Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “large scale network”

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 955 records · Page 53
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Subsurface Hydrogen and Natural Gas Storage (State of Knowledge and Research Recommendations Report)

Global efforts to decrease carbon dioxide (CO 2 ) emissions will result in increased reliance on renewable energy sources. One of the most significant obstacles to implementing this transition is the need for energy storage to compensate for varying, and sometimes intermittent, production rates and ever-increasing demand. An attractive solution is to use surplus energy to produce hydrogen gas (H 2 ), which can then be used to produce energy on demand without emitting CO 2 . This approach would require a vast H 2 network linking generators with consumers via transmission systems. Additionally, H 2 storage would be needed to support immediate needs, but larger storage reserves also will be required to ensure energy reliability and mitigate the impact of the mismatch between variable production rates and demand. Large-scale storage of H 2 may be achieved by using underground resources similar to how natural gas (NG) has been stored for the past century. While there is much experience in underground NG storage (UGS), significant technical challenges must be addressed to economically and safely store H 2 at the commercial scale. The purpose of this report is to assess the current state of knowledge and to identify technical research needs to facilitate deployment of large-scale underground hydrogen storage (UHS).

03 NATURAL GAS↗
DOE OSTI2024

Benchmarking Advanced Water Splitting Technologies: Best Practices in Materials Characterization

The high-level project goal is to create a comprehensive Best Practices benchmarking framework at the materials, component, device and systems levels for advanced water splitting technologies. All advanced water splitting pathways covered under the HydroGEN Energy Materials Network (EMN) Consortium, which include advanced high and low temperature electrolysis of water, photoelectrochemical (PEC) water splitting and solar thermochemical hydrogen (STCH) need these best practices to advance materials discovery. These practices will also aid the H2@Scale DOE initiative to accomplish their goals of large-scale H2 production.

08 HYDROGEN↗
DOE OSTI2023

Few-Shot Learning Enables Population-Scale Analysis of Leaf Traits in Populus trichocarpa

Plant phenotyping is typically a time-consuming and expensive endeavor, requiring large groups of researchers to meticulously measure biologically relevant plant traits, and is the main bottleneck in understanding plant adaptation and the genetic architecture underlying complex traits at population scale. In this work, we address these challenges by leveraging few-shot learning with convolutional neural networks to segment the leaf body and visible venation of 2,906 Populus trichocarpa leaf images obtained in the field. In contrast to previous methods, our approach (a) does not require experimental or image preprocessing, (b) uses the raw RGB images at full resolution, and (c) requires very few samples for training (e.g., just 8 images for vein segmentation). Traits relating to leaf morphology and vein topology are extracted from the resulting segmentations using traditional open-source image-processing tools, validated using real-world physical measurements, and used to conduct a genome-wide association study to identify genes controlling the traits. In this way, the current work is designed to provide the plant phenotyping community with (a) methods for fast and accurate image-based feature extraction that require minimal training data and (b) a new population-scale dataset, including 68 different leaf phenotypes, for domain scientists and machine learning researchers. All of the few-shot learning code, data, and results are made publicly available.

59 BASIC BIOLOGICAL SCIENCES↗
NASA NTRS1972

Minimizing the area required for time constants in integrated circuits

When a medium- or large-scale integrated circuit is designed, efforts are usually made to avoid the use of resistor-capacitor time constant generators. The capacitor needed for this circuit usually takes up more surface area on the chip than several resistors and transistors. When the use of this network is unavoidable, the designer usually makes an effort to see that the choice of resistor and capacitor combinations is such that a minimum amount of surface area is consumed. The optimum ratio of resistance to capacitance that will result in this minimum area is equal to the ratio of resistance to capacitance which may be obtained from a unit of surface area for the particular process being used. The minimum area required is a function of the square root of the reciprocal of the products of the resistance and capacitance per unit area. This minimum occurs when the area required by the resistor is equal to the area required by the capacitor.

Lyons, J. C.↗
DOE OSTI2023

Deep Learning Parameterization of Vertical Wind Velocity Variability via Constrained Adversarial Training

Atmospheric models with typical resolution in the tenths of kilometers cannot resolve the dynamics of air parcel ascent, which varies on scales ranging from tens to hundreds of meters. Small-scale wind fluctuations are thus characterized by a subgrid distribution of vertical wind velocity W with standard deviation σ W . The parameterization of σ W is fundamental to the representation of aerosol–cloud interactions, yet it is poorly constrained. Using a novel deep learning technique, this work develops a new parameterization for σ W merging data from global storm-resolving model simulations, high-frequency retrievals of W , and climate reanalysis products. The parameterization reproduces the observed statistics of σ W and leverages learned physical relations from the model simulations to guide extrapolation beyond the observed domain. Incorporating observational data during the training phase was found to be critical for its performance. The parameterization can be applied online within large-scale atmospheric models, or offline using output from weather forecasting and reanalysis products.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2020

