Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Differentiability”

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 1,657 records · Page 92
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

Characterization and differentiation of aluminum powders used in improvised explosive devices – Part 1: Proof of concept of the utility of particle micromorphometry

Abstract Aluminum (Al) powders are commonly used in improvised explosive devices as metallic fuels, a component of explosive mixtures. These powders can be obtained readily from industrial‐scale and consumer products, and produced using unsophisticated “kitchen chemistry” techniques. This research demonstrates the potential of automated particle micromorphometry for comparisons between known source and questioned Al powders recovered from IEDs, as well as for insight into the method of Al powder manufacture. Al powder samples were obtained from legitimate manufacturers, and 56 samples were produced “in‐house” from Al‐containing spray paints and ball‐milled Al foils. Transmitted light microscope images of Al powder particles were acquired using an automated stage with automated z‐focus; 17 size and shape parameters were measured for all particles. Approximately 37,000–2,500,000 particles/sample were analyzed using an open‐source statistical package with customized code. Dimensionality reduction was required for processing the large datasets: eight of the 17 measured variables were selected based on inspection of the correlation matrix. Data from four subsamples from each of the 56 samples produced using “in‐house” methods were analyzed using ANOVA to assess the within‐ and between‐sample variation. High within‐sample variation was noted; however, ANOVA and post‐hoc Tukey's honestly significant difference (HSD) tests demonstrated that the between‐sample variation was substantially larger than the within‐sample variation. Each sample could be differentiated from all other samples in the test set. Future experiments will focus on ways to reduce the within‐sample variation, and additional statistical and microanalytical methods to classify sources and confidently constrain the method of Al powder manufacture.

Baldaino, JenaMarie↗
DOE OSTI2019

Differentiating spatial from environmental effects on foliar fungal communities of Populus trichocarpa

Abstract Aim Foliar fungi – pathogens, endophytes, epiphytes – form taxonomically diverse communities that affect plant health and productivity. The composition of foliar fungal communities is variable at spatial scales both small (e.g. individual plants) and large (e.g. continents), yet few studies have attempted to tease apart spatial from climatic factors influencing these communities. Moreover, few studies have sampled in more than 1 year to gauge interannual variation in community structure. Location The Pacific Northwest of western North America. Taxon Foliar fungi associated with the deciduous tree Populus trichocarpa . Methods In two consecutive years, we used DNA metabarcoding to characterize foliar fungal communities of Populus trichocarpa across its geographic range, which encompasses a sharp climatic transition as it crosses the Cascade Mountain Range. We used multivariate analyses to (a) test for and differentiate spatial and environmental factors affecting community composition and (b) test for temporal variation in community composition across spatial and environmental gradients. Results In both study years, we found that foliar fungal community composition varied among sites and between regions (east vs. west of the Cascades). We found that climate explained more variation in community composition than geographic distance, although the majority of variation explained by each was shared. We also found that interannual variation in community composition depended on environmental context: communities located in the dry, eastern portion of the tree's geographic range varied more between study years than those located in the wet, western portion of the tree's range. Main conclusions Our results suggest that the environment plays a greater role in structuring foliar fungal communities than dispersal limitation.

Barge, Edward G.↗
DOE OSTI2022

Differential timing of gene expression and recruitment in independent origins of CAM in the Agavoideae (Asparagaceae)

Crassulacean acid metabolism (CAM) photosynthesis has evolved repeatedly across the plant tree of life, however our understanding of the genetic convergence across independent origins remains hampered by the lack of comparative studies. Here, we explore gene expression profiles in eight species from the Agavoideae (Asparagaceae) encompassing three independent origins of CAM. Using comparative physiology and transcriptomics, we examined the variable modes of CAM in this subfamily and the changes in gene expression across time of day and between well watered and drought-stressed treatments. We further assessed gene expression and the molecular evolution of genes encoding phosphoenolpyruvate carboxylase (PPC), an enzyme required for primary carbon fixation in CAM. Most time-of-day expression profiles are largely conserved across all eight species and suggest that large perturbations to the central clock are not required for CAM evolution. By contrast, transcriptional response to drought is highly lineage specific. Yucca and Beschorneria have CAM-like expression of PPC2 , a copy of PPC that has never been shown to be recruited for CAM in angiosperms. Together the physiological and transcriptomic comparison of closely related C 3 and CAM species reveals similar gene expression profiles, with the notable exception of differential recruitment of carboxylase enzymes for CAM function.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

