Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “linear scaling”

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 1,261 records · Page 70
All sourcesNASA NTRSDOE OSTI
NASA NTRS1989

Spaceborne radar observations: A guide for Magellan radar-image analysis

Geologic analyses of spaceborne radar images of Earth are reviewed and summarized with respect to detecting, mapping, and interpreting impact craters, volcanic landforms, eolian and subsurface features, and tectonic landforms. Interpretations are illustrated mostly with Seasat synthetic aperture radar and shuttle-imaging-radar images. Analogies are drawn for the potential interpretation of radar images of Venus, with emphasis on the effects of variation in Magellan look angle with Venusian latitude. In each landform category, differences in feature perception and interpretive capability are related to variations in imaging geometry, spatial resolution, and wavelength of the imaging radar systems. Impact craters and other radially symmetrical features may show apparent bilateral symmetry parallel to the illumination vector at low look angles. The styles of eruption and the emplacement of major and minor volcanic constructs can be interpreted from morphological features observed in images. Radar responses that are governed by small-scale surface roughness may serve to distinguish flow types, but do not provide unambiguous information. Imaging of sand dunes is rigorously constrained by specific angular relations between the illumination vector and the orientation and angle of repose of the dune faces, but is independent of radar wavelength. With a single look angle, conditions that enable shallow subsurface imaging to occur do not provide the information necessary to determine whether the radar has recorded surface or subsurface features. The topographic linearity of many tectonic landforms is enhanced on images at regional and local scales, but the detection of structural detail is a strong function of illumination direction. Nontopographic tectonic lineaments may appear in response to contrasts in small-surface roughness or dielectric constant. The breakpoint for rough surfaces will vary by about 25 percent through the Magellan viewing geometries from low to high Venusian latitudes. Examples of anomalies and system artifacts that can affect image interpretation are described.

Ford, J. P.↗
DOE OSTI2022

Modeling and implementation of vertical excursion FFA in the Zgoubi ray-tracing code

Vertical Fixed Field Accelerators (vFFAs) feature complex and highly non-linear magnetic fields, which require simulation codes allowing step-wise particle tracking. Methods to model the 3D magnetic field of scaling vFFAs have been developed in the ray-tracing code Zgoubi. The field modeling and particle tracking methods include the field non-linearities, the fringe fields, and the field superposition of neighboring magnets. The procedure implements the vFFA analytical field expressions, allowing design studies and parameter optimizations using the Zgoubi built-in fit method. The vFFA procedure has been applied to a ten-fold symmetry ring with a triplet focusing structure designed to accelerate protons from 3 MeV to 12 MeV and studied under the ISIS-II proton driver prototype project. Results from particle tracking in externally generated 3D semi-analytical field maps and the developed vFFA analytical model are shown to be in excellent agreement.

43 PARTICLE ACCELERATORS↗
NASA NTRS1989

Observations of, and sources of the spatial and temporal variability of ozone in the middle atmosphere on climatological time scales (OZMAP) and equatorial dynamics: Seasonal variations of ozone trends

The long term trends (least square linear regression with time) of ozone content at seven European, seven North American, three Japanese and two tropical stations during 21 years (1964 to 1984) are analyzed. In all regions negative trends are observed during the 1970s, but are partly compensated by limited periods of positive trends during the late 1960s and late 1970s. Solely the North American ozone data show negative trends in all 10 year periods. When the long term ozone trends are evaluated for each month of the year separately, a seasonal variation is revealed, which in Europe and North America has largest negative trends in late winter and spring. While in Europe the negative trends in winter/spring are partly compensated by positive trends in summer, in North America the summer values reach only zero, retaining the significant negative trend in annual mean values. In contrast to the antarctic ozone hole, the spring reduction of ozone in Europe and in North America is associated with stratospheric temperatures increasing in the analyzed period and therefore is consistent with the major natural ozone production and loss processes.

