SEARCH · Search NASA
Results for “Datasets”
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.
United States CMM Insights Dataset
Geo-data science applications for critical mineral analysis, development acceleration, economic impact assessment, and project efficiency. Coverage: 10 years (2014-2023), 3 geographic levels (county, state, tract), 45 features Data Categories: - census: 30 features (e.g., asian population percentage, black population percentage, citizen voting age population percentage) - ejscreen: 3 features (e.g., P_DSLPM, P_PM25, P_PWDIS) - energyexpenditure: 12 features (e.g., Housing adjustment factor, Income adjustment factor, Monthly electricity cost)
Limited Linguistic Diversity in Embodied AI Datasets
Explore the source record for details and available documents.
Analysis of Remotely Sensed Datasets to Detect Changes in Waste Sites
This work evaluated various indirect remote techniques for detecting surficial changes and structural land deformation at the Hanford site. Given the partial collapse of the PUREX tunnel, and possible surface subsidence at other burial sites containing radioactive waste and contaminated equipment, timely alert and knowledge related to land surface elevation changes has emerged as a management priority. Findings of this study suggests that space borne RADAR and multispectral based-analysis would be of significant value in detecting changes in surficial and structural land features that may be otherwise undetectable or unnoticed from field observations alone. A continuous, and potentially semi-automated, monitoring program would add critical insight to recovering from, and predicting episodic type events, such as the tunnel collapse. Furthermore, an indirect approach ultimately saves resources and time, and avoids unnecessary risk associated with ground-based monitoring. Various options for supporting such an effort are described in the final section of this document.
Using ARM-SGP Multi-Sensor Datasets to Investigate Precipitation Characteristics and Vertical Variability (Final Technical Report)
The overarching science goal of this research project was to analyze Atmospheric Radiation Mission (ARM) observations to increase our understanding of raindrop breakup, coalescence, and evaporation processes in precipitating clouds in order to improve how these processes are represented in cloud resolving models.
Blind Denoising by Self-Supervised Neural Networks in Astronomical Datasets (Noise2Self4Astro)
the same part of the sky, with different shot, read-out, cosmic ray, and detector noise. Traditionally, in order to gain denoised model for astronomical objects, researchers take the average of multiple images of the same object or apply pre-processing filters such as Gaussian filter or Median filter, feeding them into “Galfit”, which is a software for astronomical object model fitting. Instead of that, we may train a network to predict one noisy image from the other based on the Noise2Self algorithm. The resulting output will be a semi-denoised image, which would later be put into galfit to complete the denoising process.
A Dark Matter Search Using the Final SuperCDMS Soudan Dataset and the Development of a Large-Format, Highly-Multiplexed, Athermal-Phonon-Mediated Particle Detector
Over the past eighty years, numerous complementary observations of our universe have indicated that our current description of physics is far from complete. The “or- dinary matter”, such as electrons, protons, photons and neutrons, that constitutes the bulk of all human physical experiences is actually only a minority (about 16%) of the total mass of the universe. The remaining 84% is very poorly understood, but has profound effects on the dynamics and evolution of our universe. Because it does not interact with light, and is not observable in telescopes on earth, this extra mass is usually referred to as “dark matter”. Although the dark matter is poorly understood, Weakly Interacting Massive Particles (WIMPs) are a well-motivated candidate that can be directly detected via a non-gravitational interaction with nor- mal matter, potentially allowing for direct terrestrial detection and characterization of this dark matter. This dissertation is focused on this direct WIMP detect ion and will be broken into two main parts. The first part focuses on the blinded analysis of roughly three years of data collected from March 2012 to November 2015 by the SuperCDMS Soudan experiment. SuperCDMS Soudan consists of an array of 15, 0.6-kg, cryogenic, Ge iZIP particle detectors situated in a decommissioned iron mine in remote northern Minnesota. This analysis is optimized to be sensitive to theoretical WIMP masses above 10 GeV/$c^2$ . This result set the strongest limits for WIMP–germanium-nucleus interac- tions for WIMP masses greater than 12 Gev/$c^2$ and can be found here [1]. The second part focuses on the development new kind of particle detector in the style of a SuperCDMS iZIP, designed to simplify fabrication and readout, improve phonon-based position reconstruction, and help to scale to larger target arrays. These detectors replace the TES-based phonon sensors of the iZIP with Microwave Kinetic Inductance Detectors (MKIDs).
