Search NASA⌕ Search

SEARCH · Search NASA

Results for “Generalized gradient approximation”

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 631 records · Page 35

Accuracy of Gradient Reconstruction on Grids with High Aspect Ratio

Gradient approximation methods commonly used in unstructured-grid finite-volume schemes intended for solutions of high Reynolds number flow equations are studied comprehensively. The accuracy of gradients within cells and within faces is evaluated systematically for both node-centered and cell-centered formulations. Computational and analytical evaluations are made on a series of high-aspect-ratio grids with different primal elements, including quadrilateral, triangular, and mixed element grids, with and without random perturbations to the mesh. Both rectangular and cylindrical geometries are considered; the latter serves to study the effects of geometric curvature. The study shows that the accuracy of gradient reconstruction on high-aspect-ratio grids is determined by a combination of the grid and the solution. The contributors to the error are identified and approaches to reduce errors are given, including the addition of higher-order terms in the direction of larger mesh spacing. A parameter GAMMA characterizing accuracy on curved high-aspect-ratio grids is discussed and an approximate-mapped-least-square method using a commonly-available distance function is presented; the method provides accurate gradient reconstruction on general grids. The study is intended to be a reference guide accompanying the construction of accurate and efficient methods for high Reynolds number applications

Thomas, James↗

High absorption photovoltaic material and methods of making the same

A high absorption photovoltaic material and method of making the material for use in a solar cell are disclosed. The photovoltaic material includes a surface modified with a layer of repeating photonic crystal structures. The photonic crystal structures are approximately inverse conically shaped and have a curved sidewall that has an approximately Gaussian shape. The photonic crystal structures generally have a high vertical depth and sidewall angle. The structures also have a gradient refractive index profile and exhibit the parallel-to-interface refraction light trapping effect. An anti-reflective coating is disposed over the photonic crystal structure layer. The photovoltaic material exhibits near unity light absorption over a broad range of visible and near infrared wavelengths and incidence angles, even at reduced thicknesses. The photovoltaic structures are formed via a combined photolithography and reactive-ion etching method at low power with a gas mixture having a high ratio of an etchant component to a passivation component.

Kuang, Ping↗

Symmetry-Based Structured Matrices for Efficient Approximately Equivariant Networks

There has been much recent interest in designing symmetry-aware neural networks (NNs) exhibiting relaxed equivariance. Such NNs aim to interpolate between being exactly equivariant and being fully flexible, affording consistent performance benefits. In a separate line of work, certain structured parameter matrices -- those with displacement structure, characterized by low displacement rank (LDR) -- have been used to design small-footprint NNs. Displacement structure enables fast function and gradient evaluation, but permits accurate approximations via compression primarily to classical convolutional neural networks (CNNs). In this work, we propose a general framework -- based on a novel construction of symmetry-based structured matrices -- to build approximately equivariant NNs with significantly reduced parameter counts. Our framework integrates the two aforementioned lines of work via the use of so-called Group Matrices (GMs), a forgotten precursor to the modern notion of regular representations of finite groups. GMs allow the design of structured matrices -- resembling LDR matrices -- which generalize the linear operations of a classical CNN from cyclic groups to general finite groups and their homogeneous spaces. We show that GMs can be employed to extend all the elementary operations of CNNs to general discrete groups. Further, the theory of structured matrices based on GMs provides a generalization of LDR theory focussed on matrices with cyclic structure, providing a tool for implementing approximate equivariance for discrete groups. We test GM-based architectures on a variety of tasks in the presence of relaxed symmetry. We report that our framework consistently performs competitively compared to approximately equivariant NNs, and other structured matrix-based compression frameworks, sometimes with a one or two orders of magnitude lower parameter count.

Samudre, Ashwin↗

Analytical approximation of a distorted reflector surface defined by a discrete set of points

Reflector antennas on Earth orbiting spacecrafts generally cannot be described analytically. The reflector surface is subjected to a large temperature fluctuation and gradients, and is thus warped from its true geometrical shape. Aside from distortion by thermal stresses, reflector surfaces are often purposely shaped to minimize phase aberrations and scanning losses. To analyze distorted reflector antennas defined by discrete surface points, a numerical technique must be applied to compute an interpolatory surface passing through a grid of discrete points. In this paper, the distorted reflector surface points are approximated by two analytical components: an undistorted surface component and a surface error component. The undistorted surface component is a best fit paraboloid polynomial for the given set of points and the surface error component is a Fourier series expansion of the deviation of the actual surface points, from the best fit paraboloid. By applying the numerical technique to approximate the surface normals of the distorted reflector surface, the induced surface current can be obtained using physical optics technique. These surface currents are integrated to find the far field radiation pattern.

