Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “DEEP SPACE”

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 325 records · Page 18
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

DuctGPT: A Generative Transformer for Forward Screening of Ductile Refractory Multi-Principal Element Alloys

Designing ductile materials for extreme environments such as fusion reactors requires a deep understanding of the complex interplay between electronic structure, mechanical stability, and wide compositional space. Here, in this work, we introduce DuctGPT, a physics-informed, GPT-powered machine learning platform that enables rapid and accurate prediction of ductility across a wide range of refractory multi-principal element alloys (MPEAs). Trained on both experimental and high-fidelity computational data, DuctGPT integrates descriptors such as density of states at the Fermi level, elastic constants, and valence electron concentration to capture the fundamental mechanisms governing ductile versus brittle behavior. Using this framework, we screen over 1000 compositions in of body-centered cubic (BCC) MPEAs, including two new alloy classes, i.e., NbTa-rich (NbTa $>$ 50 at.%) NbTa-Ti-V and W-rich ($>$ 50 at.%) W-Ti-V MPEAs, to rapidly identify promising alloy compositions with enhanced ductility. Validation against experimental data confirms the model's ability to predict ductility with high fidelity and low uncertainty. By leveraging conversational AI and robust physical modeling, DuctGPT provides a blueprint for the next generation of alloy design assistants, enabling human-AI collaboration in the accelerated discovery of ductile, high-performance materials for fusion, aerospace, and advanced manufacturing.

AI/ML↗
DOE OSTI2020

hp -VPINNs: Variational physics-informed neural networks with domain decomposition

We formulate a general framework for hp-variational physics-informed neural networks (hp-VPINNs) based on the nonlinear approximation of shallow and deep neural networks and hp-refinement via domain decomposition and projection onto the space of high-order polynomials. The trial space is the space of neural network, which is defined globally over the entire computational domain, while the test space contains piecewise polynomials. Specifically in this study, the hp-refinement corresponds to a global approximation with a local learning algorithm that can efficiently localize the network parameter optimization. Here, we demonstrate the advantages of hp-VPINNs in both accuracy and training cost for several numerical examples of function approximation and in solving differential equations.

42 ENGINEERING↗
DOE OSTI2022

No correlation of the Lyman continuum escape fraction with spectral hardness

The properties that govern the production and escape of hydrogen-ionizing photons (Lyman continuum, LyC; with energies > 13.6 eV) in star-forming galaxies are still poorly understood, but they are key to identifying and characterizing the sources that reionized the Universe. Here we empirically explore the relation between the hardness of ionizing radiation and the LyC leakage in a large sample of low- z star-forming galaxies from the recent Hubble Space Telescope Low- z Lyman Continuum Survey. Using Sloan Digital Sky Survey stacks and deep X-shooter observations, we investigate the hardness of the ionizing spectra ( Q He + / Q H ) between 54.4 eV (He + ) and 13.6 eV (H) from the optical recombination lines He II 4686 Å and Hβ 4861 Å for galaxies with LyC escape fractions spanning a wide range, f esc (LyC) ≃ 0–90%. We find that the observed intensity of He II/Hβ is primarily driven by variations in the metallicity, but is not correlated with LyC leakage. Both very strong (< f esc (LyC)> ≃ 0.5) and nonleakers (< f esc (LyC)> ≃ 0) present similar observed intensities of He II and Hβ at comparable metallicity, between ≃0.01 and ≃0.02 for 12 + log(O/H)> 8.0 and < 8.0, respectively. Our results demonstrate that Q He + / Q H does not correlate with f esc (LyC), which implies that strong LyC emitters do not show harder ionizing spectra than nonleakers at similar metallicity.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

Anticipating Optical Availability in Hybrid RF/FSO Links Using RF Beacons and Deep Learning

