Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “edge inference”

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 217 records · Page 12
All sourcesNASA NTRSDOE OSTI
NASA NTRS2014

A Vortical Dawn Flank Boundary Layer for Near-Radial IMF: Wind Observations on 24 October 2001

We present an example of a boundary layer tailward of the dawn terminator which is entirely populated by rolled-up flow vortices. Observations were made by Wind on 24 October 2001 as the spacecraft moved across the region at the X plane approximately equal to −13 Earth radii. Interplanetary conditions were steady with a near-radial interplanetary magnetic field (IMF). Approximately 15 vortices were observed over the 1.5 hours duration of Wind's crossing, each lasting approximately 5 min. The rolling up is inferred from the presence of a hot tenuous plasma being accelerated to speeds higher than in the adjoining magnetosheath, a circumstance which has been shown to be a reliable signature of this in single-spacecraft observations. A blob of cold dense plasma was entrained in each vortex, at whose leading edge abrupt polarity changes of field and velocity components at current sheets were regularly observed. In the frame of the average boundary layer velocity, the dense blobs were moving predominantly sunward and their scale size along the X plane was approximately 7.4 Earth radii. Inquiring into the generation mechanism of the vortices, we analyze the stability of the boundary layer to sheared flows using compressible magnetohydrodynamic Kelvin-Helmholtz theory with continuous profiles for the physical quantities. We input parameters from (i) the exact theory of magnetosheath flow under aligned solar wind field and flow vectors near the terminator and (ii) the Wind data. It is shown that the configuration is indeed Kelvin-Helmholtz (KH) unstable. This is the first reported example of KH-unstable waves at the magnetopause under a radial IMF.

Magnetosheath↗
DOE OSTI2022

Investigation of core impurity transport in DIII-D diverted negative triangularity plasmas

Abstract Tokamak operation at negative triangularity has been shown to offer high energy confinement without the typical disadvantages of edge pedestals (Marinoni et al 2021 Nucl. Fusion 61 116010). In this paper, we examine impurity transport in DIII-D diverted negative triangularity experiments. Analysis of charge exchange recombination spectroscopy reveals flat or hollow carbon density profiles in the core, and impurity confinement times consistently shorter than energy confinement times. Bayesian inferences of impurity transport coefficients based on laser blow-off injections and forward modeling via the Aurora package (Sciortino et al 2021 Plasma Phys. Control. Fusion 63 112001) show core cross-field diffusion to be higher in L-mode than in H-mode. Impurity profile shapes remain flat or hollow in all cases. Inferred radial profiles of diffusion and convection are compared to neoclassical, quasilinear gyrofluid, and nonlinear gyrokinetic simulations. Heat transport is observed to be better captured by reduced turbulence models with respect to particle transport. State-of-the-art gyrokinetic modeling compares favorably with measurements across multiple transport channels. Overall, these results suggest that diverted negative triangularity discharges may offer a path to a highly-radiative L-mode scenario with high core performance.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2020

Conditional Generative Adversarial Networks for Solving Heat Transfer Problems

Generative Adversarial Networks (GANs) have been used as a deep learning approach to solving physics and engineering problems. Using deep learning for these problems is attractive in that reasonably accurate models can be inferred from only raw data, eliminating the need to define the exact physical equations governing a problem. We expand on previous work using GANs to generate steady-state solutions to the two-dimensional heat equation. Using a basic conditional GAN (cGAN), we generate accurate solutions for rectangular domains conditioned on four edge boundary conditions (MAE < 0.5%). For finding steady-state solutions over arbitrary two-dimensional domains (not constrained to rectangles), we use a cGAN designed for image-to-image translation. We train this GAN on various types of geometric domains (circles, squares, triangles, shapes with one circular or rectangular hole), achieving accurate results on test data made up of geometries similar to those in training (MAE < 1%). For both of these GANs, we experiment with different loss function terms, showing that a term using the gradients of solution images significantly improves the basic cGAN but not the image-to-image GAN. Lastly, we show that the image-to-image GAN performs poorly when applied to two-dimensional geometries that vary in structure from training data (MAE < 8% for shapes with multiple holes or different shaped holes). This demonstrates the cGAN's lack of generalizability. While the cGAN is an accurate and computationally efficient method when trained and tested on similarly structured data, it is a much less reliable method when applied to data that is slightly different in structure from the training data.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1983

