Search NASA⌕ Search

SEARCH · Search NASA

Results for “gradient methods”

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 1,351 records · Page 75

Implementation and benchmark of a long-range corrected functional in the density functional based tight-binding method

Bridging the gap between first principles methods and empirical schemes, the density functional based tight-binding method (DFTB) has become a versatile tool in predictive atomistic simulations over the past years. One of the major restrictions of this method is the limitation to local or gradient corrected exchange-correlation functionals. This excludes the important class of hybrid or long-range corrected functionals, which are advantageous in thermochemistry, as well as in the computation of vibrational, photoelectron, and optical spectra. The present work provides a detailed account of the implementation of DFTB for a long-range corrected functional in generalized Kohn-Sham theory. We apply the method to a set of organic molecules and compare ionization potentials and electron affinities with the original DFTB method and higher level theory. The new scheme cures the significant overpolarization in electric fields found for local DFTB, which parallels the functional dependence in first principles density functional theory (DFT). At the same time, the computational savings with respect to full DFT calculations are not compromised as evidenced by numerical benchmark data.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Downward continuation of the free-air gravity anomalies to the ellipsoid using the gradient solution and terrain correction: An attempt of global numerical computations

The formulas for the determination of the coefficients of the spherical harmonic expansion of the disturbing potential of the earth are defined for data given on a sphere. In order to determine the spherical harmonic coefficients, the gravity anomalies have to be analytically downward continued from the earth's surface to a sphere-at least to the ellipsoid. The goal is to continue the gravity anomalies from the earth's surface downward to the ellipsoid using recent elevation models. The basic method for the downward continuation is the gradient solution (the g sub 1 term). The terrain correction was also computed because of the role it can play as a correction term when calculating harmonic coefficients from surface gravity data. The fast Fourier transformation was applied to the computations.

Wang, Y. M.↗

Simulated Microstructural and Compositional Evolution of U-Pu-Zr Alloys Using the Potts-Phase Field Modeling Technique

U-Pu-Zr alloys are considered ideal metallic fuels for Experimental Breeder Reactors because of their superior material properties and potential for increased burnup performance. However, significant constituent redistribution has been observed in these alloys when irradiated, or subject to a thermal gradient, resulting in inhomogeneity of both composition and phase, which, in turn, alters the fuel performance. The hybrid Potts-phase field method is reformulated for ternary alloys in a thermal gradient and utilized to simulate and predict constituent redistribution and phase transformations in the U-Pu-Zr nuclear fuel system. Simulated evolution profiles for the U-16Pu-23Zr (at%) alloy show concentric zones that are compared with published experimental results; discrepancies in zone size are attributed to thermal profile differences and assumptions related to the diffusivity values used. Twenty-one alloys, over the entire ternary compositional spectrum, are also simulated to investigate the effects of alloy composition on constituent redistribution and phase transformations. The U-40Pu-20Zr (at%) alloy shows the most potential for compositional uniformity and phase homogeneity, throughout a thermal gradient, while remaining in the compositional range of feasible alloys.

36 MATERIALS SCIENCE↗

Ab Initio Thermodynamic Model for Magnesium Carbonates and Hydrates

An ab initio thermodynamic framework for predicting properties of hydrated magnesium carbonate minerals has been developed using density-functional theory linked to macroscopic thermodynamics through the experimental chemical potentials for MgO, water, and CO2. Including semiempirical dispersion via the Grimme method and small corrections to the generalized gradient approximation of Perdew, Burke, and Ernzerhof for the heat of formation yields a model with quantitative agreement for the benchmark minerals brucite, magnesite, nesquehonite, and hydromagnesite. The model shows how small differences in experimental conditions determine whether nesquehonite, hydromagnesite, or magnesite is the result of laboratory synthesis from carbonation of brucite, and what transformations are expected to occur on geological time scales. Because of the reliance on parameter-free first principles methods, the model is reliably extensible to experimental conditions not readily accessible to experiment and to any mineral composition for which the structure is known or can be hypothesized, including structures containing defects, substitutions, or transitional structures during solid state transformations induced by temperature changes or processes such as water, CO2, or O2 diffusion. Demonstrated applications of the ab initio thermodynamic framework include an independent means to evaluate differences in thermodynamic data for lansfordite, predicting the properties of Mg analogs of Ca-based hydrated carbonates monohydrocalcite and ikaite which have not been observed in nature, and an estimation of the thermodynamics of barringtonite from the stoichiometry and a single experimental observation.