Radiofrequency (RF) communications offer reliable but low data rates and energy-inefficient satellite links, while free-space optical (FSO) promises high bandwidth but struggles with disturbances imposed by atmospheric effects. A hybrid RF/FSO architecture aims to achieve optimal reliability along with high data rates for space communications. Accurate prediction of dynamic ground-to-satellite FSO link availability is critical for routing decisions in low-earth orbit constellations. In this paper, we propose a system leveraging ubiquitous RF links to proactively forecast FSO link degradation prior to signal drops below threshold levels. This enables pre-calculation of rerouting to maximally maintain high data rate FSO links throughout the duration of weather effects. We implement a supervised learning model to anticipate FSO attenuation based on the analysis of RF patterns. Through the simulation of a dense lower earth orbit (LEO) satellite constellation, we demonstrate the efficacy of our approach in a simulated satellite network, highlighting the balance between predictive accuracy and prediction duration. An emulated cloud attenuation model is proposed to provide insight into the temporal profiles of RF signals and their correlation to FSO channel dynamics. Our investigation sheds light on the trade-offs between prediction horizon and accuracy arising from RF beacon numbers and proximity.

FSO availability↗
DOE OSTI2023

DDStore: Distributed Data Store for Scalable Training of Graph Neural Networks on Large Atomistic Modeling Datasets

Graph neural networks (GNNs) are a class of Deep Learning models used in designing atomistic materials for effective screening of large chemical spaces. To ensure robust prediction, GNN models must be trained on large volumes of atomistic data on leadership class supercomputers. Even with the advent of modern architectures that consist of multiple storage layers that include node-local NVMe devices in addition to device memory for caching large datasets, extreme-scale model training faces I/O challenges at scale.We present DDStore, an in-memory distributed data store designed for GNN training on large-scale graph data. DDStore provides a hierarchical, distributed, data caching technique that combines data chunking, replication, low-latency random access, and high throughput communication. DDStore achieves near-linear scaling for training a GNN model using up to 1000 GPUs on the Summit and Perlmutter supercomputers, and reaches up to a 6.15x reduction in GNN training time compared to state-of-the-art methodologies.

Choi, Jong Youl↗
DOE OSTI2026

Data-Driven Modeling and Correction of Vehicle Dynamics

We develop a data-driven framework for learning and correcting nonautonomous vehicle dynamics. Physics-based vehicle models are often simplified for tractability and therefore exhibit inherent model-form uncertainty, motivating the need for data-driven correction. Moreover, nonautonomous dynamics are governed by time-dependent control inputs, which pose challenges in learning predictive models directly from temporal snapshot data. To address these, we reformulate the vehicle dynamics via a local parameterization of the time-dependent inputs, yielding a modified system composed ofa sequence of local parametric dynamical systems. Here, we approximate these parametric systems using two complementary approaches. First, we employ the dimension reduction and interpolation in parameter space (DRIPS) methodology to construct efficient linear surrogate models, equipped with lifted observable spaces and manifold-based operator interpolation. This enables data-efficient learning of vehicle models whose dynamics admit accurate linear representations in the lifted spaces. Second, for more strongly nonlinear systems, we employ flow map learning (FML), a deep neural network (DNN) approach that approximates the parametric evolution map without requiring special treatment of nonlinearities. We further extend FML with a transfer-learning-based model correction procedure, enabling the correction of misspecified prior models using only a sparse set of high-fidelity or experimental measurements, without assuming a prescribed form for the correction term. Through a suite of numerical experiments on unicycle, simplified bicycle, and slip-based bicycle models, we demonstrate that DRIPS offers robust and highly data-efficient learning of nonautonomous vehicle dynamics, while FML provides expressive nonlinear modeling and effective correction of model-form errors under severe data scarcity.

data-driven modeling↗
DOE OSTI2021

Disentangling Rotational Dynamics and Ordering Transitions in a System of Self-Organizing Protein Nanorods via Rotationally Invariant Latent Representations

The dynamics of complex ordering systems with active rotational degrees of freedom exemplified by protein self-assembly is explored using a machine learning workflow that combines deep learning-based semantic segmentation and rotationally invariant variational autoencoder-based analysis of orientation and shape evolution. The latter allows for disentanglement of the particle orientation from other degrees of freedom and compensates for lateral shifts. The disentangled representations in the latent space encode the rich spectrum of local transitions that can now be visualized and explored via continuous variables. The time dependence of ensemble averages allows insight into the time dynamics of the system and, in particular, illustrates the presence of the potential ordering transition. Finally, analysis of the latent variables along the single-particle trajectory allows tracing these parameters on a single-particle level. The proposed approach is expected to be universally applicable for the description of the imaging data in optical, scanning probe, and electron microscopy seeking to understand the dynamics of complex systems where rotations are a significant part of the process.

