SEARCH · Engineering Papers

Results for “function mapping”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 451 records · Page 25

All sources NASA NTRS DOE OSTI

DOE OSTI2023

On the influence of over-parameterization in manifold based surrogates and deep neural operators

Constructing accurate and generalizable approximators (surrogate models) for complex physico-chemical processes exhibiting highly non-smooth dynamics is challenging. The main question is what type of surrogate models we should construct and should these models be under-parameterized or over-parameterized. In this work, we propose new developments and perform comparisons for two promising approaches: manifold-based polynomial chaos expansion (m-PCE) and the deep neural operator (DeepONet), and we examine the effect of over-parameterization on generalization. While m-PCE enables the construction of a mapping by first identifying low-dimensional embeddings of the input functions, parameters, and quantities of interest (QoIs), a neural operator learns the nonlinear mapping via the use of deep neural networks. Here, we demonstrate the performance of these methods in terms of generalization accuracy by solving the 2D time-dependent Brusselator reaction-diffusion system with uncertainty sources, modeling an autocatalytic chemical reaction between two species. We first propose an extension of the m-PCE by constructing a mapping between latent spaces formed by two separate embeddings of the input functions and the output QoIs. To further enhance the accuracy of the DeepONet, we introduce weight self-adaptivity in the loss function. We demonstrate that the performance of m-PCE and DeepONet is comparable for cases of relatively smooth input-output mappings. However, when highly non-smooth dynamics is considered, DeepONet shows higher approximation accuracy. We also find that for m-PCE, modest over-parameterization leads to better generalization, both within and outside of distribution, whereas aggressive over-parameterization leads to over-fitting. In contrast, an even highly over-parameterized DeepONet leads to better generalization for both smooth and non-smooth dynamics. Furthermore, we compare the performance of the above models with another recently proposed operator learning model, the Fourier Neural Operator, and show that its over-parameterization also leads to better generalization. Taken together, our studies show that m-PCE can provide very good accuracy at very low training cost, whereas a highly over-parameterized DeepONet can provide better accuracy and robustness to noise but at higher training cost. In both methods, the inference cost is negligible.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2015

The climatological distribution of extreme Arctic winds and implications for ocean and sea ice processes

Abstract Some of the strongest near‐surface winds on Earth form in the Arctic and sub‐Arctic due to intense midlatitude cyclones and mesoscale processes, and these strong surface winds have important impacts on ocean and sea ice processes. We examine the climatological distribution of over‐ocean, near‐surface wind speeds within a Pan‐Arctic domain for 18 years (1990–2007) in four gridded data sets: the European Centre for Medium‐Range Weather Forecasts Interim reanalysis (ERA‐I), the Climate Forecast System Reanalysis, version 2 of the Common Ocean‐Ice Reference Experiment data set, and a regional climate simulation generated using the Weather Research and Forecasting (WRF) model run at 50 km (WRF50) horizontal resolution with ERA Interim as lateral boundary conditions. We estimate probability density functions, the annual cycle, and map the 50th and 99th percentile winds. We then perform the same statistical analysis of winds for 2 years when 10 km WRF data are available (June 2005 to May 2007); despite the much shorter time period, the Pan‐Arctic statistics are very similar to those from the 18 year analysis. We repeat the wind speed statistical analysis within a subdomain surrounding Greenland and find that WRF10 has consistently larger maximum wind speeds, but this difference only appears at wind speed percentiles higher than 99%. Differences in the 99th percentile wind speeds are spatially heterogeneous. An investigation of surface fluxes within WRF50 and WRF10 reveals unrealistically large sensible heat fluxes along the sea ice edge, and the geographic distribution and magnitude of these fluxes is shown to be sensitive to sea ice representation in WRF.

Hughes, Mimi↗

DOE OSTI2014

Universal Design Principles for Cascade Heterojunction Solar Cells with High Fill Factors and Internal Quantum Efficiencies Approaching 100%

Cascade heterojunction (CHJ) organic solar cells have recently emerged as an alternative to conventional bulk heterojunctions and series-connected tandems due to their significant promise for high internal quantum efficiency (IQE) and broad spectral coverage. However, CHJ devices thus far have also exhibited poor fill factor (FF), resulting in minimal enhancements (or even decreases) in power conversion efficiency (PCE) when compared with single heterojunction (SHJ) cells. In this study, the major variables controlling the CHJ maximum power point and FF are determined using a combinatorial approach. By matching the maximum power point voltage (VMPP) of the constituent parallel-connected heterojunctions (subjunctions) and minimizing the injection barriers intrinsic to CHJs, high FF and PCE can be achieved. Optimized CHJ devices are demonstrated with >99% IQE in the interlayer and a 46% increase in PCE compared to a SHJ reference (4.1% versus 2.8%). Devices with a transparent exciton dissociation layer (EDL)/interlayer/acceptor structure are employed, such that each CHJ has absorption efficiency identical to its interlayer/acceptor SHJ counterpart. Using these results, a clear map of performance as a function of material parameters is developed, providing straightforward, universal design rules to guide future engineering of molecules and layer architectures for CHJ organic photovoltaic devices.