Environmental Molecular Sciences Laboratory↗

Method and apparatus for producing a carbon based foam article having a desired thermal-conductivity gradient

A carbon based foam article is made by heating the surface of a carbon foam block to a temperature above its graphitizing temperature, which is the temperature sufficient to graphitize the carbon foam. In one embodiment, the surface is heated with infrared pulses until heat is transferred from the surface into the core of the foam article such that the graphitizing temperature penetrates into the core to a desired depth below the surface. The graphitizing temperature is maintained for a time sufficient to substantially entirely graphitize the portion of the foam article from the surface to the desired depth below the surface. Thus, the foam article is an integral monolithic material that has a desired conductivity gradient with a relatively high thermal conductivity in the portion of the core that was graphitized and a relatively low thermal conductivity in the remaining portion of the foam article.

Klett, James W↗

A Meshfree Cell-based Smoothed Point Interpolation Method for Solid Mechanics Problems

In the framework of a weakened weak (W{sup 2}) formulation using a generalized gradient smoothing operation, this paper introduces a novel meshfree cell-based smoothed point interpolation method (CS-PIM) for solid mechanics problems. The W{sup 2} formulation seeks solutions from a normed G space which includes both continuous and discontinuous functions and allows the use of much more types of methods to create shape functions for numerical methods. When PIM shape functions are used, the functions constructed are in general not continuous over the entire problem domain and hence are not compatible. Such an interpolation is not in a traditional H{sup 1} space, but in a G{sup 1} space. By introducing the generalized gradient smoothing operation properly, the requirement on function is now further weakened upon the already weakened requirement for functions in a H{sup 1} space and G{sup 1} space can be viewed as a space of functions with weakened weak (W{sup 2}) requirement on continuity. The cell-based smoothed point interpolation method (CS-PIM) is formulated based on the W{sup 2} formulation, in which displacement field is approximated using the PIM shape functions, which possess the Kronecker delta property facilitating the enforcement of essential boundary conditions [3]. The gradient (strain) field is constructed by the generalized gradient smoothing operation within the cell-based smoothing domains, which are exactly the triangular background cells. A W{sup 2} formulation of generalized smoothed Galerkin (GS-Galerkin) weak form is used to derive the discretized system equations. It was found that the CS-PIM possesses the following attractive properties: (1) It is very easy to implement and works well with the simplest linear triangular mesh without introducing additional degrees of freedom; (2) it is at least linearly conforming; (3) this method is temporally stable and works well for dynamic analysis; (4) it possesses a close-to-exact stiffness, which is much softer than the overly-stiff FEM model and much stiffer than the overly-soft node-based smoothed point interpolation method (NS-PIM); (5) the results of the present method are of better accuracy and higher convergence rate than the linear FEM model using the same set of triangular meshes.

97 MATHEMATICS AND COMPUTING↗

A Differentiable Hydrology Approach for Modeling With Time-Varying Parameters

In the analysis of observational data with mathematical models, it is often desirable to let one or more model parameters vary with time to account for changing environmental conditions or to offer more flexibility. A standard approach for dynamic models is to allow for the number of parameters to grow as a function of time, leading to a high-dimensional inference problem for long data records. We propose using Hamiltonian Monte Carlo, a gradient-based Markov chain Monte Carlo (MCMC) method, to perform Bayesian inference for time-varying parameters (TVP) in hydrology models. As derivatives of model error functions with regard to parameters are not available in closed form, we implemented the GR4J rainfall-runoff models in Theano/PyMC3, and Jax/NumPyro, allowing for the application of automatic differentiation techniques. A simulation experiment assessing the viability of these methods for recovering underlying temporal variation in parameters indicates that a discrete Gaussian random walk-based prior appears to be best suited across a range of scenarios while the continuous-time Gaussian process compares relatively poorly. In an analysis of 20 years of daily real-world streamflow records from the Model Parameter Estimation Experiment (MOPEX) hydrology data set with increasing data sparsity and at multiple temporal resolutions, we find that TVP inference with gradient-based MCMC is a flexible and reliable approach for analyzing hydrology models with dynamic parameter sets.

