Search NASA⌕ Search

SEARCH · Search NASA

Results for “data generation”

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 127 records · Page 7
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Enhanced Oblique Decision Tree Enabled Policy Extraction for Deep Reinforcement Learning in Power System Emergency Control

Deep reinforcement learning (DRL) algorithms have successfully solved many challenging problems in various power system control scenarios. However, their decision-making process is usually regarded as black-boxes. Furthermore, how DRL models interact with human intelligence remains an open problem. Thus, this paper proposes a policy extraction framework to extract a complex DRL model into an explainable policy. This framework includes three parts: 1) DRL training and data generation. We train an agent for a specific control task and generate data, which contains the control policy of the agent. 2) Policy extraction. We propose an information gain rate based weighted oblique decision tree (IGR-WODT) for DRL policy extraction. 3) Policy evaluation. We define three metrics to evaluate the performance of the proposed approach. A case study for the under-voltage load shedding problem shows that the IGR-WODT presents a performance enhancement compared with DRL, weighted oblique decision tree, and univariate decision tree. The proposed policy extraction method could provide an intuitive explanation of the neural network decision-making process to the dispatchers when making final decisions on power grid operation. Also, the resulted rule-based controller could replace the deep neural network-based controller in many field edge devices with limited computing resources, providing comparable performance.

deep reinforcement learning↗
DOE OSTI2016

Landrush: Rethinking In-Situ Analysis for GPGPU Workflows

In-situ analysis on the output data of scientific simulations has been made necessary by ever-growing output data volumes and increasing costs of data movement as supercomputing is moving towards exascale. With hardware accelerators like GPUs becoming increasingly common in high end machines, new opportunities arise to co-locate scientific simulations and online analysis performed on the scientific data generated by the simulations. However, the asynchronous nature of GPGPU programming models and the limited context-switching capabilities on the GPU pose challenges to co-locating the scientific simulation and analysis on the same GPU. This paper dives deeper into these challenges to understand how best to co-locate analysis with scientific simulations on the GPUs in HPC clusters. Specifically, our 'Landrush' approach to GPU sharing proposes a solution that utilizes idle cycles on the GPU to provide an improved time-to-answer, that is, the total time to run the scientific simulation and analysis of the generated data. Landrush is demonstrated with experimental results obtained from leadership high-end applications on ORNL's Titan supercomputer, which show that (i) GPU-based scientific simulations have varying degrees of idle cycles to afford useful analysis task co-location, and (ii) the inability to context switch on the GPU at instruction granularity can be overcome by careful control of the analysis kernel launches and software-controlled early completion of analysis kernel executions. Results show that Landrush is superior in terms of time-to-answer compared to serially running simulations followed by analysis or by relying on the GPU driver and hardwired thread dispatcher to run analysis concurrently on a single GPU.

Computer Science↗
DOE OSTI2023

Machine Learning-Based Regression Models for Ironmaking Blast Furnace Automation

Computational fluid dynamics (CFD)-based simulation has been the traditional way to model complex industrial systems and processes. One very large and complex industrial system that has benefited from CFD-based simulations is the steel blast furnace system. The problem with the CFD-based simulation approach is that it tends to be very slow for generating data. The CFD-only approach may not be fast enough for use in real-time decisionmaking. To address this issue, in this work, the authors propose the use of machine learning techniques to train and test models based on data generated via CFD simulation. Regression models based on neural networks are compared with tree-boosting models. In particular, several areas (tuyere, raceway, and shaft) of the blast furnace are modeled using these approaches. The results of the model training and testing are presented and discussed. The obtained R 2 metrics are, in general, very high. The results appear promising and may help to improve the efficiency of operator and process engineer decisionmaking when running a blast furnace.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1987

Self-Identifying Emergency Radio Beacons

Rescue teams aided by knowledge of vehicle in distress. Similar to conventional emergency transmitters except contains additional timing and modulating circuits. Additions to standard emergency transmitter enable transmitter to send rescuers identifying signal in addition to conventional distress signal created by sweep generator. Data generator contains identifying code.