Diagnosis of convective organization and cold pools using ARM datasets and evaluation of a unified convection parameterization (UNICON)
Tropical thunderstorms often cluster together. Studies have suggested that the degree to which the tropical thunderstorms are clustered impacts Earth's energy balance and water cycle, as well as extreme precipitation events. However, the processes controlling the spatial distribution of the thunderstorms are poorly understood and are not properly represented in most computer models for weather and climate prediction. Under the goals of better understanding how convection organizes at the mesoscale and advancing the representation of mesoscale convective organization in global climate models, we i) objectively quantified the degrees of convective organization and diagnosed cold pool processes using ARM field campaign observations, ii) examined the organization processes in storm-resolving model simulations, and iii) evaluated the impacts of mesoscale convective organization in global model simulations. The project yielded a firm reference against which the global model representation of mesoscale convective organization and cold pools can be evaluated against and shed new light into the role of parameterized convective organization in the global model simulation of the basic state and variability. Our results revealed two distinct phases of convective clustering during the two-day rain episodes (Cheng et al. 2018) and a new mechanism through which vertical wind shear in the low-troposphere can aid convective organization over tropical oceans (Cheng et al. 2020). It was demonstrated that the interactive representation of cold pools and mesoscale convective organization is key for global models to successfully simulate both the mean state and intraseasonal variability in the tropics (Ahn et al. 2019; 2020).
Process Discovery through Assimilation of Complex Biogeochemical Datasets
This white paper addresses two of three focal areas identified in the white paper call: 1) Biogeochemical data acquisition and assimilation enabled by machine learning and 3) Insight gleaned from complex data using AI. We focus on AI application to complex biogeochemistry (BGC) data (e.g. laboratory experimental data, field manipulation data, literature data), which is an untapped source of information for improving Earth System Predictability (ESP).
AI4ESP: Process Discovery through Assimilation of Complex Geochemical Datasets
This white paper addresses two of three focal areas identified in the white paper call: 1) Biogeochemical data acquisition and assimilation enabled by machine learning and 3) Insight gleaned from complex data using AI. We focus on AI application to complex biogeochemistry (BGC) data (e.g. laboratory experimental data, field manipulation data, literature data), which is an untapped source of information for improving Earth System Predictability (ESP).
Transient Deformation in Additively Manufactured 316L Stainless Steel Lattices Characterized with in-situ X-ray Phase Contrast Imaging: The Complete Dataset for Three Geometrical Lattices
Metallic lattice structures are being considered for shock mitigation applications due to their superior mechanical properties, energy absorption capability and lightweight characteristics inherent of the additive manufacturing process. In this study, shock compression experiments coupled to x-ray phase contrast imaging (PCI) were conducted on 316L stainless steel lattices. Meso-scale simulations incorporating the as-built lattice structure characterized by computed tomography were used to simulate PCI radiographs in CTH for direct comparison to experimental data. The methodology presented here offers robust validation for constitutive properties to further our understanding of lattice compaction at application-relevant strain rates.
LANL-UW Collaboration on LC-QTOF Datasets (Annual Report)
Prior to the application of non-targeted chemometric techniques (i.e., Fisher ratio analysis), pre-processing steps must be carried out on the data collected using liquid chromatography-quadrupole time of flight (LC-QTOF) mass spectrometry. Data compression is the most important pre-processing step due to the vast number of high-resolution mass channels (m/z; 0.0001 Da) collected. Our first approach was to bin the mass spectral dimension to 0.01 Da; however, this approach was computationally expensive and did not preserve the high-resolution data.
