Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “data locality”

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 1,585 records · Page 88
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Federated Quantum Machine Learning

Distributed training across several quantum computers could significantly improve the training time and if we could share the learned model, not the data, it could potentially improve the data privacy as the training would happen where the data is located. One of the potential schemes to achieve this property is the federated learning (FL), which consists of several clients or local nodes learning on their own data and a central node to aggregate the models collected from those local nodes. However, to the best of our knowledge, no work has been done in quantum machine learning (QML) in federation setting yet. In this work, we present the federated training on hybrid quantum-classical machine learning models although our framework could be generalized to pure quantum machine learning model. Specifically, we consider the quantum neural network (QNN) coupled with classical pre-trained convolutional model. Our distributed federated learning scheme demonstrated almost the same level of trained model accuracies and yet significantly faster distributed training. It demonstrates a promising future research direction for scaling and privacy aspects.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Similarity Metric for Data Optimization and Efficient Training of Reactive Machine Learning Force Fields for Hydrocarbon Radiolysis

Radiolysis is a common approach to sterilize polymers, chemically modify them for upcycling, and accelerate their decomposition for recycling purposes. Reactive molecular dynamics (MD) simulations provide a powerful tool to generate atomic-level trajectories of the reactive processes and quantify radiolytic chemical degradation pathways. For this, machine learning (ML) surrogate models for reactive force fields with quantum mechanical accuracy are now widely used, which require ML training data sets that can provide information on atomic environments for target chemical systems. However, radiolysis chemistry can be highly complex and diverse, which poses significant challenges for generating training data to parametrize ML models. In this regard, we developed a method for optimizing the training data set using a cosine similarity metric to help guide training set selection for radiolysis of polyethylene, a model hydrocarbon polymer, as well as to enhance the transferability of our reactive ML force field (MLFF) to a variety of molecular and polymeric systems. Our approach performs atom-by-atom comparisons between local atomic environments to pinpoint important data points associated with rare and localized events, such as radiolysis damage within structures. We apply this approach to train the Chebyshev Interaction Model for Efficient Simulation (ChIMES) MLFF model, which expresses the atomic interaction potentials in terms of linear combinations of many-body Chebyshev polynomials. We first show that our method can reduce our training set size by ∼70% while improving overall accuracy compared to more standard MD model fitting approaches. We then validate our optimum model against diverse hydrocarbon simulation data, including simple alkanes and systems with unsaturated carbon bonds, over a wide range of thermodynamic conditions. Finally, we use our ChIMES model to perform MD simulations of radiolytic damage with large-scale systems that help avoid system size effects. Overall, our approach yields an MD force field that retains most of the accuracy of the underlying quantum method while yielding many orders of improvement in computational efficiency. In conclusion, our efforts will have impact on future hydrocarbon polymer radiolysis studies, where the chemical details of the polymer–radiation interactions can have a strong effect on the resulting products observed in experiments.

Hydrocarbons↗
DOE OSTI2020

Striped Data Analysis Framework

A columnar data representation is known to be an efficient way for data storage, specifically in cases when the analysis is often done based only on a small fragment of the available data structures. A data representation like Apache Parquet is a step forward from a columnar representation, which splits data horizontally to allow for easy parallelization of data analysis. Based on the general idea of columnar data storage, working on the [LDRD Project], we have developed a striped data representation, which, we believe, is better suited to the needs of High Energy Physics data analysis. A traditional columnar approach allows for efficient data analysis of complex structures. While keeping all the benefits of columnar data representations, the striped mechanism goes further by enabling easy parallelization of computations without requiring special hardware. We will present an implementation and some performance characteristics of such a data representation mechanism using a distributed no-SQL database or a local file system, unified under the same API and data representation model. The representation is efficient and at the same time simple so that it allows for a common data model and APIs for wide range of underlying storage mechanisms such as distributed no-SQL databases and local file systems. Striped storage adopts Numpy arrays as its basic data representation format, which makes it easy and efficient to use in Python applications. The Striped Data Server is a web service, which allows to hide the server implementation details from the end user, easily exposes data to WAN users, and allows to utilize well known and developed data caching solutions to further increase data access efficiency. We are considering the Striped Data Server as the core of an enterprise scale data analysis platform for High Energy Physics and similar areas of data processing. We have been testing this architecture with a 2TB dataset from a CMS dark matter search and plan to expand it to multiple 100 TB or even PB scale. We will present the striped format, Striped Data Server architecture and performance test results.