Solar radius change between 1925 and 1979

From an analysis of numerous reports from different locations on the duration of totality of the solar eclipses on January 24, 1925, and February 26, 1979, it is found that the solar radius at the earlier date was 0.5 arcsec (or 375 km) larger than at the later date. The correction to the standard solar radius found for each eclipse is different when different subsets of the observations are used (for example, edge of path of totality timings compared with central timings). This is seen as suggesting the existence of systematic inaccuracies in our knowledge of the lunar figure. The differences between the corrections for both eclipses, however, are very similar for all subsets considered, indicating that changes of the solar size may be reliably inferred despite the existence of the lunar figure errors so long as there is proper consideration of the distribution of the observations. These results are regarded as strong evidence in support of the occurrence of solar radius changes on shorter than evolutionary time scales.

Sofia, S.↗
NASA NTRS1993

Sharp edges to neutral hydrogen disks in galaxies and the extragalactic radiation field

It is shown that the very sharp truncation of the neutral hydrogen distribution seen in NGC 3198 (and probably M33) is well modeled as the result of ionization of the atomic gas by the extragalactic radiation field. Below a critical column density of about a few times 10 exp 19/sq cm the gas is dominantly ionized and undetectable in the 21-cm line. It is inferred from the photoionization models that the total disk gas distribution in NGC 3198 is actually fairly axisymmetric. The critical column density for ionization is not a strong function of galaxy mass or mass distribution; thus, all galaxies should show a cutoff at approximately the same column density. Specific models of 3198 suggest that the extragalactic ionizing photon flux is 5000-10,000 photons/sq cm s.

Maloney, Philip↗
NASA NTRS1988

Geodetic measurement of deformation east of the San Andreas Fault in Central California

The shear strain rates in the Diablo Range of California have been calculated, and the slip rate along the Calaveras and Paicines faults in Central California have been estimated, on the basis of triangulation and trilateration data from two geodetic networks located between the western edge of the Great Valley and the San Andreas Fault. The orientation of the principal compressive strain predicted from the azimuth of the major structures in the region is N 25 deg E, leading to an average shear strain value that corresponds to a relative shortening rate of 4.5 + or - 2.4 mm/yr. It is inferred that the measured strain is due to compression across the fold of this area. The hypothesized uniform, fault-normal compression within the Coast Ranges is not supported by these results.

Sauber, Jeanne↗
NASA NTRS1999

Pulse Profiles, Accretion Column and a flare in Gx 1+4 During a Faint State

The Rossi X-ray Timing Explorer (RXTE) spacecraft observed the X-ray pulsar GX 1+4 for a period of 34 hours on July 19/20 1996. The source faded from an intensity of approx. 20 mcrab to a minimum of less than or equal to 0.7 mcrab and then partially recovered towards the end of the observation. This extended minimum lasted approx. 40,000 seconds. Phase folded light curves at a barycentric rotation period of 124.36568 +/- 0.00020 seconds show that near the center of the extended minimum the source stopped pulsing in the traditional sense but retained a weak dip feature at the rotation period. Away from the extended minimum the dips are progressively narrower at higher energies and may be interpreted as obscurations or eclipses of the hot spot by the accretion column. The pulse profile changed from leading-edge bright before the extended minimum to trailing-edge bright after it. Data from the Burst and Transz'ent Source Experiment (BATSE) show that a torque reversal occurred less than 10 days after our observation. Our data indicate that the observed rotation departs from a constant period with a P/P value off approx. -5% per year at a 4.5(sigma) significance. We infer that we may have serendipitously obtained data. with high sensitivity and temporal resolution about the time of an accretion disk spin reversal. We also observed a rapid flare which had some precursor activity close to the center of the extended minimum.

