Search NASA⌕ Search

SEARCH · Search NASA

Results for “embedding model”

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 253 records · Page 14
All sourcesNASA NTRSDOE OSTI
DOE OSTI2015

Implementation of density functional embedding theory within the projector-augmented-wave method and applications to semiconductor defect states

We report a new implementation of the density functional embedding theory (DFET) in the VASP code, using the projector-augmented-wave (PAW) formalism. Newly developed algorithms allow us to efficiently perform optimized effective potential optimizations within PAW. The new algorithm generates robust and physically correct embedding potentials, as we verified using several test systems including a covalently bound molecule, a metal surface, and bulk semiconductors. We show that with the resulting embedding potential, embedded cluster models can reproduce the electronic structure of point defects in bulk semiconductors, thereby demonstrating the validity of DFET in semiconductors for the first time. Compared to our previous version, the new implementation of DFET within VASP affords use of all features of VASP (e.g., a systematic PAW library, a wide selection of functionals, a more flexible choice of U correction formalisms, and faster computational speed) with DFET. Furthermore, our results are fairly robust with respect to both plane-wave and Gaussian type orbital basis sets in the embedded cluster calculations. This suggests that the density functional embedding method is potentially an accurate and efficient way to study properties of isolated defects in semiconductors.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

Frame invariant neural network closures for Kraichnan turbulence

Numerical simulations of geophysical and atmospheric flows have to rely on parameterizations of subgrid scale processes due to their limited spatial resolution. Despite substantial progress in developing parameterization (or closure) models for subgrid scale (SGS) processes using physical insights and mathematical approximations, they remain imperfect and can lead to inaccurate predictions. In recent years, machine learning has been successful in extracting complex patterns from high-resolution spatio-temporal data, leading to improved parameterization models, and ultimately better coarse grid prediction. However, the inability to satisfy known physics and poor generalization hinders the application of these models for real-world problems. In this work, we put forth a frame invariant closure approach to improve the accuracy and generalizability of deep learning-based subgrid scale closure models by embedding physical symmetries directly into the structure of the neural network. Specifically, we utilized specialized layers within the convolutional neural network in such a way that desired constraints are theoretically guaranteed without the need for any regularization terms. We demonstrate our framework for a two-dimensional decaying turbulence test case mostly characterized by the forward enstrophy cascade. We show that our frame invariant SGS model (i) accurately predicts the subgrid scale source term, (ii) respects the physical symmetries such as translation, Galilean, and rotation invariance, and (iii) is numerically stable when implemented in coarse-grid simulation with generalization to different initial conditions and Reynolds number. This work opens up a possibility of connecting physics-based theories and data-driven modeling paradigms, and thus represents a promising step towards the development of physically consistent data-driven turbulence closure models.

42 ENGINEERING↗
DOE OSTI2023

Physics-embedded inverse analysis with algorithmic differentiation for the earth’s subsurface

Abstract Inverse analysis has been utilized to understand unknown underground geological properties by matching the observational data with simulators. To overcome the underconstrained nature of inverse problems and achieve good performance, an approach is presented with embedded physics and a technique known as algorithmic differentiation. We use a physics-embedded generative model, which takes statistically simple parameters as input and outputs subsurface properties (e.g., permeability or P-wave velocity), that embeds physical knowledge of the subsurface properties into inverse analysis and improves its performance. We tested the application of this approach on four geologic problems: two heterogeneous hydraulic conductivity fields, a hydraulic fracture network, and a seismic inversion for P-wave velocity. This physics-embedded inverse analysis approach consistently characterizes these geological problems accurately. Furthermore, the excellent performance in matching the observational data demonstrates the reliability of the proposed method. Moreover, the application of algorithmic differentiation makes this an easy and fast approach to inverse analysis when dealing with complicated geological structures.

