this correlation, whether intrinsic or observationally induced accounted for, we find that there is a strong luminosity evolution with redshift in both wavebands, with significantly higher radio than optical evolution. We conclude that the luminosity evolution obtained by arbitrarily separating the sources into radio-loud (R > 10) and radio-quiet (R < 10) populations introduces significant biases that skew the result considerably. We also construct the local radio and optical luminosity functions and the density evolution. Finally, we consider the distribution of the radio-loudness parameter R obtained from careful treatment of the selection effects and luminosity evolutions with that obtained from the raw data without such considerations. We find a significant difference between the two distributions and no clear sign of bi-modality in the true distribution for the range of R values considered. Our results indicate therefore, somewhat surprisingly, that there is no critical switch in the efficiency of the production of disk outflows/jets between very radio-quiet and very radio-loud quasars, but rather a smooth transition. Also, this efficiency seems higher for the high-redshift and more luminous sources in the sample considered.

79 ASTRONOMY AND ASTROPHYSICS↗

Neural Network‐Based Methods for Ocean Surface Wave Measurement Using Submarine Distributed Acoustic Sensing (DAS)

Two new data-driven models for estimating ocean surface waves from distributed acoustic sensing (DAS) submarine cable strain rate are developed using supervised machine learning on a 10-day data set collected offshore of Oliktok Point, Alaska. The new models were trained on target data from seafloor pressure moorings at three sites spaced evenly along 27.1 km of cable and were benchmarked against an empirical transfer function method previously used to estimate waves from DAS. A model which uses convolutional neural networks to transform 2-km frequency-wavenumber strain spectra to seafloor pressure spectra outperforms the benchmark in wave height prediction (RMSE of 0.15 vs. 0.41 m) and period prediction (0.29 vs. 0.37 s) when evaluated on a held-out test data set. When applied to a DAS data set collected on the same cable 2 years prior, the CNN-based model maintained similar significant wave height performance (RMSE = 0.23 m) relative to available satellite altimetry data. A two-hidden-layer, fully connected neural network which transforms 1-D strain spectra to seafloor pressure spectra also outperforms the benchmark in wave height prediction (RMSE of 0.19 vs. 0.41 m), but does not generalize as well to the prior data. Regression-based machine learning is useful for estimating waves from DAS data when the pressure-strain relationship varies temporally and spatially across different wave conditions. Models can be applied to DAS data to measure waves with higher spatial resolution and longer temporal coverage than traditional methods, which often measure waves only at a single point.

Davis, Jacob R. [Univ. of Washington, Seattle, WA ↗

A machine-learning-driven data labeling pipeline for scientific analysis in MLExchange

This study introduces a novel labeling pipeline to accelerate the labeling process of scientific data sets by using artificial intelligence (AI)-guided tagging techniques. This pipeline includes a set of interconnected web-based graphical user interfaces (GUIs), where Data Clinic and MLCoach enable the preparation of machine learning (ML) models for data reduction and classification, respectively, while Label Maker is used for label assignment. Throughout this pipeline, data can be accessed through a direct connection to a file system or through Tiled for access through Hypertext Transfer Protocol (HTTP). Our experimental results present three use cases where this labeling pipeline has been instrumental for the study of large X-ray scattering data sets in the area of pattern recognition, the remote analysis of resonant soft X-ray scattering data and the fine-tuning process of foundation models. These use cases highlight the labeling capabilities of this pipeline, including the ability to label large data sets in a short period of time, to perform remote data analysis while minimizing data movement and to enhance the fine-tuning process of complex ML models with human involvement.

Chavez, Tanny (ORCID:0000000193172896)↗

SPRUCE Phospholipid Fatty Acid (PLFA) Abundances, August 2021-June 2022

This data set provides the results for phospholipid fatty acid analysis (PLFA) of peat samples collected from ambient plots in the SPRUCE experiment site, with experimental plot samples still being processed. The samples used to generate this data set were collected on 2021-08-23 and 2022-06-21. This data set includes abundances for groups of lipids indicating total biomass, fungi, Gram-positive bacteria, Gram-negative bacteria, actinomycetes, and anaerobic bacteria. This data set contains one file in comma separate (*.csv) format. Samples: On each sampling date, cores extending to 200 cm were collected and subsampled at 10 cm intervals from 0 and 100 cm and 25 cm intervals below 100 cm depth in Plot 7 and 21. During the lipid extraction procedure, samples were pooled into the following depth increments in order to obtain sufficient material to achieve adequate lipid yield: 0-10, 10-20, 20-30, 30-50, 50-100, and 100-150, and 150-200 cm.

actinomycetes↗