representation learning↗
DOE OSTI2023

Gumby: Quantifying multi-modal model resiliency

With the rise of cheap data and sensors, more use cases are emerging for multi-input models. Research has shown that including multiple data modalities can improve performance, suggesting that deep learning models can successfully learn to leverage complementary information from different modalities. However, this improved predictive power comes with unanticipated costs: additional inputs change model resiliency and expand the threat space for adversarial attacks. We first provide theoretical underpinnings for how adversarial success scales with input dimension. We then characterize the performance of a suite of multispectral deep learning models with different fusion approaches, quantify their relative reliance on different input bands, and evaluate their robustness to naturalistic and adversarial image corruptions.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

Deep Learning and Structural Imaging of Materials

Deep learning has had a transformative effect on numerous domains and is actively utilized by many scientists in data-intensive fields such as high-energy physics and cosmology. Materials science, and in particular, the structural imaging of materials with electrons and X-rays are projected to enter the age of scientific data torrents, positioning them as new application spaces for modern artificial intelligence. In this contribution, we provide a synopsis on the foundations and latest progress in deep learning and present an in-depth application of utilizing modern deep artificial neural networks in scanning transmission electron microscopy to extract structural material properties. We use this case study to expose the strengths of deep learning-based models and discuss their current limitations, in the process highlighting their potential use in other data-intensive structural imaging modalities.

Laanait, Nouamane↗
DOE OSTI2022

Deep learning the properties of inorganic perovskites

We report the ability to accurately and quickly predict the stability of materials and their structural and electronic properties remains a grand challenge in materials science. Density functional theory is widely used as a means of predicting these material properties, but is known to be computationally expensive and scales as the cube of the number of electrons in the material’s unit cell. In this article, for a previously published dataset of inorganic perovskites, we show that a single neural network model using only the elemental properties of the compounds’ constituents can predict lattice constants to within 0.1 Å, heat of formation to within 0.2 eV, and band gaps to within 0.7 eV RMSE. We also compare the performance of the trained network to two widely used regression techniques, namely random forest and Kernel ridge regression, and find that the neural network’s predictions are more accurate for each of the properties. The simultaneous accurate prediction of multiple key properties of technologically relevant materials is promising for rational design and optimization in known and novel chemical spaces.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Cataloging accreted stars within Gaia DR2 using deep learning

The goal of this study is to present the development of a machine learning based approach that utilizes phase space alone to separate the Gaia DR2 stars into two categories: those accreted onto the Milky Way from those that are in situ. Traditional selection methods that have been used to identify accreted stars typically rely on full 3D velocity, metallicity information, or both, which significantly reduces the number of classifiable stars. The approach advocated here is applicable to a much larger portion of Gaia DR2. Methods. A method known as “transfer learning” is shown to be effective through extensive testing on a set of mock Gaia catalogs that are based on the F IRE cosmological zoom-in hydrodynamic simulations of Milky Way-mass galaxies. The machine is first trained on simulated data using only 5D kinematics as inputs and is then further trained on a cross-matched Gaia /RAVE data set, which improves sensitivity to properties of the real Milky Way. Results. The result is a catalog that identifies ~767 000 accreted stars within Gaia DR2. This catalog can yield empirical insights into the merger history of the Milky Way and could be used to infer properties of the dark matter distribution.

