Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “generative diffusion”

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 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
DOE OSTI2016

INTERSTELLAR PICKUP ION ACCELERATION IN THE TURBULENT MAGNETIC FIELD AT THE SOLAR WIND TERMINATION SHOCK USING A FOCUSED TRANSPORT APPROACH

We study the physics of locally born interstellar pickup proton acceleration at the nearly perpendicular solar wind termination shock (SWTS) in the presence of a random magnetic field spiral angle using a focused transport model. Guided by Voyager 2 observations, the spiral angle is modeled with a q -Gaussian distribution. The spiral angle fluctuations, which are used to generate the perpendicular diffusion of pickup protons across the SWTS, play a key role in enabling efficient injection and rapid diffusive shock acceleration (DSA) when these particles follow field lines. Our simulations suggest that variation of both the shape ( q -value) and the standard deviation ( σ -value) of the q -Gaussian distribution significantly affect the injection speed, pitch-angle anisotropy, radial distribution, and the efficiency of the DSA of pickup protons at the SWTS. For example, increasing q and especially reducing σ enhances the DSA rate.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

Neural Network Code MATLAB

SAND2022-12606 O Neural Network Code MATLAB reads in an excel file, processes the data, trains a neural network, and tests the neural network. The code generates predictions of diffusion constants based on the data in the excel file. Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.

Allers, Joshua↗
DOE OSTI2017

Laser-induced generation of surface periodic structures in media with nonlinear diffusion

A model of fast formation of high-contrast periodic structure appearing on a semiconductor surface under action of laser radiation is proposed. The process of growing a surface structure due to the interaction surface plasmon- polaritons excited on nonequilibrium electrons with incident laser radiation are considered in the framework of a medium with nonlinear diffusion of nonequilibrium carriers (defects). A resonance effect of superfast pico- and subpicosecond amplification of the plasmon-polariton structure generated on the surface, the realization of which can result in a high-contrast defect lattice.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2019

Tokamak parameter regimes with low toroidal momentum diffusivity

The strong, sharp flow structures that are seen frequently in tokamak cores, and large amplitude spontaneous global toroidal rotation are both surprising in light of current theories where toroidal flow evolution is dominantly diffusive. Mechanisms for spontaneously generating strong poloidal shear flows have been extensively investigated, but these processes were thought not to apply to toroidal flows. We confirm, however, that there is a regime with near-zero toroidal momentum diffusivity, where toroidal flow structures are spontaneously generated, as shown in earlier global gyrokinetic simulations. This allows strong rotation with negligible applied external torque, and the regime where this occurs is also favourable for strong turbulence stabilisation by rotation. In conclusion, the transition to low momentum diffusivity occurs for tight aspect ratio, low safety factor and at low levels of heat flux, in agreement with a simple zero-dimensional model for the flow–turbulence interaction.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2023

Flash Diffusivity Measurement of Semi-Porous Insulation Material Intended for Radioisotope Thermoelectric Generators

This research involves the analysis of flash-heating data pursuant to find the thermal diffusivity of a semi-porous insulation material (Min-K) that is being considered for use in a radioisotope thermoelectric generator (RTG). An RTG uses radioactive nuclear fuel to produce electricity through a temperature difference imposed on a bimetallic thermocouple. Insulation is required to protect sensitive equipment from high temperatures and to conserve heat in the fuel. Using flash diffusivity temperature data, various simulations were fitted in order to find the most appropriate heat transfer model for the experiments. Four models were allowed to compete and the standard deviation of the residuals were compared for each model in evaluating model performance. The residuals are simply the difference in temperatures between the measurements and the mathematical models at each measurement point.

McMasters, Robert L.↗
DOE OSTI2017

Pulsed Gas Lasers Pumped by a Runaway Electron Initiated Discharge

The generation parameters are investigated in a runaway electron preionized diffuse discharge (REP DD). Laser generation is produced in different spectral bands from the IR to VUV range. New modes of the nitrogen laser operation are obtained. Ultimate efficiencies of N{sub 2}- and nonchain HF(DF)-lasers are achieved. A possibility of increasing the pulse durations of XeF-, KrF-, ArF- and VUV F{sub 2}- lasers (157 nm) in an oscillating REP DD is shown. The efficiencies of VUV- and UV-generation comparable with that of a laser pumped by a self-sustained volume discharge with preionization are gained.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2025

Enhancing Unknown Waveform Detection by Learning Intra and Inter-domain Dependencies with Advanced Attention Fusion Mechanisms