54 ENVIRONMENTAL SCIENCES↗
NASA NTRS2016

Performance of the Goddard Multiscale Modeling Framework with Goddard Ice Microphysical Schemes

The multiscale modeling framework (MMF), which replaces traditional cloud parameterizations with cloud-resolving models (CRMs) within a host atmospheric general circulation model (GCM), has become a new approach for climate modeling. The embedded CRMs make it possible to apply CRM-based cloud microphysics directly within a GCM. However, most such schemes have never been tested in a global environment for long-term climate simulation. The benefits of using an MMF to evaluate rigorously and improve microphysics schemes are here demonstrated. Four one-moment microphysical schemes are implemented into the Goddard MMF and their results validated against three CloudSat/CALIPSO cloud ice products and other satellite data. The new four-class (cloud ice, snow, graupel, and frozen drops/hail) ice scheme produces a better overall spatial distribution of cloud ice amount, total cloud fractions, net radiation, and total cloud radiative forcing than earlier three-class ice schemes, with biases within the observational uncertainties. Sensitivity experiments are conducted to examine the impact of recently upgraded microphysical processes on global hydrometeor distributions. Five processes dominate the global distributions of cloud ice and snow amount in long-term simulations: (1) allowing for ice supersaturation in the saturation adjustment, (2) three additional correction terms in the depositional growth of cloud ice to snow, (3) accounting for cloud ice fall speeds, (4) limiting cloud ice particle size, and (5) new size-mapping schemes for snow and graupel. Despite the cloud microphysics improvements, systematic errors associated with subgrid processes, cyclic lateral boundaries in the embedded CRMs, and momentum transport remain and will require future improvement.

snow↗
DOE OSTI2025

Modeling low cycle fatigue (LCF) of additively manufactured Hastelloy X using An accelerated crystal plasticity fatigue damage model

This paper presents a microstructure-based model for low cycle fatigue (LCF) behavior and life of Nickel-based alloy Hastelloy X manufactured using laser-powder bed fusion (L-PBF) additive manufacturing (AM). AM Hastelloy X, a solution-strengthened alloy, is tested at elevated temperature under fully reversed LCF conditions at different strain levels. A generalized plane strain finite element model is generated from electron backscatter diffraction (EBSD) characterization. The constitutive behavior of the material under fatigue is modeled using crystal plasticity and calibrated with both monotonic tensile and cyclic stress–strain data. The fatigue micro-crack initiation and propagation in the microstructure is modeled using a modified Chaboche fatigue damage model. An embedded boundary condition with a homogenous medium is used to apply the cyclic deformation and prevent numerically introduced over-constraints during fatigue simulation. A ‘cycle-jump’ method is used to accelerate the fatigue simulation and reduce the computational cost. The simulation results are compared to LCF experiments, showing satisfactory matches in cyclic stress behavior and number of cycles to macro-crack initiation for all applied strain ranges. In addition, the model illustrates the potential for quantifying microscale fatigue life impacting factors such as microstructure and surface roughness, which is needed to accurately quantify the reliability of AM components in service.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Modeling and simulation of vascular tumors embedded in evolving capillary networks

Here in this work, we present a coupled 3D–1D model of solid tumor growth within a dynamically changing vascular network to facilitate realistic simulations of angiogenesis. Additionally, the model includes erosion of the extracellular matrix, interstitial flow, and coupled flow in blood vessels and tissue. We employ continuum mixture theory with stochastic Cahn–Hilliard type phase-field models of tumor growth. The interstitial flow is governed by a mesoscale version of Darcy’s law. The flow in the blood vessels is controlled by Poiseuille flow, and Starling’s law is applied to model the mass transfer in and out of blood vessels. The evolution of the network of blood vessels is orchestrated by the concentration of the tumor angiogenesis factors (TAFs); blood vessels grow towards the increasing TAFs concentrations. This process is not deterministic, allowing random growth of blood vessels and, therefore, due to the coupling of nutrients in tissue and vessels, makes the growth of tumors stochastic. We demonstrate the performance of the model by applying it to a variety of scenarios. Numerical experiments illustrate the flexibility of the model and its ability to generate satellite tumors. Simulations of the effects of angiogenesis on tumor growth are presented as well as sample-independent features of cancer.

3D–1D coupled blood flow models↗
DOE OSTI2016

Reynolds averaged turbulence modelling using deep neural networks with embedded invariance

There exists significant demand for improved Reynolds-averaged Navier–Stokes (RANS) turbulence models that are informed by and can represent a richer set of turbulence physics. This paper presents a method of using deep neural networks to learn a model for the Reynolds stress anisotropy tensor from high-fidelity simulation data. A novel neural network architecture is proposed which uses a multiplicative layer with an invariant tensor basis to embed Galilean invariance into the predicted anisotropy tensor. It is demonstrated that this neural network architecture provides improved prediction accuracy compared with a generic neural network architecture that does not embed this invariance property. Furthermore, the Reynolds stress anisotropy predictions of this invariant neural network are propagated through to the velocity field for two test cases. For both test cases, significant improvement versus baseline RANS linear eddy viscosity and nonlinear eddy viscosity models is demonstrated.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
NASA NTRS1983