A Neural Differential Equation Formulation for Modeling Atmospheric Effects in Hyperspectral Images

Atmospheric correction is the process for removing atmospheric effects from spectral data; a necessary step for recovering salient spectral properties. The complex interactions between the atmosphere and light are dominated by absorbance and scattering physics. Existing methods for modeling atmospheric interactions typically rely on deep knowledge of relevant environmental conditions and high-fidelity numerical simulations of the governing physics in order to obtain accurate estimates of these effects. Additionally, existing approaches often require a subject matter expert for pre/post-processing of the data. Model-based approaches for removing atmospheric effects struggle in situations where such domain expertise is not available, and require significant human effort and computational power even when that expertise is available. In contrast, we propose a data-driven approach the uses Neural Differential Equations (NDEs) to accurately learn the interactions between electromagnetic radiation and the atmospheric without access to location specific environmental information. Once trained, the NDE can be applied bi-directionally; to apply or remove atmospheric effects. We demonstrate the effectiveness and utility of these techniques on an example multi-spectral scene.

Koch, James V.↗
DOE OSTI2026

DP-TwoLevel: two-stage gradient subspace learning for differentially private federated learning

Federated learning (FL) enables collaborative model training across distributed data sources without sharing raw data, but faces fundamental challenges in communication efficiency and privacy. Differentially private (DP) training mitigates information leakage but introduces noise that degrades model performance, especially in high-dimensional settings. We propose DP-TwoLevel, a hierarchical gradient projection method that improves utility under fixed DP constraints by exploiting low-dimensional structure in model updates. Our approach learns a two-level PCA-based representation of gradients and applies DP noise in a reduced-dimensional subspace, thereby lowering the effective noise magnitude while preserving dominant signal components. We evaluate the method across three datasets (MNIST, Fashion-MNIST, CIFAR-10) and three privacy regimes (ϵ∈0.5, 1.0, 2.0). Across nine experimental settings, DP-TwoLevel consistently outperforms DP-FedAvg, achieving an average accuracy improvement of 9.44%, with larger gains observed in lower ϵ(higher-noise) regimes (up to +22.31%). We further analyze scalability across models ranging from 100K to 1.49M parameters and identify a variance-based success criterion: performance remains strong when the projection preserves more than 75% of gradient variance, degrades in a marginal regime (65–75%), and fails below this threshold. Our results demonstrate that structure-aware dimensionality reduction can significantly improve the privacy–utility tradeoff in FL without modifying formal privacy guarantees. We also provide empirical evidence of scaling limitations for global projections and motivate per-layer extensions for larger models.

Kotevska, Olivera [ORNL] (ORCID:0000000316772243)↗
DOE OSTI2016

Whole-remnant and maximum-voxel SPECT/CT dosimetry in {sup 131}I-NaI treatments of differentiated thyroid cancer

Purpose: To investigate the possible differences between SPECT/CT based whole-remnant and maximum-voxel dosimetry in patients receiving radio-iodine ablation treatment of differentiated thyroid cancer (DTC). Methods: Eighteen DTC patients were administered 1.11 GBq of {sup 131}I-NaI after near-total thyroidectomy and rhTSH stimulation. Two patients had two remnants, so in total dosimetry was performed for 20 sites. Three SPECT/CT scans were performed for each patient at 1, 2, and 3–7 days after administration. The activity, the remnant mass, and the maximum-voxel activity were determined from these images and from a recovery-coefficient curve derived from experimental phantom measurements. The cumulated activity was estimated using trapezoidal-exponential integration. Finally, the absorbed dose was calculated using S-values for unit-density spheres in whole-remnant dosimetry and S-values for voxels in maximum-voxel dosimetry. Results: The mean absorbed dose obtained from whole-remnant dosimetry was 40 Gy (range 2–176 Gy) and from maximum-voxel dosimetry 34 Gy (range 2–145 Gy). For any given patient, the activity concentrations for each of the three time-points were approximately the same for the two methods. The effective half-lives varied (R = 0.865), mainly due to discrepancies in estimation of the longer effective half-lives. On average, absorbed doses obtained from whole-remnant dosimetry were 1.2 ± 0.2 (1 SD) higher than for maximum-voxel dosimetry, mainly due to differences in the S-values. The method-related differences were however small in comparison to the wide range of absorbed doses obtained in patients. Conclusions: Simple and consistent procedures for SPECT/CT based whole-volume and maximum-voxel dosimetry have been described, both based on experimentally determined recovery coefficients. Generally the results from the two approaches are consistent, although there is a small, systematic difference in the absorbed dose due to differences in the S-values, and some variability due to differences in the estimated effective half-lives, especially when the effective half-life is long. Irrespective of the method used, the patient absorbed doses obtained span over two orders of magnitude.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2020