Detection of unknown waveforms in mission-critical communications is a crucial area of interest for the Department of Energy (DoE). Traditional methods and recent deep learning-based approaches often assume that the training set includes all possible classes, which is impractical for detecting new waveforms. This limitation gives rise to the problem of open-set recognition (OSR), which involves correctly identifying known classes while detecting and rejecting unknown or unseen classes. To address this limitation, we propose a novel dual-domain complex-valued neural architecture that jointly processes time-domain and frequency-domain signal representations using transformer mechanisms. A transformer model is a deep learning architecture that uses self-attention mechanisms to process and learn relationships in sequential data. Our model employs a cosine similarity loss to extract domain-specific features and incorporates a transformer architecture in the latent space to weigh the importance of different features from the time and frequency domains. The transformer layer includes stacked self-attention and cross-attention modules to learn intra-domain and inter-domain dependencies, creating a more holistic signal representation. An attention-based fusion module intelligently combines the time and frequency-domain features using multi-head attention, enabling the network to learn the optimal feature for each domain in each input signal. Quantitative results demonstrate the impact of these architectural choices on overall performance, showing significant improvement after incorporating self and cross-attention modules and using complex attention fusion over simple weighted fusion. Our ongoing work will focus on addressing the limitations of threshold-based OSR methods by developing a novel generative framework that integrates a conditional diffusion probabilistic model (DPM). DPM is a generative framework that learns to synthesize complex data by reversing a gradual noising process using a neural network trained to denoise step-by-step. Our goal is to leverage the inherent strengths of DPMs for identifying unknown signals more robustly. One primary advantage of using a DPM is its ability to provide a more reliable anomaly score based on the model's reconstruction error, rather than relying solely on classifier confidence. Additionally, the iterative denoising process of DPMs makes this approach naturally resilient to low Signal-to-Noise Ratio (SNR) conditions, where traditional methods often fail. By implementing this generative framework, we aim to enhance the model's capability to accurately detect unknown waveforms and maintain performance in challenging environments.

99 - GENERAL AND MISCELLANEOUS↗
DOE OSTI2014

COSMOLOGICAL SIMULATIONS OF GALAXY FORMATION WITH COSMIC RAYS

We investigate the dynamical impact of cosmic rays (CR) in cosmological simulations of galaxy formation using adaptive-mesh refinement simulations of a 10{sup 12} M {sub ☉} halo. In agreement with previous work, a run with only our standard thermal energy feedback model results in a massive spheroid and unrealistically peaked rotation curves. However, the addition of a simple two-fluid model for CRs drastically changes the morphology of the forming disk. We include an isotropic diffusive term and a source term tied to star formation due to (unresolved) supernova-driven shocks. Over a wide range of diffusion coefficients, the CRs generate thin, extended disks with a significantly more realistic (although still not flat) rotation curve. We find that the diffusion of CRs is key to this process, as they escape dense star-forming clumps and drive outflows within the more diffuse interstellar medium.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2016

Modeling the formation of methane hydrate-bearing intervals in fine-grained sediments

Sediment grain size exerts a fundamental control on how methane hydrates are distributed within the pore space. Fine-grained muds are the predominant sediments in continental margins, and hydrates in these sediments have often been observed in semi-vertical veins and fractures. In several instances, these hydrate veins/fractures are found in discrete depth intervals a few tens meters thick within the gas hydrate stability zone (GHSZ) surrounded by hydrate-free sediments above and below. As they are not obviously connected with free gas occurring beneath the base of the GHSZ, these isolated hydrate-bearing intervals have been interpreted as formed by microbial methane generated in situ. To investigate further the formation of these hydrate deposits, we applied a time-dependent advection-diffusion-reaction model that includes the effects of sedimentation, solute diffusion, and microbial methane generation. The microbial methane generation term depends on the amount of metabolizable organic carbon deposited at the seafloor, which is degraded at a prescribed rate resulting in methane formation beneath the sulfate reduction zone. In the model, methane hydrate precipitates once the dissolved methane concentration is greater than solubility, or hydrate dissolves if concentration goes below solubility. If the deposition of organic carbon at the seafloor is kept constant in time, we found that the predicted amounts of hydrate formed in discrete intervals within the GHSZ are much less than those estimated from observations. We then investigated the effect of temporal variations in the deposition of organic carbon. If greater amounts of organic carbon are deposited during some time interval, methane generation is enhanced during burial in the corresponding sediment interval. With variations in organic carbon deposition that are consistent with observations in continental margin sediments, we were able to reproduce the methane hydrate contents estimated in discrete depth intervals. Our results support the suggestion that in situ microbial methane generation is the source for hydrates within fine-grained sediments.

