Search NASA⌕ Search

SEARCH · Search NASA

Results for “Data reduction methods”

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 937 records · Page 52

Improving Thermal Model Prediction Through Statistical Analysis of Irradiation and Post-Irradiation Data from AGR Experiments

As part of the Research and Development program for Next Generation High Temperature Reactors (HTR), a series of irradiation tests, designated as Advanced Gas-cooled Reactor (AGR), have been defined to support development and qualification of fuel design, fabrication process, and fuel performance under normal operation and accident conditions. The AGR tests employ fuel compacts placed in a graphite cylinder shrouded by a steel capsule and instrumented with thermocouples (TC) embedded in graphite blocks enabling temperature control. The data representing the crucial test fuel conditions (e.g., temperature, neutron fast fluence, and burnup) while impossible to obtain from direct measurements are calculated by physics and thermal models. The irradiation and post-irradiation examination (PIE) experimental data are used in model calibration effort to reduce the inherent uncertainty of simulation results. This paper is focused on fuel temperature predicted by the ABAQUS code’s finite element-based thermal models. The work follows up on a previous study, in which several statistical analysis methods were adapted, implemented in the NGNP Data Management and Analysis System (NDMAS), and applied for improving qualification of AGR-1 thermocouple data. The present work exercises the idea that the abnormal trends of measured data observed from statistical analysis may be caused by either measuring instrument deterioration or physical mechanisms in capsules that may have shifted the system thermal response. As an example, the uneven reduction of the control gas gap in Capsule 5 revealed by the capsule metrology measurements in PIE helps justify the reduction in TC readings instead of TC drift. This in turn prompts modification of thermal model to better fit with experimental data, thus help increase confidence, and in other word reduce model uncertainties in thermal simulation results of the AGR-1 test.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Diagnostic methods for a high efficiency exhaust aftertreatment system

A method includes interpreting NOx data indicative of an amount of NOx exiting an engine and an amount of NOx exiting an exhaust aftertreatment system coupled to the engine, determining a NOx conversion efficiency fault is present based on the amount of NOx exiting the engine and the amount of NOx exiting the exhaust aftertreatment system, and determining at least one of a selective catalytic reduction (SCR) catalyst of the exhaust aftertreatment system and a diesel particulate filter having a coating of a SCR reaction catalyst (DPF-SCR) of the exhaust aftertreatment system are responsible for the NOx conversion efficiency fault based on at least one of a reductant slip amount and a NOx conversion value across at least one of the SCR catalyst and the DPF-SCR.

42 ENGINEERING↗

Data for Monitoring Fracture Saturation with Internal Transportable Seismic Sources and Twin Neural Networks

Seismic coda-wave analysis is a well-developed method for detecting subtle physical changes in complex media by measuring arrival times in the late-arriving energy from multiply-scattered or reflected waves. However, a challenge arises when multiply-scattered waves are not sufficiently separated in time from the direct arrivals to provide a clear coda wave train. Additional complications for monitoring changes in fracture systems arise when the signals originate from unsynchronized internal sources, such as natural induced seismicity, from acoustic emission, or from transportable intra-fracture sources (chattering dust), that generate uncontrolled signals that vary in arrival time, amplitude and frequency content. Here, we use a twin neural network (also known as a Siamese neural network) for dimensionality reduction to analyze signals from chattering dust to classify the fluid saturation state of a synthetic fracture system. The twin neural network with shared weights generates a low-dimensional representation of the data input by minimizing contrastive loss, serving as the input to a multiclass classifier that accurately classifies whether multiple fractures in a fracture system are fully saturated or partially saturated, or whether a change in saturation has occurred in different fractures in the system. These results show that information buried in unresolved codas from uncontrolled sources can be extracted using machine learning to monitor the evolution of fracture systems caused by physical and chemical processes even when the scattered and direct wave fields overlap.