Gutsche, Oliver↗
DOE OSTI2018

Striped Data Server for Scalable Parallel Data Analysis

A columnar data representation is known to be an efficient way for data storage, specifically in cases when the analysis is often done based only on a small fragment of the available data structures. A data representation like Apache Parquet is a step forward from a columnar representation, which splits data horizontally to allow for easy parallelization of data analysis. Based on the general idea of columnar data storage, working on the [LDRD Project], we have developed a striped data representation, which, we believe, is better suited to the needs of High Energy Physics data analysis. A traditional columnar approach allows for efficient data analysis of complex structures. While keeping all the benefits of columnar data representations, the striped mechanism goes further by enabling easy parallelization of computations without requiring special hardware. We will present an implementation and some performance characteristics of such a data representation mechanism using a distributed no-SQL database or a local file system, unified under the same API and data representation model. The representation is efficient and at the same time simple so that it allows for a common data model and APIs for wide range of underlying storage mechanisms such as distributed no-SQL databases and local file systems. Striped storage adopts Numpy arrays as its basic data representation format, which makes it easy and efficient to use in Python applications. The Striped Data Server is a web service, which allows to hide the server implementation details from the end user, easily exposes data to WAN users, and allows to utilize well known and developed data caching solutions to further increase data access efficiency. We are considering the Striped Data Server as the core of an enterprise scale data analysis platform for High Energy Physics and similar areas of data processing. As a result, we have been testing this architecture with a 2TB dataset from a CMS dark matter search and plan to expand it to multiple 100 TB or even PB scale. We will present the striped format, Striped Data Server architecture and performance test results.

96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗
DOE OSTI2020

Quantifying redox heterogeneity in single-crystalline LiCoO 2 cathode particles

Active cathode particles are fundamental architectural units for the composite electrode of Li-ion batteries. The microstructure of the particles has a profound impact on their behavior and, consequently, on the cell-level electrochemical performance. LiCoO 2 (LCO, a dominant cathode material) is often in the form of well-shaped particles, a few micrometres in size, with good crystallinity. In contrast to secondary particles (an agglomeration of many fine primary grains), which are the other common form of battery particles populated with structural and chemical defects, it is often anticipated that good particle crystallinity leads to superior mechanical robustness and suppressed charge heterogeneity. Yet, sub-particle level charge inhomogeneity in LCO particles has been widely reported in the literature, posing a frontier challenge in this field. Herein, this topic is revisited and it is demonstrated that X-ray absorption spectra on single-crystalline particles with highly anisotropic lattice structures are sensitive to the polarization configuration of the incident X-rays, causing some degree of ambiguity in analyzing the local spectroscopic fingerprint. To tackle this issue, a methodology is developed that extracts the white-line peak energy in the X-ray absorption near-edge structure spectra as a key data attribute for representing the local state of charge in the LCO crystal. This method demonstrates significantly improved accuracy and reveals the mesoscale chemical complexity in LCO particles with better fidelity. In addition to the implications on the importance of particle engineering for LCO cathodes, the method developed herein also has significant impact on spectro-microscopic studies of single-crystalline materials at synchrotron facilities, which is broadly applicable to a wide range of scientific disciplines well beyond battery research.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
NASA NTRS1997