Malinverno, Alberto↗
DOE OSTI2026

Enabling probabilistic learning on manifolds through double diffusion maps

Here, we present a generative learning framework for probabilistic sampling that extends Probabilistic Learning on Manifolds (PLoM), which is designed to generate statistically consistent realizations of a random vector in a finite-dimensional Euclidean space, informed by a (representative) set of observations. In its original form, PLoM constructs a reduced-order probabilistic model by combining three main components: (a) kernel density estimation to approximate the underlying probability measure, (b) Diffusion Maps to characterize the manifold of the data, and (c) a reduced-order Itô Stochastic Differential Equation (ISDE) to sample from the learned distribution. However, its sampling dynamics are posed in the ambient space and the retained number of reduced coordinates is chosen by projection-reconstruction error. In practice, this often (i) requires more coordinates than the data’s intrinsic dimension to achieve stable sampling and (ii) lacks a smooth, basis-independent lifting back to the data domain; moreover, standard Diffusion Maps emphasize harmonic eigenfunctions and can miss non-harmonic latent structure. We address these limitations by decoupling geometry learning from sampling: a first Diffusion Maps pass identifies non-harmonic coordinates on which we formulate a full-order ISDE directly in the latent space, while Double Diffusion Maps captures multiscale geometric features and Geometric Harmonics (GH) learns a smooth lifting map to the ambient variables that is independent of the particular diffusion basis. This hybrid design preserves the system’s dynamical richness with a compact geometric representation and enables principled out-of-sample inference. The effectiveness and robustness of the proposed method are illustrated through two numerical studies: one based on data generated from two-dimensional Hermite polynomial functions and another based on high-fidelity simulations of a detonation wave in a reactive flow.

Double diffusion maps↗
DOE OSTI2018

Generation and healing of porosity in high purity copper by high-pressure torsion

Highlights: • Direct characterization of severe plastic deformation induced and healed porosities. • HPT process of disk with pre-fabricated holes reveals healing of porosity. • Ultrafine grains coexisting with shear bands and pores exhibit high stability. • Excess vacancies and pores cause drag effect on the grain boundary motion. • Accelerated vacancy diffusion leads to the generation and growth of pores. - Abstract: Generation and healing of pores in metals under loading or during processing have long been important research topics, and their mechanisms are still debatable. In recent years, unexpected percolating networks of pores acting as ultra-fast diffusion paths were found in ultrafine-grained metals processed by severe plastic deformation. Herein, we conducted a systematic study of the mechanisms of generation and healing of pores in ultrahigh-purity copper (99.9995 wt%) during high-pressure torsion (HPT). In addition to a solid Cu disk, the disk with holes that were drilled before processing, resulting in nominal initial porosity of 4.48%, was processed by HPT. After HPT, the percolating porosities of 0.07 ± 0.01% and 0.04 ± 0.01% were found in the solid disk, and in the disk with pre-fabricated holes, respectively. Thus, higher initial porosity resulted in lower percolating porosity in the HPT-processed disk. Systemic characterization using scanning electron microscopy revealed that in both cases shear bands and ultra-fine grains are located in the regions with induced and healed pores. We discussed the mechanisms of porosity evolution in both cases. In the end, to verify the possibility of pores nucleation and growth by vacancy agglomeration, vacancy diffusivity during HPT processing was estimated. This work sheds a new light on the mechanisms of pores nucleation and evolution during severe plastic deformation.

36 MATERIALS SCIENCE↗
DOE OSTI2013

A new indirect measure of diffusion model error

We define a new indirect measure of the diffusion model error called the diffusion model error source. When this model error source is added to the diffusion equation, the transport solution for the angular-integrated intensity is obtained. This source represents a means by which a transport code can be used to generate information relating to the adequacy of diffusion theory for any given problem without actually solving the diffusion equation. The generation of this source does not relate in any way to acceleration of the iterative convergence of transport solutions. Perhaps the most well-known indirect measure of the diffusion model error is the variable-Eddington tensor. This tensor provides a great deal of information about the angular dependence of the angular intensity solution, but it is not always simple to interpret. In contrast, our diffusion model error source is a scalar that is conceptually easy to understand. In addition to defining the diffusion model error source analytically, we show how to generate this source numerically relative to the S{sub n} radiative transfer equations with linear-discontinuous spatial discretization. This numerical source is computationally tested and shown to reproduce the Sn solution for a Marshak-wave problem. (authors)

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2016

Effects of size and concentration on diffusion-induced stress in lithium-ion batteries