Giles, A. B.↗
NASA NTRS1993

Surface winds over West Antarctica

Five winter months (April-August 1988) of thermal infrared satellite images were examined to investigate the occurrence of dark (warm) signatures across the Ross Ice Shelf in the Antarctic continent. These features are inferred to be generated by katabatic winds that descend from southern Marie Byrd Land and then blow horizontally across the ice shelf. Significant mass is added to this airstream by katabatic winds blowing from the major glaciers that flow through the Transantarctic Mountains from East Antarctica. These negatively buoyant katabatic winds can reach the northwestern edge of the shelf - a horizontal propagation distance of up to 1,000 km - 14 percent of the time. Where the airstream crosses from the ice shelf to the ice-covered Ross Sea, a prominent coastal polynya is formed. Because the downslope buoyancy force is near zero over the Ross Ice Shelf, the northwestward propagation of the katabatic air mass requires pressure gradient support. The study shows that the extended horizontal propagation of this atmospheric density current occurred in conjunction with the passage of synoptic cyclones over the southern Amundsen Sea. These cyclones can strengthen the pressure gradient in the interior of West Antarctica and make the pressure field favorable for northwestward movement of the katabatic winds from West Antarctica across the ice shelf in a geostrophic direction. The glacier winds from East Antarctica are further accelerated by the synoptic pressure gradient, usually undergo abrupt adjustment beyond the exit to the glacier valley, and merge into the mountain-parallel katabatic air mass.

Bromwich, David↗
NASA NTRS1994

The high-state/low-state transition in V794 Aquilae

We present a V-magnitude light curve of the cataclysmic binary V794 Quilae covering an 800 day time span. The system shows variations of ups to 3 mag. In particular, there are two dips of approximately 1 mag which last for about 50 days, and are followed by a rapid recovery to the original brightness, and a third dip of approximately 3 mag lasting about 100 days. These fluctuations are thought to be caused by the response of the accretion disk in the system to the cessation of mass transfer from the mass-losing secondary star. We present computations using a time-dependent accretion disk code to delineate a range of allowed behavior for th accretion disk. To model the observed light curve, we require that the model parameters alpha(sub cold) and alpha(sub hot), which characterize the degree of coupling of the viscous dissipation to the orbital shear in the low and high states of the accretion disk, be smaller than they have been inferred to be in the dwarf novae, and that the ratio alpha(sub hot)/alpha(sub cold) also be smaller. The fact that alpha seems to vary with the rate of mass transfer may provide support for the model of Vishniac and Diamond, in which the impact of the mass transfer stream onto the outer edge of the accretion disk excites internal waves which transport angular momentum outward and provide the viscous dissipation.

Honeycutt, R. Kent↗
DOE OSTI2026

Peering down the barrel with DESI DR2: 10 000+ inflows at $z$ < 0.6 reveal how galaxies accrete cold gas

Direct observational constraints on how galaxies acquire their gas remain remarkably limited, hindering our understanding of the baryon cycle. We present a search for down-the-barrel NaI D absorption towards 15.6 million galaxies at $z < 0.6$ in DESI Data Release 2. We use Bayesian evidence ratios to assess whether the absorption requires additional components tracing interstellar gas distinct from the systemic component of the galaxy. We construct a catalogue of 50 088 (27 420) galaxies with moderate (strong) evidence for down-the-barrel absorption. The inferred absorption components are broadly distributed in velocity, with approximately 50% at $v_{\rm flow} < -50$ km/s, 30% within 50 km/s of the systemic velocity and the remaining 20% at $v_{\rm flow} > 50$ km/s. We find strong evidence for a large population of low-velocity, infalling absorbers with velocities $\sim$20 km/s in edge-on galaxies, consistent with radial inflows predicted in simulations. The stronger correlation in early-type galaxies between inflow velocity and stellar velocity dispersion, compared to that with stellar mass, suggests that a portion of these inflows may be associated with accreting satellites. These results reveal the multiple pathways in which galaxies accrete gas at redshift $z < 0.6$ for the first time in a statistically significant sample.

