Search NASA⌕ Search

SEARCH · Search NASA

Results for “deep”

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 91 records · Page 5
All sourcesNASA NTRSDOE OSTI
DOE OSTI2017

Deep Full-sky Coadds from Three Years of WISE and NEOWISE Observations

We have reprocessed over 100 terabytes of single-exposure Wide-field Infrared Survey Explorer ( WISE )/NEOWISE images to create the deepest ever full-sky maps at 3–5 microns. We include all publicly available W1 and W2 imaging—a total of ∼8 million exposures in each band—from ∼37 months of observations spanning 2010 January to 2015 December. Our coadds preserve the native WISE resolution and typically incorporate ∼3× more input frames than those of the AllWISE Atlas stacks. Our coadds are designed to enable deep forced photometry, in particular for the Dark Energy Camera Legacy Survey (DECaLS) and Mayall z-Band Legacy Survey (MzLS), both of which are being used to select targets for the Dark Energy Spectroscopic Instrument. We describe newly introduced processing steps aimed at leveraging added redundancy to remove artifacts, with the intent of facilitating uniform target selection and searches for rare/exotic objects (e.g., high-redshift quasars and distant galaxy clusters). Forced photometry depths achieved with these coadds extend 0.56 (0.46) magnitudes deeper in W1 (W2) than is possible with only pre-hibernation WISE imaging.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS1982

The Deep Space Network. An instrument for radio navigation of deep space probes

The Deep Space Network (DSN) network configurations used to generate the navigation observables and the basic process of deep space spacecraft navigation, from data generation through flight path determination and correction are described. Special emphasis is placed on the DSN Systems which generate the navigation data: the DSN Tracking and VLBI Systems. In addition, auxiliary navigational support functions are described.

Renzetti, N. A.↗
NASA NTRS1983

The Deep Space Network: A Radio Communications Instrument for Deep Space Exploration

The primary purpose of the Deep Space Network (DSN) is to serve as a communications instrument for deep space exploration, providing communications between the spacecraft and the ground facilities. The uplink communications channel provides instructions or commands to the spacecraft. The downlink communications channel provides command verification and spacecraft engineering and science instrument payload data.

Renzetti, N. A.↗
NASA NTRS1991

Interference prediction for deep space spacecraft from ground stations - Interference to the Galileo deep space spacecraft from the GOMS geostationary network

A technique for predicting the interference power spectral density from a ground station to a spacecraft in deep space is described. The interference prediction throughout the complete mission is simulated, and distance between the ground station and the deep space spacecraft is calculated. A power spectral density from the GOMS satellite network to the Galileo spacecraft receiver is computed using the thermal noise density level at the Galileo spacecraft as the interference criterion. Data obtained indicate that all three GOMS stations cause at least 30 minutes of interference during Galileo earth flyby number one, and two GOMS ground stations cause at least four minutes of interference during Galileo earth flyby number.

Bishop, Dennis F.↗
NASA NTRS2012

Preliminary Results of a U.S. Deep South Warm Season Deep Convective Initiation Modeling Experiment using NASA SPoRT Initialization Datasets for Operational National Weather Service Local Model Runs

The initiation of deep convection during the warm season is a forecast challenge in the relative high instability and low wind shear environment of the U.S. Deep South. Despite improved knowledge of the character of well known mesoscale features such as local sea-, bay- and land-breezes, observations show the evolution of these features fall well short in fully describing the location of first initiates. A joint collaborative modeling effort among the NWS offices in Mobile, AL, and Houston, TX, and NASA s Short-term Prediction Research and Transition (SPoRT) Center was undertaken during the 2012 warm season to examine the impact of certain NASA produced products on the Weather Research and Forecasting Environmental Modeling System. The NASA products were: a 4-km Land Information System data, a 1-km sea surface temperature analysis, and a 4-km greenness vegetation fraction analysis. Similar domains were established over the southeast Texas and Alabama coastlines, each with a 9 km outer grid spacing and a 3 km inner nest spacing. The model was run at each NWS office once per day out to 24 hours from 0600 UTC, using the NCEP Global Forecast System for initial and boundary conditions. Control runs without the NASA products were made at the NASA SPoRT Center. The NCAR Model Evaluation Tools verification package was used to evaluate both the forecast timing and location of the first initiates, with a focus on the impacts of the NASA products on the model forecasts. Select case studies will be presented to highlight the influence of the products.