36 MATERIALS SCIENCE↗

DOE OSTI2025

Kernel Manifolds: Nonlinear‐Augmentation Dimensionality Reduction Using Reproducing Kernel Hilbert Spaces

This paper generalizes recent advances on quadratic manifold (QM) dimensionality reduction by developing kernel methods-based nonlinear-augmentation dimensionality reduction. QMs, and more generally feature map-based nonlinear corrections, augment linear dimensionality reduction with a nonlinear correction term in the reconstruction map to overcome approximation accuracy limitations of purely linear approaches. While feature map-based approaches typically learn a least squares optimal polynomial correction term, we generalize this approach by learning an optimal nonlinear correction from a user-defined reproducing kernel Hilbert space. Our approach allows one to impose arbitrary nonlinear structure on the correction term, including polynomial structure, and includes feature map and radial basis function-based corrections as special cases. Furthermore, our method has relatively low training cost and has monotonically decreasing error as the latent space dimension increases. In conclusion, we compare our approach to proper orthogonal decomposition and several recent QM approaches on data from several example problems.

kernel methods↗

DOE OSTI2023

Grain boundary softening from stress assisted helium cavity coalescence in ultrafine-grained tungsten

The formation of helium cavities in coarse-grained materials produces hardening proportional to the number density and size of the cavities and due to the interaction of dislocations with intragranular helium defects. In nanostructured metals containing a high density of interfacial sinks, preferential cavity formation in the grain boundaries instead produces softening that is often attributed to enhanced interfacial plasticity. Here, employing two grades of ultrafine-grained tungsten, we explore this effect using targeted implantation studies to map cavity evolution as a function of the irradiation conditions and quantify its impact on the mechanical response through nanoindentation. Softening is reported at implantation temperatures above the threshold for preferential grain boundary cavity formation but at a sufficiently low fluence prior to the growth of intragranular cavities. Collective changes in the mean cavity size, density, and morphology beneath a residual impression on an implanted surface indicate that cavity coalescence accompanied the reduction in hardness. Complementary atomistic simulations demonstrate that, in tungsten grain structures exhibiting softening, grain boundary bubble coalescence is driven by stress concentrations that further act to localize strain in the grain boundaries through cooperative deformation processes involving local atomic shuffling and sliding, dislocation emission, and even the nucleation of unstable twinning events.

36 MATERIALS SCIENCE↗

DOE OSTI2023

Reliable extrapolation of deep neural operators informed by physics or sparse observations

Deep neural operators can learn nonlinear mappings between infinite-dimensional function spaces via deep neural networks. As promising surrogate solvers of partial differential equations (PDEs) for real-time prediction, deep neural operators such as deep operator networks (DeepONets) provide a new simulation paradigm in science and engineering. Pure data-driven neural operators and deep learning models, in general, are usually limited to interpolation scenarios, where new predictions utilize inputs within the support of the training set. However, in the inference stage of real-world applications, the input may lie outside the support, i.e., extrapolation is required, which may result to large errors and unavoidable failure of deep learning models. Here, we address this challenge of extrapolation for deep neural operators. First, we systematically investigate the extrapolation behavior of DeepONets by quantifying the extrapolation complexity, via the 2-Wasserstein distance between two function spaces and propose a new strategy of bias–variance trade-off for extrapolation with respect to model capacity. Subsequently, we develop a complete workflow, including extrapolation determination, and we propose five reliable learning methods that guarantee a safe prediction under extrapolation by requiring additional information—the governing PDEs of the system or sparse new observations. The proposed methods are based on either fine-tuning a pre-trained DeepONet or multifidelity learning. We demonstrate the effectiveness of the proposed framework for various types of parametric PDEs. Furthermore, our systematic comparisons provide practical guidelines for selecting a proper extrapolation method depending on the available information, desired accuracy, and required inference speed.

