Search NASA⌕ Search

SEARCH · Search NASA

Results for “scientific method”

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 235 records · Page 13
All sourcesNASA NTRSDOE OSTI
NASA NTRSDate not supplied

Enabling Intelligent Data Downlink Prioritization of In-Situ Observations through Generalizable and Computationally Inexpensive Anomaly Detection

High-fidelity measurements of magnetic fields and other observed properties, such as energetic particle fluxes, are a necessary component to our understanding of the highly dynamic near-Earth space environment. As our desire to study smaller-scale phenomena such as shocks and dipolorizations has increased, we have been driven to take and telemeter measurements at higher cadences. Unfortunately, many missions are unable to downlink all their captured data due to the well-known data transmission bottleneck at the DSN. These missions must then prioritize their high-cadence data such that the most scientifically useful intervals are transmitted. One simple prioritization technique uses the spacecraft position to telemeter data from only the region of interest. Although easy to implement, this method does not leverage the available scientific data and can omit intervals of useful scientific data when they lie outside the region of interest. The Magnetospheric Multiscale Mission (MMS) uses mission-specific parameterization of several data products to automatically prioritize scientifically useful intervals. Then, MMS verifies the automatically selected intervals by having a domain expert manually select intervals for downlink. The overall complexity required by this technique make it prohibitive for deployment on low-cost platforms (i.e., CubeSats) or on future missions featuring large constellations of satellites such as the Geospace Dynamics Constellation (GDC). We present preliminary results for a simple, generic, and data-driven method of downlink prioritization for magnetic field (and other) measurements. Specifically, Principal Components Analysis (PCA) and One-Class Support Vector Machines (OC-SVMs) are used to detect intervals containing anomalous activity, which can then be prioritized for subsequent downlink. The computational simplicity of this algorithm makes it an excellent candidate for implementation on spaceflight hardware, as well as provide generalizability to a broad range of missions and data products. Initial analysis of this technique has been performed using magnetic field measurements from the Magnetospheric Multiscale Mission and CASSIOP, where it automatically identified scientifically interesting intervals containing Alfvén waves and EMIC activity.

Matthew G. Finley↗
DOE OSTI2023

Machine-learning-based spectral methods for partial differential equations

Spectral methods are an important part of scientific computing’s arsenal for solving partial differential equations (PDEs). However, their applicability and effectiveness depend crucially on the choice of basis functions used to expand the solution of a PDE. The last decade has seen the emergence of deep learning as a strong contender in providing efficient representations of complex functions. In the current work, we present an approach for combining deep neural networks with spectral methods to solve PDEs. In particular, we use a deep learning technique known as the Deep Operator Network (DeepONet) to identify candidate functions on which to expand the solution of PDEs. We have devised an approach that uses the candidate functions provided by the DeepONet as a starting point to construct a set of functions that have the following properties: (1) they constitute a basis, (2) they are orthonormal, and (3) they are hierarchical, i.e., akin to Fourier series or orthogonal polynomials. We have exploited the favorable properties of our custom-made basis functions to both study their approximation capability and use them to expand the solution of linear and nonlinear time-dependent PDEs. The proposed approach advances the state of the art and versatility of spectral methods and, more generally, promotes the synergy between traditional scientific computing and machine learning.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

Affine Transformations to Enable Machine Learning for Semi-Quantitative EDS Analysis

Energy Dispersive X-ray Spectroscopy (EDS) is an essential technique for determining elemental concentrations and distributions within microstructures, critical for materials discovery, optimization, and qualification. However, most published EDS data is qualitative because current quantitative EDS analysis methods require extensive calibration and post-processing, limiting their practicality and widespread adoption. This work seeks to establish a framework for accelerated EDS characterization and spectrum analysis that can leverage ML to analyze correlations between various elemental compositions and resulting EDS spectra. The complex physics and data result in a high-dimensional problem that grows exponentially with the number of elements in the system and the complexity of the spectrum analysis. ML provides a way to compute and optimize the results of this highly dimensional problem in a flexible way to tailor it to the user’s specific needs and material system. However, the framework emphasizes transparency through a strictly mathematical affine transformation, so the analysis remains understandable and reviewable to facilitate adoption by the scientific community. While currently implemented methods are simplistic and unvalidated, further development and demonstration of this framework could enable high-throughput, accurate, and accessible EDS characterization.