54 ENVIRONMENTAL SCIENCES↗

Nonequilibrium solvent effects in Born-Oppenheimer molecular dynamics for ground and excited electronic states

The effects of solvent on molecular processes such as excited state relaxation and photochemical reaction often occurs in a nonequilibrium regime. Dynamic processes such as these can be simulated using excited state molecular dynamics. In this work, we describe methods of simulating nonequilibrium solvent effects in excited state molecular dynamics using linear-response time-dependent density functional theory and apparent surface charge methods. These developments include a propagation method for solvent degrees of freedom and analytical energy gradients for the calculation of forces. Molecular dynamics of acetaldehyde in water or acetonitrile are demonstrated where the solute-solvent system is out of equilibrium due to photoexcitation and emission.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Gradient of molecular Hartree–Fock–Bogoliubov energy with a linear combination of atomic orbital quasiparticle wave functions

The analytical gradient for the atomic-orbital-based Hartree–Fock–Bogoliubov (HFB) energy functional, the modified form of which was proposed by Staroverov and Scuseria to account for the static electron correlation [J. Chem. Phys. 117, 11107 (2002)], is derived. Interestingly, the Pulay force for the HFB energy is expressed with the same formula as that for the Hartree–Fock method. The efficiency of the present HFB energy gradient is demonstrated in the geometry optimizations of conjugated and biradical systems. The geometries optimized by using the HFB method with the appropriate factor ζ, which controls the degree of static correlation included, are found to show good agreement with those obtained by using a complete active-space self-consistent field method, although they are significantly dependent on ζ.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Nuclear–Electronic Orbital QM/MM Approach: Geometry Optimizations and Molecular Dynamics

Hybrid quantum mechanical/molecular mechanical (QM/MM) methods allow simulations of chemical reactions in atomistic solvent and heterogeneous environments such as proteins. Herein, the nuclear–electronic orbital (NEO) QM/MM approach is introduced to enable the quantization of specified nuclei, typically protons, in the QM region using a method such as NEO-density functional theory (NEO-DFT). This approach includes proton delocalization, polarization, anharmonicity, and zero-point energy in geometry optimizations and dynamics. Expressions for the energies and analytical gradients associated with the NEO-QM/MM method, as well as the previously developed polarizable continuum model (NEO-PCM), are provided. Geometry optimizations of small organic molecules hydrogen bonded to water in either dielectric continuum solvent or explicit atomistic solvent illustrate that aqueous solvation can strengthen hydrogen-bonding interactions for the systems studied, as indicated by shorter intermolecular distances at the hydrogen-bond interface. We then performed a real-time direct dynamics simulation of a phenol molecule in explicit water using the NEO-QM/MM method. Furthermore, these developments and initial examples provide the foundation for future studies of nuclear–electronic quantum dynamics in complex chemical and biological environments.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Application of static and dynamic enclosures for determining dimethyl sulfide and carbonyl sulfide exchange in Sphagnum peatlands: Implications for the magnitude and direction of flux

A static enclosure method was applied to determine the exchange of dimethyl sulfide (DMS) and carbonyl sulfide (OCS) between the surface of Sphagnum peatlands and the atmosphere. Measurements were performed concurrently with dynamic (flow through) enclosure measurements with sulfur-free air used as sweep gas. This latter technique has been used to acquire the majority of available data on the exchange of S gases between the atmosphere and the continental surfaces and has been criticized because it is thought to overestimate the true flux of gases by disrupting natural S gas gradients. DMS emission rates determined by both methods were not statistically different between 4 and greater than 400 nmol/sq m/h, indicating that previous data on emissions of at least DMS are probably valid. However, the increase in DMS in static enclosures was not linear, indicating the potential for a negative feedback of enlosure DMS concentrations on efflux. The dynamic enclosure method measured positive OCS flux rates (emission) at all sites, while data using static enclosures indicated that OCS was consumed from the atmosphere at these same sites at rates of 3.7 to 55 nmol/sq m/h. Measurements using both enclosure techniques at a site devoid of vegetation showed that peat was a source of both DMS and OCS. However, the rate of OCS efflux from decomposing peat was more than counterbalanced by OCS consumption by vegetation, including Sphagnum mosses, and net OCS uptake occurred at all sites. We propose that all wetlands are net sinks for OCS.