A conservative type-dependent full potential method for the treatment of supersonic flows with embedded subsonic regions

A nonlinear method based on the full potential equation in conservation form, cast in an arbitrary coordinate system, has been developed to treat predominantly supersonic flows with embedded subsonic regions. This type of flow field occurs frequently near the fuselage-canopy junction area and wing leading edge regions for a moderately swept fighter configuration. The method uses the theory of characteristics to accurately monitor the type-dependent flow field. A conservative switching scheme is developed to transition from the supersonic marching algorithm to a subsonic relaxation procedure, and vice versa. An implicit approximate factorization scheme is employed to solve the finite-differenced equation. Results are shown for a few configurations, including a wing-body-wake realistic fighter model having embedded subsonic regions.

Shankar, V.↗
DOE OSTI2023

Scaling Containment on a Large Centrifuge

Containment science concerns the trapping or leakage of radioactive cavity gases from underground nuclear explosions (UNEs). The most physically relevant data for validation of containment science comes from legacy UNE testing or large field-scale chemical explosions. Field-scale tests can inform on the degree of residual stress imparted to host geology and possible formation of an accompanying gas-containing “stress cage”, for example, as a function of chemical explosion yield. However, field testing can be expensive, difficult to conduct, and challenging to instrument with limitations on data coverage (i.e., the number of boreholes; difficult coring through damage zones). Small-scale laboratory tests are typically simpler to conduct and more thoroughly characterize, but may not include the right scale of containment processes for stress-cage or chimney formation. This report presents the theory and fundamentals of large centrifuge physical modeling, including a history of explosive or similar testing relevant to containment science investigations on centrifuges. Subscale models with embedded chemical explosives in the enhanced gravity of a large centrifuge can represent hundreds of meters of depth and large explosions not otherwise attainable in the laboratory as based on scaling of length, energy, and other processes by the g-factor of the centrifuge (i.e., the number of times larger the centrifugal force is than the gravitational force on the Earth’s surface). The centrifuge uniquely joins the simplicity and exhaustiveness of laboratory-scale characterization with the physics of field-scale processes. This report is part of Sandia National Laboratories’ (SNL’s) effort to prepare for physical modeling with approximately one-meter tall geologic models of stress cage and/or chimney formation on its 29-foot [8.84 m] radius centrifuge with a load capacity of 1.6 million 𝑔-pounds [726 𝑔-ton]. The scaling relationships herein will inform model design decisions and performance requirements for in-flight sensors.

42 ENGINEERING↗
DOE OSTI2023

An finite element analysis surrogate model with boundary oriented graph embedding approach for rapid design

Abstract In this work, we present a boundary oriented graph embedding (BOGE) approach for the graph neural network to assist in rapid design and digital prototyping. The cantilever beam problem has been solved as an example to validate its potential of providing physical field results and optimized designs using only 10 ms. Providing shortcuts for both boundary elements and local neighbor elements, the BOGE approach can embed unstructured mesh elements into the graph and performs an efficient regression on large-scale triangular-mesh-based finite element analysis (FEA) results, which cannot be realized by other machine-learning-based surrogate methods. It has the potential to serve as a surrogate model for other boundary value problems. Focusing on the cantilever beam problem, the BOGE approach with 3-layer DeepGCN model achieves the regression with mean square error (MSE) of 0.011 706 (2.41% mean absolute percentage error) for stress field prediction and 0.002 735 MSE (with 1.58% elements having error larger than 0.01) for topological optimization. The overall concept of the BOGE approach paves the way for a general and efficient deep-learning-based FEA simulator that will benefit both industry and Computer Aided Design (CAD) design-related areas.

42 ENGINEERING↗
DOE OSTI2018

Ensemble Grouping Strategies for Embedded Stochastic Collocation Methods Applied to Anisotropic Diffusion Problems

Previous work has demonstrated that propagating groups of samples, called ensembles, together through forward simulations can dramatically reduce the aggregate cost of sampling-based uncertainty propagation methods [E. Phipps, M. D'Elia, H. C. Edwards, M. Hoemmen, J. Hu, and S. Rajamanickam, SIAM J. Sci. Comput., 39 (2017), pp. C162--C193]. However, critical to the success of this approach when applied to challenging problems of scientific interest is the grouping of samples into ensembles to minimize the total computational work. For example, the total number of linear solver iterations for ensemble systems may be strongly influenced by which samples form the ensemble when applying iterative linear solvers to parameterized and stochastic linear systems. In this paper we explore sample grouping strategies for local adaptive stochastic collocation methods applied to PDEs with uncertain input data, in particular canonical anisotropic diffusion problems where the diffusion coefficient is modeled by truncated Karhunen--Loève expansions. Finally, we demonstrate that a measure of the total anisotropy of the diffusion coefficient is a good surrogate for the number of linear solver iterations for each sample and therefore provides a simple and effective metric for grouping samples.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2014