Three dimensions, two microscopes, one code: Automatic differentiation for x-ray nanotomography beyond the depth of focus limit

Conventional tomographic reconstruction algorithms assume that one has obtained pure projection images, involving no within-specimen diffraction effects nor multiple scattering. Advances in x-ray nanotomography are leading toward the violation of these assumptions, by combining the high penetration power of x-rays, which enables thick specimens to be imaged, with improved spatial resolution that decreases the depth of focus of the imaging system. We describe a reconstruction method where multiple scattering and diffraction effects in thick samples are modeled by multislice propagation and the 3D object function is retrieved through iterative optimization. We show that the same proposed method works for both full-field microscopy and for coherent scanning techniques like ptychography. Our implementation uses the optimization toolbox and the automatic differentiation capability of the open-source deep learning package TensorFlow, demonstrating a straightforward way to solve optimization problems in computational imaging with flexibility and portability.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2021

Learning the solution operator of parametric partial differential equations with physics-informed DeepONets

Partial differential equations (PDEs) play a central role in the mathematical analysis and modeling of complex dynamic processes across all corners of science and engineering. Their solution often requires laborious analytical or computational tools, associated with a cost that is markedly amplified when different scenarios need to be investigated, for example, corresponding to different initial or boundary conditions, different inputs, etc. In this work, we introduce physics-informed DeepONets, a deep learning framework for learning the solution operator of arbitrary PDEs, even in the absence of any paired input-output training data. We illustrate the effectiveness of the proposed framework in rapidly predicting the solution of various types of parametric PDEs up to three orders of magnitude faster compared to conventional PDE solvers, setting a previously unexplored paradigm for modeling and simulation of nonlinear and nonequilibrium processes in science and engineering.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Differential absorption of circularly polarized light by a centrosymmetric crystal

Crystalline solids are governed by universal structure-property relationships derived from their crystal symmetry, leading to paradigmatic rules on what properties they can and cannot exhibit. A long-held structure-property relationship is that centrosymmetric crystals cannot differentially absorb circularly polarized light. In this study, we demonstrate the design, synthesis, and characterization of the centrosymmetric material Li 2 Co 3 (SeO 3 ) 4 , which violates this relationship not by defying symmetry-imposed selection rules but by invoking a photophysical process not previously characterized for crystalline solids. This process originates from an interference between linear dichroism and linear birefringence, referred to as LD-LB, and involves strong chiroptical signals that invert upon sample flipping. Here, in addition to enabling a chiroptical response under centrosymmetry, this process opens up photonic engineering opportunities based on crystalline solids.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Machine Learning of Bacterial Transcriptomes Reveals Responses Underlying Differential Antibiotic Susceptibility

In vitro antibiotic susceptibility testing often fails to accurately predict in vivo drug efficacies, in part due to differences in the molecular composition between standardized bacteriologic media and physiological environments within the body. Here, we investigate the interrelationship between antibiotic susceptibility and medium composition in Escherichia coli K-12 MG1655 as contextualized through machine learning of transcriptomics data. Application of independent component analysis, a signal separation algorithm, shows that complex phenotypic changes induced by environmental conditions or antibiotic treatment are directly traced to the action of a few key transcriptional regulators, including RpoS, Fur, and Fnr. Integrating machine learning results with biochemical knowledge of transcription factor activation reveals medium-dependent shifts in respiration and iron availability that drive differential antibiotic susceptibility. By extension, the data generation and data analytics workflow used here can interrogate the regulatory state of a pathogen under any measured condition and can be applied to any strain or organism for which sufficient transcriptomics data are available.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2017