De Mello, William Z.↗

Enhanced magnetic flux density mapping using coherent steady state equilibrium signal in MREIT

Measuring the z-component of magnetic flux density B = (B{sub x}, B{sub y}, B{sub z}) induced by transversally injected current, magnetic resonance electrical impedance tomography (MREIT) aims to visualize electrical property (current density and/or conductivity distribution) in a three-dimensional imaging object. For practical implementations of MREIT technique, it is critical to reduce injection of current pulse within safety requirements. With the goal of minimizing the noise level in measured B{sub z} data, we propose a new method to enhance the measure B{sub z} data using steady-state coherent gradient multi-echo (SSC-GME) MR pulse sequence combining with injection current nonlinear encoding (ICNE) method in MREIT, where the ICNE technique injects current during a readout gradient to maximize the signal intensity of phase signal including B{sub z}. The total phase offset in SSC-GME includes additional magnetic flux density due to the injected current, which is different from the phase signal for the conventional spoiled MR pulse sequence. We decompose the magnetization precession phase from the total phase offset including B{sub z} and optimize B{sub z} data using the steady-state equilibrium signal. Results from a real phantom experiment including different kinds of anomalies demonstrated that the proposed method enhanced B{sub z} comparing to a conventional spoiled pulse sequence.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Adjoint approach to calculating shape gradients for three-dimensional magnetic confinement equilibria

The shape gradient quantifies the change in some figure of merit resulting from differential perturbations to a shape. Shape gradients can be applied to gradient-based optimization, sensitivity analysis and tolerance calculation. An efficient technique for computing the shape gradient for toroidal three-dimensional magnetohydrodynamic (MHD) equilibria is introduced. The method is based on the self-adjoint property of the equations for driven perturbations of MHD equilibria and is similar to the Onsager symmetry of transport coefficients. Two versions of the shape gradient are considered. One describes the change in a figure of merit due to an arbitrary displacement of the outer flux surface; the other describes the change in the figure of merit due to the displacement of a coil. The method is implemented for several example figures of merit and compared with direct calculation of the shape gradient. In these examples the adjoint method reduces the number of equilibrium computations by factors of $O(N)$ , where $N$ is the number of parameters used to describe the outer flux surface or coil shapes.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

SU-E-T-78: A Study of Dose Falloff Gradient in RapidArc Planning of Lung SBRT

Purpose: Rapid dose falloff beyond PTV is an important criterion for normal tissue sparing in SBRT. RTOG protocols use D2cm and R50% for plan quality evaluation. This study is aimed at analyzing the dose falloff gradient beyond the PTV extending into normal tissue structures and to ascertain the impact of PTV geometry and location on the dose falloff gradient in RapidArc planning of lung SBRT Methods: In this retrospective study, we analyzed 39 clinical RapidArc lung SBRT treatment plans that met RTOG-0915 criteria. Planning was done on Eclipse 8.9 for delivery on either Novalis NTx or TrueBeam STx equipped with HD MLCs. PTV volumes ranged between 5.3 and 113 cc (2.2 to 6 cm sphere equivalent diameter respectively) and their geographic locations were distributed in both lungs. 6X, 6X-FFF, 10X, and 10X-FFF energies were used for planning. All of these SBRT plans were planned using either 2 or 3 full or hemi arcs, with moderate couch kicks. Dose falloff gradients were obtained by generating 7 concentric 5 mm rings beyond PTV surface. Mean dose in each ring is used to evaluate percentage dose falloff gradient as a function of distance from the PTV surface. Results: The mean percentage dose falloff beyond PTV surface in all plans followed an exponential decay and the data was modeled with double exponential decay fit. Photon energy selection in the plan had a minimal impact on the mean percentage dose fall off beyond PTV surface. Conclusion: Dose falloff beyond PTV surface as a function of distance can be ascertained by the use of the double exponential decay fit coefficients in RapidArc planning of lung SBRT. This will help also in plan quality evaluation in addition to D2cm and R50% defined by RTOG.

60 APPLIED LIFE SCIENCES↗

Fronts propagating with curvature dependent speed: Algorithms based on Hamilton-Jacobi formulations

New numerical algorithms are devised (PSC algorithms) for following fronts propagating with curvature-dependent speed. The speed may be an arbitrary function of curvature, and the front can also be passively advected by an underlying flow. These algorithms approximate the equations of motion, which resemble Hamilton-Jacobi equations with parabolic right-hand-sides, by using techniques from the hyperbolic conservation laws. Non-oscillatory schemes of various orders of accuracy are used to solve the equations, providing methods that accurately capture the formation of sharp gradients and cusps in the moving fronts. The algorithms handle topological merging and breaking naturally, work in any number of space dimensions, and do not require that the moving surface be written as a function. The methods can be used also for more general Hamilton-Jacobi-type problems. The algorithms are demonstrated by computing the solution to a variety of surface motion problems.