Acosta, Roberto J.↗

Plateau Phenomenon in Gradient Descent Training of RELU Networks: Explanation, Quantification, and Avoidance

The ability of neural networks to provide ‘best in class’ approximation across a wide range of applications is well-documented. Nevertheless, the powerful expressivity of neural networks comes to naught if one is unable to effectively train (choose) the parameters defining the network. In general, neural networks are trained by gradient descent type optimization methods,a stochastic variant thereof. In practice, such methods result in the loss function decreases rapidly at the beginning of training but then, after a relatively small number of steps, significantly slow down. The loss may even appear to stagnate over the period of a large number of epochs, only to then suddenly start to decrease fast again for no apparent reason. This so-called plateau phenomenon manifests itself in many learning tasks. The present work aims to identify and quantify the root causes of plateau phenomenon.analysis is carried out in the setting of univariate ReLU networks. No assumptions are made on the number of neurons relative to the number of training data, and our results hold for both the lazy and adaptive regimes. Here, the main findings are: plateaux correspond to periods during which activation patterns remain constant, where activation pattern refers to the number of data points that activate a given neuron; quantification of convergence of the gradient flow dynamics; and, characterization stationary points in terms solutions of local least squares regression lines over subsets of the training data. Based on these conclusions, we propose a new iterative training method, the Active Neuron Least Squares (ANLS), characterised by the explicit adjustment of the activation pattern at each step, which is designed to enable a quick exit from a plateau. Illustrative numerical examples are included throughout.

97 MATHEMATICS AND COMPUTING↗

Modeling error PDF optimization based wavelet neural network modeling of dynamic system and its application in blast furnace ironmaking

In general, the modeling errors of dynamic system model are a set of random variables. The traditional performance index of modeling such as means square error (MSE) and root means square error (RMSE) can not fully express the connotation of modeling errors with stochastic characteristics both in the dimension of time domain and space domain. Therefore, the probability density function (PDF) is introduced to completely describe the modeling errors in both time scales and space scales. Based on it, a novel wavelet neural network (WNN) modeling method is proposed by minimizing the two-dimensional (2D) PDF shaping of modeling errors. First, the modeling error PDF by the tradional WNN is estimated using data-driven kernel density estimation (KDE) technique. Then, the quadratic sum of 2D deviation between the modeling error PDF and the target PDF is utilized as performance index to optimize the WNN model parameters by gradient descent method. Since the WNN has strong nonlinear approximation and adaptive capability, and all the parameters are well optimized by the proposed method, the developed WNN model can make the modeling error PDF track the target PDF, eventually. Simulation example and application in a blast furnace ironmaking process show that the proposed method has a higher modeling precision and better generalization ability compared with the conventional WNN modeling based on MSE criteria. Furthermore, the proposed method has more desirable estimation for modeling error PDF that approximates to a Gaussian distribution whose shape is high and narrow.

Modelling↗

Convective instability when the temperature gradient and rotation vector are oblique to gravity. I - Fluids without diffusion

A linear stability analysis of fluid layers under uniform rotation (generally oblique to gravity) which possess both vertical and horizontal temperature gradients is made by considering ideal fluids without diffusion within a Boussinesq approximation. This simplified configuration is used to assess the preferred convective modes as a function of latitude on a planet like Jupiter. The tilted rotation vector introduces a preference for roll-like disturbances with north-south orientations, while the horizontal temperature gradient produces a thermal wind shear which favors convective rolls oriented parallel to the flow in an east-west direction. It is found that the horizontal temperature gradient needed to produce a preference for the axisymmetric or east-west rolls increases with an increasing rotation rate and a decreasing latitude. The parameter values for Jupiter are estimated with the use of a simple radiative convective model, indicating a preference for axisymmetric rolls at nearly all latitudes if the convection zone depth is greater than about 200 km below the one atmosphere pressure level and convective roll characteristics which contribute to an equatorial acceleration.