Understanding the Spatial Organization of Simultaneous Heavy Precipitation Events Over the Conterminous United States

We introduce the idea of simultaneous heavy precipitation events (SHPEs) to understand whether extreme precipitation has a spatial organization manifested as specified tracks or contiguous fields with inherent scaling relationships. For this purpose, we created a database of SHPEs using ground-based precipitation observations recorded by the daily Global Historical Climatology Network across the conterminous United States during 1900–2014. SHPEs are examined for their seasonality, spatial manifestation, orientation, and areal extent. We quantified the spatial distribution of the centroids and principal axes of SHPEs and their quasi-elliptical manifestations, azimuthal orientations, and areal extents on the ground. Four seasons, December-January-February (DJF), March-April-May (MAM),June-July-August (JJA), and September-October-November (SON) are considered to examine the spatial patterns and associated large-scale atmospheric circulations. Results indicate that there are 54, 58, 103, and204 SHPEs in DJF, MAM, JJA, and SON seasons, and their longest stretches range on average between 650and 1,600, between 850 and 1,500, between 950 and 1,550, and between 750 and 1,450 km, respectively. SHPEs in the DJF, MAM, and JJA seasons occur mostly over the Pacific coast and central and midwestern United States, respectively. The atmospheric circulation patterns and mechanisms of precipitable water vapor and moisture transport in the atmosphere are also discussed in relation to these SHPEs. Power laws explain SHPEs' underlying area scaling behavior in all the four seasons, with stronger evidence in DJF and MAM. A seasonal spatial risk model is developed to predict the likelihood of SHPEs. Quantifying the characteristics of SHPEs and modeling their footprints can improve the projections of flood risk and understanding of damages to interconnected infrastructure systems.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

Ultra-low-density digitally architected carbon with a strutted tube-in-tube structure

Porous materials with engineered stretching-dominated lattice designs, which offer attractive mechanical properties with ultra-light weight and large surface area for wide-ranging applications, have recently achieved near-ideal linear scaling between stiffness and density. Here, rather than optimizing the microlattice topology, we explore a different approach to strengthen low-density structural materials by designing tube-in-tube beam structures. We develop a process to transform fully dense, three-dimensional printed polymeric beams into graphitic carbon hollow tube-in-tube sandwich morphologies, where, similar to grass stems, the inner and outer tubes are connected through a network of struts. Compression tests and computational modelling show that this change in beam morphology dramatically slows down the decrease in stiffness with decreasing density. In situ pillar compression experiments further demonstrate large deformation recovery after 30–50% compression and high specific damping merit index. Lastly, our strutted tube-in-tube design opens up the space and realizes highly desirable high modulus–low density and high modulus–high damping material structures.

36 MATERIALS SCIENCE↗
DOE OSTI2023

EPiC-GAN: Equivariant point cloud generation for particle jets

With the vast data-collecting capabilities of current and future high-energy collider experiments, there is an increasing demand for computationally efficient simulations. Generative machine learning models enable fast event generation, yet so far these approaches are largely constrained to fixed data structures and rigid detector geometries. In this paper, we introduce EPiC-GAN - equivariant point cloud generative adversarial network - which can produce point clouds of variable multiplicity. This flexible framework is based on deep sets and is well suited for simulating sprays of particles called jets. The generator and discriminator utilize multiple EPiC layers with an interpretable global latent vector. Crucially, the EPiC layers do not rely on pairwise information sharing between particles, which leads to a significant speed-up over graph- and transformer-based approaches with more complex relation diagrams. We demonstrate that EPiC-GAN scales well to large particle multiplicities and achieves high generation fidelity on benchmark jet generation tasks.

Buhmann, Erik↗
DOE OSTI2021

ForSE: A GAN-based Algorithm for Extending CMB Foreground Models to Subdegree Angular Scales

We present ForSE (Foreground Scale Extender), a novel Python package that aims to overcome the current limitations in the simulation of diffuse Galactic radiation, in the context of cosmic microwave background (CMB) experiments. ForSE exploits the ability of generative adversarial neural networks (GANs) to learn and reproduce complex features present in a set of images, with the goal of simulating realistic and non-Gaussian foreground radiation at subdegree angular scales. This is of great importance in order to estimate the foreground contamination to lensing reconstruction, delensing, and primordial B-modes for future CMB experiments. We applied this algorithm to Galactic thermal dust emission in both total intensity and polarization. Our results show how ForSE is able to generate small-scale features (at 12') having as input the large-scale ones (80'). The injected structures have statistical properties, evaluated by means of the Minkowski functionals, in good agreement with those of the real sky and which show the correct amplitude scaling as a function of the angular dimension. Furthermore, the obtained thermal dust Stokes Q and U full-sky maps as well as the ForSE package are publicly available for download.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2023