Entzian, G.↗
DOE OSTI2026

HARMONY: Large-Scale Architecture Search for Efficient Hybrid Language Models

As large language models scale to trillions of parameters, their computational and memory requirements present critical challenges for efficient training and deployment. While Mixture of Experts (MoE) architectures enable efficient scaling through sparse parameter activation, and state-space models like Mamba offer linear-time complexity, principled methods for combining these paradigms remain undeveloped. We introduce HARMONY (Hybrid Architecture Research for Mamba, Optimized with Neural efficiencY), a multi-objective evolutionary neural architecture search framework for discovering efficient hybrid language models that integrate Transformer attention mechanisms, Mixture-of-Experts routing, and Mamba state-space components. Through large-scale distributed search using 16,384 MI250X GPUs on the Frontier supercomputer, HARMONY explores a comprehensive design space encompassing six attention variants (MHA, MQA, GQA, MLA, SWA, and Mamba-2), variable MoE configurations with both routed and shared experts, and extensive Mamba hyperparameters. Our framework discovers heterogeneous architectures that balance training performance with computational efficiency through multi-objective optimization incorporating latency penalties and fitness-based selection. Analysis of discovered architectures reveals that optimal hybrid designs favor heterogeneous component mixing rather than homogeneous patterns, with Mamba-2 and Multi-Head Latent Attention (MLA) emerging as preferred mechanisms. Discovered architectures demonstrate superior training efficiency: our best configuration achieves a final perplexity of 1.0874 with 2.38B parameters while processing 4,320 tokens/second, outperforming significantly larger manually designed models. Full-scale evaluation shows HARMONY's top architectures achieve better loss trajectories than equivalently-sized models using state-of-the-art configurations including Mixtral, Jamba, and Samba. Additionally, we demonstrate 91% weak scaling efficiency when training discovered 36B-parameter models across 1,024 GPUs. HARMONY is released as an open framework with comprehensive tools for building and training hybrid models using expert-data-pipeline parallelism, democratizing access to automated architecture design for next-generation language models.

Herron, Emily [ORNL] (ORCID:0000000273008172)↗
DOE OSTI2022

Non‐Linear Dimensionality Reduction With a Variational Encoder Decoder to Understand Convective Processes in Climate Models

Abstract Deep learning can accurately represent sub‐grid‐scale convective processes in climate models, learning from high resolution simulations. However, deep learning methods usually lack interpretability due to large internal dimensionality, resulting in reduced trustworthiness in these methods. Here, we use Variational Encoder Decoder structures (VED), a non‐linear dimensionality reduction technique, to learn and understand convective processes in an aquaplanet superparameterized climate model simulation, where deep convective processes are simulated explicitly. We show that similar to previous deep learning studies based on feed‐forward neural nets, the VED is capable of learning and accurately reproducing convective processes. In contrast to past work, we show this can be achieved by compressing the original information into only five latent nodes. As a result, the VED can be used to understand convective processes and delineate modes of convection through the exploration of its latent dimensions. A close investigation of the latent space enables the identification of different convective regimes: (a) stable conditions are clearly distinguished from deep convection with low outgoing longwave radiation and strong precipitation; (b) high optically thin cirrus‐like clouds are separated from low optically thick cumulus clouds; and (c) shallow convective processes are associated with large‐scale moisture content and surface diabatic heating. Our results demonstrate that VEDs can accurately represent convective processes in climate models, while enabling interpretability and better understanding of sub‐grid‐scale physical processes, paving the way to increasingly interpretable machine learning parameterizations with promising generative properties.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Microphysics of shock-grain interaction for inertial confinement fusion ablators in a fluid approach

Ablator materials used for inertial confinement fusion, such as high-density carbon (HDC) and beryllium, have grain structure which may lead to small-scale density nonuniformity and the generation of perturbations when the materials are shocked and compressed. Here, we use a combination of a linear theory of shock interaction with density nonuniformity [Velikovich et al., Phys. Plasmas 14, 072706 (2007)] and numerical simulations to study shock interaction with a model representation of HDC grains. While the shock-grain interaction is nonlinear, the linear theory shows some key features of the shock-grain interaction, which also hold for the (nonlinear) simulations. The postshock perturbations are made up of sonic reflections off of grain boundaries and vorticity deposition along them, with the latter dominating the perturbed energy content. The mean (per mass) postshock perturbed kinetic energy decreases with increasing grain size, but energy will be deposited at increasing spatial scale. From the perspective of the postshock perturbed energy, the detailed linear theory largely supports a proposed method [S. Davidovits et al., Phys. Plasmas 29, 112708 (2022)] for deresolving the grains (in a similar grains model) that treats the grains statistically. Finally, our simulation results highlight the influence of thermal conduction on the perturbation dynamics at grain scales.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2023