Hathaway, D. H.↗

Experimental study of vibration isolation in thin-walled structural assemblies with embedded total-internal-reflection metasurfaces

The concept of total-internal-reflection elastic metasurface (TIR-MS) was recently proposed [1] and employed within flexible planar waveguides in order to create highly subwavelength sound-hard barriers impenetrable to low frequency elastic waves. The underlying physical mechanism relies on the design of engineered interfaces exhibiting extreme phase gradients such that any incoming wave at, approximately, any incidence will experience total-internal-reflection conditions. At the design frequency, the metasurface exhibits a large phase gradient such that, in accordance with the generalized Snell's law, the first critical angle is virtually always exceeded. It is worth noting that in practical realizations, the actual total reflection performance might vary depending on the angle of incidence. This dependence is due to the discrete implementation of the metasurface which results in diffraction effects. This paper presents the results of an experimental study that explores the vibration isolation performance of TIR-MS when applied to structures made of complex combinations of different elastic waveguides (e.g. bolted assemblies of beams, plates, and shells). Such system can be seen as a prototypical structure emulating mechanical assemblies of practical interest for many engineering applications. Experimental results confirm that, when the TIR-MS is embedded in the host waveguide, significant vibration isolation capabilities are achieved under quasi-omnidirectional incidence and highly subwavelength excitation conditions (i.e. the ratio of the operating wavelength to the width of the TIR-MS is approximately 5.25). Furthermore, these experimental results suggest new interesting directions to achieve vibration isolation and mechanical energy filtering for practical engineering systems.

42 ENGINEERING↗

The Velikhov and anti-Velikhov effects in the theory of magnetorotational instability

A theory of magnetorotational instability (MRI) allowing an equilibrium plasma pressure gradient and nonaxisymmetry of perturbations is developed. This approach reveals that in addition to the Velikhov effect driving the MRI due to negative rotation frequency profile, d{theta}{sup 2}/dr < 0, there is an opposite effect (the anti-Velikhov effect) weakening this driving (here, {theta} is the rotation frequency and r is the radial coordinate). It is shown that in addition to the Velikhov mechanism, two new mechanisms of MRI driving are possible, one of which is due to the pressure gradient squared and the other is due to the product of the pressure and density gradients. The analysis includes both the one-fluid magnetohydrodynamic plasma model and the kinetics allowing collisionless effects. In addition to the pure plasma containing ions and electrons, the dusty plasma is considered. The charged dust effect on stability is analyzed using the approximation of immobile dust. In the presence of dust, a term with the electric field appears in the one-fluid equation of plasma motion. This electric field affects the equilibrium plasma rotation and also gives rise to a family of instabilities of the rotating plasma, called the dust-induced rotational instabilities.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

An Approximate Solution for the Circumsolar Flow Field of a Sun Moving Through the Local Interstellar Medium (LISM)

The solar system is in relative motion with respect to the ambient interstellar medium. The supersonic solar wind is expected to pass through the termination shock, thus the solar wind plasma eventually has to enter into an asymptotic outflow geometry appropriately adopted to this counterflow situation. Many attempts have been done to simulate the interaction between the solar wind and the LISM numerically. In this paper we generalize a Parker type analytical solution of the counterflow. The idea is to introduce a special kind of compressibility of the solar wind flow. With the assumption that only a transversal component of the density gradient normal to the flow lines exists we are able to calculate a full set of hydrodynamical quantities describing the circumsolar flow field of a Sun moving through the LISM. The equations governing the velocity and density fields lead to analytical solutions which can be taken as good approximations to the more general case of compressible plasma flows.

Ratkiewicz, Romana E.↗

Low-frequency modes and nonbarotropic effects in pseudo-Newtonian accretion disks