A GPU-Accelerated Population Generation, Sorting, and Mutation Kernel for an Optimization-Based Causal Inference Model

We develop a GPU-accelerated machine learning generative adversarial network model that can be used with observational data for the purpose of constructing causal inferences. The theoretical basis of our machine learning model is novel and is conceptualized to be operable and scalable for high performance computing platforms. Our GPU-accelerated code enables large-scale parallelization of the computation within a common and accessible computing environment. This will expand the reach of our model and empower research in new substantive domains while maintaining the underlying theoretical properties.

Cho, Wendy K. Tam↗
DOE OSTI2021

Data for Adoption of Perennial Energy Crops in the U.S. Midwest, Causal and Heterogeneous Determinants

The success of a bioenergy policy relies largely on the wide adoption of perennial energy crops at the farm scale. This study uses survey data to examine potential adoption decisions by farmers in the U.S. Midwest and the causal effects of various direct and indirect influencing factors, especially heterogeneous preferences of farmers. A Bayesian network (BN) model is developed to delineate the causal relationship between farmers adoption decisions and the influencing factors. We find a dominating role of economic factors and a non-negligible impact of non-economic factors, such as the perceived environmental benefits and the extent of familiarity with perennial energy crops. To examine the effect of heterogeneity in farmer preferences, we classify the surveyed farmers into four categories based on their attitudes toward the economic, social, and environmental dimensions of perennial energy crops. We identified statistically significant between-group differences in the responses of the four types of farmers to the various influencing factors. Our findings contribute to disentangling the complicated motivations that will influence perennial energy crop adoption decisions and provide implications for more targeted policy development that need to consider the heterogeneous drivers of farmer decisions about land use.

Modeling↗
DOE OSTI2024

Addressing GPU memory limitations for Graph Neural Networks in High-Energy Physics applications

Introduction Reconstructing low-level particle tracks in neutrino physics can address some of the most fundamental questions about the universe. However, processing petabytes of raw data using deep learning techniques poses a challenging problem in the field of High Energy Physics (HEP). In the Exa.TrkX Project, an illustrative HEP application, preprocessed simulation data is fed into a state-of-art Graph Neural Network (GNN) model, accelerated by GPUs. However, limited GPU memory often leads to Out-of-Memory (OOM) exceptions during training, due to the large size of models and datasets. This problem is exacerbated when deploying models on High-Performance Computing (HPC) systems designed for large-scale applications. Methods We observe a high workload imbalance issue during GNN model training caused by the irregular sizes of input graph samples in HEP datasets, contributing to OOM exceptions. We aim to scale GNNs on HPC systems, by prioritizing workload balance in graph inputs while maintaining model accuracy. Our paper introduces diverse balancing strategies aimed at decreasing the maximum GPU memory footprint and avoiding the OOM exception, across various datasets. Results Our experiments showcase memory reduction of up to 32.14% compared to the baseline. We also demonstrate the proposed strategies can avoid OOM in application. Additionally, we create a distributed multi-GPU implementation using these samplers to demonstrate the scalability of these techniques on the HEP dataset. Discussion By assessing the performance of these strategies as data loading samplers across multiple datasets, we can gauge their effectiveness in both single-GPU and distributed environments. Our experiments, conducted on datasets of varying sizes and across multiple GPUs, broaden the applicability of our work to various GNN applications that handle input datasets with irregular graph sizes.

Lee, Claire Songhyun↗
NASA NTRS1986

The spacecraft encounters of Comet Halley

The characteristics of the Comet Halley spacecraft 'fleet' (VEGA 1 and VEGA 2, Giotto, Suisei, and Sakigake) are presented. The major aims of these missions were (1) to discover and characterize the nucleus, (2) to characterize the atmosphere and ionosphere, (3) to characterize the dust, and (4) to characterize the nature of the large-scale comet-solar wind interaction. While the VEGA and Giotto missions were designed to study all four areas, Suisei addressed the second and fourth. Sakigake was designed to study the solar wind conditions upstream of the comet. It is noted that NASA's Deep Space Network played an important role in spacecraft tracking.

Asoka Mendis, D.↗
NASA NTRS1992

VME rollback hardware for time warp multiprocessor systems

The purpose of the research effort is to develop and demonstrate innovative hardware to implement specific rollback and timing functions required for efficient queue management and precision timekeeping in multiprocessor discrete event simulations. The previously completed phase 1 effort demonstrated the technical feasibility of building hardware modules which eliminate the state saving overhead of the Time Warp paradigm used in distributed simulations on multiprocessor systems. The current phase 2 effort will build multiple pre-production rollback hardware modules integrated with a network of Sun workstations, and the integrated system will be tested by executing a Time Warp simulation. The rollback hardware will be designed to interface with the greatest number of multiprocessor systems possible. The authors believe that the rollback hardware will provide for significant speedup of large scale discrete event simulation problems and allow multiprocessors using Time Warp to dramatically increase performance.