Correction of Single Frequency Altimeter Measurements for Ionosphere Delay

This study is a preliminary analysis of the accuracy of various ionosphere models to correct single frequency altimeter height measurements for Ionospheric path delay. In particular, research focused on adjusting empirical and parameterized ionosphere models in the parameterized real-time ionospheric specification model (PRISM) 1.2 using total electron content (TEC) data from the global positioning system (GPS). The types of GPS data used to adjust PRISM included GPS line-of-sight (LOS) TEC data mapped to the vertical, and a grid of GPS derived TEC data in a sun-fixed longitude frame. The adjusted PRISM TEC values, as well as predictions by IRI-90, a climatotogical model, were compared to TOPEX/Poseidon (T/P) TEC measurements from the dual-frequency altimeter for a number of T/P tracks. When adjusted with GPS LOS data, the PRISM empirical model predicted TEC over 24 1 h data sets for a given local time to with in a global error of 8.60 TECU rms during a midnight centered ionosphere and 9.74 TECU rms during a noon centered ionosphere. Using GPS derived sun-fixed TEC data, the PRISM parameterized model predicted TEC within an error of 8.47 TECU rms centered at midnight and 12.83 TECU rms centered at noon. From these best results, it is clear that the proposed requirement of 3-4 TECU global rms for TOPEX/Poseidon Follow-On will be very difficult to meet, even with a substantial increase in the number of GPS ground stations, with any realizable combination of the aforementioned models or data assimilation schemes.

Schreiner, William S.↗
NASA NTRS2021

Landslide Hazard and Exposure Modelling in Data‐Poor Regions: The Example of the Rohingya Refugee Camps in Bangladesh

Landslide hazards significantly affect economies and populations around the world, but locations where the greatest proportional losses occur are in data‐poor regions where capacity to estimate and prepare for these hazards is most limited. Earth observation (EO) data can fill key knowledge gaps, and can be rapidly used in settings with lower analytical capacity. In this study, we describe a novel series of methods designed to analyze landslide susceptibility, hazard and exposure in the region in and around the Rohingya refugee camps in Bangladesh, where limited data is juxtaposed with a major humanitarian crisis. We demonstrate that a high degree of accuracy is possible even when estimating susceptibility of relatively small landslides. In the context of this example, we also explore how estimates of landslide hazard and exposure are most beneficial to decisions made by humanitarian stakeholders relevant to natural hazards and risk. The unique opportunity to work alongside humanitarian end‐users has allowed us to produce focused products that can be tested while in development. In particular, we stress the importance of communicating the difference between a landslide “early warning system”—for which satellite data may be unsuitable at local scales—and a model that provides relative hazard estimates, where EO may be valuable. The toolbox of methods presented here could be used to generate landslide hazard and exposure maps in other data‐poor regions around the globe.

R A Emberson↗
DOE OSTI2019

Designing a Site Closure Strategy for Uranium in Groundwater and Surface Water - 19534