Medlin, Jeffrey M.↗
DOE OSTI2023

Deep Learning Potential For Investigation of Chemical Conversion of CO2 at Air-Reactive Deep Eutectic Solvent Interface

Capturing CO2 from air is urgent to battle against the climate change. Traditional CO2 capture by alkaline solution or aqueous amine has high regeneration cost, equipment corrosion, and amine leakage. Deep eutectic solvents (DESs) are promising alternatives with low corrosion, non-toxicity, and biodegradable nature. Functionalized DESs can capture and convert CO2 at low partial pressures. Developing reactive DESs for efficient CO2 capture needs a foundational understanding of the reaction-transport coupling of CO2 at the air-DES interface. In this study, we develop a deep learning potential (DLP) to investigate the absorption and reaction of CO2 at the air-DES interface (DES: 1-ethyl-3-methylimidazolium 2-cyanopyrrolide ([Emim][2-CNpyr]) and ethylene glycol (EG)). The DLP model allows us to simulate chemical reactions at a lower computational cost with ab initio accuracy. By analyzing reaction free energy surfaces, molecular interactions among CO2, DES components and reaction products, and transport of CO2 through the interface, we identify three main reaction pathways of CO2 at the interface: forming carboxylate with [Emim], carbamate with [2-CNpyr], and carbonate with EG. The mechanistic understanding of CO2 chemisorption at the interface will facilitate the development of novel direct air capture technology using DESs.

Nguyen, Manh Tien↗
DOE OSTI2022

Improved Architectures and Training Algorithms for Deep Operator Networks

Operator learning techniques have recently emerged as a powerful tool for learning maps between infinite-dimensional Banach spaces. Trained under appropriate constraints, they can also be effective in learning the solution operator of partial differential equations (PDEs) in an entirely self-supervised manner. In this work we analyze the training dynamics of deep operator networks (DeepONets) through the lens of Neural Tangent Kernel theory, and reveal a bias that favors the approximation of functions with larger magnitudes. To correct this bias we propose to adaptively re-weight the importance of each training example, and demonstrate how this procedure can effectively balance the magnitude of back-propagated gradients during training via gradient descent. We also propose a novel network architecture that is more resilient to vanishing gradient pathologies. Taken together, our developments provide new insights into the training of DeepONets and consistently improve their predictive accuracy by a factor of 10-50x, demonstrated in the challenging setting of learning PDE solution operators in the absence of paired input-output observations.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2019

Model reduction of dynamical systems on nonlinear manifolds using deep convolutional autoencoders

Nearly all model-reduction techniques project the governing equations onto a linear subspace of the original state space. Such subspaces are typically computed using methods such as balanced truncation, rational interpolation, the reduced-basis method, and (balanced) proper orthogonal decomposition (POD). Unfortunately, restricting the state to evolve in a linear subspace imposes a fundamental limitation to the accuracy of the resulting reduced-order model (ROM). In particular, linear-subspace ROMs can be expected to produce low-dimensional models with high accuracy only if the problem admits a fast decaying Kolmogorov n -width (e.g., diffusion-dominated problems). Unfortunately, many problems of interest exhibit a slowly decaying Kolmogorov n-width (e.g., advection-dominated problems). To address this, we propose a novel framework for projecting dynamical systems onto nonlinear manifolds using minimum-residual formulations at the time-continuous and time-discrete levels; the former leads to manifold Galerkin projection, while the latter leads to manifold least-squares Petrov–Galerkin (LSPG) projection. We perform analyses that provide insight into the relationship between these proposed approaches and classical linear-subspace reduced-order models; we also derive a posteriori discrete-time error bounds for the proposed approaches. In addition, we propose a computationally practical approach for computing the nonlinear manifold, which is based on convolutional autoencoders from deep learning. Lastly, we demonstrate the ability of the method to significantly outperform even the optimal linear-subspace ROM on benchmark advection-dominated problems, thereby demonstrating the method's ability to overcome the intrinsic n -width limitations of linear subspaces.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2016

Comparison of gain degradation and deep level transient spectroscopy in pnp Si bipolar junction transistors irradiated with different ion species

Here, we studied the effect of light ion and heavy ion irradiations on pnp Si BJTs. A mismatch in DLTS deep peak amplitude for devices with same final gain but irradiated with different ion species was observed. Also, different ions cause different gain degradation when the DLTS spectra are matched. Pre-dosed ion-irradiated samples show that ion induced ionization does not account for the differences in DLTS peak height but isochronal annealing studies suggest that light ions produce more VP defects than heavy ions to compensate for the lack of clusters that heavy ions produce. The creation of defect clusters by heavy ions is evident by the higher content of E4 and V* 2 defects compared to light ions.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
NASA NTRS1997

Deep Space One: NASA's First Deep-Space Technology Validation Mission

Under development for launch in July 1998, Deep Space One (DS1) is the first flight of NASA's New Millennium Program, Chartered to validate selected technologies required for future low-cost space science programs. Advanced technologies chosen for validation on DS1 include solar electric propulsion, high-power solar concentrator arrays, autonomous on-board optical navigation, two low-mass science instrument packages, and several telecommunications and microelectronics devices.