Reconstructing cosmological initial conditions from late-time structure with convolutional neural networks

ABSTRACT We present a method to reconstruct the initial linear-regime matter density field from the late-time non-linearly evolved density field in which we channel the output of standard first-order reconstruction to a convolutional neural network (CNN). Our method shows dramatic improvement over the reconstruction of either component alone. We show why CNNs are not well-suited for reconstructing the initial density directly from the late-time density: CNNs are local models, but the relationship between initial and late-time density is not local. Our method leverages standard reconstruction as a preprocessing step, which inverts bulk gravitational flows sourced over very large scales, transforming the residual reconstruction problem from long-range to local and making it ideally suited for a CNN. We develop additional techniques to account for redshift distortions, which warp the density fields measured by galaxy surveys. Our method improves the range of scales of high-fidelity reconstruction by a factor of 2 in wavenumber above standard reconstruction, corresponding to a factor of 8 increase in the number of well-reconstructed modes. In addition, our method almost completely eliminates the anisotropy caused by redshift distortions. As galaxy surveys continue to map the Universe in increasingly greater detail, our results demonstrate the opportunity offered by CNNs to untangle the non-linear clustering at intermediate scales more accurately than ever before.

Astronomy & Astrophysics↗
DOE OSTI2025

Quantum Algorithm for Linear Non-unitary Dynamics with Near-Optimal Dependence on All Parameters

We introduce a family of identities that express general linear non-unitary evolution operators as a linear combination of unitary evolution operators, each solving a Hamiltonian simulation problem. This formulation can exponentially enhance the accuracy of the recently introduced linear combination of Hamiltonian simulation (LCHS) method [An, Liu, and Lin, Physical Review Letters, 2023]. For the first time, this approach enables quantum algorithms to solve linear differential equations with both optimal state preparation cost and near-optimal scaling in matrix queries on all parameters.

Applied Dynamical Systems↗
NASA NTRS1993

Spanwise scale selection in plane mixing layers

Direct numerical simulations of temporally evolving plane mixing layers undergoing as many as three pairings have been examined for evidence of spanwise scale change. All simulations were begun from a few low-wavenumber disturbances, usually derived from linear stability theory, in addition to the mean velocity. The amplitude of the initial three-dimensional disturbances varied from infinitesimal to large enough to trigger a rapid transition to turbulence. The time required for a change of characteristic spanwise scale with infinitesimal three-dimensional disturbances was found to be very long, requiring three or more pairings to complete a doubling of the spanwise scale. Stronger three-dimensionality can produce more rapid scale changes, but it is also likely to trigger transition to turbulence. No evidence was found for a change from an organized array of rib vortices at one spanwise scale to a similar array at a larger spanwise scale.

Rogers, Michael M.↗
DOE OSTI2023

New linear stability parameter to describe low-β electromagnetic microinstabilities driven by passing electrons in axisymmetric toroidal geometry

In magnetic confinement fusion devices, the ratio of the plasma pressure to the magnetic field energy, β, can become sufficiently large that electromagnetic microinstabilities become unstable, driving turbulence that distorts or reconnects the equilibrium magnetic field. In this paper, a theory is proposed for electromagnetic, electron-driven linear instabilities that have current layers localised to mode-rational surfaces and binormal wavelengths comparable to the ion gyroradius. The model retains axisymmetric toroidal geometry with arbitrary shaping, and consists of orbit-averaged equations for the mode-rational surface layer, with a ballooning space kinetic matching condition for passing electrons. The matching condition connects the current layer to the large scale electromagnetic fluctuations, and is derived in the limit that β is comparable to the square root of the electron-to-ion-mass ratio. Electromagnetic fluctuations only enter through the matching condition, allowing for the identification of an effective β that includes the effects of equilibrium flux surface shaping. The scaling predictions made by the asymptotic theory are tested with comparisons to results from linear simulations of micro-tearing and electrostatic microinstabilities in MAST discharge #6252, showing excellent agreement. In particular, it is demonstrated that the effective β can explain the dependence of the local micro-tearing mode (MTM) growth rate on the ballooning parameter θ 0 – possibly providing a route to optimise local flux surfaces for reduced MTM-driven transport.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2022