42 ENGINEERING↗

DOE OSTI2024

Separable physics-informed DeepONet: Breaking the curse of dimensionality in physics-informed machine learning

The deep operator network (DeepONet) has shown remarkable potential in solving partial differential equations (PDEs) by mapping between infinite-dimensional function spaces using labeled datasets. However, in scenarios lacking labeled data, the physics-informed DeepONet (PI-DeepONet) approach, which utilizes the residual loss of the governing PDE to optimize the network parameters, faces significant computational challenges, particularly due to the curse of dimensionality. This limitation has hindered its application to high-dimensional problems, making even standard 3D spatial with 1D temporal problems computationally prohibitive. Additionally, the computational requirement increases exponentially with the discretization density of the domain. Here, to address these challenges and enhance scalability for high-dimensional PDEs, we introduce the Separable physics-informed DeepONet (Sep-PI-DeepONet). This framework employs a factorization technique, utilizing sub-networks for individual one-dimensional coordinates, thereby reducing the number of forward passes and the size of the Jacobian matrix required for gradient computations. By incorporating forward-mode automatic differentiation (AD), we further optimize computational efficiency, achieving linear scaling of computational cost with discretization density and dimensionality, making our approach highly suitable for high-dimensional PDEs. We demonstrate the effectiveness of Sep-PI-DeepONet through three benchmark PDE models: the viscous Burgers’ equation, Biot’s consolidation theory, and a parameterized heat equation. Our framework maintains accuracy comparable to the conventional PI-DeepONet while reducing training time by two orders of magnitude. Notably, for the heat equation solved as a 4D problem, the conventional PI-DeepONet was computationally infeasible (estimated 289.35 h), while the Sep-PI-DeepONet completed training in just 2.5 h. These results underscore the potential of Sep-PI-DeepONet in efficiently solving complex, high-dimensional PDEs, marking a significant advancement in physics-informed machine learning.

Neural operator↗

DOE OSTI2025

Dual-season thermal energy storage-integrated heat pump system for load shifting: experiments, modeling and nationwide analysis

Here, this paper proposes and validates a dual-season thermal energy storage-integrated heat pump (TES-HP) system that shifts building electric load in both cooling and heating modes using a single phase-change TES. The design uses the TES as a heat sink for the condenser in summer and a heat source for the evaporator in winter, thereby reducing on-peak electric demand while ensuring compatible with existing air-distribution systems. A 14-kW TES-HP prototype with six operating modes and novel refrigerant charge management strategies was experimentally tested under varied conditions. Subsequently, data-driven polynomial performance curves were developed and validated against experiments, then coupled with a phase change material (PCM) model and a calibrated building model within a rule-based controller. Single-day case studies demonstrated effective on-peak demand reduction while maintaining thermal comfort. During cooling, hourly power was lowered by up to 1.5 kW; in heating, average hourly power decreased from over 3.5 kW to 1.8 kW, compared to an HP-only system, achieving 64.5 % load shifting, and the need for resistance heating was eliminated. Seasonal simulations showed typical on-peak electricity savings of 25–35 % in cooling and 40–65 % in heating, with the largest benefits on the hottest and coldest days. Extended response-surface analysis and nationwide mapping quantified load shifting as functions of ambient and TES temperatures, revealing a seasonal trade-off in TES phase-change temperature. These results demonstrate the TES-HP system as a practical and scalable solution for grid-interactive buildings that reduces on-peak demand and electricity use while maintaining thermal comfort and enhancing grid flexibility and reliability.

Grid-interactive buildings↗

DOE OSTI2014

The Chlamydomonas genome project: a decade on

The green alga Chlamydomonas reinhardtii is a popular unicellular organism for studying photosynthesis, cilia biogenesis, and micronutrient homeostasis. Ten years since its genome project was initiated an iterative process of improvements to the genome and gene predictions has propelled this organism to the forefront of the omics era. Housed at Phytozome, the plant genomics portal of the Joint Genome Institute (JGI), the most up-to-date genomic data include a genome arranged on chromosomes and high-quality gene models with alternative splice forms supported by an abundance of whole transcriptome sequencing (RNA-Seq) data. Here, we present here the past, present, and future of Chlamydomonas genomics. Specifically, we detail progress on genome assembly and gene model refinement, discuss resources for gene annotations, functional predictions, and locus ID mapping between versions and, importantly, outline a standardized framework for naming genes.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2016

Complexation and phase evolution at dimethylformamide-Ag(111) interfaces

The interaction of solvent molecules with metallic surfaces impacts many interfacial chemical processes. We investigate the chemical and structure evolution that follows adsorption of the polar solvent dimethylformamide (DMF) on Ag(111). An Ag(DMF) 2 coordination complex forms spontaneously by DMF etching of Ag(111), yielding mixed films of the complexes and DMF. Utilizing ultrahigh vacuum scanning tunneling microscopy (UHV-STM), in combination with X-ray photoelectron spectroscopy (XPS) and density functional theory (DFT) computations, we map monolayer phases from the 2-D gas regime, consisting of a binary mixture of DMF and Ag(DMF) 2 , through the saturation monolayer limit, in which these two chemical species phase separate into ordered islands. Structural models for the near-square DMF phase and the chain-like Ag(DMF) 2 phase are presented and supported by DFT computation. Interface evolution is summarized in a surface pressure-composition phase diagram, which allows structure prediction over arbitrary experimental conditions. In conclusion, this work reveals new surface coordination chemistry for an important electrolyte-electrode system, and illustrates how surface pressure can be used to tune monolayer phases.