Astronomy & Astrophysics↗
DOE OSTI2022

Curving the space by non-Hermiticity

Quantum systems are often classified into Hermitian and non-Hermitian ones. Extraordinary non-Hermitian phenomena, ranging from the non-Hermitian skin effect to the supersensitivity to boundary conditions, have been widely explored. Whereas these intriguing phenomena have been considered peculiar to non-Hermitian systems, we show that they can be naturally explained by a duality between non-Hermitian models in flat spaces and their counterparts, which could be Hermitian, in curved spaces. For instance, prototypical one-dimensional (1D) chains with uniform chiral tunnelings are equivalent to their duals in two-dimensional (2D) hyperbolic spaces with or without magnetic fields, and non-uniform tunnelings could further tailor local curvatures. Such a duality unfolds deep geometric roots of non-Hermitian phenomena, delivers an unprecedented routine connecting Hermitian and non-Hermitian physics, and gives rise to a theoretical perspective reformulating our understandings of curvatures and distance. In practice, it provides experimentalists with a powerful two-fold application, using non-Hermiticity to engineer curvatures or implementing synthetic curved spaces to explore non-Hermitian quantum physics.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2022

Accelerating the discovery of battery electrode materials through data mining and deep learning models

The availability of crystalline materials databases allows for building accurate machine learning (ML) models that can accelerate the exploration of materials chemical space for energy storage applications. In this work, we screen all inorganic materials included in the Materials Project and AFLOW databases as potential metal-ion battery electrodes. We develop an efficient protocol to mine and screen raw data in current databases and provide a new database of electrode materials by considering pairs of charged and discharged electrodes. This effort leads to a new database with over 190,000 instances, in contrast to the original battery database which contains about 5000. The expanded battery data set is then used to build regression-based deep neural network models for predicting average voltages and percentage volume changes upon charging and discharging, which present improvements of at least 28% for target properties with respect to previous models, and are now able to predict anode electrodes (low voltage region) as well as electrodes that will not work in electrochemical cells (negative voltages), overcoming the challenges identified in previous ML models for battery electrodes. Additionally, a further screening of the expanded database itself allowed us to identify 35 novel electrode candidates with excellent battery performance metrics.

25 ENERGY STORAGE↗
DOE OSTI2021

Learning nonlinear operators via DeepONet based on the universal approximation theorem of operators

It is widely known that neural networks (NNs) are universal approximators of continuous functions. However, a less known but powerful result is that a NN with a single hidden layer can accurately approximate any nonlinear continuous operator. This universal approximation theorem of operators is suggestive of the structure and potential of deep neural networks (DNNs) in learning continuous operators or complex systems from streams of scattered data. Here, in this work, we thus extend this theorem to DNNs. We design a new network with small generalization error, the deep operator network (DeepONet), which consists of a DNN for encoding the discrete input function space (branch net) and another DNN for encoding the domain of the output functions (trunk net). We demonstrate that DeepONet can learn various explicit operators, such as integrals and fractional Laplacians, as well as implicit operators that represent deterministic and stochastic differential equations. We study different formulations of the input function space and its effect on the generalization error for 16 different diverse applications.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Diffusion power spectra as a window into dynamic materials architecture

Chemical recycling of commodity and specialty polymers presents a multifaceted challenge for industrial societies. On one hand, macromolecular architectures must be engineered to yield durable products that, on the other hand, rapidly deconstruct to recyclable monomers under pre-determined conditions. Polymer deconstruction is a chemical process that requires deep understanding of molecular reactivity in heterogeneous media, where porous material architectures evolve in both space and time. To build this understanding, we develop herein experimental and analytical methods describing sets of diffusive eigenmodes that exist within time-varying, non-Euclidean boundary conditions, a situation commonly encountered in the reactive deconstruction of polymers where chain fragments splay, alter their local dynamics, and evolve in their confinement of reacting media. Diffusion power spectra, discerned experimentally by NMR, yield polymer and solvent frequency-domain velocity autocorrelation functions that are analyzed in the context of physical models for chemical reactions parameterized with fractal mathematics. The results connect local motion in polymers to chemical reactivity during acidolysis of circular elastomers.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2019

Anatomy of a Cooling Flow: The Feedback Response to Pure Cooling in the Core of the Phoenix Cluster