Osher, Stanley↗

Accelerated fast iterative shrinkage thresholding algorithms for sparsity-regularized cone-beam CT image reconstruction

Purpose: The development of iterative image reconstruction algorithms for cone-beam computed tomography (CBCT) remains an active and important research area. Even with hardware acceleration, the overwhelming majority of the available 3D iterative algorithms that implement nonsmooth regularizers remain computationally burdensome and have not been translated for routine use in time-sensitive applications such as image-guided radiation therapy (IGRT). In this work, two variants of the fast iterative shrinkage thresholding algorithm (FISTA) are proposed and investigated for accelerated iterative image reconstruction in CBCT. Methods: Algorithm acceleration was achieved by replacing the original gradient-descent step in the FISTAs by a subproblem that is solved by use of the ordered subset simultaneous algebraic reconstruction technique (OS-SART). Due to the preconditioning matrix adopted in the OS-SART method, two new weighted proximal problems were introduced and corresponding fast gradient projection-type algorithms were developed for solving them. We also provided efficient numerical implementations of the proposed algorithms that exploit the massive data parallelism of multiple graphics processing units. Results: The improved rates of convergence of the proposed algorithms were quantified in computer-simulation studies and by use of clinical projection data corresponding to an IGRT study. The accelerated FISTAs were shown to possess dramatically improved convergence properties as compared to the standard FISTAs. For example, the number of iterations to achieve a specified reconstruction error could be reduced by an order of magnitude. Volumetric images reconstructed from clinical data were produced in under 4 min. Conclusions: The FISTA achieves a quadratic convergence rate and can therefore potentially reduce the number of iterations required to produce an image of a specified image quality as compared to first-order methods. We have proposed and investigated accelerated FISTAs for use with two nonsmooth penalty functions that will lead to further reductions in image reconstruction times while preserving image quality. Moreover, with the help of a mixed sparsity-regularization, better preservation of soft-tissue structures can be potentially obtained. The algorithms were systematically evaluated by use of computer-simulated and clinical data sets.

60 APPLIED LIFE SCIENCES↗

Approximate Bayesian Model Inversion for PDEs with Heterogeneous and State-Dependent Coefficients

We present two approximate Bayesian inference methods for parameter estimation in partial differential equation (PDE) models with space-dependent and state-dependent parameters. We demonstrate that these methods provide accurate and cost-effective alternatives to Markov Chain Monte Carlo simulation. We assume a parameterized Gaussian prior on the unknown functions, and approximate the posterior density by a parameterized multivariate Gaussian density. The parameters of the prior and posterior are estimated from sparse observations of the PDE model's states and the unknown functions themselves by maximizing the evidence lower bound (ELBO), a lower bound on the log marginal likelihood of the observations. The first method, Laplace-EM, employs the expectation maximization algorithm to maximize the ELBO, with a Laplace approximation of the posterior on the E-step, and minimization of a Kullback-Leibler divergence on the M-step. The second method, DSVI-EB, employs the doubly stochastic variational inference (DSVI) algorithm, in which the ELBO is maximized via gradient-based stochastic optimization, with nosiy gradients computed via simple Monte Carlo sampling and Gaussian backpropagation. We apply these methods to identifying diffusion coefficients in linear and nonlinear diffusion equations, and we find that both methods provide accurate estimates of posterior densities and the hyperparameters of Gaussian priors. While the Laplace-EM method is more accurate, it requires computing Hessians of the physics model. The DSVI-EB method is found to be less accurate but only requires gradients of the physics model.

approximate Bayesian inference, model inversion, v↗

Efficient kinetic thermal inverse modeling for organic material decomposition

The prevalent use of organic materials in manufacturing is a fire safety concern, and motivates the need for predictive thermal decomposition models. A critical component of predictive modeling is numerical inference of kinetic parameters from bench scale data. Currently, an active area of computational pyrolysis research focuses on identifying efficient, robust methods for optimization. This paper demonstrates that kinetic parameter calibration problems can successfully be solved using classical gradient-based optimization. We explore calibration examples that exhibit characteristics of concern: high nonlinearity, high dimensionality, complicated schemes, overlapping reactions, noisy data, and poor initial guesses. The examples demonstrate that a simple, non-invasive change to the problem formulation can simultaneously avoid local minima, avoid computation of derivative matrices, achieve a computational efficiency speedup of 10x, and make optimization robust to perturbations of parameter components. Techniques from the mathematical optimization and inverse problem communities are employed. By re-examining gradient-based algorithms, we highlight opportunities to develop kinetic parameter calibration methods that should outperform current methods.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