2-D phases↗

DOE OSTI2025

Gapless dispersive continuum in a modulated quantum kagome antiferromagnet

The pursuit of quantum spin liquid (QSL) states in condensed matter physics has drawn attention to kagome antiferromagnets (AFM) where a two-dimensional corner-sharing network of triangles frustrates conventional magnetic orders. While quantum kagome AFMs based on Cu 2+ (3d 9 , s = 1/2) ions have been extensively studied, there is so far little work beyond copper-based systems. Here we present our bulk magnetization, specific heat and neutron scattering studies on single crystals of a new titanium fluorides Cs 8 RbK 3 Ti 12 F 48 where Ti 3+ (3d 1 , s = 1/2) ions form a modulated quantum kagome antiferromagnet that does not order magnetically down to 1.5 K. Our comprehensive map of the dynamic response function $S$($Q$,$\hslash$$\omega$) acquired at 1.5 K where the heat capacity is T-linear reveals a dispersive continuum emanating from soft lines that extend along (100). The data indicate fractionalized spinon-like excitations with quasi-one-dimensional dispersion within a quasi-two-dimensional spin system.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

DOE OSTI2023

Transient pattern formation in an active matter contact poisoning model

Abstract One of the most notable features in repulsive particle based active matter systems is motility-induced-phase separation (MIPS) where a dense, often crystalline phase and low density fluid coexist. Most active matter studies involve time-dependent activity; however, there are many active systems where individual particles transition from living or moving to dead or nonmotile due to lack of fuel, infection, or poisoning. Here we consider an active matter particle system at densities where MIPS does not occur. When we add a small number of infected particles that can poison other particles, rendering them nonmotile, we find a rich variety of time dependent pattern formation, including MIPS, a wetting phase, and a fragmented state formed when mobile particles plow through a nonmotile packing. We map the patterns as a function of time scaled by epidemic duration, and show that the pattern formation is robust for a wide range of poisoning rates and activity levels. We also show that pattern formation does not occur in a random death model, but requires the promotion of nucleation by contact poisoning. Our results should be relevant to biological and active matter systems where there is some form of poisoning, death, or transition to nonmotility.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

DOE OSTI2014

On the photoelectron velocity-map imaging of lutetium monoxide anion LuO{sup −}

We report a combined photoelectron velocity-map imaging spectroscopy and density functional theory investigation on lutetium monoxide anion. Transition between the X {sup 1}Σ{sup +} anion electronic ground state and the neutral X {sup 2}Σ{sup +} electronic ground state is observed. Vibrationally resolved spectra were obtained at four different photon energies, providing a wealth of spectroscopic information for the electronic ground states of the anionic lutetium monoxide and corresponding neutral species. Franck-Condon simulations of the ground-state transition are performed to assign vibrational structure in the spectra and to assist in identifying the observed spectral bands. The electronic ground state of LuO{sup −} is found to have a vibrational frequency of 743 ± 10 cm{sup −1} and an equilibrium bond length of 1.841 Å. The electron affinity of LuO is measured to be 1.624 ± 0.002 eV. The fundamental frequency of ground-state LuO is estimated to be 839 ± 10 cm{sup −1}.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2014

Time-domain analysis of spin-torque induced switching paths in nanoscale CoFeB/MgO/CoFeB magnetic tunnel junction devices

We performed thousands of single-shot, real-time measurements of spin-transfer-torque induced switching in nanoscale CoFeB/MgO/CoFeB magnetic tunnel junctions having in-plane magnetizations. Our investigation discovered a variety of switching paths occurring in consecutive, nominally identical switching trials of a single device. By mapping the voltage as a function of time to an effective magnetization angle, we determined that reversal of a single device occurs via a variety of thermally activated paths. Our results show a complex switching behavior that has not been captured by previous observations and cannot be fully explained within the simple macrospin model.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

