Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Edge computing”

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 1,639 records · Page 91

Dark counts in optical superconducting transition-edge sensors for rare-event searches

Superconducting transition-edge sensors (TESs) are a type of quantum sensor known for their high single-photon detection efficiency and low background. This makes TESs ideal for particle-physics experiments searching for rare events. In this work, we present a comprehensive characterization of the background in optical TESs, distinguishing three types of events: electrical-noise, high-energy, and photonlike events. We introduce computational methods to automate the classification of events. We experimentally verify and simulate the source of the high-energy events. We also isolate the photonlike events, the expected signal in dielectric haloscopes searching for dark-matter dark photons, and achieve a photonlike dark-count rate of 3.6× 10 − 4 in the 0.8–3.2 eV energy range.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Revolutionizing Turbine Cooling with Micro-Architectures Enabled by Direct Metal Laser Sintering

The objective of this research effort was to explore innovative cooling architectures enabled by additive manufacturing techniques for improved turbine cooling performance. The ability to create complex internal geometries was leveraged to better distribute coolant as well as to integrate inherently unstable flow devices to enhance internal and external heat transfer. This was accomplished with a multi-faceted approach including analytical, experimental, and computational components. This final report documents progress during the total 4.25 year effort that was extended (at no additional cost) from the original 3-year cooperative agreement. Prior to starting the project, significant effort was invested in investigating innovative cooling designs from the open literature. When the project started, we quickly down-selected to 4 configurations: impingement jet fluidic oscillators (for internal leading edge cooling), reverse film cooling (for pressure surface), sweeping film cooling jets (for suction surface), and trailing edge slot cooling with microchannels. During the 2nd year, the reverse film cooling design was shown to be sub-optimal and work on that topic was halted. The three remaining technologies were integrated into a large scale nozzle guide vane and installed into a low-speed linear cascade facility for further interrogation. Initial results showed the sweeping film cooling jets (on the suction surface) to be the most promising technology compared to a baseline diffusion shaped film hole (777 design). The benefit was particularly evident at high blowing ratios (>1.5) when the 777 coolant film separated from the downstream surface. Benefits were also evident at elevated freestream turbulence levels. The leading edge cooling with unsteady jets was less effective in terms of peak or average cooling – however it was superior to round hole impingement cooling in terms of spatial uniformity. Finally, the trailing edge cooling design with micro-channels was scrapped in favor of a pinned arrangement with centerbody. During the 3rd year, the same three technologies were integrated into a transonic linear cascade for an assessment of compressibility effects. Again, the sweeping film cooling jets proved superior at high blowing ratios while the unsteady leading edge impingement and trailing edge pin-fins with centerbody designs yielded mixed results. Finally, during the 4th and final year, these 3 technologies were integrated into a direct metal laser sintered (DMLS) nozzle guide vane for testing in a high temperature, transonic annular vane cascade. The facility matches the flow temperature, Mach number, and coolant pressure ratios of an actual gas turbine. Before doing this, an additional series of tests were completed to validate a redesigned trailing edge cooling architecture. This final design included a centerbody with triangular pins between the centerbody and the vane external skin. The majority of these triangular pins are fabricated with a 30% gap to the centerbody – reducing their pressure drop considerably while still providing excellent heat transfer augmentation. This additional test campaign required an additional 3-month extension request to complete testing in the high temperature NGV test facility (Turbine Reacting Flow Rig – TuRFR). Testing of the DMLS vane in TuRFR pitted the 3 innovative cooling technologies against more traditional technologies (shaped 777 film hole, round impingement jet, and full pin arrangement in the trailing edge. The sweeping film cooling jets (on the suction surface) showed approximately 15% improvement in overall effectiveness compared to the 777 film hole while the trailing edge showed up to 20% improvement. The leading edge impingement was disappointing with a substantially lower effectiveness than traditional direct impingement. In summary, the study was successful in demonstrating that DMLS-enabled cooling technologies can yield significant gains in cooling performance.