Accretion outbursts in self-gravitating protoplanetary disks

We improve on our previous treatments of the long-term evolution of protostellar disks by explicitly solving disk self-gravity in two dimensions. The current model is an extension of the one-dimensional layered accretion disk model of Bae et al. We find that gravitational instability (GI)-induced spiral density waves heat disks via compressional heating (i.e., PdV work), and can trigger accretion outbursts by activating the magnetorotational instability (MRI) in the magnetically inert disk dead zone. The GI-induced spiral waves propagate well inside of the gravitationally unstable region before they trigger outbursts at R ≲ 1 AU where GI cannot be sustained. This long-range propagation of waves cannot be reproduced with the previously used local α treatments for GI. In our standard model where zero dead-zone residual viscosity (α{sub rd}) is assumed, the GI-induced stress measured at the onset of outbursts is locally as large as 0.01 in terms of the generic α parameter. However, as suggested in our previous one-dimensional calculations, we confirm that the presence of a small but finite α{sub rd} triggers thermally driven bursts of accretion instead of the GI + MRI-driven outbursts that are observed when α{sub rd} = 0. The inclusion of non-zero residual viscosity in the dead zone decreases the importance of GI soon after mass feeding from the envelope cloud ceases. During the infall phase while the central protostar is still embedded, our models stay in a 'quiescent' accretion phase with M-dot {sub acc}∼10{sup −8}--10{sup −7} M{sub ⊙} yr{sup −1} over 60% of the time and spend less than 15% of the infall phase in accretion outbursts. While our models indicate that episodic mass accretion during protostellar evolution can qualitatively help explain the low accretion luminosities seen in most low-mass protostars, detailed tests of the mechanism will require model calculations for a range of protostellar masses with some constraint on the initial core angular momentum, which affects the length of time spent in a quasi-steady disk accretion phase.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2018

The Vertical Momentum Budget of Shallow Cumulus Convection: Insights From a Lagrangian Perspective

A Lagrangian Particle Dispersion Model is embedded into large eddy simulations to diagnose the responses of shallow cumulus convection to a small-amplitude large-scale temperature perturbation. The Lagrangian framework allows for a decomposition of the vertical momentum budget and diagnosis of the forces that regulate cloudy updrafts. The results are used to shed light on the parameterization of vertical velocity in convective schemes, where the treatment of the effects of entrainment as well as buoyancy-induced and mechanically induced pressure gradients remains highly uncertain. We show that both buoyancy-induced and mechanically induced pressure gradients are important for the vertical momentum budget of cloudy updrafts, whereas the entrainment dilution term is relatively less important. Based on the analysis of the dominant force balance, we propose a simple model to derive the perturbation pressure gradient forces. We further illustrate that the effective buoyancy and dynamic perturbation pressure can be approximated to a good extent using a simple cylindrical updraft model given the cloud radius. This finding has the potential for improving the parameterization of vertical velocity in convective schemes and the development of a unified scheme for cumulus convection.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2019

Twisted fracton models in three dimensions

We study novel three-dimensional gapped quantum phases of matter which support quasiparticles with restricted mobility, including immobile “fracton” excitations. So far, most existing fracton models may be instructively viewed as generalized Abelian lattice gauge theories. Here, by analogy with Dijkgraaf-Witten topological gauge theories, we discover a natural generalization of fracton models, obtained by twisting the gauge symmetries. Introducing generalized gauge transformation operators carrying an extra phase factor depending on local configurations, we construct a plethora of exactly solvable three-dimensional models, which we dub “twisted fracton models.” A key result of our approach is to demonstrate the existence of rich non-Abelian fracton phases of distinct varieties in a three-dimensional system with finite-range interactions. For an accurate characterization of these novel phases, the notion of being inextricably non-Abelian is introduced for fractons and quasiparticles with one-dimensional mobility, referring to their new behavior of displaying braiding statistics that is, and remains, non-Abelian regardless of which quasiparticles with higher mobility are added to or removed from them. We also analyze these models by embedding them on a 3-torus and computing their ground-state degeneracies, which exhibit a surprising and novel dependence on the system size in the non-Abelian fracton phases. Furthermore, as an important advance in the study of fracton order, we develop a general mathematical framework which systematically captures the fusion and braiding properties of fractons and other quasiparticles with restricted mobility.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2018