DOE OSTI2024

Fluctuating charge-density-wave correlations in the three-band Hubbard model

The high-temperature superconducting cuprates host unidirectional spin- and charge-density-wave orders that can intertwine with superconductivity in nontrivial ways. While the charge components of these stripes have now been observed in nearly all cuprate families, their detailed evolution with doping varies across different materials and at high and low temperatures. We address this problem using nonperturbative determinant quantum Monte Carlo calculations for the three-band Hubbard model. Using an efficient implementation, we can resolve the model’s fluctuating spin and charge modulations and map their evolution as a function of the charge transfer energy and doping. We find that the incommensurability of the charge modulations is decoupled from the spin modulations and decreases with hole doping, consistent with experimental measurements at high temperatures. These findings support the proposal that the high-temperature charge correlations are distinct from the intertwined stripe order observed at low-temperature and in the single-band Hubbard model.

42 ENGINEERING↗

DOE OSTI2011

WHAT DRIVES THE EXPANSION OF GIANT H II REGIONS?: A STUDY OF STELLAR FEEDBACK IN 30 DORADUS

Observations show that star formation is an inefficient and slow process. This result can be attributed to the injection of energy and momentum by stars that prevents free-fall collapse of molecular clouds. The mechanism of this stellar feedback is debated theoretically; possible sources of pressure include the classical warm H II gas, the hot gas generated by shock heating from stellar winds and supernovae, direct radiation of stars, and the dust-processed radiation field trapped inside the H II shell. In this paper, we measure observationally the pressures associated with each component listed above across the giant H II region 30 Doradus in the Large Magellanic Cloud. We exploit high-resolution, multi-wavelength images (radio, infrared, optical, ultraviolet, and X-ray) to map these pressures as a function of position. We find that radiation pressure dominates within 75 pc of the central star cluster, R136, while the H II gas pressure dominates at larger radii. By contrast, the dust-processed radiation pressure and hot gas pressure are generally weak and not dynamically important, although the hot gas pressure may have played a more significant role at early times. Based on the low X-ray gas pressures, we demonstrate that the hot gas is only partially confined and must be leaking out the H II shell. Additionally, we consider the implications of a dominant radiation pressure on the early dynamics of 30 Doradus.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2012

A PER-BASELINE, DELAY-SPECTRUM TECHNIQUE FOR ACCESSING THE 21 cm COSMIC REIONIZATION SIGNATURE

A critical challenge in measuring the power spectrum of 21 cm emission from cosmic reionization is compensating for the frequency dependence of an interferometer's sampling pattern, which can cause smooth-spectrum foregrounds to appear unsmooth and degrade the separation between foregrounds and the target signal. In this paper, we present an approach to foreground removal that explicitly accounts for this frequency dependence. We apply the delay transformation introduced in Parsons and Backer to each baseline of an interferometer to concentrate smooth-spectrum foregrounds within the bounds of the maximum geometric delays physically realizable on that baseline. By focusing on delay modes that correspond to image-domain regions beyond the horizon, we show that it is possible to avoid the bulk of smooth-spectrum foregrounds. We map the point-spread function of delay modes to k-space, showing that delay modes that are uncorrupted by foregrounds also represent samples of the three-dimensional power spectrum, and can be used to constrain cosmic reionization. Because it uses only spectral smoothness to differentiate foregrounds from the targeted 21 cm signature, this per-baseline analysis approach relies on spectrally and spatially smooth instrumental responses for foreground removal. For sufficient levels of instrumental smoothness relative to the brightness of interfering foregrounds, this technique substantially reduces the level of calibration previously thought necessary to detect 21 cm reionization. As a result, this approach places fewer constraints on antenna configuration within an array, and in particular, facilitates the adoption of configurations that are optimized for power-spectrum sensitivity. Under these assumptions, we demonstrate the potential for the Precision Array for Probing the Epoch of Reionization (PAPER) to detect 21 cm reionization at an amplitude of 10 mK{sup 2} near k {approx} 0.2 h Mpc{sup -1} with 132 dipoles in 7 months of observing.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2022

Directed motion of liquid crystal skyrmions with oscillating fields

Abstract Using continuum simulations, we show that under a sinusoidal electric field, liquid crystal skyrmions undergo periodic shape oscillations which produce controlled directed motion. The speed of the skyrmion is non-monotonic in the frequency of the applied field, and exhibits multiple reversals of the motion as a function of changing frequency. We map out the dynamical regime diagram of the forward and reverse motion for two superimposed ac driving frequencies, and show that the reversals and directed motion can occur even when only a single ac driving frequency is present. Using pulsed ac driving, we demonstrate that the motion arises due to an asymmetry in the relaxation times of the skyrmion shape. We discuss the connection between our results and ratchet effects observed in systems without asymmetric substrates.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