Weng, S. [Marseille, Lab. Astrophys.]↗
DOE OSTI2026

The Double-edged Sword of Data-driven Super-Resolution: Adversarial Super-resolution Models

Data-driven super-resolution (SR) methods are often integrated into imaging pipelines as preprocessing steps to improve downstream tasks such as classification and detection. However, these SR models introduce a previously unexplored attack surface into imaging pipelines. In this paper, we present AdvSR, a framework demonstrating that adversarial behavior can be embedded directly into SR model weights during training, requiring no access to inputs at inference time. Unlike prior attacks that perturb inputs or rely on backdoor triggers, AdvSR operates entirely at the model level. By jointly optimizing for reconstruction quality and targeted adversarial outcomes, AdvSR produces models that appear benign under standard image quality metrics while inducing downstream misclassification. We evaluate AdvSR on three SR architectures (SRCNN, EDSR, SwinIR) paired with a YOLOv11 classifier and demonstrate that AdvSR models can achieve high attack success rates with minimal quality degradation. These findings highlight a new model-level threat for imaging pipelines, with implications for how practitioners source and validate models in safety-critical applications.

Sullivan, Haley [ORNL] (ORCID:0000000274069217)↗
NASA NTRS2010

Understanding the Scalability of Bayesian Network Inference Using Clique Tree Growth Curves

One of the main approaches to performing computation in Bayesian networks (BNs) is clique tree clustering and propagation. The clique tree approach consists of propagation in a clique tree compiled from a Bayesian network, and while it was introduced in the 1980s, there is still a lack of understanding of how clique tree computation time depends on variations in BN size and structure. In this article, we improve this understanding by developing an approach to characterizing clique tree growth as a function of parameters that can be computed in polynomial time from BNs, specifically: (i) the ratio of the number of a BN s non-root nodes to the number of root nodes, and (ii) the expected number of moral edges in their moral graphs. Analytically, we partition the set of cliques in a clique tree into different sets, and introduce a growth curve for the total size of each set. For the special case of bipartite BNs, there are two sets and two growth curves, a mixed clique growth curve and a root clique growth curve. In experiments, where random bipartite BNs generated using the BPART algorithm are studied, we systematically increase the out-degree of the root nodes in bipartite Bayesian networks, by increasing the number of leaf nodes. Surprisingly, root clique growth is well-approximated by Gompertz growth curves, an S-shaped family of curves that has previously been used to describe growth processes in biology, medicine, and neuroscience. We believe that this research improves the understanding of the scaling behavior of clique tree clustering for a certain class of Bayesian networks; presents an aid for trade-off studies of clique tree clustering using growth curves; and ultimately provides a foundation for benchmarking and developing improved BN inference and machine learning algorithms.

Mengshoel, Ole J.↗
DOE OSTI2024

Effects of Injected Current Streams on MHD Equilibrium Reconstruction of Local Helicity Injection Plasmas in a Spherical Tokamak

Open field line currents are intrinsic to DC helicity injection plasma startup and pose a challenge for inferring the plasma equilibrium with standard reconstruction analysis. Local helicity injection (LHI) is a type of DC helicity injection which uses small, modular current sources to drive force-free current along helical field lines to produce tokamak plasmas. MHD modeling and magnetic measurements during LHI indicate the injected current streams remain coherent as helical structures on the outboard edge of a core toroidal plasma that is tokamak-like in a toroidally averaged sense. To extract core plasma equilibrium properties, external magnetic diagnostics corrected for contributions from the injected current streams are fitted by a standard Grad-Shafranov equilibrium code. An iterative approach for estimating and subtracting the stream contributions from the diagnostic signals is described and applied to a model equilibrium database to reduce systematic errors introduced by the streams. Convergence is usually attained with 2 to 4 iterations, with derived equilibrium parameters matching the prescribed axisymmetric core values to within estimated experimental uncertainties. Accurate recovery of core parameters occurs when the ratio of the net toroidal windup current from the streams to the core plasma current is less than 0.2, which is typically satisfied in most experiments.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
NASA NTRS1980