03 NATURAL GAS↗

Radiation-Induced Noise Resilience of Neuromorphic Architectures

Neuromorphic event-based networks use asynchronous time-dependent information to extract features from input data that can allow for edge-based distributed applications such as object recognition. The noise resilience properties of such networks, especially in the context of space applications, are yet to be explored. In this paper, we use the hierarchy of time surfaces (HOTS) algorithm, which is one of the neuromorphic algorithms, to understand the least and most resilient modules in a neuromorphic network. The HOTS algorithm relies on the computing of time surfaces that maps the temporal delays between neighboring pixels into normalized features that involve many computations that are also found in other neuromorphic networks such as exponential decays, distance computations, etcetera. We implemented HOTS on a Digilent PYNQ board with a Xilinx Zynq 7020 system on a chip, and we subjected the boards running the HOTS network inference to neutron radiation at the Los Alamos Neutron Science Center. Furthermore, we used simulation models from our previous similar experiments on the event-based sensor to create a neutron induced noise model to quantify the effect of this noise on the overall performance of the network. This experiment provides the preliminary measurements of the reliability of the HOTS algorithm and proposes methods to create a more reliable HOTS architecture in future spacecraft missions.

Engineering↗

Conformal refinement of unstructured quadrilateral meshes

We present a multilevel adaptive refinement technique for unstructured quadrilateral meshes in which the mesh is kept conformal at all times. This means that the refined mesh, like the original, is formed of only quadrilateral elements that intersect strictly along edges or at vertices, i.e., vertices of one quadrilateral element do not lie in an edge of another quadrilateral. Elements are refined using templates based on 1:3 refinement of edges. We demonstrate that by careful design of the refinement and coarsening strategy, we can maintain high quality elements in the refined mesh. We demonstrate the method on a number of examples with dynamically changing refinement regions.

99 GENERAL AND MISCELLANEOUS↗

An Efficient Sweep-Based Solver for the S N Equations on High-Order Meshes

In this study, we propose a graph-based sweep algorithm for solving the steady-state, monoenergetic discrete ordinates on meshes of high-order (HO) curved mesh elements. Our spatial discretization consists of arbitrarily HO discontinuous Galerkin finite elements using upwinding at mesh element faces. To determine mesh element sweep ordering, we define a directed, weighted graph whose vertices correspond to mesh elements and whose edges correspond to mesh element upwind dependencies. This graph is made acyclic by removing select edges in a way that approximately minimizes the sum of removed edge weights. Once the set of removed edges is determined, transport sweeps are performed by lagging the upwind dependency associated with the removed edges. The proposed algorithm is tested on several two-dimensional and three-dimensional meshes composed of HO curved mesh elements.

97 MATHEMATICS AND COMPUTING↗

A Generalization of Threshold-Based and Probability-Based Models of Information Diffusion

Diffusion of information through complex networks is of interest in studies such as propagation prediction and influence maximization, both of which have applications in viral marketing and rumor controlling. There are a variety of information diffusion models, all of which simulate the adoption and spread of information over time. However, there is a lack of understanding of whether, despite their conceptual differences, these models represent the same underlying generative structures. For instance, if two different models utilize different conceptual mechanisms, but generate the same results, does the choice of model matter? A classification of diffusion of information models is developed based on the neighbor knowledge of the model infection requirement and the stochasticity of the model. This classification allows for the identification of models that fall into each respective category. The study involves the analysis of the following agent-based models on directed scale-free networks: (1) a linear absolute threshold model (LATM), (2) a linear fractional threshold model (LTFM), (3) the independent cascade model (ICM), (4) Bass-Rand-Rust model (BRRM) (5) a stochastic linear absolute threshold model (SLATM) (6) a stochastic fractional threshold model (SLFTM), and (7) Dodds–Watts model (DWM). Through the execution of simulations and analysis of the experimental results, the distinctive properties of each model are identified. Our analysis reveals that similarity in conceptual design does not imply similarity in behavior concerning speed, final state of nodes and edges, and sensitivity to parameters. Therefore, we highlight the importance of considering the unique behavioral characteristics of each model when selecting a suitable information diffusion model for a particular application.