Application of Backward Differentiation Formulas to Neutron Kinetics Problems

The neutron kinetics equations belong to the class of stiff equations for numerical time integration schemes. In this work, the accuracy and speed of algorithms based on backward differentiation formulas (BDFs) are studied with regard to point and spatial neutron kinetics problems. Using a BDF algorithm with the automatic selection of time step and order, solutions to a number of model problems with both a positive and negative reactivity insertion are analyzed. Plots of numerical cost dependences on local and global errors are presented. The results indicate that algorithms based on BDFs are highly efficient and that their application to nuclear reactor simulation is justified.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2017

On fast solid-body rotation of the solar core and differential (liquid-like) rotation of the solar surface

On the basis of a two-component (two-fluid) hydrodynamic model, it is shown that the probable phenomenon of solar core rotation with a velocity higher than the average velocity of global rotation of the Sun, discovered by the SOHO mission, can be related to fast solid-body rotation of the light hydrogen component of the solar plasma, which is caused by thermonuclear fusion of hydrogen into helium inside the hot dense solar core. Thermonuclear fusion of four protons into a helium nucleus (α-particle) creates a large free specific volume per unit particle due to the large difference between the densities of the solar plasma and nuclear matter. As a result, an efficient volumetric sink of one of the components of the solar substance—hydrogen—forms inside the solar core. Therefore, a steady-state radial proton flux converging to the center should exist inside the Sun, which maintains a constant concentration of hydrogen as it burns out in the solar core. It is demonstrated that such a converging flux of hydrogen plasma with the radial velocity v{sub r}(r) = −βr creates a convective, v{sub r}∂v{sub φ}/∂r, and a local Coriolis, v{sub r}v{sub φ}/r,φ nonlinear hydrodynamic forces in the solar plasma, rotating with the azimuthal velocity v{sub φ}. In the absence of dissipation, these forces should cause an exponential growth of the solid-body rotation velocity of the hydrogen component inside the solar core. However, friction between the hydrogen and helium components of the solar plasma due to Coulomb collisions of protons with α-particles results in a steady-state regime of rotation of the hydrogen component in the solar core with an angular velocity substantially exceeding the global rotational velocity of the Sun. It is suggested that the observed differential (liquid-like) rotation of the visible surface of the Sun (photosphere) with the maximum angular velocity at the equator is caused by sold-body rotation of the solar plasma in the radiation zone and strong turbulence in the tachocline layer, where the turbulent viscosity reaches its maximum value at the equator. There, the tachocline layer exerts the most efficient drag on the less dense outer layers of the solar plasma, which are slowed down due to the interaction with the ambient space plasma (solar wind).

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2018

Stochastic basis adaptation and spatial domain decomposition for partial differential equations with random coefficients

We present a novel approach for high-dimensional stochastic partial differential equations that decomposes the computational domain into smaller disjoint subdomains and reduces the random dimension in each subdomain to save computational costs. The proposed approach combines spatial domain decomposition and stochastic basis adaptation in each subdomain. We adapt the stochastic basis in each subdomain such that the local solution has a low-dimensional random space representation. To compute the stochastic solution in each subdomain, we use the Neumann-Neumann algorithm, where we first estimate the interface solution and then evaluate the interior solution in each subdomain using the interface solution as a boundary condition. The interior solutions in each subdomain are computed independently of each other, which reduces the operation count from $\approx O(N^\alpha)$ to $\approx O(M^\alpha),$ where $N$ is the total number of degrees of freedom, $M$ is the number of degrees of freedom in each subdomain, and the exponent $\alpha>1$ depends on the uncertainty quantification method used. In addition, the localized nature of solutions makes the proposed approach highly parallelizable. We illustrate the accuracy and efficiency of the approach for a steady-state diffusion equation with a random space-dependent diffusion coefficient.

uncertainty quantification, basis adaptation↗
DOE OSTI2026

Hierarchical Gaussian Random Field Sampling for Multilevel Markov Chain Monte Carlo: Coupling Stochastic Partial Differential Equation and the Karhunen–Loève Decomposition