Friedman, Morton L.↗
NASA NTRS1990

High rate science data handling on Space Station Freedom

A study by NASA's User Information System Working Group for Space Station Freedom (SSF) has determined that the proposed onboard Data Management System, as initially configured, will be incapable of handling the data-generation rates typical of numerous scientific sensor payloads; many of these generate data at rates in excess of 10 Mbps, and there are at least four cases of rates in excess of 300 Mbps. The SSF Working Group has accordingly suggested an alternative conceptual architecture based on technology expected to achieve space-qualified status by 1995. The architecture encompasses recorders with rapid data-ingest capabilities and massive storage capabilities, optical delay lines allowing the recording of only the phenomena of interest, and data flow-compressing image processors.

Handley, Thomas H., Jr.↗
NASA NTRS1993

Eddy mixing coefficient values derived from simulations with the NASA Ames Mars GCM

Values of eddy mixing coefficients, especially that for vertical mixing, are of particular importance for various photochemical models of the Mars atmosphere, which are either globally or zonally averaged. These models represent atmospheric transport processes in terms of eddy diffusion. While this is not appropriate for the advective transport by the mean meridional circulation (in general, a part of the eddy diffusion tensor actually does represent advection--a 'correction' to the Eulerian-mean circulation), it can be applied to simulate both the small- and larger-scale eddy transports of an atmospheric constituent. An effort is underway to estimate values of the eddy mixing coefficients for the Mars atmosphere using circulation data generated with the NASA Ames Mars GCM. This model simulates the three-dimensional winds in the atmosphere, and these are then used as inputs for a tracer transport model. The latter model has previously been used in interactive dust transport experiments with the Ames GCM. Idealized tracer transport experiments, with a conservative or nearly conservative hypothetical tracer, are performed to generate data from which the eddy diffusion coefficients can be estimated. Carrying out matched pairs of tracer experiments, using two very different initial states, permits all four components of the diffusion tensor to be determined. A large number of GCM experiments have been conducted, spanning a range of seasons and atmospheric dust loading, allowing the eddy mixing coefficients to be estimated for a variety of atmospheric conditions.

Barnes, J. R.↗
NASA NTRS1994

Continued reduction and analysis of data from the Dynamics Explorer Plasma Wave Instrument

The plasma wave instrument on the Dynamics Explorer 1 spacecraft provided measurements of the electric and magnetic components of plasma waves in the Earth's magnetosphere. Four receiver systems processed signals from five antennas. Sixty-seven theses, scientific papers and reports were prepared from the data generated. Data processing activities and techniques used to analyze the data are described and highlights of discoveries made and research undertaken are tabulated.

Gurnett, Donald A.↗
NASA NTRS2020

CyberGAN: Generating High-fidelity Cybersecurity Data With Generative Adversarial Networks

Machine learning for cyber defense offers the promise of detecting adversarial activity against the ground data systems managing critical space assets. A fundamental challenge facing machine learning research in cybersecurity is the lack of high-fidelity, shareable datasets for robust evaluation and testing of machine learning-based solutions. High-fidelity, real-world datasets are necessary for reliable benchmarking of nominal system behavior and malicious activity. Unfortunately, such realistic datasets of both nominal and adversarial activity are rarely shared publicly by data owners due to security and privacy concerns. Besides, the available adversarial data is sparse, which makes training models on malicious activity much harder. This situation has impeded and continues to impede the research and successful adoption of machine learning methods for cyber defense. Researchers have dealt with this problem by generating data within a low-fidelity lab environment, using classified and thus unshareable datasets, or downloading low-fidelity public datasets made available by others. We propose an innovative solution to the problem by employing machine learning methods to generate high-fidelity data. Specifically, we propose the use of Generative Adversarial Networks (GANs) to generate high-fidelity data for cybersecurity purposes. GANs have found successful image processing and natural language applications, but have not yet been investigated for cyber data generation. Our proposed approach first involves training the `discriminator' network of the GAN with a sample of real-world data consisting of malicious and nominal samples. We then use the `generator' network to generate new high-fidelity data samples consisting of an appropriate mix of malicious and nominal activity. We demonstrate applications of our architecture by generating high-fidelity cybersecurity data containing both malicious and nominal samples. We thoroughly evaluate the fidelity of our generated data using heuristics and evaluate its usefulness for machine learning applications using three different datasets. Overall, our approach results in high-fidelity, shareable datasets.