Deep↗
DOE OSTI2018

Weak Electron Phonon Coupling and Deep Level Impurity for High Thermoelectric Performance Pb 1–x Ga x Te

High ZT of 1.34 at 766 K and a record high average ZT above 1 in the temperature range of 300-864 K are attained in n-type PbTe by engineering the temperature-dependent carrier concentration and weakening electron–phonon coupling upon Ga doping. The experimental studies and first principles band structure calculations show that doping with Ga introduces a shallow level impurity contributing extrinsic carriers and imparts a deeper impurity level that ionizes at higher temperatures. This adjusts the carrier concentration closer to the temperature-dependent optimum and thus maximizes the power factor in a wide temperature range. The maximum power factor of 35 μW cm -1 K -2 is achieved for the Pb 0.98 Ga 0.02 Te compound, and is maintained over 20 μWcm-1 K-2 from 300 to 767 K. Band structure calculations and X-ray photoelectron spectroscopy corroborate the amphoteric role of Ga in PbTe as the origin of shallow and deep levels. Additionally, Ga doping weakens the electron–phonon coupling, leading to high carrier mobilities in excess of 1200 cm 2 V -1 s -1 . Enhanced point defect phonon scattering yields a reduced lattice thermal conductivity. Here, we provide a new avenue, beyond the conventional shallow level doping, for further improving the average ZT in thermoelectric materials.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2019

Direct Nanoscale Characterization of Deep Levels in AgCuInGaSe 2 Using Electron Energy–Loss Spectroscopy in the Scanning Transmission Electron Microscope

A new experimental framework for the characterization of defects in semiconductors is demonstrated. Through the direct, energy-resolved correlation of three analytical techniques spanning six orders of magnitude in spatial resolution, a critical mid-bandgap electronic trap level (E V + 0.56 eV) within Ag 0.2 Cu 0.8 In 1–x Ga x Se 2 is traced to its nanoscale physical location and chemical source. This is achieved through a stepwise, site-specific correlated characterization workflow consisting of device-scale (≈1 mm 2 ) deep level transient spectroscopy (DLTS) to survey the traps present, scanning probe–based DLTS (scanning-DLTS) for mesoscale-resolved (hundreds of nanometers) mapping of the target trap state’s spatial distribution, and scanning transmission electron microscope based electron energy-loss spectroscopy (STEM-EELS) and X-ray energy-dispersive spectroscopy for nanoscale energy-, structure, and chemical-resolved investigation of the defect source. Furthermore, this first demonstration of the direct observation of sub-bandgap defect levels via STEM-EELS, combined with the DLTS methods, provides strong evidence that the longsuspected Cu In/Ga substitutional defects are indeed the most likely source of the E V + 0.56 eV trap state and serves as a key example of this approach for the fundamental identification of defects within semiconductors, in general.

14 SOLAR ENERGY↗
DOE OSTI2016

High thermoelectric performance of n-type PbTe 1-y S y due to deep lying states induced by indium doping and spinodal decomposition

Good thermoelectric materials should have high engineering figure-of-merit (ZT) eng , not peak ZT, to achieve high conversion efficiency. In this work, we achieved a good (ZT) eng by optimizing the carrier concentration to improve the room temperature ZT using deep lying dopant, indium, in PbTe 1-y S y . It was found that a room temperature ZT as high as ~0.5 and a peak ZT ~1.1 at about 673 K were obtained in Pb 0.98 In 0.02 Te 0.8 S 0.2 due to a lower thermal conductivity by alloy scattering and Spinodal decomposition. The calculated efficiency could be as high as ~12% at cold side 323 K and hot side 773 K. The approach is expected to work in other materials systems too.

77 NANOSCIENCE AND NANOTECHNOLOGY↗
DOE OSTI2021

Comparative evaluation of deep learning workloads for leadership-class systems

Deep learning (DL) workloads and their performance at scale are becoming important factors to consider as we design, develop and deploy next-generation high-performance computing systems. Since DL applications rely heavily on DL frameworks and underlying compute (CPU/GPU) stacks, it is essential to gain a holistic understanding from compute kernels, models, and frameworks of popular DL stacks, and to assess their impact on science-driven, mission-critical applications. At Oak Ridge Leadership Computing Facility (OLCF), we employ a set of micro and macro DL benchmarks established through the Collaboration of Oak Ridge, Argonne, and Livermore (CORAL) to evaluate the AI readiness of our next-generation supercomputers. In this paper, we present our early observations and performance benchmark comparisons between the Nvidia V100 based Summit system with its CUDA stack and an AMD MI100 based testbed system with its ROCm stack. We take a layered perspective on DL benchmarking and point to opportunities for future optimizations in the technologies that we consider.

Yin, Junqi↗