11 - NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2026

Affine Transformations to Correlate Experimental and Simulated EDS Spectra for Multi-Element Systems

Energy Dispersive X-ray Spectroscopy (EDS) is an essential technique for determining elemental concentrations and distributions within microstructures, critical for materials discovery, optimization, and qualification. However, most published EDS data is qualitative because current quantitative EDS analysis methods require extensive calibration and post-processing, limiting their practicality and widespread adoption. This work seeks to establish a framework for accelerated EDS characterization and spectrum analysis that can leverage ML to analyze correlations between various elemental compositions and resulting EDS spectra. The complex physics and data result in a high-dimensional problem that grows exponentially with the number of elements in the system and the complexity of the spectrum analysis. ML provides a way to compute and optimize the results of this highly dimensional problem in a flexible way to tailor it to the user’s specific needs and material system. However, the framework emphasizes transparency through a strictly mathematical affine transformation, so the analysis remains understandable and reviewable to facilitate adoption by the scientific community. While currently implemented methods are simplistic and unvalidated, further development and demonstration of this framework could enable high-throughput, accurate, and accessible EDS characterization.

11 - NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2023

WfCommons: Data Collection and Runtime Experiments using Multiple Workflow Systems

Scientific workflows have become ubiquitous across scientific fields, and their execution methods and systems continue to be the subject of research and development. Most experimental evaluations of these workflows rely on workflow instances, which can be either real-world or synthetic, to ensure relevance to current application domains or explore hypothetical/future scenarios. The WfCommons project addresses this need by providing data and tools to support such evaluations. In this paper, we present an overview of WfCommons and describe two recent developments. Firstly, we introduce a workflow execution "tracer" for Nextflow, which significantly enhances the set of real-world instances available in WfCommons. Secondly, we describe a workflow instance "translator" that enables the execution of any real-world or synthetic WfCommons workflow instance using Dask. Our contributions aim to provide researchers and practitioners with more comprehensive resources for evaluating scientific workflows.

Casanova, Henri↗
DOE OSTI2021

Accelerating Multigrid-based Hierarchical Scientific Data Refactoring on GPUs

Rapid growth in scientific data and a widening gap between computational speed and I/O bandwidth make it increasingly infeasible to store and share all data produced by scientific simulations. Instead, we need methods for reducing data volumes: ideally, methods that can scale data volumes adaptively so as to enable negotiation of performance and fidelity tradeoffs in different situations. Multigrid-based hierarchical data representations hold promise as a solution to this problem, allowing for flexible conversion between different fidelities so that, for example, data can be created at high fidelity and then transferred or stored at lower fidelity via logically simple and mathematically sound operations. However, the effective use of such representations has been hindered until now by the relatively high costs of creating, accessing, reducing, and otherwise operating on such representations. We describe here highly optimized data refactoring kernels for GPU accelerators that enable efficient creation and manipulation of data in multigrid-based hierarchical forms. We demonstrate that our optimized design can achieve up to 250 TB/s aggregated data refactoring throughput—83% of theoretical peak—on 1024 nodes of the Summit supercomputer. We showcase our optimized design by applying it to a large-scale scientific visualization workflow and the MGARD lossy compression software.

Data refactoring↗
DOE OSTI2021

Accelerating Multigrid-based Hierarchical Scientific Data Refactoring on GPUs

Rapid growth in scientific data and a widening gap between computational speed and I/O bandwidth make it increasingly infeasible to store and share all data produced by scientific simulations. Instead, we need methods for reducing data volumes: ideally, methods that can scale data volumes adaptively so as to enable negotiation of performance and fidelity tradeoffs in different situations. Multigrid-based hierarchical data representations hold promise as a solution to this problem, allowing for flexible conversion between different fidelities so that, for example, data can be created at high fidelity and then transferred or stored at lower fidelity via logically simple and mathematically sound operations. However, the effective use of such representations has been hindered until now by the relatively high costs of creating, accessing, reducing, and otherwise operating on such representations. We describe here highly optimized data refactoring kernels for GPU accelerators that enable efficient creation and manipulation of data in multigrid-based hierarchical forms. We demonstrate that our optimized design can achieve up to 250 TB/s aggregated data refactoring throughput—83% of theoretical peak—on 1024 nodes of the Summit supercomputer. We showcase our optimized design by applying it to a large-scale scientific visualization workflow and the MGARD lossy compression software.