Zhang, Yuening↗
NASA NTRS1989

Optical archiving for scientific data

Write-once-read-many (WORM) optical disk technology is now practical for use as (1) a near online storage media, (2) long-term storage in a passive repository, and (3) under some circumstances, transporting from and to data generators and data users at remote sites. Optical technology provides several significant advantages over magnetic tapes as a data storage media because it is more compact, provides random access, has a longer lifetime, at least 30 years, and is less sensitive to environmental factors such as temperature and humidity. Two major archival projects using optical technology underway are discussed: the Coastal Zone Color Scanner data and the Dynamics Explorer data. Both projects have used the software for optical archive and retrieval (SOAR) to transfer data in a volume format that can be read transparently by VAX/VMS. Through the use of this software technique, (1) data generators and data users do not have to rewrite extant software and (2) the VAX/VMS Files-11 volume format is in wide use, so it can serve as an interim standard and will be accessible in the long-term.

Lowrey, Barbara E.↗
DOE OSTI2024

Data Efficiency Assessment of Generative Adversarial Networks for Critical Heat Flux Synthetic Data Generation

This study investigates the application of generative artificial intelligence techniques, particularly conditional generative adversarial networks (cGAN), in real-world engineering contexts, with a specific focus on synthetic data generation for critical heat flux (CHF). Utilizing a dataset comprising more than 20,000 real experimental CHF measurements, we conduct a series of experiments to examine cGAN’s behavior. These experiments encompass varying sizes of the training dataset, training cGAN on data from diverse experimental sources to generate new data on unseen experimental setups, and assessing the impact of excluding various input features on cGAN’s data generation accuracy. Our findings underscore the pronounced data dependency of cGAN for reliable performance, with decreased efficacy observed with smaller training dataset sizes. Notably, cGAN exhibits varying performance when trained on data from different experiments, with superior predictive capabilities observed for certain experiment sources compared to others. For instance, when cGAN was trained on data from Smolin et al.’s experiments or Zenkevich et al., it exhibited relatively good performance in generating the data from Becker et al., Kirillov et al., and Alekseev et al. experiments. In contrast, when trained with Alekseev et al.’s data and tasked with generating other experimental setups, cGAN showed notably poor performance. In both scenarios, cGAN’s performance was inferior compared to training on samples from all experiments concurrently. A feature importance analysis highlights the significant influence of parameters such as mass flux and heated length on accurate CHF generation, while other parameters like diameter and pressure have less impact. Inlet temperature is identified as a moderating factor by cGAN.

22 - GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2012

PHOTOMETRIC REDSHIFTS AND QUASAR PROBABILITIES FROM A SINGLE, DATA-DRIVEN GENERATIVE MODEL

We describe a technique for simultaneously classifying and estimating the redshift of quasars. It can separate quasars from stars in arbitrary redshift ranges, estimate full posterior distribution functions for the redshift, and naturally incorporate flux uncertainties, missing data, and multi-wavelength photometry. We build models of quasars in flux-redshift space by applying the extreme deconvolution technique to estimate the underlying density. By integrating this density over redshift, one can obtain quasar flux densities in different redshift ranges. This approach allows for efficient, consistent, and fast classification and photometric redshift estimation. This is achieved by combining the speed obtained by choosing simple analytical forms as the basis of our density model with the flexibility of non-parametric models through the use of many simple components with many parameters. We show that this technique is competitive with the best photometric quasar classification techniques-which are limited to fixed, broad redshift ranges and high signal-to-noise ratio data-and with the best photometric redshift techniques when applied to broadband optical data. We demonstrate that the inclusion of UV and NIR data significantly improves photometric quasar-star separation and essentially resolves all of the redshift degeneracies for quasars inherent to the ugriz filter system, even when included data have a low signal-to-noise ratio. For quasars spectroscopically confirmed by the SDSS 84% and 97% of the objects with Galaxy Evolution Explorer UV and UKIDSS NIR data have photometric redshifts within 0.1 and 0.3, respectively, of the spectroscopic redshift; this amounts to about a factor of three improvement over ugriz-only photometric redshifts. Our code to calculate quasar probabilities and redshift probability distributions is publicly available.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2018