Capacity fade of lithium-ion batteries induced by chemo-mechanical degradation during charge-discharge cycles is the bottleneck in design of high-performance batteries, especially high-capacity electrode materials. Stress generated due to diffusion-mechanical coupling in lithium-ion intercalation and deintercalation is accompanied by swelling, shrinking, and even micro-cracking. In this paper, we propose a theoretical model for a cylindrical nanowire electrode by combining the bond-order-length-strength and diffusion theories. It is shown that size and concentration have a significant influence on the stress fields in radial, hoop, and axial directions. This can explain why a smaller electrode with a huge volume change survives in the lithiation/delithiation process.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2024

ForceGen: End-to-end de novo protein generation based on nonlinear mechanical unfolding responses using a language diffusion model

Through evolution, nature has presented a set of remarkable protein materials, including elastins, silks, keratins and collagens with superior mechanical performances that play crucial roles in mechanobiology. However, going beyond natural designs to discover proteins that meet specified mechanical properties remains challenging. Here, we report a generative model that predicts protein designs to meet complex nonlinear mechanical property-design objectives. Our model leverages deep knowledge on protein sequences from a pretrained protein language model and maps mechanical unfolding responses to create proteins. Via full-atom molecular simulations for direct validation, we demonstrate that the designed proteins are de novo, and fulfill the targeted mechanical properties, including unfolding energy and mechanical strength, as well as the detailed unfolding force-separation curves. Our model offers rapid pathways to explore the enormous mechanobiological protein sequence space unconstrained by biological synthesis, using mechanical features as the target to enable the discovery of protein materials with superior mechanical properties.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Realistic galaxy image simulation via score-based generative models

ABSTRACT We show that a denoising diffusion probabilistic model (DDPM), a class of score-based generative model, can be used to produce realistic mock images that mimic observations of galaxies. Our method is tested with Dark Energy Spectroscopic Instrument (DESI) grz imaging of galaxies from the Photometry and Rotation curve OBservations from Extragalactic Surveys (PROBES) sample and galaxies selected from the Sloan Digital Sky Survey. Subjectively, the generated galaxies are highly realistic when compared with samples from the real data set. We quantify the similarity by borrowing from the deep generative learning literature, using the ‘Fréchet inception distance’ to test for subjective and morphological similarity. We also introduce the ‘synthetic galaxy distance’ metric to compare the emergent physical properties (such as total magnitude, colour, and half-light radius) of a ground truth parent and synthesized child data set. We argue that the DDPM approach produces sharper and more realistic images than other generative methods such as adversarial networks (with the downside of more costly inference), and could be used to produce large samples of synthetic observations tailored to a specific imaging survey. We demonstrate two potential uses of the DDPM: (1) accurate inpainting of occluded data, such as satellite trails, and (2) domain transfer, where new input images can be processed to mimic the properties of the DDPM training set. Here we ‘DESI-fy’ cartoon images as a proof of concept for domain transfer. Finally, we suggest potential applications for score-based approaches that could motivate further research on this topic within the astronomical community.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

Direct Visualization of Defect‐Controlled Diffusion in van der Waals Gaps

Abstract Diffusion processes govern fundamental phenomena such as phase transformations, doping, and intercalation in van der Waals (vdW) bonded materials. Here, the diffusion dynamics of W atoms by visualizing the motion of individual atoms at three different vdW interfaces: hexagonal boron nitride (BN)/vacuum, BN/BN, and BN/WSe 2 , by recording scanning transmission electron microscopy movies is quantified. Supported by density functional theory (DFT) calculations, it is inferred that in all cases diffusion is governed by intermittent trapping at electron beam‐generated defect sites. This leads to diffusion properties that depend strongly on the number of defects. These results suggest that diffusion and intercalation processes in vdW materials are highly tunable and sensitive to crystal quality. The demonstration of imaging, with high spatial and temporal resolution, of layers and individual atoms inside vdW heterostructures offers possibilities for direct visualization of diffusion and atomic interactions, as well as for experiments exploring atomic structures, their in situ modification, and electrical property measurements of active devices combined with atomic resolution imaging.

Chemistry↗
DOE OSTI2019

High Performance Gas Diffusion Layer (Phase I Final Report)

Hydrogen fuel cells are of interest for automotive applications because they produce zero emissions. However, fuel cells currently suffer from the high cost associated with the fuel cell stack. Power is limited by water management in the cathode. In this project, pH Matter, LLC synthesized and demonstrated the performance of next generation fuel cell Gas Diffusion Layers (GDLs). The materials out-performed current GDLs in terms of power density and cost per power. The materials were shown to meet the application requirements for key properties. At the commercial scale, these materials will be less expensive than GDLs currently used in hydrogen fuel cells, and enable higher power density, thus reducing fuel cell stack costs. Fuel cell system applications that will benefit most from the technology include automotive, materials handling (i.e. forklifts), fleet vehicles, and stationary back-up power.

25 ENERGY STORAGE↗