58 GEOSCIENCES↗

Evaluation of U.S. Building Energy Benchmarking and Transparency Programs: Attributes, Impacts, and Best Practices

In the last decade, a new policy area has emerged to boost energy efficiency in buildings that focuses on the simple action of measuring energy use as compared to buildings of similar type and size, and making that data publicly available. These efforts, referred to as benchmarking and transparency (B&T) policies, seek to unlock new energy efficiency opportunities in the country’s existing buildings by promoting data-driven decision-making and creating stronger market signals. This report focuses on the 24 state and local jurisdictions that (as of December 31, 2016) require owners of privately owned commercial buildings, multifamily buildings, or both to comply with a B&T policy. The report provides a summary of U.S. B&T policy design and implementation characteristics, reports results and impacts for jurisdictions with B&T policies, and discusses opportunities for increasing the efficacy of B&T policies, as well as suggested areas for further research. Among the findings, all but one of the B&T policy evaluation studies reviewed indicate some reduction (from 1.6% to 14%) in energy use, energy costs, or energy intensity over the two- to four-year period of the analyses. More specifically, most of the studies reviewed indicate 3% to 8% reductions in gross energy consumption or energy use intensity over a two- to four-year period of B&T policy implementation. Two additional evaluation studies indicate that there is a causal relationship between B&T policies and energy savings or energy cost savings. These documented impacts should be reviewed with some caution. While consistently showing energy savings benefits associated with B&T policies, these savings estimates should be considered preliminary because of the limited period of analyses and inconsistencies in analysis methods for the various studies. A nationally standardized method for data collection, reporting, and evaluation of B&T policies—developed with an advisory group of state and local jurisdictions, energy efficiency and evaluation experts, building owner and real estate associations, and other stakeholders—could improve the consistency and quality of B&T impact studies, providing policymakers and others with a more complete understanding of the present and future impacts of these policies.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Converging Stereotactic Radiotherapy Using Kilovoltage X-Rays: Experimental Irradiation of Normal Rabbit Lung and Dose-Volume Analysis With Monte Carlo Simulation

Purpose: To validate the feasibility of developing a radiotherapy unit with kilovoltage X-rays through actual irradiation of live rabbit lungs, and to explore the practical issues anticipated in future clinical application to humans through Monte Carlo dose simulation. Methods and Materials: A converging stereotactic irradiation unit was developed, consisting of a modified diagnostic computed tomography (CT) scanner. A tiny cylindrical volume in 13 normal rabbit lungs was individually irradiated with single fractional absorbed doses of 15, 30, 45, and 60 Gy. Observational CT scanning of the whole lung was performed every 2 weeks for 30 weeks after irradiation. After 30 weeks, histopathologic specimens of the lungs were examined. Dose distribution was simulated using the Monte Carlo method, and dose-volume histograms were calculated according to the data. A trial estimation of the effect of respiratory movement on dose distribution was made. Results: A localized hypodense change and subsequent reticular opacity around the planning target volume (PTV) were observed in CT images of rabbit lungs. Dose-volume histograms of the PTVs and organs at risk showed a focused dose distribution to the target and sufficient dose lowering in the organs at risk. Our estimate of the dose distribution, taking respiratory movement into account, revealed dose reduction in the PTV. Conclusions: A converging stereotactic irradiation unit using kilovoltage X-rays was able to generate a focused radiobiologic reaction in rabbit lungs. Dose-volume histogram analysis and estimated sagittal dose distribution, considering respiratory movement, clarified the characteristics of the irradiation received from this type of unit.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Validation of hydrogen gas stratification and mixing models

Two validation benchmarks confirm that the BMIX++ code is capable of simulating unintended hydrogen release scenarios efficiently. The BMIX++ (UC Berkeley mechanistic MIXing code in C++) code has been developed to accurately and efficiently predict the fluid mixture distribution and heat transfer in large stratified enclosures for accident analyses and design optimizations. The BMIX++ code uses a scaling based one-dimensional method to achieve large reduction in computational effort compared to a 3-D computational fluid dynamics (CFD) simulation. Two BMIX++ benchmark models have been developed. One is for a single buoyant jet in an open space and another is for a large sealed enclosure with both a jet source and a vent near the floor. Both of them have been validated by comparisons with experimental data. Excellent agreements are observed. The entrainment coefficients of 0.09 and 0.08 are found to fit the experimental data for hydrogen leaks with the Froude number of 99 and 268 best, respectively. In addition, the BIX++ simulation results of the average helium concentration for an enclosure with a vent and a single jet agree with the experimental data within a margin of about 10% for jet flow rates ranging from 1.21 × 10⁻⁴ to 3.29 × 10⁻⁴ m³/s. In conclusion, computing time for each BMIX++ model with a normal desktop computer is less than 5 min.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Quality Assurance Analysis of a Large Multicenter Practice: Does Increased Complexity of Intensity-Modulated Radiotherapy Lead to Increased Error Frequency?