Do Better Quality Embedding Potentials Accelerate the Convergence of QM/MM Models? The Case of Solvated Acid Clusters

This study examines whether the use of more accurate embedding potentials improves the convergence of quantum mechanics/molecular mechanics (QM/MM) models with respect to the size of the QM region. In conjunction with density functional theory calculations using the ωB97X-D functional, various embedding potentials including the TIP3P water model, the effective fragment potential (EFP), and semi-empirical methods (PM6, PM7, and DFTB) were used to simulate the deprotonation energies of solvated acid clusters. The calculations were performed on solvated neutral (HA) and cationic (HB+) acids clusters containing 160 and 480 water molecules using configurations sampled from molecular dynamics simulations. Consistently, the ωB97X-D/EFP model performed the best when using a minimal QM region size. The performance for the other potentials appears to be highly sensitive to the charge character of the acid/base pair. Neutral acids display the expected trend that semi-empirical methods generally perform better than TIP3P; however, an opposite trend was observed for the cationic acids. Additionally, electronic embedding provided an improvement over mechanical embedding for the cationic systems, but not the neutral acids. For the best performing ωB97X-D/EFP model, a QM region containing about 6% of the total number of solvent molecules is needed to approach within 10 kJ mol−1 of the pure QM result if the QM region was chosen based on the distance from the reaction centre.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2015

Polarizable embedding with a multiconfiguration short-range density functional theory linear response method

We present here the coupling of a polarizable embedding (PE) model to the recently developed multiconfiguration short-range density functional theory method (MC-srDFT), which can treat multiconfigurational systems with a simultaneous account for dynamical and static correlation effects. PE-MC-srDFT is designed to combine efficient treatment of complicated electronic structures with inclusion of effects from the surrounding environment. The environmental effects encompass classical electrostatic interactions as well as polarization of both the quantum region and the environment. Using response theory, molecular properties such as excitation energies and oscillator strengths can be obtained. The PE-MC-srDFT method and the additional terms required for linear response have been implemented in a development version of DALTON. To benchmark the PE-MC-srDFT approach against the literature data, we have investigated the low-lying electronic excitations of acetone and uracil, both immersed in water solution. The PE-MC-srDFT results are consistent and accurate, both in terms of the calculated solvent shift and, unlike regular PE-MCSCF, also with respect to the individual absolute excitation energies. To demonstrate the capabilities of PE-MC-srDFT, we also investigated the retinylidene Schiff base chromophore embedded in the channelrhodopsin protein. While using a much more compact reference wave function in terms of active space, our PE-MC-srDFT approach yields excitation energies comparable in quality to CASSCF/CASPT2 benchmarks.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2015

First-principles assessment of hole transport in pure and Li-doped NiO

Alloying nickel oxide (NiO) with lithium oxide (Li 2 O) at high Li concentrations may reduce NiO's band gap and expand its use as a light absorber in photocatalytic and tandem dye-sensitized solar cell technologies. In this work, we evaluate the viability of this alloy as a p-type hole transport material. We use embedded cluster models, along with unrestricted Hartree–Fock and complete active space self-consistent field theories, to study the impact of alloying on polaronic transport of holes. Our calculated energy barrier for hole transfer in undoped NiO is in excellent agreement with the experimental value of ~0.1 eV. We predict that hole transport in NiO is anisotropic and mostly confined parallel to the (111) ferromagnetic planes. Applying the same model to Li-doped NiO indicates that isolated Li ions do not introduce free holes into NiO samples. However, free holes can be created in the homogeneous Li 0.125 Ni 0.875 O alloy, in which the Li concentration is very high. Our kinetic Monte Carlo calculations show that hole mobility in this alloy is lower than in undoped NiO. However, the additional free holes and the predicted lower band gap of Li 0.125 Ni 0.875 O should increase hole conductivity compared to NiO upon alloy formation. Therefore, Li 0.125 Ni 0.875 O alloys have potential for use as a hole transporter, as well as a sunlight absorber, in a variety of solar energy applications.

25 ENERGY STORAGE↗