Spatial distribution of plasma in the Io torus

In situ measurements of ion densities and temperatures have been analyzed to produce profiles of these plasma parameters along the Voyager 1 inbound trajectory between 7 and 5 Jupiter radii. The temperature profile shows a sharp decrease by a factor of 50 between 5.8 and 5.2 Jupiter radii corresponding to a temperature gradient of 7 x 10 to the 5th per Jupiter radius. The electron density profile, inferred from the ion density measurements, has two maxima at 5.7 and 5.3 Jupiter radii. A two-dimensional model of the spatial distribution of various ionic species in the Io plasma torus has been constructed. Using this model a contour map of electron density in a meridional plane has been made, it exhibits a well-defined inner edge to the torus at 5.6 Jupiter radii. The contour map of S(+) ion density indicates that most of the S(+) ions are concentrated close to the centrifugal symmetry surface and radially inward of the larger electron density maximum near 5.7 Jupiter radii.

Bagenal, F.↗
NASA NTRS2021

Pretest Computational Assessment of Boundary Layer Transition in the NASA Juncture Flow Model with an NACA 0015-Based Wing

The first two phases of the NASA Juncture Flow experiment were carried out on a DLR-F6 swept-wing model and were designed to provide “CFD validation-quality” data toward the assessment and improvement of existing CFD turbulence models in predicting onset and extent of three-dimensional separated flow near the wing-juncture trailing-edge region. The next phase of experiments will involve an NACA 0015-based swept wing, as prior risk reduction experiments had indicated that this wing shape resulted in reduced separation near the juncture region than the DLR-F6 wing, thus providing a better option to evaluate the ability of CFD models to predict incipient turbulent separation. The NACA 0015 measurements will also include IR thermography to infer the variation of transition front with respect to an increasing angle of attack. The primary objective of this work is to computationally make a preliminary assessment of the transition front on both surfaces of the NACA 0015 wing at a crank-chord-based Reynolds number of 2.4 x 106 for four different angles of attack, (0°, 2.5°, 5°, and 7.5°) and to determine the dominant mechanisms responsible for transition. This assessment includes both RANS-based transition models from NASA’s OVERFLOW 2.3b flow solver and linear parabolized stability equations (PSE) stability analysis based on the Langley Stability and Transition Analysis code, LASTRAC. Linear PSE results indicate that the upper surface of the wing is dominated by Tollmien- Schlichting (TS) instabilities, and that the laminar flow region shrinks from about 50% chord to a very small region just downstream of the attachment line as the angle of attack is increased from 0° to 7.5°. Consequently, the transition fronts predicted by the Spalart- Allmaras-based amplification factor transport (AFT-2017b) equation model (which accounts for the TS instabilities alone) and the Menter’s shear-stress transport equation (SST2003)- based Langtry-Menter transition model with ability to account for both TS and crossflow effects (LM2015) compare well with those predicted using linear PSE. On the lower surface of the wing, stationary crossflow (CF) instabilities begin to appear on the inboard portion of the wing in addition to the TS-instabilities for the larger angles of attack (5° and 7.5°), further reducing the laminar flow extent within the inboard region. The LM2015 model that accounts for CF effects is able to replicate this trend but appears to predict a slightly earlier transition. The outcome of this effort will inform the experiment and, when the actual experimental data become available, provide further opportunity to assess and improve the various transition models.

CFD modeling↗
DOE OSTI2024

Resolving Nonequilibrium Shape Variations among Millions of Gold Nanoparticles

Nanoparticles, exhibiting functionally relevant structural heterogeneity, are at the forefront of cutting-edge research. Now, high-throughput single-particle imaging (SPI) with X-ray free-electron lasers (XFELs) creates opportunities for recovering the shape distributions of millions of particles that exhibit functionally relevant structural heterogeneity. To realize this potential, three challenges have to be overcome: (1) simultaneous parametrization of structural variability in real and reciprocal spaces; (2) efficiently inferring the latent parameters of each SPI measurement; (3) scaling up comparisons between 10 5 structural models and 10 6 XFEL-SPI measurements. Here, we describe how we overcame these three challenges to resolve the nonequilibrium shape distributions within millions of gold nanoparticles imaged at the European XFEL. These shape distributions allowed us to quantify the degree of asymmetry in these particles, discover a relatively stable “shape envelope” among nanoparticles, discern finite-size effects related to shape-controlling surfactants, and extrapolate nanoparticles’ shapes to their idealized thermodynamic limit. Ultimately, these demonstrations show that XFEL SPI can help transform nanoparticle shape characterization from anecdotally interesting to statistically meaningful.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Stabilization of U 5 f 2 configuration in UTe 2 through U 6 d dimers in the presence of Te2 chains