Purpose: Error reduction is an important concern in clinical medicine. Intensity-modulated radiotherapy (IMRT) is an important advancement in radiation oncology that increases the complexity of treatment, potentially increasing the error risk. We studied the frequency and severity of errors in a large multicenter practice to ascertain the impact of quality improvement interventions over time, IMRT, and type of practice. Methods and Materials: We analyzed prospective data from three academic and 16 community practice sites with 24,775 courses of radiotherapy (9,210 IMRT courses and 15,565 non-IMRT) between January 2006 and December 2009. All IMRT treatment was performed using one centralized dose planning center for all sites. Results: We prospectively identified various errors or potential errors in 0.14 % vs. 0.40 % of the IMRT vs. non-IMRT courses (13/9,210 vs. 62/15,565, p = 0.0004) and excluding potential errors: 0.03 % for IMRT vs. 0.21% for non-IMRT. We developed the Clinical Radiotherapy Error Severity Scale (CRESS) to classify error severity from 1 to 10, with 1 to 3 for potential or completely correctable errors, 4 to 5 for dose variations <5%, and 6 to 10 for dose variations >5%. Multivariate analyses of CRESS values, severity >4, and any error (including potential) correlated significantly reduced errors with IMRT (p = 0.0001-0.0024) but found no significant difference between the academic and community practice sites and no change in error frequency over time despite implementation of 39 system-wide policy changes by the centralized quality improvement committee. Conclusions: Despite the increase in complexity with IMRT compared with conventional radiotherapy, it can be delivered with reduced error frequency.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

γ-Ray spectroscopy using a binned likelihood approach

The measurement of a reaction cross section from a pulse height spectrum is a ubiquitous problem in experimental nuclear physics. In γ-ray spectroscopy, this is accomplished frequently by measuring the intensity of full-energy primary transition peaks and correcting the intensities for experimental artifacts, such as detection efficiencies and angular correlations. Implicit in this procedure is the assumption that full-energy peaks do not overlap with any secondary peaks, escape peaks, or environmental back- grounds. However, for complex γ-ray cascades, this is often not the case. Furthermore, this technique is difficult to adapt for coincidence spectroscopy, where intensities depend not only on the detection efficiency, but also the detailed decay scheme. We present a method that incorporates the intensities of the entire spectrum (e.g., primary and secondary transition peaks, escape peaks, Compton continua, etc.) into a statistical model, where the transition intensities and branching ratios can be determined using Bayesian statistical inference. This new method provides an elegant solution to the difficulties associated with analyzing coincidence spectra. Finally, we describe it in detail and examine its efficacy in the analysis of 18 O(p,γ) 19 F and 25 Mg(p,γ) 26 Al resonance data. For the 18 O(p,γ) 19 F reaction, the measured branching ratios improve upon the literature values, with a factor of 3 reduction in the uncertainties.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Framework of compressive sensing and data compression for 4D-STEM

Four-dimensional Scanning Transmission Electron Microscopy (4D-STEM) is a powerful technique for high-resolution and high-precision materials characterization at multiple length scales, including the characterization of beam-sensitive materials. However, the field of view of 4D-STEM is relatively small, which in absence of live processing is limited by the data size required for storage. Furthermore, the rectilinear scan approach currently employed in 4D-STEM places a resolution- and signal-dependent dose limit for the study of beam sensitive materials. Improving 4D-STEM data and dose efficiency, by keeping the data size manageable while limiting the amount of electron dose, is thus critical for broader applications. Here we introduce a general method for reconstructing 4D-STEM data with subsampling in both real and reciprocal spaces at high fidelity. The approach is first tested on the subsampled datasets created from a full 4D-STEM dataset, and then demonstrated experimentally using random scan in real-space. The same reconstruction algorithm can also be used for compression of 4D-STEM datasets, leading to a large reduction (100 times or more) in data size, while retaining the fine features of 4D-STEM imaging, for crystalline samples.

4D-STEM↗

Leveraging unlabeled SEM datasets with self-supervised learning for enhanced particle segmentation

Scanning Electron Microscopes (SEMs) are widely used in experimental science laboratories, often requiring cumbersome and repetitive user analysis. Automating SEM image analysis processes is highly desirable to address this challenge. In particle sample analysis, Machine Learning (ML) has emerged as the most effective approach for particle segmentation. However, the time-intensive process of manually annotating thousands of SEM images limits the applicability of supervised learning approaches. Self-Supervised Learning (SSL) offers a promising alternative by enabling knowledge extraction from raw, unlabeled data. This study presents a framework for evaluating SSL techniques in SEM image analysis, focusing on novel methods leveraging the ConvNeXtV2 architecture for particle detection. A dataset comprising 25,000 SEM images is curated to benchmark these proposed SSL methods. The results demonstrate that ConvNeXtV2 models, with varying parameter counts, consistently outperform other techniques in particle detection across different length scales, achieving up to a 34% reduction in relative error compared to established SSL methods. Furthermore, an ablation study explores the relationship between dataset size and SSL performance, providing actionable insights for practitioners regarding model selection and resource efficiency. This research advances the integration of SSL into autonomous analysis pipelines and supports its application in accelerating materials science discovery.