A site closure strategy was designed and mapped for residual uranium present in groundwater and surface water at the US DOE Office of Legacy Management (LM) Monticello Mill Tailings Site (MMTS) Operable Unit (OU) III, which is regulated under CERCLA. The MMTS is a former mill site that was used to process uranium and vanadium ore in southeastern Utah until 1960. OU III includes surface water in Montezuma Creek and groundwater present in a shallow alluvial aquifer along the creek valley for several kilometers downgradient of the former mill site. The current remedy consists of an operating pump-and-treat system, monitored natural attenuation (MNA), and institutional controls (ICs). Performance evaluations indicate that a prolonged restoration period is likely, as well as an asymptotic trend in uranium concentrations that still exceed the remediation goal. Multiple strategies and scenarios were evaluated before selecting a preferred site closure strategy. Recommendations were developed to support advancing the site towards closure over the next few years and guide the process of regulatory interactions and decision making at the MMTS OU III. Site closure is an important milestone that reflects successful remedial progress at LM sites. Post-closure site management would be maintained to ensure conditions are protective of human health and the environment. The site closure strategy was considered in the context of site remedial progress. The primary sources of groundwater contamination have been removed, including mill tailings and contaminated soils, sediment, and debris. Five-year reviews have regularly concluded that the remedy is protective of human health and environment and that ICs are effective in preventing exposure to contaminated groundwater. The affected aquifer is low-yielding and has no history of domestic use. In 2004, the final remedy selected in the Record of Decision (ROD) was MNA with ICs. However, a pump-and-treat system was installed and continues to operate as a contingency remedy in the area with the highest residual uranium concentrations. The need for a contingency pump-and-treat system was triggered by a departure of the long-term monitoring data from model projections that were used as the basis for remedy performance metrics specified in the ROD. In 2015, the pump-and-treat system was expanded. Average uranium concentrations have declined locally over time. However, monitoring data and column studies indicate that the system effectiveness will eventually become ineffective in reducing uranium concentrations, likely due to the reequilibration of uranium present in the solid matrix in various states (i.e., sorbed, precipitated). Currently, there are no agreed-upon metrics for discontinuing the contingency pump-and-treat system. Three closure strategies effectively applied at other CERCLA sites were considered at MMTS OU III: (1) restoring natural attenuation conditions and conducting MNA; (2) applying alternate concentration limits (ACLs); and (3) waiving a remedial goal based on technical impracticability (TI). Each closure strategy was evaluated in the following manner: (1) identifying the criteria and requirements associated with current regulations, policy, and guidance, (2) assessing the ability to meet those criteria based on the current site knowledge and the conceptual site model, and (3) identifying data gaps that need to be filled to further assess the closure strategy. Following this evaluation, a preferred closure strategy was selected: to transition to a remedy based on MNA and maintenance of ICs, without active remediation. Several scenarios for transitioning to the preferred closure strategy were identified. Each scenario consisted of 'if/then logic' associated with a larger decision diagram. The decision diagram considered the current conceptual site model (CSM) and potential outcomes of studies that will be conducted to fill data gaps, as well as the performance of the existing treatment system and long-term monitoring data. Two scenarios for transitioning to the preferred closure strategy were highlighted for future planning and were described in detail. For both scenarios, the preferred closure strategy includes demonstrating criteria for MNA acceptance, pump-and-treat termination, long-term monitoring, and eventual site delisting. Recommended actions over the next few years were also developed, including updating the CSM, conducting numerical modeling to predict remedial performance and time frames, conducting geochemical analyses to assess uranium fate and transport, and summarizing the lines of evidence and basis for MNA. This detailed example of the development of a site-specific closure strategy provides the thought process and decision logic used to assess uncertainty using an if/then framework; identify common elements and recommended actions to advance the site towards closure; and navigate the inherent uncertainty in the timing of future events, outcome of regulatory discussions, and remedial progress over time that is common to many complex US DOE facilities. In addition, the national regulations, policy, guidance, and state of practice regarding commonly used site closure strategies was summarized for consideration, including MNA, ACLs, and TI. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
DOE OSTI2015

Non-local gravity and comparison with observational datasets

We study the cosmological predictions of two recently proposed non-local modifications of General Relativity. Both models have the same number of parameters as ΛCDM, with a mass parameter m replacing the cosmological constant. We implement the cosmological perturbations of the non-local models into a modification of the CLASS Boltzmann code, and we make a full comparison to CMB, BAO and supernova data. We find that the non-local models fit these datasets very well, at the same level as ΛCDM. Among the vast literature on modified gravity models, this is, to our knowledge, the only example which fits data as well as ΛCDM without requiring any additional parameter. For both non-local models parameter estimation using Planck +JLA+BAO data gives a value of H{sub 0} slightly higher than in ΛCDM.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

Does Age, Residency, or Feeding Guild Coupled with a Drought Index Predict Avian Health during Fall Migration?

Birds are good indicators of environmental change and are often studied for responses to climate. Many studies focus on breeding birds, while fewer look at the migration period, which is a critical time for many birds. Birds are more susceptible to unusual climatic events during their migration due to the metabolic stress of long-distance movements. In the fall of 2020, an unusual cold weather event coupled with drought and wildfire smoke led to a large avian mortality event in New Mexico. Later analysis pointed to the mortality being largely due to starvation. This was the impetus for our research. We used 11 years of fall bird banding data from two locations, along with local drought indices, to determine what predicts avian health during the migration period. We used fat score data from over 15,000 individual birds to assess whether drought indices, age, diet, or residency influenced avian health using multiple logistic regression. We found that the probability of positive fat scores decreased as drought severity increased for younger, insectivorous, migratory birds. Insectivores had a higher probability of receiving a fat score greater than zero relative to local drought conditions, which is important, since many North American insectivores are in steep decline. Migratory birds showed a greater response than year-round residents, and older birds showed a lower but significant response compared to hatch-year birds. Our results suggest that migratory insectivores in the southwestern United States may be less resilient to drought-related climate change.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Cavity evolution and void swelling in dual ion irradiated tempered martensitic steels