This work introduces structure preserving hierarchical decompositions for sampling Gaussian random fields (GRFs) within the context of multilevel Bayesian inference in high-dimensional space. Existing scalable hierarchical sampling methods, such as those based on stochastic partial differential equations (SPDEs), often reduce the dimensionality of the sample space at the cost of accuracy of inference. Other approaches, such that those based on Karhunen-Loève (KL) expansions, offer sample space dimensionality reduction but sacrifice GRF representation accuracy and ergodicity of the Markov chain Monte Carlo (MCMC) sampler and are computationally expensive for high-dimensional problems. The proposed method integrates the dimensionality reduction capabilities of KL expansions with the scalability of SPDE-based sampling, thereby providing a robust, unified framework for high-dimensional uncertainty quantification (UQ) that is scalable and accurate, preserves ergodicity, and offers dimensionality reduction of the sample space. The hierarchy in our multilevel algorithm is derived from the geometric multigrid hierarchy. By constructing a hierarchical decomposition that maintains the covariance structure across the levels in the hierarchy, the approach enables efficient coarse-to-fine sampling while ensuring that all samples are drawn from the desired distribution. The effectiveness of the proposed method is demonstrated on a benchmark subsurface flow problem, demonstrating its effectiveness in improving computational efficiency and statistical accuracy. Furthermore, our proposed technique is more efficient and accurate and displays better convergence properties than existing methods for high-dimensional Bayesian inference problems.

Gaussian random fields↗
DOE OSTI2021

$J/\psi $-p scattering length from the total and differential photoproduction cross sections

The J/ψ-p scattering length, αJ/ψp, can be extracted from the J/ψ photoproduction cross section near threshold using the Vector Meson Dominance (VMD) model to relate the reaction γp→J/ψp to J/ψp→J/ψp. Such estimates based on experimental data result in values for |αJ/ψp|, which are much lower than most of the theoretical predictions. In this work, we study the relations between the different results, depending on the use of the total or the differential cross sections, and the method of extrapolating the data to threshold in the case of a low-statistics data sample, such as the near threshold J/ψ photoproduction dataset. We estimate a range for |αJ/ψp| of 0.003–0.025 fm as extracted from experimental data within the VMD model and discuss possible reasons for such lower values compared to the theoretical results.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2018

Negative differential conductivity in liquid aluminum from real-time quantum simulations

The conduction of electricity in materials is usually described by Ohm’s law, which is a first order approximation to a more complex and non-linear behavior. It is well known that in some semiconductors, the conductivity, the constant that relates voltage and current, changes for high enough currents. In this work we predict for the first time that a metal, liquid aluminum, exhibits negative-differential conductivity, a non-linear effect where the current decreases as the applied voltage is increased. We observe this change in the conductivity for very high current densities of the order of 10 12 - 10 13 A/cm 2 . Our predictions are based on a computational approach that can atomistically model, for the first time, non-linear effects in the conductivity from first principles by following in real-time the quantum dynamics of the electrons. From our simulations, we find that the change in the non-linear conductivity emerges from a competition between the currentinduced accumulation of charge around the nuclei, which increases the scattering of the conduction electrons, and a decreasing ion-scattering cross-section at high currents. Our results illustrate how normal matter behaves under extreme fields that will become available from free electron lasers and other future experiments.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2016

Measurement of the double-differential inclusive jet cross section in proton–proton collisions at $\sqrt{s} = 13\,\text {TeV} $

A measurement of the double-differential inclusive jet cross section as a function of jet transverse momentum $p_{\mathrm {T}} $ and absolute jet rapidity $|y |$ is presented. The analysis is based on proton–proton collisions collected by the CMS experiment at the LHC at a centre-of-mass energy of 13 $\,\text {TeV}$ . The data samples correspond to integrated luminosities of 71 and 44 $\,\text {pb}^\text {-1}$ for $|y |<3$ and $3.2<|y |<4.7$ , respectively. Jets are reconstructed with the anti- $k_{\mathrm {t}} $ clustering algorithm for two jet sizes, R, of 0.7 and 0.4, in a phase space region covering jet $p_{\mathrm {T}} $ up to 2 $\,\text {TeV}$ and jet rapidity up to $|y |$ = 4.7. Predictions of perturbative quantum chromodynamics at next-to-leading order precision, complemented with electroweak and nonperturbative corrections, are used to compute the absolute scale and the shape of the inclusive jet cross section. The cross section difference in R, when going to a smaller jet size of 0.4, is best described by Monte Carlo event generators with next-to-leading order predictions matched to parton showering, hadronisation, and multiparton interactions. In the phase space accessible with the new data, this measurement provides a first indication that jet physics is as well understood at $\sqrt{s}=13\,\text {TeV} $ as at smaller centre-of-mass energies.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