Scalable single-mode surface-emitting laser via open-Dirac singularities

Single-aperture cavities are a key component of lasers that are instrumental for the amplification and emission of a single light mode. However, the appearance of high-order transverse modes as the size of the cavities increases has frustrated efforts to scale-up cavities while preserving single-mode operation since the invention of the laser six decades ago. A suitable physical mechanism that allows single-mode lasing irrespective of the cavity size-a 'scale invariant' cavity or laser-has not been identified yet. Here we propose and demonstrate experimentally that open-Dirac electromagnetic cavities with linear dispersion-which in our devices are realized by a truncated photonic crystal arranged in a hexagonal pattern-exhibit unconventional scaling of losses in reciprocal space, leading to single-mode lasing that is maintained as the cavity is scaled up in size. The physical origin of this phenomenon lies in the convergence of the complex part of the free spectral range in open-Dirac cavities towards a constant governed by the loss rates of distinct Bloch bands, whereas for common cavities it converges to zero as the size grows, leading to inevitable multimode emission. An unconventional flat-envelope fundamental mode locks all unit cells in the cavity in phase, leading to single-mode lasing. We name such sources Berkeley surface-emitting lasers (BerkSELs) and demonstrate that their far-field corresponds to a topological singularity of charge two, in agreement with our theory. Open-Dirac cavities unlock avenues for light-matter interaction and cavity quantum electrodynamics.

42 ENGINEERING↗
NASA NTRS1994

Power spectrum, correlation function, and tests for luminosity bias in the CfA redshift survey

We describe and apply a method for directly computing the power spectrum for the galaxy distribution in the extension of the Center for Astrophysics Redshift Survey. Tests show that our technique accurately reproduces the true power spectrum for k greater than 0.03 h Mpc(exp -1). The dense sampling and large spatial coverage of this survey allow accurate measurement of the redshift-space power spectrum on scales from 5 to approximately 200 h(exp -1) Mpc. The power spectrum has slope n approximately equal -2.1 on small scales (lambda less than or equal 25 h(exp -1) Mpc) and n approximately -1.1 on scales 30 less than lambda less than 120 h(exp -1) Mpc. On larger scales the power spectrum flattens somewhat, but we do not detect a turnover. Comparison with N-body simulations of cosmological models shows that an unbiased, open universe CDM model (OMEGA h = 0.2) and a nonzero cosmological constant (CDM) model (OMEGA h = 0.24, lambda(sub zero) = 0.6, b = 1.3) match the CfA power spectrum over the wavelength range we explore. The standard biased CDM model (OMEGA h = 0.5, b = 1.5) fails (99% significance level) because it has insufficient power on scales lambda greater than 30 h(exp -1) Mpc. Biased CDM with a normalization that matches the Cosmic Microwave Background (CMB) anisotropy (OMEGA h = 0.5, b = 1.4, sigma(sub 8) (mass) = 1) has too much power on small scales to match the observed galaxy power spectrum. This model with b = 1 matches both Cosmic Background Explorer Satellite (COBE) and the small-scale power spect rum but has insufficient power on scales lambda approximately 100 h(exp -1) Mpc. We derive a formula for the effect of small-scale peculiar velocities on the power spectrum and combine this formula with the linear-regime amplification described by Kaiser to compute an estimate of the real-space power spectrum. Two tests reveal luminosity bias in the galaxy distribution: First, the amplitude of the pwer spectrum is approximately 40% larger for the brightest 50% of galaxies in volume-limited samples that have M(sub lim) greater than M*. This bias in the power spectrum is independent of scale, consistent with the peaks-bias paradigm for galaxy formation. Second, the distribution of local density around galaxies shows that regions of moderate and high density contain both very bright (M less than M* = -19.2 + 5 log h) and fainter galaxies, but that voids preferentially harbor fainter galaxies (approximately 2 sigma significance level).