97 MATHEMATICS AND COMPUTING↗

Quantum Random Number Generator (QRNG)

The Los Alamos Quantum Random Number Generator (QRNG) is a hardware-based, high-performance Random Number Generator capable of generating 200 Mbit/s or more of true random numbers. Like flipping a coin, it is very much random and essential for information security like encrypting data on the internet, checking email, or purchasing something from an online vendor. The device harvests entropy from fluctuations in an optical source that arise from quantum mechanical properties of light. Qrypt, Inc., a company launched in 2017, began making strategic investments and developing partnerships to advance cutting-edge quantum hardware solutions. One of those key investments was licensing QRNG from Los Alamos and subsequently collaborating with advanced quantum materials and technology researcher Dr. Raymond Newell to create high-quality random keys at scale.

97 MATHEMATICS AND COMPUTING↗

Intelligent Experiments through Real-Time AI: Fast Data Processing and Autonomous Detector Control for High-Energy Nuclear Experiments

The aim of this project is to develop software and hardware for fast real-time data processing and autonomous detector control and calibration for the sPHENIX and the future EIC experiments. Below summarizes Georgia Tech team efforts in the past year: 1. We developed a real-time clustering algorithm and FPGA-based pipeline architecture for processing fired pixel data from ALPIDE sensors in sPHENIX experiments. Our Columnar Clustering Co-Design introduces a hardware-aware, stream-friendly approach that segments pixel data by column pairs using a Column Pair Clustering (CPC) strategy, followed by Cluster Stitching to merge adjacent subclusters. Implemented in Vitis HLS, the pipeline comprises five stages—read-in, subclustering, stitching, analysis, and write-out—connected by tagged HLS streams with custom end-of-event signaling for robust synchronization. We designed a pipelined dataflow model optimized for throughput, low latency, and minimal buffering, enabling scalable clustering across events of arbitrary size. Our system maintains spatial precision via center-of-mass and shape key extraction and efficiently handles edge cases such as fragmented or nested clusters. Compared against DBSCAN in both software and hardware, our approach demonstrates competitive performance under FPGA constraints. 2. We also conducted a comprehensive algorithm-to-hardware co-design of connected component analysis tailored for sPHENIX experiments, focusing on real-time, low-latency processing using FPGAs and High-Level Synthesis (HLS). Starting from a Python-based particle tracking pipeline, the team translated the core logic—graph traversal via DFS and Union-Find—into an HLS-compatible C++ model, replacing dynamic memory and recursion with static arrays and pipelined control flow. The final design includes a fully streamed and dataflow-compatible Union-Find kernel optimized across five iterations, incorporating loop pipelining, array partitioning, AXI/FIFO interface tuning, and function flattening. Experimental results show up to 14.8× speedup over the CPU baseline, reducing per-graph latency to 1.58 μs and demonstrating strong resource efficiency with only ~7k LUTs and zero BRAM usage. The design maintains functional correctness against the Python reference using a Python-based C-simulation framework and Mean Squared Error metrics. This work validates the potential of HLS-driven FPGA designs for edge-level HEP data acquisition, laying a scalable foundation for future integration with real-time detector pipelines and multi-graph processing systems.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Image registration method for medical image sequences

Image registration of low contrast image sequences is provided. In one aspect, a desired region of an image is automatically segmented and only the desired region is registered. Active contours and adaptive thresholding of intensity or edge information may be used to segment the desired regions. A transform function is defined to register the segmented region, and sub-pixel information may be determined using one or more interpolation methods.

97 MATHEMATICS AND COMPUTING↗

A front-end readout Detector Board for the OpenPET electronics system

Here, we present a 16-channel front-end readout board for the OpenPET electronics system. A major task in developing a nuclear medical imaging system, such as a positron emission computed tomograph (PET) or a single-photon emission computed tomograph (SPECT), is the electronics system. While there are a wide variety of detector and camera design concepts, the relatively simple nature of the acquired data allows for a common set of electronics requirements that can be met by a flexible, scalable, and high-performance OpenPET electronics system. The analog signals from the different types of detectors used in medical imaging share similar characteristics, which allows for a common analog signal processing. The OpenPET electronics processes the analog signals with Detector Boards. Here we report on the development of a 16-channel Detector Board. Each signal is digitized by a continuously sampled analog-to-digital converter (ADC), which is processed by a field programmable gate array (FPGA) to extract pulse height information. A leading edge discriminator creates a timing edge that is "time stamped" by a time-to-digital converter (TDC) implemented inside the FPGA. In conclusion, this digital information from each channel is sent to an FPGA that services 16 analog channels, and then information from multiple channels is processed by this FPGA to perform logic for crystal lookup, DOI calculation, calibration, etc.