Microstructural evolutions in two variants of the F82H 8 wt%Cr tempered martensitic steel under dual (He + plus Fe 3+ ) ion irradiations at 500 °C have been characterized by transmission electron microscopy. Emphasis is on understanding the combined effects of displacements per atom (dpa) and the helium/dpa ratio (He/dpa) on cavity evolution and void swelling (f v ). The dual ion database includes 231 alloy-dpa-He/dpa conditions for irradiations up to ~ 82 dpa and 3700 appm helium. Here, we hypothesize that the large “scatter” observed in the f v data is primarily due to local microstructural variations, especially those that shield regions at shallower depths from the effects of injected interstitial atoms, which would otherwise reduce f v . More generally, our hypothesis is that the highest f v data, at a specified dpa and He/dpa, provides the most appropriate estimate of swelling for fusion neutron irradiation conditions. The dose dependence of f v can be described by an incubation dpa i , marking the onset of significant void growth, and a post-incubation swelling rate, f v ’ = Δf v /Δdpa. Both dpa i and f v ’ decrease with increasing He/dpa. The f v ’ data are quantitatively consistent with a simple model for defect partitioning between evolving bubble, void, and dislocation sinks strengths and biases. Most notably, the model correctly predicts both the incubation dpa i , and the subsequent f v ’, in spite of the fact that they reflect somewhat different physics. Our analysis also suggests that defect recombination is modest at 500 °C. Along with other considerations, this supports the hypothesis that, even at high dpa rates, dual ion irradiations can be used to reasonably emulate void nucleation and growth for a range of fusion-relevant neutron irradiation service conditions. A companion paper incorporates the results of this study in a reduced-order, data-driven void swelling model.

36 MATERIALS SCIENCE↗
NASA NTRS2021

Landslide Hazard and Exposure Modelling in Data-Poor Regions: The Example of the Rohingya Refugee Camps in Bangladesh

Landslide hazards significantly affect economies and populations around the world, but locations where the greatest proportional losses occur are in data-poor regions where capacity to estimate and prepare for these hazards is most limited. Earth observation (EO) data can fill key knowledge gaps, and can be rapidly used in settings with lower analytical capacity. In this study, we describe a novel series of methods designed to analyze landslide susceptibility, hazard and exposure in the region in and around the Rohingya refugee camps in Bangladesh, where limited data is juxtaposed with a major humanitarian crisis. We demonstrate that a high degree of accuracy is possible even when estimating susceptibility of relatively small landslides. In the context of this example, we also explore how estimates of landslide hazard and exposure are most beneficial to decisions made by humanitarian stakeholders relevant to natural hazards and risk. The unique opportunity to work alongside humanitarian end-users has allowed us to produce focused products that can be tested while in development. In particular, we stress the importance of communicating the difference between a landslide ‘early warning system’ –for which satellite data may be unsuitable at local scales –and a model that provides relative hazard estimates, where EO may be valuable. The toolbox of methods presented here could be used to generate landslide hazard and exposure maps in other data-poor regions around the globe.

R A Emberson↗
DOE OSTI2021

RISE: Reducing I/O Contention in Staging-based Extreme-Scale In-situ Workflows

While in-situ workflow formulations have addressed some of the data-related challenges associated with extreme-scale scientific workflows, these workflows involve complex interactions and different modes of data exchange. In the context of increasing system complexity, such workflows present significant resource management challenges, requiring complex cost-performance tradeoffs. This paper presents RISE, an intelligent staging-based data management middleware, which builds on the DataSpaces framework and performs intelligent scheduling of data management operations to reduce I/O contention. In RISE, data are always written immediately to local buffers to reduce the effect of the transfer impact upon application performance. RISE identifies applications’ data access patterns and moves data towards data consumers only when the network is expected to be idle, reducing the impact of asynchronous background data movement upon critical data read/write requests. Here, we experimentally demonstrate that RISE can take advantage of staging nodes to offload data during writes without degrading application data movement performance.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2019