Fusing Diverse Datasets with Machine Learning to Inform System State Conditions: An Application for Photovoltaic Plants.
Abstract not provided.
LLNL International Working Group on Gamma Spectrometry - Preparation of IDB Dataset
The spectral information and associated metadata, including source information, detector setup, and shielding configuration is described in human readable, comma-separated text files (csv). Five csv files are associated with each material type and include the following information (where [Isotope] is U, Pu, and MOX): [Isotope]_measurement_metadata - Metadata of the measurement configuration, including material composition at time of acquisition (decay corrected from certificate date). Data that is unique to a measurement configuration e.g. detector geometry, source-detector distance, analyzer and other electronics used, attenuating materials are provided in this file. [Isotope]_spectrum_metadata - The information that is unique to the spectrum. This includes detector specifications (e.g. date of acquisition, number of channels, gain (keV/channel), energy range, real and live time, full width half-maximum (FWHM) of the 186 or 208 peak for uranium or plutonium measurements, respectively. [Isotope]_source_metadata - The source material used for the measurement, this includes the ratio and uncertainties of isotopes present and processing information of the material. In addition, the certificate information for the source material is listed. [Isotope]_spectrum_counts_metadata - The counts per channel for each spectrum. This data is linked via the UID.spectrum index to the measurement_metadata file. [Isotope]_spectrum_checksum_metadata The MD5 checksum for the original spectrum file. This data is linked via the UID.spectrum index to the measurement_metadata file. This information is provided to prevent duplication of entries in the database.
Experimental datasets on soot formation in turbulent non-premixed ethylene and JP-8 jet flames.
Abstract not provided.
Marine Shallow Cloud Adjustments to the Presence of Shortwave-Absorbing Aerosols: Advancing Understanding Through a Combined Analysis of Lasic Datasets and Process Modeling (Final Report)
This is a final report for this award, whose purpose was to support the analysis of data emanating from the LASIC (Layered Atlantic Smoke Interaction with Clouds) campaign. (Layered Atlantic Smoke Interactions with Clouds) is a strategy to improve our understanding of aged carbonaceous aerosol, its seasonal evolution, and the mechanisms by which clouds adjust to the presence of the aerosol. The observational strategy centers on deploying the AMF1 cloud, aerosol, and atmospheric profiling instrumentation to Ascension Island, located within the trade-wind shallow cumulus regime (14.50W, 80S) 3000 km offshore of continental Africa. The location is within the latitude zone of the maximum outflow of aerosol, with the deepening boundary layer known to entrain free-tropospheric smoke. The primary activities for LASIC are: 1) to improve current knowledge on aged biomass burning aerosol and its radiative properties as a function of the seasonal cycle; 2) to use surface-based remote sensing to sensitively interrogate the atmosphere for the relative vertical location of aerosol and clouds; 3) to improve our understanding of the cloud adjustments to the presence of shortwave-absorbing aerosol within the vertical column, both through aerosol-radiation and through aerosol-cloud interactions; 4) to aid low cloud parameterization efforts for climate models. This award serves to further these activities and includes a sub-award to Dr. Pablo Saide at UCLA for activities 3) and 4).
Safeguards-Informed Hybrid Imagery Dataset [Poster]
Deep Learning computer vision models require many thousands of properly labelled images for training, which is especially challenging for safeguards and nonproliferation, given that safeguards-relevant images are typically rare due to the sensitivity and limited availability of the technologies. Creating relevant images through real-world staging is costly and limiting in scope. Expert-labeling is expensive, time consuming, and error prone. We aim to develop a data set of both realworld and synthetic images that are relevant to the nuclear safeguards domain that can be used to support multiple data science research questions. In the process of developing this data, we aim to develop a novel workflow to validate synthetic images using machine learning explainability methods, testing among multiple computer vision algorithms, and iterative synthetic data rendering. We will deliver one million images – both real-world and synthetically rendered – of two types uranium storage and transportation containers with labelled ground truth and associated adversarial examples.