47 OTHER INSTRUMENTATION↗

Effect of pre-cold rolling on the evolution of microstructure, microtexture, and mechanical properties of the mushy state rolled in-situ Al-4.5Cu-5TiB2 composite

Highlights: • Microstructure-Microtexture evolution study of the mushy state rolled composite (MRC) and MRC with pre-cold rolling (PCMRC). • The Microstructure and microtexture were characterized using SEM, EBSD, and TEM. • Exactly opposite trends in micro-hardness variation across the thickness of MRC and PCMRC were discussed. • The PCMRC exhibited the best combination of hardness, strength, and ductility. • A proper microstructure-microtexture-property correlation was established with experimental evidence. The mushy state rolling has been found as a promising technique for near-net forming of the cast Al-4.5 wt% Cu-5 wt% TiB{sub 2} in-situ composite to obtain microstructural refinement and moderate mechanical properties. Cold rolling with 30% reduction in thickness was carried out before mushy state rolling, for achieving a relatively more homogeneous distribution of CuAl{sub 2}-TiB{sub 2} particles as well as an appropriate microstructure to enhance the mechanical properties. A comparative investigation on the microstructural and microtextural evolution, along with mechanical properties of both the mushy state rolled composite (MRC) and pre-cold rolled mushy-state rolled composite (PCMRC) was performed, and the microstructure-property correlation was established. The PCMRC exhibited the lowest homogenized lengthscale parameter (L{sub H} 29 μm) indicating the maximum uniformity of particle distribution, computed with the help of multi-scalar analysis of area fractions (MSAAF) technique. Electron back-scattered diffraction (EBSD) and transmission electron microscopy (TEM) studies revealed the presence of fine sub-grains within centrally located deformed Al grains due to dynamic recovery, as well as the particle stimulated nucleation of equiaxed fine Al grains by dynamic recrystallization or rapid solidification of liquid near edges for both the investigated composites. The PCMRC displayed a superior combination of hardness (96.82 ± 4.5 HV{sub 5kgf}), 40% higher ultimate tensile strength (301.4 ± 1.8 MPa), and 68% more uniform elongation (30.8% ± 0.5%), compared to the MRC. The improvement in mechanical properties is attributed to the uniform particle distribution and presence of both higher density of low angle boundaries and recrystallized grains adjacent to the particles, obtained from prior cold rolling.

36 MATERIALS SCIENCE↗

The Tail Wagging the Dog: Insights into Catalysis in R67 Dihydrofolate Reductase

Plasmid-encoded R67 dihydrofolate reductase (DHFR) catalyzes a hydride transfer reaction between substrate dihydrofolate (DHF) and its cofactor, nicotinamide adenine dinucleotide phosphate (NADPH). R67 DHFR is a homotetramer that exhibits numerous characteristics of a primitive enzyme, including promiscuity in binding of substrate and cofactor, formation of nonproductive complexes, and the absence of a conserved acid in its active site. Furthermore, R67's active site is a pore, which is mostly accessible by bulk solvent. This study uses a computational approach to characterize the mechanism of hydride transfer. Not surprisingly, NADPH remains fixed in one-half of the active site pore using numerous interactions with R67. Also, stacking between the nicotinamide ring of the cofactor and the pteridine ring of the substrate, DHF, at the hourglass center of the pore, holds the reactants in place. However, large movements of the p-aminobenzoylglutamate tail of DHF occur in the other half of the pore because of ion pair switching between symmetry-related K32 residues from two subunits. This computational result is supported by experimental results that the loss of these ion pair interactions (located >13 {angstrom} from the center of the pore) by addition of salt or in asymmetric K32M mutants leads to altered enzyme kinetics [Hicks, S. N., et al. (2003) Biochemistry 42, 10569-10578; Hicks, S. N., et al. (2004) J. Biol. Chem. 279, 46995?47002]. The tail movement at the edge of the active site, coupled with the fixed position of the pteridine ring in the center of the pore, leads to puckering of the pteridine ring and promotes formation of the transition state. Flexibility coupled to R67 function is unusual as it contrasts with the paradigm that enzymes use increased rigidity to facilitate attainment of their transition states. A comparison with chromosomal DHFR indicates a number of similarities, including puckering of the nicotinamide ring and changes in the DHF tail angle, accomplished by different elements of the dissimilar protein folds.