Rettenberger, Luca↗

The maximum extent of the filaments and sheets in the cosmic web: an analysis of the SDSS DR17

Filaments and sheets are striking visual patterns in cosmic web. The maximum extent of these large-scale structures are difficult to determine due to their structural variety and complexity. We construct a volume-limited sample of galaxies in a cubic region from the SDSS, divide it into smaller subcubes and shuffle them around. We quantify the average filamentarity and planarity in the 3D galaxy distribution as a function of the density threshold and compare them with those from the shuffled realizations of the original data. The analysis is repeated for different shuffling lengths by varying the size of the subcubes. The average filamentarity and planarity in the shuffled data show a significant reduction when the shuffling scales are smaller than the maximum size of the genuine filaments and sheets. We observe a statistically significant reduction in these statistical measures even at a shuffling scale of $\sim 130 \, {{\, \rm Mpc}}$, indicating that the filaments and sheets in three dimensions can extend up to this length scale. They may extend to somewhat larger length scales that are missed by our analysis due to the limited size of the SDSS data cube. We expect to determine these length scales by applying this method to deeper and larger surveys in future.

79 ASTRONOMY AND ASTROPHYSICS↗

Privacy-Preserving Artificial Intelligence on Edge Devices: A Homomorphic Encryption Approach

Recent advancements in privacy-preserving artificial intelligence (AI) have paved the way for enhanced privacy in computational processes. A standing challenge, however, is the robust privacy preservation in AI algorithms, especially when integrated into edge devices and Internet-of-Thing (IoT) infrastructures. Most prevailing solutions have adopted traditional encryption methods which, though secure, often introduce significant overhead and potential dips in accuracy. In this study, we put forth an innovative approach, utilizing the CKKS encryption scheme, aiming to harmoniously balance computational efficiency with stringent data privacy. By harnessing the capabilities of Full Homomorphic Encryption (FHE) under the CKKS scheme, we ensure the preservation of privacy, successfully curbing the inherent noise traditionally linked with accuracy reductions in similar encryption-oriented solutions. Through comprehensive experiments, our approach showcased its potential as a strong contender for privacy preservation, demonstrating commendable performance across all tests, affirming that FHE is indeed viable for devices with constrained computational power and energy resources.

Khan, Muhammad Jahanzeb↗

Compressed basis GMRES on high-performance graphics processing units

Krylov methods provide a fast and highly parallel numerical tool for the iterative solution of many large-scale sparse linear systems. To a large extent, the performance of practical realizations of these methods is constrained by the communication bandwidth in current computer architectures, motivating the investigation of sophisticated techniques to avoid, reduce, and/or hide the message-passing costs (in distributed platforms) and the memory accesses (in all architectures). This article leverages Ginkgo’s memory accessor in order to integrate a communication-reduction strategy into the (Krylov) GMRES solver that decouples the storage format (i.e., the data representation in memory) of the orthogonal basis from the arithmetic precision that is employed during the operations with that basis. Given that the execution time of the GMRES solver is largely determined by the memory accesses, the cost of the datatype transforms can be mostly hidden, resulting in the acceleration of the iterative step via a decrease in the volume of bits being retrieved from memory. Together with the special properties of the orthonormal basis (whose elements are all bounded by 1), this paves the road toward the aggressive customization of the storage format, which includes some floating-point as well as fixed-point formats with mild impact on the convergence of the iterative process. We develop a high-performance implementation of the “compressed basis GMRES” solver in the Ginkgo sparse linear algebra library using a large set of test problems from the SuiteSparse Matrix Collection. We demonstrate robustness and performance advantages on a modern NVIDIA V100 graphics processing unit (GPU) of up to 50% over the standard GMRES solver that stores all data in IEEE double-precision.