Chen, Jieyang↗
NASA NTRS2004

Summary of the key features of seven biomathematical models of human fatigue and performance

BACKGROUND: Biomathematical models that quantify the effects of circadian and sleep/wake processes on the regulation of alertness and performance have been developed in an effort to predict the magnitude and timing of fatigue-related responses in a variety of contexts (e.g., transmeridian travel, sustained operations, shift work). This paper summarizes key features of seven biomathematical models reviewed as part of the Fatigue and Performance Modeling Workshop held in Seattle, WA, on June 13-14, 2002. The Workshop was jointly sponsored by the National Aeronautics and Space Administration, U.S. Department of Defense, U.S. Army Medical Research and Materiel Command, Office of Naval Research, Air Force Office of Scientific Research, and U.S. Department of Transportation. METHODS: An invitation was sent to developers of seven biomathematical models that were commonly cited in scientific literature and/or supported by government funding. On acceptance of the invitation to attend the Workshop, developers were asked to complete a survey of the goals, capabilities, inputs, and outputs of their biomathematical models of alertness and performance. Data from the completed surveys were summarized and juxtaposed to provide a framework for comparing features of the seven models. RESULTS: Survey responses revealed that models varied greatly relative to their reported goals and capabilities. While all modelers reported that circadian factors were key components of their capabilities, they differed markedly with regard to the roles of sleep and work times as input factors for prediction: four of the seven models had work time as their sole input variable(s), while the other three models relied on various aspects of sleep timing for model input. Models also differed relative to outputs: five sought to predict results from laboratory experiments, field, and operational data, while two models were developed without regard to predicting laboratory experimental results. All modelers provided published papers describing their models, with three of the models being proprietary. CONCLUSIONS: Although all models appear to have been fundamentally influenced by the two-process model of sleep regulation by Borbely, there is considerable diversity among them in the number and type of input and output variables, and their stated goals and capabilities.

Fatigue/physiopathology↗
NASA NTRS1989

A language comparison for scientific computing on MIMD architectures

Choleski's method for solving banded symmetric, positive definite systems is implemented on a multiprocessor computer using three FORTRAN based parallel programming languages, the Force, PISCES and Concurrent FORTRAN. The capabilities of the language for expressing parallelism and their user friendliness are discussed, including readability of the code, debugging assistance offered, and expressiveness of the languages. The performance of the different implementations is compared. It is argued that PISCES, using the Force for medium-grained parallelism, is the appropriate choice for programming Choleski's method on the multiprocessor computer, Flex/32.

Jones, Mark T.↗
DOE OSTI2024

Accelerating Multivariate Functional Approximation Computation with Domain Decomposition Techniques⋆

Modeling large datasets through Multivariate Functional Approximations (MFA) provide an elegant way to handle many visualization and scientific analysis workflows. The process necessitates scalable data partitioning methods to compute MFA representations efficiently without compromising the accuracy or continuity of the reconstructed solution. We propose a domain -decomposed method for computing the MFA with B -spline bases, which reduces the total work per task and uses a restricted Additive Schwarz (RAS) method to converge the control point data degrees -of -freedom along subdomain boundaries. We provide an in-depth analysis of the parallel approach with domain decomposition solvers, aiming to minimize local subdomain error residuals and recover high -order continuity at subdomain interfaces with appropriate choices of knot overlaps. The communication cost, determined by the overlap regions in the RAS implementation, is optimized to recover the numerical error profile of the single subdomain case. Our proposed method stands in contrast to previous methods, which typically only recover either C 0 or at best C 1 continuity for arbitrary B -spline degree expansions, or those that require post -processing to blend discontinuities in the reconstructed data. We demonstrate the effectiveness of our approach using analytical and real -world datasets in 1D, 2D, and 3D through both strong and weak scaling studies. The performance results indicate that the overall cost of computing the approximation is directly proportional to the underlying nearest -neighbor communication implementation, and is only weakly dependent on the overlap region size that determines the size of the messages. This finding underscores the efficiency and scalability of our proposed method, making it a promising solution for handling large datasets in scientific workflows.