Park, Changbom↗
DOE OSTI2021

Localizing Transformations of the Galaxy-Galaxy Lensing Observable

Modern cosmological analyses of galaxy-galaxy lensing face a theoretical systematic effect arising from the nonlocality of the observed galaxy-galaxy lensing signal. Because the predicted tangential shear signal at a given separation depends on the physical modeling on all scales internal to that separation, systematic uncertainties in the modeling of nonlinear small scales are propagated outward to larger scales. Even in the absence of other limiting factors, this systematic effect alone can necessitate conservative small-scale cuts, resulting in significant losses of information in the tangential shear data vector. We develop a simple linear transformation of the standard galaxy-galaxy observable that removes this nonlocality, which ensures that the cosmological signal contained within the transformed observable is exclusively drawn from well-understood physical scales. This new observable, through its robustness against nonlocality, also enables a significant extension in the range of usable scales in galaxy-galaxy lensing compared to the standard approach in current cosmological analyses.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS1997

Applications of Data Assimilation to Analysis of the Ocean on Large Scales

It is commonplace to begin talks on this topic by noting that oceanographic data are too scarce and sparse to provide complete initial and boundary conditions for large-scale ocean models. Even considering the availability of remotely-sensed data such as radar altimetry from the TOPEX and ERS-1 satellites, a glance at a map of available subsurface data should convince most observers that this is still the case. Data are still too sparse for comprehensive treatment of interannual to interdecadal climate change through the use of models, since the new data sets have not been around for very long. In view of the dearth of data, we must note that the overall picture is changing rapidly. Recently, there have been a number of large scale ocean analysis and prediction efforts, some of which now run on an operational or at least quasi-operational basis, most notably the model based analyses of the tropical oceans. These programs are modeled on numerical weather prediction. Aside from the success of the global tide models, assimilation of data in the tropics, in support of prediction and analysis of seasonal to interannual climate change, is probably the area of large scale ocean modeling and data assimilation in which the most progress has been made. Climate change is a problem which is particularly suited to advanced data assimilation methods. Linear models are useful, and the linear theory can be exploited. For the most part, the data are sufficiently sparse that implementation of advanced methods is worthwhile. As an example of a large scale data assimilation experiment with a recent extensive data set, we present results of a tropical ocean experiment in which the Kalman filter was used to assimilate three years of altimetric data from Geosat into a coarsely resolved linearized long wave shallow water model. Since nonlinear processes dominate the local dynamic signal outside the tropics, subsurface dynamical quantities cannot be reliably inferred from surface height anomalies. Because of its potential for large scale synoptic coverage of the deep ocean, acoustic travel time data should be a natural complement to satellite altimetry. Satellite data give us vertical integrals associated with thermodynamic and dynamic processes.

Miller, Robert N.↗
NASA NTRS1988

Compact radio cores in three Seyfert galaxies - High resolution studies of Mkn 3, Mkn 231, and Mkn 463

The small-scale morphologies of Mkn 231 and Mkn 463 have been investigated using the European VLBI Network (EVN) at 1.6 GHz and the VLA at 15 and 22 GHz. Mkn 3 has been studied using the EVN at 1.4 and 1.6 GHz. All of the galaxies have a substantial fraction of their core radio flux in linear or elongated structures, typically tens of parsecs in extent. Agreement between the parsec- and kiloparsec-scale position angles of continuum radio emission suggests collimated outflow from the galaxy centers that continues over several orders of magnitude in scale. Supernova rates of order one hundred per year on scales of tens of parsecs are required to explain the nuclear radio emission as the result of bursts of star formation. The luminosities, small extents, and linear morphologies of the radio sources all support the idea that the nuclear radio sources are energized by active galactic nuclei and therefore support the more general picture of Seyfert galaxy nuclei being low-luminosity quasars. All of the galaxies appear to have undergone violent interactions in the recent past.