97 MATHEMATICS AND COMPUTING↗

A Multi-Objective, Hub-and-Spoke Supply Chain Design Model for Densified Biomass

In this paper we propose a model to design the supply chain for densified biomass. Rail is typically used for longhaul, high-volume shipment of densified biomass. This is the reason why a hub-and-spoke network structure is used to model this supply chain. The model is formulated as a multi-objective, mixed-integer programing problem under economic, environmental, and social criteria. The goal is to identify the feasibility of meeting the Renewable Fuel Standard (RFS) by using biomass for production of cellulosic ethanol. The focus is not just on the costs associated with meeting these standards, but also exploring the social and environmental benefits that biomass production and processing offers by creating new jobs and reducing greenhouse gas (GHG) emissions. We develop an augmented ?-constraint method to find the exact Pareto solution to this optimization problem. We develop a case study using data from the Mid-West. The model identifies the number, capacity and location of biorefineries needed to make use of the biomass available in the region. The model estimates the delivery cost of cellulosic ethanol under different scenario, the number new jobs created and the GHG emission reductions in the supply chain.

09 BIOMASS FUELS↗

SU-E-T-08: A Beam Source Model for Monte Carlo Simulations of a Double-Scattering Proton Beam Delivery System Using Beam Current Modulation

Purpose: To reconstruct phase-space information upstream of patient specific collimators for Monte Carlo simulations using only radiotherapy planning system data. Methods: The proton energies are calculated based on residual ranges, e.g., sum of prescribed ranges in a patient and SSD. The Kapchinskij and Vladimirskij (KV) distribution was applied to sample proton’s x-y positions and momentum direction and the beam shape was assumed to be a circle. Free parameters, e.g., the initial energy spread and the emittance of KV distribution were estimated from the benchmarking with commissioning data in a commercial treatment planning system for an operational proton therapy center. The number of histories, which defines the height of individual pristine Bragg peaks (BP) of Spread-out Bragg peak (SOBP), are weighted based on beam current modulation and a correction factor is applied to take into account the fluence reduction as the residual range decreases due to the rotation of the range modulator wheel. The timedependent behaviors, e.g., the changes of the residual range and histories per a pristine BP, are realized by utilizing TOPAS (Tool for Particle Simulation). Results: Benchmarking simulations for selected SOBPs ranging 7.5 cm to 15.5 cm matched within 2 mm in range and up to 5 mm in SOBP width against measurement data in water phantom. We found this model tends to underestimate entrance dose by about 5 % in comparison to measurement. This was attributed to the situation that the energy distribution used in the model was limited in its granularity at the limit of single energy spectrum for the narrow angle modulator steps used in the proximal pull back region of the SOBPs. Conclusion: Within these limitations the source modeling method proved itself an acceptable alternative of a full treatment head simulation when the machine geometry and materials information are not available.

60 APPLIED LIFE SCIENCES↗

Data-driven approach to parameterize SCAN + U for an accurate description of 3 d transition metal oxide thermochemistry

Semilocal density-functional theory (DFT) methods exhibit significant errors for the phase diagrams of transition-metal oxides that are caused by an incorrect description of molecular oxygen and the large self-interaction error in materials with strongly localized electronic orbitals. Empirical and semiempirical corrections based on the DFT + U method can reduce these errors, but the parameterization and validation of the correction terms remains an on-going challenge. Here we develop a systematic methodology to determine the parameters and to statistically assess the results by considering interlinked thermochemical data across a set of transition metal compounds. We consider three interconnected levels of correction terms: (1) a constant oxygen binding correction, (2) Hubbard-U correction, and (3) DFT/DFT + U compatibility correction. The parameterization is expressed as a unified optimization problem. We demonstrate this approach for 3d transition metal oxides, considering a target set of binary and ternary oxides. With a total of 37 measured formation enthalpies taken from the literature, the dataset is augmented by the reaction energies of 1710 unique reactions that were derived from the formation energies by systematic enumeration. To ensure a balanced dataset across the available data, the reactions were grouped by their similarity using clustering and suitably weighted. The parameterization is validated using leave-one-out cross validation (CV), a standard technique for the validation of statistical models. We apply the methodology to the strongly constrained and appropriately normed (SCAN) density functional. Based on the CV score, the error of binary (ternary) oxide formation energies is reduced by 40% (75%) to 0.10 (0.03) eV/atom. A simplified correction scheme that does not involve SCAN/SCAN + U compatibility terms still achieves an error reduction of 30% (25%). The method and tools demonstrated here can be applied to other classes of materials or to parameterize the corrections to optimize DFT + U performance for other target physical properties.

36 MATERIALS SCIENCE↗

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↗