additive Schwarz solvers↗
NASA NTRS2001

Design of Unstructured Adaptive (UA) NAS Parallel Benchmark Featuring Irregular, Dynamic Memory Accesses

We describe the design of a new method for the measurement of the performance of modern computer systems when solving scientific problems featuring irregular, dynamic memory accesses. The method involves the solution of a stylized heat transfer problem on an unstructured, adaptive grid. A Spectral Element Method (SEM) with an adaptive, nonconforming mesh is selected to discretize the transport equation. The relatively high order of the SEM lowers the fraction of wall clock time spent on inter-processor communication, which eases the load balancing task and allows us to concentrate on the memory accesses. The benchmark is designed to be three-dimensional. Parallelization and load balance issues of a reference implementation will be described in detail in future reports.

Feng, Hui-Yu↗
DOE OSTI2023

A Data Deposition Platform for Sharing Nuclear Magnetic Resonance Data

Nuclear magnetic resonance (NMR) data are rarely deposited in open databases, leading to loss of critical scientific knowledge. Existing data reporting methods (images, tables, lists of values) contain less information than raw data, and are poorly standardized. Together, these issues limit FAIR (findable, accessible, interoperable, reusable) access to these data, which in turn creates barriers for compound dereplication and the development of new data-driven discovery tools. Existing NMR databases are either not designed for natural products data, or employ complex deposition interfaces that disincentivize deposition. Journals, including the Journal of Natural Products (JNP), are now requiring data submission as part of the publication process, creating the need for a streamlined, user-friendly mechanism to deposit and distribute NMR data. Recently, our team reported the development of the Natural Products Magnetic Resonance Database (NP-MRD; www.np-mrd.org). Here in this paper we present a new data deposition platform for the NP-MRD project that is designed to enable users to deposit NMR data for published or submitted manuscripts in under five minutes. This platform includes a suite of automated data extraction and standardization tools, together with a simple-to-use web-based interface and detailed error reporting to simplify the data deposition process and is available at www.np-mrd.org/submissions.

96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗
DOE OSTI2023

Heterogenous electromediated depolymerization of highly crystalline polyoxymethylene

Abstract Post-consumer plastic waste in the environment has driven the scientific community to develop deconstruction methods that yield valued substances from these synthetic macromolecules. Electrocatalysis is a well-established method for achieving challenging transformations in small molecule synthesis. Here we present the first electro-chemical depolymerization of polyoxymethylene—a highly crystalline engineering thermoplastic (Delrin®)—into its repolymerizable monomer, formaldehyde/1,3,5-trioxane, under ambient conditions. We investigate this electrochemical deconstruction by employing solvent screening, cyclic voltammetry, divided cell studies, electrolysis with redox mediators, small molecule model studies, and control experiments. Our findings determine that the reaction proceeds via a heterogeneous electro-mediated acid depolymerization mechanism. The bifunctional role of the co-solvent 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) is also revealed. This study demonstrates the potential of electromediated depolymerization serving as an important role in sustainable chemistry by merging the concepts of renewable energy and circular plastic economy.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

A rule-free workflow for the automated generation of databases from scientific literature

Abstract In recent times, transformer networks have achieved state-of-the-art performance in a wide range of natural language processing tasks. Here we present a workflow based on the fine-tuning of BERT models for different downstream tasks, which results in the automated extraction of structured information from unstructured natural language in scientific literature. Contrary to existing methods for the automated extraction of structured compound-property relations from similar sources, our workflow does not rely on the definition of intricate grammar rules. Hence, it can be adapted to a new task without requiring extensive implementation efforts and knowledge. We test our data-extraction workflow by automatically generating a database for Curie temperatures and one for band gaps. These are then compared with manually curated datasets and with those obtained with a state-of-the-art rule-based method. Furthermore, in order to showcase the practical utility of the automatically extracted data in a material-design workflow, we employ them to construct machine-learning models to predict Curie temperatures and band gaps. In general, we find that, although more noisy, automatically extracted datasets can grow fast in volume and that such volume partially compensates for the inaccuracy in downstream tasks.