Mapping Benthic Habitat Conditions and Seafloor Deposits using Sediment Profile Imaging and a Semiautomated Image Processing System

Objectives/Scope: Sediment profile imaging (SPI) technology characterizes in situ physical, geochemical, and biological seafloor features. SPI was required by Mexico’s Agency for Safety, Energy, and Environment in 2017 for environmental baseline surveys (EBS) of oil lease blocks in the southern Gulf of Mexico. Because of its ability to provide information on benthic community health and the distribution of very thin (down to the centimeter scale) layers of deposited materials, SPI technology is highly effective for mapping drill cuttings or drilling muds released around wellheads during exploration or production and documenting their ecological impacts. SPI technology has not been widely used in the oil and gas industry for EBS or other monitoring activities. A primary objective of this work is to improve the transparency and consistency of the SPI data generation and data management process. Methods, Procedures, Process: The SPI camera works like an inverted periscope and obtains an undisturbed 21x15-cm cross-sectional image of the upper sediment column. The camera is internally powered and can be deployed rapidly from a standard winch in depths to 4,000 m. Many stations can be sampled in a single day by “pogo-sticking” across a survey area. Sediment grain size, penetration depth, surface boundary roughness, natural and anthropogenic depositional layers, depth of the oxidized surface sediment layer, maximum biogenic mixing depth, and infaunal successional stage can be directly measured at sea or immediately following the cruise. Final SPI data sets can be provided within a few weeks of the survey. Results, Observations, Conclusions: Details on the features measured in SPI images and the underlying interpretive paradigms are presented. To standardize the SPI data generation process, Integral Consulting Inc. has developed 1) a semiautomated image analysis platform, and 2) a SPI data-specific database architecture that allows both numerical and non-numerical metrics to be incorporated into a standard database structure. An integrated, software-based SPI analysis platform has been developed that imports image files and metadata and provides a graphical user interface. The software automatically stores the data, which can then be reviewed for quality assurance, plotted, statistically analyzed, and mapped or exported to other platforms (e.g., Esri ArcGIS©) for further evaluation. Image processing algorithms have been developed using a combination of open-source and commercially available software packages (e.g., MATLAB® and OpenCV) to automatically quantify key parameters. Novel/Additive Information: SPI technology’s underutilization in the oil and gas industry may be in part due to a lack of standardization in the measurement of basic features in SPI images. A primary objective of this work is to develop a streamlined, standardized, and transparent process for generating and managing SPI data.

02 PETROLEUM↗
DOE OSTI2015

Data from: Generation of transcript assemblies and identification of single nucleotide polymorphisms from seven lowland and upland cultivars of switchgrass

Switchgrass (Panicum virgatum) is a North American perennial prairie species that has been used as a rangeland and forage crop and has recently been targeted as a potential biofuel feedstock species. Switchgrass, which occurs as tetraploid and octoploid forms, is classified into lowland or upland ecotypes that differ in growth phenotypes and adaptation to distinct habitats. Using RNA-sequencing reads derived from crown, young shoot and leaf tissues, we generated sequence data from seven switchgrass cultivars, three lowland and four upland, to enable comparative analyses between switchgrass cultivars and to identify single nucleotide polymorphisms (SNPs) for use in breeding and genetic analysis. We also generated individual transcript assemblies for each of the cultivars. Transcript data indicate that subgenomes of octoploid switchgrass are not substantially different from subgenomes of tetraploids as expected for an autopolyploid origin of switchgrass octoploids. Using RNA-sequencing reads aligned to the switchgrass Release 0 AP13 reference genome, we identified 1,305,976 high-confidence SNPs. Of these SNPs, 438,464 were unique to lowland cultivars, but only 12,002 were found in all lowlands. Conversely, 723,678 SNPs were unique to upland cultivars with only 34,665 observed in all uplands. Comparison of our high-confidence transcriptome-derived SNPs with SNPs previously identified in a genotyping-by-sequencing study of an association panel revealed limited overlap between the two methods, highlighting the utility of transcriptome-based SNP discovery in augmenting genome diversity polymorphism datasets. The transcript and SNP data described here provide a useful resource for switchgrass gene annotation and marker-based analyses of the switchgrass genome.