We present new, deep observations of the Phoenix cluster from the Chandra X-ray Observatory, the Hubble Space Telescope, and the Karl Jansky Very Large Array. These data provide an order of magnitude improvement in depth and/or angular resolution at X-ray, optical, and radio wavelengths, yielding an unprecedented view of the core of the Phoenix cluster. We find that the one-dimensional temperature and entropy profiles are consistent with expectations for pure-cooling hydrodynamic simulations and analytic descriptions of homogeneous, steady-state cooling flow models. In the inner ~10 kpc, the cooling time is shorter by an order of magnitude than any other known cluster, while the ratio of the cooling time to freefall time approaches unity, signaling that the ICM is unable to resist multiphase condensation on kpc scales. When we consider the thermodynamic profiles in two dimensions, we find that the cooling is highly asymmetric. The bulk of the cooling in the inner ~20 kpc is confined to a low-entropy filament extending northward from the central galaxy. We detect a substantial reservoir of cool (10^4 K) gas (as traced by the [OII] doublet), which is coincident with the low-entropy filament. The bulk of this cool gas is draped around and behind a pair of X-ray cavities, presumably bubbles that have been inflated by radio jets, which are detected for the first time on kpc scales. These data support a picture in which AGN feedback is promoting the formation of a multiphase medium via a combination of ordered buoyant uplift and locally enhanced turbulence. These processes ought to counteract the tendency for buoyancy to suppress condensation, leading to rapid cooling along the jet axis. The recent mechanical outburst has sufficient energy to offset cooling, and appears to be coupling to the ICM via a cocoon shock, raising the entropy in the direction orthogonal to the radio jets.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

The High Latitude Spectroscopic Survey on the Nancy Grace Roman Space Telescope

The Nancy Grace Roman Space Telescope will conduct a High Latitude Spectroscopic Survey (HLSS) over a large volume at high redshift, using the near-IR grism (1.0–1.93 μm, R = 435–865) and the 0.28 deg 2 wide-field camera. We present a reference HLSS that maps 2000 deg 2 and achieves an emission-line flux limit of 10 -16 erg s -1 cm -2 at 6.5σ, requiring ~0.6 yr of observing time. We summarize the flowdown of the Roman science objectives to the science and technical requirements of the HLSS. We construct a mock redshift survey over the full HLSS volume by applying a semianalytic galaxy formation model to a cosmological N-body simulation and use this mock survey to create pixel-level simulations of 4 deg 2 of HLSS grism spectroscopy. We find that the reference HLSS would measure ~10 million H α galaxy redshifts that densely map large-scale structure at z = 1–2 and 2 million [O III] galaxy redshifts that sparsely map structures at z = 2–3. We forecast the performance of this survey for measurements of the cosmic expansion history with baryon acoustic oscillations and the growth of large-scale structure with redshift-space distortions. We also study possible deviations from the reference design and find that a deep HLSS at f line > 7 × 10 -17 erg s -1 cm -2 over 4000 deg 2 (requiring ~1.5 yr of observing time) provides the most compelling stand-alone constraints on dark energy from Roman alone. This provides a useful reference for future optimizations. The reference survey, simulated data sets, and forecasts presented here will inform community decisions on the final scope and design of the Roman HLSS.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2020

Deep learning for characterizing the self-assembly of three-dimensional colloidal systems

Creating a systematic framework to characterize the structural states of colloidal self-assembly systems is crucial for unraveling the fundamental understanding of these systems’ stochastic and non-linear behavior. The most accurate characterization methods create high-dimensional neighborhood graphs that may not provide useful information about structures unless these are well-defined reference crystalline structures. Dimensionality reduction methods are thus required to translate the neighborhood graphs into a low-dimensional space that can be easily interpreted and used to characterize non-reference structures. We investigate a framework for colloidal system state characterization that employs deep learning methods to reduce the dimensionality of neighborhood graphs. The framework next uses agglomerative hierarchical clustering techniques to partition the low-dimensional space and assign physically meaningful classifications to the resulting partitions. In this work, we first demonstrate the proposed colloidal self-assembly state characterization framework on a three-dimensional in silico system of 500 multi-flavored colloids that self-assemble under isothermal conditions. We next investigate the generalizability of the characterization framework by applying the framework to several independent self-assembly trajectories, including a three-dimensional in silico system of 2052 colloidal particles that undergo evaporation-induced self-assembly.

36 MATERIALS SCIENCE↗