We investigate the topological superconductor candidate UTe 2 using high-resolution valence-band resonant inelastic x-ray scattering at the U M 4 , 5 edges. We observe atomiclike low-energy excitations that support the correlated nature of this unconventional superconductor. These excitations originate from the U 5 f 2 configuration, which is unexpected since the short Te2-Te2 distances exclude Te2 being 2 − . By utilizing the photoionization cross-section dependence of the photoemission spectra in combination with band structure calculations, we infer that the stabilization of the U 5 f 2 configuration is due to the U 6 d bonding states in the U dimers acting as a charge reservoir. Our results emphasize that the description of the physical properties should commence with a 5 f 2 ansatz. Published by the American Physical Society 2024

Christovam, Denise S. (ORCID:0000000250698107)↗
DOE OSTI2025

Even Higher-Level Synthesis: An Exploration of AI Hardware Accelerators using HLS4ML

With the rise of artificial intelligence, the popularization of deep learning, and a constantly evolving industry, the demand for flexible and efficient tools has never been greater. As algorithms grow more complex, their runtime and energy consumption increase exponentially. Customized hardware accelerators, long used for specific mathematical operations, remain essential for managing modern applications' computational and power demands. Hardware accelerators can speed up complex computations by orders of magnitude, but their manual design and verification processes are often challenging and time-consuming. High-Level Synthesis (HLS) provides a solution by transforming high-level algorithm descriptions, typically written in C++ or SystemC, into synthesizable RTL suitable for hardware implementation. This approach reduces development time for RTL engineers while offering flexibility beyond what traditional handwritten RTL can provide. We extended this capability to the machine-learning domain with the open-source framework hls4ml, which allows neural networks trained in Python frameworks like Tensorflow or PyTorch to be synthesized into efficient hardware representations for the traditional FPGA and ASIC flows. This breakthrough addresses the growing need for reduced design turnaround and easy verification of ML hardware accelerators with low latency and power efficiency constraints. During this tutorial, we will demonstrate how Python complements HLS by simplifying the ML design process, bridging the gap between software and hardware development. Attendees will explore how we translate neural networks modeled in Python into fixed-point C++ models suitable for HLS workflows. We will dive into strategies like Value-Range Analysis and Quantization-Aware Training, which optimize these designs for deployment and evaluate their accuracy, power consumption, and energy efficiency. To exemplify these concepts, experts from Fermilab will share their experiences applying this technology to high-energy physics experiments, where real-time, low-latency processing is critical. Over the years, Fermilab engineers have demonstrated how deep neural networks, optimized for hardware using hls4ml, can meet the stringent requirements of trigger systems at the CERN Large Hadron Collider. These systems rely on rapid decision-making to process immense data volumes while retaining only the most relevant events for further analysis. The application of hls4ml has also been extended to innovative technologies like smart pixel arrays. These smart pixels integrate ML inference capabilities directly into sensor devices, enabling localized data processing at the pixel level. This approach drastically reduces the need to transmit raw data to external processing units, significantly decreasing power consumption and latency. By embedding neural networks within the pixel architecture, the smart pixels can identify and prioritize relevant data in real time, providing a highly efficient solution for edge computing in scenarios such as particle detectors and imaging systems. Fermilab's work highlights the potential of hardware-accelerated ML in scenarios where both speed and power efficiency are mission-critical. Through this tutorial, attendees will gain valuable insights into the challenges and solutions of deploying ML in hardware. Understanding how HLS and hls4ml streamline the development of neural network-based hardware accelerators is fundamental for the industry's future. Participants will learn how these technologies are shaping the future of AI and scientific computing.

Di Guglielmo, Giuseppe [Fermilab]↗