09 BIOMASS FUELS↗
DOE OSTI2020

Methodology for Generating Covariance Data of Thermal Neutron Scattering Cross Sections

This paper details and implements a framework for evaluating thermal neutron scattering cross sections that provide S(α,β) data and covariance data for hydrogen in light water. This methodology involves perturbing model parameters of molecular dynamics potentials and fitting the simulation results to experimental data. The framework is general and can be applied to any material or simulation method. The fit is made using the Unified Monte Carlo method to experimentally measure double-differential scattering cross sections of light water at the Spallation Neutron Source at Oak Ridge National Laboratory. Mean values and covariance data were generated for model parameters, phonon density of states, double-differential cross sections, and total scattering cross sections. These posterior parameter values were very similar to their prior values with a maximum relative error of 0.54%. This falls within in the Unified Monte Carlo–calculated uncertainties on the order of 2.7%. Additionally, posterior double-differential cross sections agree favorably with ENDF/B-VIII.0 cross sections. The new thermal scattering law was tested by comparing it against benchmarks from the International Criticality Safety Benchmark Evaluation Project Handbook, which showed a slight improvement over the ENDF/B-VIII.0 library. Additionally, the covariance matrix of the phonon density of states was validated to confirm that the spread of k eff from the density of states used to generate the covariance matrix was similar to the spread of k eff from the density of states of the sampled covariance matrix.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
NASA NTRS2022

Unique Observational Constraints on the Seasonal and Longitudinal Variability of the Earth’s Planetary Albedo and Cloud Distribution Inferred From EPIC Measurements

Thorough comparison to observations is key to developing a credible climate model forecasting capability. Deep Space Climate Observatory (DSCOVR) measurements of Earth’s reflected solar and emitted thermal radiation provide a unique observational perspective that permits a more reliable model/data comparison than is possible with the otherwise available satellite data. The uniqueness is in the DSCOVR satellite’s viewing geometry, which enables continuous viewing of the Earth’s sunlit hemisphere from its Lissajous orbit around the Lagrangian L1 point. The key instrument is the Earth Polychromatic Imaging Camera (EPIC), which views the Earth’s sunlit hemisphere with 1024-by-1024-pixel imagery in 10 narrow spectral bands from 317 to 780 nm, acquiring up to 22 high spatial resolution images per day. The additional feature is that the frequency of EPIC image acquisition is nearly identical to that of the climate GCM data generation scheme where climate data for the entire globe are ‘instantaneously’ calculated at 1-h radiation time-step intervals. Implementation of the SHS (Sunlit Hemisphere Sampling) EPIC-view geometry for the in-line GCM output data sampling establishes a precise self-consistency in the space-time data sampling between EPIC observational and GCM output data generation and sampling. The remaining problem is that the GCM generated data are radiative fluxes, while the EPIC measurements are backscatter-dependent radiances. Radiance to flux conversion is a complex problem with no simple way to convert GCM radiative fluxes into spectral radiances. The more expedient approach is to convert the EPIC spectral radiances into broadband radiances by MODIS/CERES-based regression relationships and then into solar radiative fluxes using the CERES angular distribution models. Averaging over the sunlit hemisphere suppresses the meteorological weather noise, but preserves the intra-seasonal larger scale variability. Longitudinal slicing by the Earth’s rotation permits a self-consistent model/data comparison of the longitudinal model/data differences in the variability of the reflected solar radiation. Ancillary EPIC Composite data provide additional cloud property information for climate model diagnostics. Comparison of EPIC-derived seasonal and longitudinal variability of the Earth’s planetary albedo with the GISS ModelE2 results shows systematic overestimate of cloud reflectivity over the Pacific Ocean with corresponding underestimates over continental land areas.

DSCOVR↗
NASA NTRS1973

Sampling command generator corrects for noise and dropouts in recorded data

Generator measures period between zero crossings of reference signal and accepts as correct timing points only those zero crossings which occur acceptably close to nominal time predicted from last accepted command. Unidirectional crossover points are used exclusively so errors from analog nonsymmetry of crossover detector are avoided.

Anderson, T. O.↗