Robb, Michael J.↗
NASA NTRS1995

Neural Networks For Demodulation Of Phase-Modulated Signals

Hopfield neural networks proposed for demodulating quadrature phase-shift-keyed (QPSK) signals carrying digital information. Networks solve nonlinear integral equations prior demodulation circuits cannot solve. Consists of set of N operational amplifiers connected in parallel, with weighted feedback from output terminal of each amplifier to input terminals of other amplifiers. Used to solve signal processing problems. Implemented as analog very-large-scale integrated circuit that achieves rapid convergence. Alternatively, implemented as digital simulation of such circuit. Also used to improve phase estimation performance over that of phase-locked loop.

Altes, Richard A.↗
NASA NTRS1998

The Locations of Gamma-Ray Bursts Measured by Comptel

The COMPTEL instrument on the Compton Gamma Ray Observatory is used to measure the locations of gamma-ray bursts through direct imaging of MeV photons. In a comprehensive search, we have detected and localized 29 bursts observed between 1991 April 19 and 1995 May 31. The average location accuracy of these events is 1.25 deg (1 sigma), including a systematic error of approx. 0.5 deg, which is verified through comparison with Interplanetary Network (IPN) timing annuli. The combination of COMPTEL and IPN measurements results in locations for 26 of the bursts with an average "error box" area of only approx. 0.3 deg (1 sigma). We find that the angular distribution of COMPTEL burst locations is consistent with large-scale isotropy and that there is no statistically significant evidence of small-angle autocorrelations. We conclude that there is no compelling evidence for burst repetition since no more than two of the events (or approx. 7% of the 29 bursts) could possibly have come from the same source. We also find that there is no significant correlation between the burst locations and either Abell clusters of galaxies or radio-quiet quasars. Agreement between individual COMPTEL locations and IPN annuli places a lower limit of approx. 100 AU (95% confidence) on the distance to the stronger bursts.

Kippen, R. Marc↗
NASA NTRS2003

A Framework for Intelligent Rocket Test Facilities with Smart Sensors

A long-term center goal at the John C. Stennis Space Center (SSC) is the formulation and implementation of a framework for an Intelligent Rocket Test Facility (IRTF), which incorporates distributed smart sensor elements. The IRTF is to provide reliable, high-confident measurements. Specific objectives include: 1. Definition of a framework and architecture that supports implementation of highly autonomous methodologies founded on basic physical principles and embedded knowledge. 2. Modeling of autonomous sensors and processes as self-sufficient, evolutionary elements. 3. Development of appropriate communications protocols to enable the complex interactions that must take place to allow timely and high-quality flow of of information among all the autonomous elements of the system. 4. Development of lab-scale prototypes of key system elements. Though our application is next-generation rocket test facilities, applications for the approach are much wider and include monitoring of shuttle launch operations, air and spacecraft operations and health monitoring, and other large-scale industrial system operations such as found in processing and manufacturing plans. Elements of prototype IRTF have been implemented in preparation for advanced development and validation using rocket test stand facilities as SSC. This work has identified issues that are important to further development of complex network and should be of interest to other working with sensor networks.

Figueroa, Fernando↗
NASA NTRS2023

A Machine Learning-Based Approach to Time-Series Wave Identification in the Solar Wind

The Wind spacecraft has yielded several decades of high-resolution magnetic field data, a large fraction of which displays small-scale structures. In particular, the solar wind is full of wavelike fluctuations that appear in both the field magnitude and its components. The nature of these fluctuations can be tied to the properties of other structures in the solar wind, such as shocks, that have implications for the time evolution of the solar wind. As such, having a large collection of wave events would facilitate further study of the effects that these fluctuations have on solar wind evolution. Given the large volume of magnetic field data available, machine learning is the most practical approach to classifying the myriad small-scale structures observed. To this end, a subset of Wind data is labeled and used as a training set for a multi-branch 1D convolutional neural network aimed at classifying circularly polarized wave modes. Using this algorithm, a preliminary statistical study of one year of data is performed, yielding about 300,000 wave intervals out of about 5,000,000 solar wind intervals. The wave intervals come about more often in the fast solar wind and at higher temperatures, and the number of waves per day is highly periodic. This machine learning-based approach to wave detection has the potential to be a powerful, inexpensive way to catalog waves throughout decades of spacecraft data.

Samuel Fordin↗