Extreme-bandwidth scalable performance-per-watt GPU architecture

A technique for accessing memory in an accelerated processing device coupled to stacked memory dies is provided herein. The technique includes receiving a memory access request from an execution unit and identifying whether the memory access request corresponds to memory cells of the stacked dies that are considered local to the execution unit or non-local. For local accesses, the access is made “directly”, that is, without using a bus. A control die coordinates operations for such local accesses, activating particular through-silicon-vias associated with the memory cells that include the data for the access. Non-local accesses are made via a distributed cache fabric and an interconnect bus in the control die. Various other features and details are provided below.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1994

Expert system technologies for Space Shuttle decision support: Two case studies

This paper addresses the issue of integrating the C Language Integrated Production System (CLIPS) into distributed data acquisition environments. In particular, it presents preliminary results of some ongoing software development projects aimed at exploiting CLIPS technology in the new mission control center (MCC) being built at NASA Johnson Space Center. One interesting aspect of the control center is its distributed architecture; it consists of networked workstations which acquire and share data through the NASA/JSC-developed information sharing protocol (ISP). This paper outlines some approaches taken to integrate CLIPS and ISP in order to permit the development of intelligent data analysis applications which can be used in the MCC. Three approaches to CLIPS/IPS integration are discussed. The initial approach involves clearly separating CLIPS from ISP using user-defined functions for gathering and sending data to and from a local storage buffer. Memory and performance drawbacks of this design are summarized. The second approach involves taking full advantage of CLIPS and the CLIPS Object-Oriented Language (COOL) by using objects to directly transmit data and state changes from ISP to COOL. Any changes within the object slots eliminate the need for both a data structure and external function call thus taking advantage of the object matching capabilities within CLIPS 6.0. The final approach is to treat CLIPS and ISP as peer toolkits. Neither is embedded in the other; rather the application interweaves calls to each directly in the application source code.

Ortiz, Christopher J.↗
NASA NTRS2005

Constrained spectral clustering under a local proximity structure assumption

This work focuses on incorporating pairwise constraints into a spectral clustering algorithm. A new constrained spectral clustering method is proposed, as well as an active constraint acquisition technique and a heuristic for parameter selection. We demonstrate that our constrained spectral clustering method, CSC, works well when the data exhibits what we term local proximity structure.

domain knowledge↗
NASA NTRS1981

Flame structure in a swirl stabilized combustor inferred by radiant emission measurements

Results of measurements of time-averaged chemiluminescent emissions from CH, OH, and CO2 and of Na tracer emissions along lateral lines-of-sight through a cylindrical premixed, swirl-stabilized combustor are reported. Assuming axial symmetry and small optical depth, raw data are inverted to obtain local emission levels from these species as a function of radius. The chemiluminescent emissions are interpreted as signatures of chemical reaction and used in determining the regions of reactions and heat release in the combustor. The data are compared with composition and velocity data obtained in the combustor for identical operating conditions. The results demonstrate that reaction occurs in a relatively narrow, turbulent flame-like combustion zone which begins upstream of the time-averaged location of the swirl-induced recirculation zone and propagates around and laterally away from the recirculation zone into the unburned gas.

Beyler, C. L.↗
NASA NTRS1993

Determining torques over the ocean and their role in the planetary momentum budget

The transfer of angular momentum between the atmosphere and oceans was examined using a surface wind data set. The data set combines special sensor microwave imager wind speed estimates and operational wind analyses from the European Centre for Medium Range Weather Forecasts for estimation of zonal torques over the global ocean from July 1987 to June 1989. An attempt was made to determine the global space-time structure of ocean torques. The wind data sets are described, the general statistics of local and global torques are presented, and their characteristics were evaluated. Consideration is given to the impact of satellite data on the calculation of local and global torques. The contribution of ocean torques to fluctuations in atmospheric angular momentum was analyzed as a function of time scale and geographical location.

Ponte, Rui M.↗