59 BASIC BIOLOGICAL SCIENCES↗

De novo design of small beta barrel proteins

Small beta barrel proteins are attractive targets for computational design because of their considerable functional diversity despite their very small size (<70 amino acids). However, there are considerable challenges to designing such structures, and there has been little success thus far. Because of the small size, the hydrophobic core stabilizing the fold is necessarily very small, and the conformational strain of barrel closure can oppose folding; also intermolecular aggregation through free beta strand edges can compete with proper monomer folding. Here, we explore the de novo design of small beta barrel topologies using both Rosetta energy–based methods and deep learning approaches to design four small beta barrel folds: Src homology 3 (SH3) and oligonucleotide/oligosaccharide-binding (OB) topologies found in nature and five and six up-and-down-stranded barrels rarely if ever seen in nature. Both approaches yielded successful designs with high thermal stability and experimentally determined structures with less than 2.4 Å rmsd from the designed models. Using deep learning for backbone generation and Rosetta for sequence design yielded higher design success rates and increased structural diversity than Rosetta alone. The ability to design a large and structurally diverse set of small beta barrel proteins greatly increases the protein shape space available for designing binders to protein targets of interest.

59 BASIC BIOLOGICAL SCIENCES↗

SU-E-J-176: Characterization of Inter-Fraction Breast Variability and the Implications On Delivered Dose

Purpose: To systematically characterize inter-fraction breast variability and determine implications on delivered dose. Methods: Weekly port films were used to characterize breast setup variability. Five evenly spaced representative positions across the contour of each breast were chosen on the electronic port film in reference to graticule, and window and level was set such that the skin surface of the breast was visible. Measurements from the skin surface to treatment field edge were taken on each port film at each position and compared to the planning DRR, quantifying the variability. The systematic measurement technique was repeated for all port films for 20 recently treated breast cancer patients. Measured setup variability for each patient was modeled as a normal distribution. The distribution was randomly sampled from the model and applied as isocentric shifts in the treatment planning computer, representing setup variability for each fraction. Dose was calculated for each shifted fraction and summed to obtain DVHs and BEDs that modeled the dose with daily setup variability. Patients were categorized in to relevant groupings that were chosen to investigate the rigorousness of immobilization types, treatment techniques, and inherent anatomical difficulties. Mean position differences and dosimetric differences were evaluated between planned and delivered doses. Results: The setup variability was found to follow a normal distribution with mean position differences between the DRR and port film between − 8.6–3.5 mm with sigma range of 5.3–9.8 mm. Setup position was not found to be significantly different than zero. The mean seroma or whole breast PTV dosimetric difference, calculated as BED, ranged from a −0.23 to +1.13Gy. Conclusion: A systematic technique to quantify and model setup variability was used to calculate the dose in 20 breast cancer patients including variable setup. No statistically significant PTV or OAR BED differences were found between immobilization, treatment or anatomical differences investigated.

60 APPLIED LIFE SCIENCES↗

Enriched immersed finite element and isogeometric analysis: algorithms and data structures

Immersed finite element methods provide a convenient analysis framework for problems involving geometrically complex domains, such as those found in topology optimization and microstructures for engineered materials. However, their implementation remains a major challenge due to, among other things, the need to apply nontrivial stabilization schemes and generate custom quadrature rules. This article introduces the robust and computationally efficient algorithms and data structures comprising an immersed finite element preprocessing framework. The input to the preprocessor consists of a background mesh and one or more geometries defined on its domain. The output is structured into groups of elements with custom quadrature rules formatted such that common finite element assembly routines may be used without or with only minimal modifications. The key to the preprocessing framework is the construction of material topology information, concurrently with the generation of a quadrature rule, which is then used to perform enrichment and generate stabilization rules. While the algorithmic framework applies to a wide range of immersed finite element methods using different types of meshes, integration, and stabilization schemes, the preprocessor is presented within the context of the extended isogeometric analysis. This method utilizes a structured B-spline mesh, a generalized Heaviside enrichment strategy considering the material layout within individual basis functions’ supports, and face-oriented ghost stabilization. Using a set of examples, the effectiveness of the enrichment and stabilization strategies is demonstrated alongside the preprocessor’s robustness in geometric edge cases. Additionally, the performance and parallel scalability of the implementation are evaluated.