Neff, S. G.↗
DOE OSTI2024

Performance Testing of an Integrated Magnetic Power Take-Off

A wave energy converter (WEC) and the power take-off (PTO) generator system can be represented by using a mass-spring-dampener model. By incorporating a negative stiffness spring within the PTO, the overall stiffness of the PTO can be lowered allowing the impedance of the PTO to be more closely matched with the WEC. By designing for impedance matching the WEC can greatly enhance its power generation capability. This project has involved the design, fabrication and testing of a new type of linear-stroke length and rotary stroke length adjustable negative stiffness magnetic springs for use within a wave energy converter (WEC). The magnetic springs were studied by using 3-D finite element analysis with the objective of creating high energy density and a long linear stroke length. After the construction and testing of both a proof-of-principle adjustable linear and rotary magnetic spring prototype the rotary (torsional) magnetic spring was selected for scaling up analysis. The selected scaled-up proof-of-principle magnetic spring had a peak torque of 850 Nm with a ±45 degree stroke length. By translating the inner rotor relative to the outer rotor, the stiffness could be adjusted to be either negative or positive stiffness. During this project a magnetic lead screw was also studied, and it was shown that by combining the magnetic lead screw with a linearly translating magnetic spring a very long rotary stroke length could be attained. However, as the magnetic lead screw increased complexity and reduced overall energy density relative to a rotary (torsional) spring, this design approach was not further pursed. Dynamometer testing of the variable stiffness magnetic springs was first completed by Portland State University and following this the magnetic spring performance was independently verified by Sandia National Laboratory (Sandia). A WEC analysis when using a variable stiffness magnetic spring was completed by using the WecOptTool. WecOptTool is an open-source WEC optimization software developed by Sandia that supports efficient power take-off (PTO) and control optimization. Wave condition data from the Oregon PacWave test-site was used in this study. The analysis showed that a WEC with a tunable stiffness value could consistently achieve about 80% of its maximum theoretical power production. The use of a tunable stiffness WEC, rather than zero-stiffness or a constant stiffness WEC was also shown to lead to a smaller maximum PTO force requirement. The variable stiffness magnetic spring was integrated into an experimental WEC developed by Sandia, called a Wave-Bot. Sandia successfully completed water-tank testing of the Wave-Bot at the Navy’s Carderock, Maryland, wave-basin test site. The wave-basin testing helped to experimentally demonstrate the operating capabilities and increased power generation capability of a WEC containing an integrated variable stiffness magnetic spring. A WEC capacity factor analysis was also completed. The capacity factor was defined as the ratio of annual average WEC generator power to the maximum (RMS) generated power. This capacity factor provided a means of identifying the ratio of potential revenue to cost. It was calculated that if a tunable variable stiffness magnetic spring has RMS constraints it could operate with a capacity factor above 30%. This is comparable to a wind turbine’s capacity factor.

16 TIDAL AND WAVE POWER↗
DOE OSTI2022

Neural network acceleration of large-scale structure theory calculations

Here, we make use of neural networks to accelerate the calculation of power spectra required for the analysis of galaxy clustering and weak gravitational lensing data. For modern perturbation theory codes, evaluation time for a single cosmology and redshift can take on the order of two seconds. In combination with the comparable time required to compute linear predictions using a Boltzmann solver, these calculations are the bottleneck for many contemporary large-scale structure analyses. Here, we construct neural network-based surrogate models for Lagrangian perturbation theory (LPT) predictions of matter power spectra, real and redshift space galaxy power spectra, and galaxy-matter cross power spectra that attain ~ 0.1% (at one sigma) accuracy over a broad range of scales in a ωCDM parameter space. The neural network surrogates can be evaluated in approximately one millisecond, a factor of 1000 times faster than the full Boltzmann code and LPT computations. In a simulated full-shape redshift space galaxy power spectrum analysis, we demonstrate that the posteriors obtained using our surrogates are accurate compared to those obtained using the full LPT model. We make our surrogate models public at https://github.com/sfschen/EmulateLSS, so that others may take advantage of the speed gains they provide to enable rapid iteration on analysis settings, something that is essential in complex contemporary large-scale structure analyses.

79 ASTRONOMY AND ASTROPHYSICS↗