36 MATERIALS SCIENCE↗
DOE OSTI2023

The Zap Energy approach to commercial fusion

Zap Energy is a private fusion energy company developing the sheared-flow-stabilized (SFS) Z-pinch concept for commercial energy production. Spun out from the University of Washington, these experimental and computational efforts have resulted in devices with quasi-steady DD fusion yields above 109 per pulse. These devices support scaling toward energy breakeven on existing devices as well as beyond to commercially relevant engineering fusion gains. This article discusses the strategy behind Zap's development path, which is derived directly from the engineering and scientific elegance of the confinement method. Without need for external confinement or heating technologies, the SFS Z pinch relies on plasma self-organization. This compact magnetic confinement technology could, in turn, provide the basis for a cost-effective fusion power plant, vastly reduced in complexity from its competitors.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2023

Characterization of D-T Generator Neutron Flux Spectrum for Cyclic Neutron Activation Analysis Experiments

Improving nuclear data for short-lived fission product yields will further our fundamental understanding of fission, which is needed across various scientific fields and applications. One method of attaining the needed product yield data is through cyclic neutron activation, which allows a target to be irradiated in a neutron environment and then transported for counting of the radionuclides produced, typically via g spectroscopy. Recently, such a system has been constructed and commissioned at Pacific Northwest National Laboratory. Targets are shuttled between the head of a D-T neutron generator and a counting station with a transit time of 2 seconds. As part of the characterization of this system, the neutron flux was studied using two activation targets. The neutron flux from the deuterium-tritium fusion generator was determined to be 9.95(33) x10^8 n/cm2-s with a peak energy of 14.9 MeV and a spread of approximately 4 decades between the epithermal and 14 MeV peak group flux.

cyclic neutron activation, gamma-ray spectroscopy,↗
DOE OSTI2024

Toward a Holistic Performance Evaluation of Large Language Models Across Diverse AI Accelerators

Artificial intelligence (AI) methods have become critical in scientific applications to help accelerate scientific discovery. Large language models (LLMs) are being considered a promising approach to address some challenging problems because of their superior generalization capabilities across domains. The effectiveness of the models and the accuracy of the applications are contingent upon their efficient execution on the underlying hardware infrastructure. Specialized Al accelerator hardware systems have recently become available for accelerating Al applications. However, the comparative performance of these AI accelerators on large language models has not been previously studied. In this paper, we systematically study LLMs on multiple AI accelerators and GPUs and evaluate their performance characteristics for these models. We evaluate these systems with (i) a micro-benchmark using a core transformer block, (ii) a GPT-2 model, and (iii) an 1,I,M-driven science use case, GenSLM. We present our findings and analyses of the models' performance to better understand the intrinsic capabilities of AI accelerators. Furthermore, our analysis takes into account key factors such as sequence lengths, scaling behavior, and sensitivity to gradient accumulation steps.

Emani, Murali↗
DOE OSTI2021

Characterizing Machine Learning I/O Workloads on Leadership Scale HPC Systems

High performance computing (HPC) is no longer solely limited to traditional workloads such as simulation and modeling. With the increase in the popularity of machine learning (ML) and deep learning (DL) technologies, we are observing that an increasing number of HPC users are incorporating ML methods into their workflow and scientific discovery processes, across a wide spectrum of science domains such as biology, earth science, and physics. This gives rise to a diverse set of I/O patterns than the traditional checkpoint/restart-based HPC I/O behavior. The details of the I/O characteristics of such ML I/O workloads have not been studied extensively for large-scale leadership HPC systems. This paper aims to fill that gap by providing an in-depth analysis to gain an understanding of the I/O behavior of ML I/O workloads using darshan - an I/O characterization tool designed for lightweight tracing and profiling. We study the darshan logs of more than 23, 000 HPC ML I/O jobs over a time period of one year running on Summit - the second-fastest supercomputer in the world. This paper provides a systematic I/O characterization of ML I/O jobs running on a leadership scale supercomputer to understand how the I/O behavior differs across science domains and the scale of workloads, and analyze the usage of parallel file system and burst buffer by ML I/O workloads.

Paul, Arnab↗