Computer implementation↗

Automated defect identification in electroluminescence images of solar modules

Solar photovoltaic (PV) modules are susceptible to manufacturing defects, mishandling problems or extreme weather events that can limit energy production or cause early device failure. Trained professionals use electroluminescence (EL) images to identify defects in modules, however, field surveys or inline image acquisition can generate millions of EL images, which are infeasible to analyze by rote inspection. Here, we develop a rapid automatic computer vision pipeline (~0.5 seconds/module) to analyze EL images and identify defects including cracks, intra-cell defects, oxygen-induced defects, and solder disconnections. Defect identification is achieved with a machine learning model (Random Forest, ResNet models and YOLO) trained on 762 manually-labeled EL images of PV modules. We compare model performance on an imbalanced real-world validation set containing 134 EL images and determine that ResNet18 and YOLO are the optimal models; we next evaluated these models on a dedicated testing set (129 module images) with resulting macro F1 scores of 0.83 (ResNet18) and 0.78 (YOLO). Using a field EL survey of a PV power plant damaged in a vegetation fire, we analyze 18,954 EL images (2.4 million cells) and inspect the spatial distribution of defects on the solar modules. The results find increased frequency of ‘crack’, ‘solder’ and ‘intra-cell’ defects on the edges of the solar module closest to the ground after fire. We also find an abnormal increase of striation rings on cells which were assumed to be caused mainly in fabrication process. Our methods are published as open-source software. It can also be used to identify other kinds of defects or process different types of solar cells with minor modification on models by transfer learning.

14 SOLAR ENERGY↗

Semi-Analytic Integration of Hypersingular Galerkin BIEs for 3D Potential Problems

An accurate and efficient semi-analytic integration technique is developed for three-dimensional hypersingular boundary integral equations of potential theory. Investigated in the context of a Galerkin approach, surface integrals are defined as limits to the boundary and linear surface elements are employed to approximate the geometry and field variables on the boundary. In the inner integration procedure, all singular and non-singular integrals over a triangular boundary element are expressed exactly as analytic formulae over the edges of the integration triangle. In the outer integration scheme, closed-form expressions are obtained for the coincident case, wherein the divergent terms are identified explicitly and are shown to cancel with corresponding terms from the edge-adjacent case. The remaining surface integrals, containing only weak singularities, are carried out successfully by use of standard numerical cubatures. Sample problems are included to illustrate the performance and validity of the proposed algorithm.

97 MATHEMATICS AND COMPUTING↗

Anomaly analysis and visualization for dynamic networks through spatiotemporal graph segmentations

With recent technology advance in Internet of Things (IoT) that involves human, sensors, and mobile devices, networks are not only growing much larger but more complex and dynamic in nature. The spatiotemporal dynamics of networks are represented by both topological changes and temporal shifts of attribute information associated with network components. Understanding the pattern and trend of dynamic networks is increasingly important. While data mining approaches are generally useful in analyzing the statistical properties of networks, there has been recent trend to consider bringing human into the loop, and to examine how feedback from visualizations of large-scale dynamic networks can further improve data mining and machine learning. Traditional visualization methods based on animation and sequences of snapshot graphs are also limited by human cognitive capability. Here, we present a dynamic network analysis and visualization (DNAV) tool which explores the spatiotemporal dimensions of graph components. In particular, nodes and edges are augmented with spatiotemporal segmentation based on both topological and attribute dynamics (e.g., time and locations of connectivity). To further facilitate analysis of large dynamic networks, DNAV includes statistical dynamic overviews alongside graph views, as well as data filtering modules for scalable analysis. Using case studies on public datasets, we demonstrate the effectiveness of DNAV in understanding and analyzing anomalies in dynamic networks such as computer communication networks.

97 MATHEMATICS AND COMPUTING↗