A recently developed formalism is used to reexamine the question of the existence of hydrodynamical modes that pulsate with very low frequencies in the inner regions of accretion disks. The formalism is valid in an exact sense for the adiabatic pulsations of rotating Newtonian fluids that are generally nonbarotropic (such as those with 'nonadiabatic temperature gradients,' for example), and hence its application in the present context represents an improvement over previous analyses that are more approximate. The formalism is applied to thin non-self-gravitating disks, with the gravitational potential of the central source modified in the usual way in order to simulate relativistic effects. In the barotropic limit, the analyses indicate that in many cases nearly Keplerian disks exhibit nonaxisymmetric modes of pulsation that are trapped in the inner disk regions, with pulsation periods much longer than the dynamical timescale. These results are similar to those of earlier calculations that assume disks pulsate without changing the temperature distribution. A method is developed for including lowest order nonbarotropic effects. Previous analyses have been incapable of accurately treating the nonbarotropic regime. The application of the present method to the low-frequency modes reveals that, due to unexpected cancellations among terms, the nonbarotropic correction to the pusation frequency omega is only of order tilde-omega(sub BV exp 2) omega, where tilde-omega(sub BV) is the appropriate dimensionless Brunt-Vaisala frequency. This correction is much smaller than the expected correction of order tilde-omega(sub BV) Omega, where Omega is the rotation angular velocity. The important conclusion drawn from this is that nonbarotropic corrections are generally small and hence that low-frequency modes persist into the nonbarotropic regime. For disk temperatures appropriate to X-ray emission, the adiabatic frequencies of trapped modes are of the same order as the frequencies associated with quasi-periodic oscillations (QPOs) of certain low-mass X-ray binaries and other X-ray sources. It is suggested that further consideration be given to the possiblity that modulation of emission processes by these modes can explain at least some of the QPO phenomena.

Ipser, James R.↗

Dynamics of stratified liquids in the presence of space charge.

The dispersion characteristics of small signal electrohydrodynamic space charge and polarization waves in continuously and discretely stratified liquids are developed. The liquids, modeled as incompressible, inviscid, and perfectly insulating have distributions of mass density, charge density, permittivity, and electric field. A general set of relations for perturbation field and flow variables defined for a prototype layer are used to describe the dynamics of many superposed planar layers. It is shown that the general relations can be used to represent systems of coupled layers which approximate continuous distributions by a series of step functions. Specific examples of weak gradient and exponential distributions are presented showing that the solution found directly from the distributed theory is approached by the system of coupled layers, if the limit is taken in which the number of layers become large while each layer thickness approaches zero.

Zahn, M.↗

Effects of pressure gradients on turbulent premixed flames

The influence of a constant acceleration on a turbulent premixed flame is studied by direct numerical simulation. This acceleration induces a mean pressure gradient across the flame brush, leading to a modification of the turbulent flame structure due to differential buoyancy mechanisms between heavy cold fresh and light hot burnt gases. Such a pressure gradient may be encountered in practical applications in ducted flames. A favorable pressure gradient, i.e. the pressure decreases from unburnt to burnt gases, is found to decrease the flame wrinkling, the flame brush thickness, and the turbulent flame speed. A favorable pressure gradient also promotes counter-gradient turbulent transport. On the other hand, adverse pressure gradients tend to increase the flame brush thickness and turbulent flame speed, and promote classical gradient turbulent transport. The balance equation for the turbulent flux of the Favre averaged progress variable is also analyzed. The first results show that the fluctuating pressure term, cannot be neglected as generally assumed in models. Simple models assuming that a high mean pressure gradient may only be balanced by the cross-dissipation term seem too approximate. This analysis has to be continued to compare simulation data and closure schemes proposed for the transport equation. The analysis developed by Veynante et al.(1995) has been extended to imposed acceleration and mean pressure gradients. A simple model for the turbulent flux is proposed and validated from simulation data. Then, a modified criterion is derived to delineate between counter-gradient and gradient turbulent diffusion. In fact, counter-gradient diffusion may occur in most practical applications, especially for ducted flames.

Veynante, D.↗

General Purpose Fortran Program for Discrete-Ordinate-Method Radiative Transfer in Scattering and Emitting Layered Media: An Update of DISORT

This update reports a state-of-the-art discrete ordinate algorithm for monochromatic unpolarized radiative transfer in non-isothermal, vertically inhomogeneous, but horizontally homogeneous media. The physical processes included are Planckian thermal emission, scattering with arbitrary phase function, absorption, and surface bidirectional reflection. The system may be driven by parallel or isotropic diffuse radiation incident at the top boundary, as well as by internal thermal sources and thermal emission from the boundaries. Radiances, fluxes, and mean intensities are returned at user-specified angles and levels. DISORT has enjoyed considerable popularity in the atmospheric science and other communities since its introduction in 1988. Several new DISORT features are described in this update: intensity correction algorithms designed to compensate for the 8-M forward-peak scaling and obtain accurate intensities even in low orders of approximation; a more general surface bidirectional reflection option; and an exponential-linear approximation of the Planck function allowing more accurate solutions in the presence of large temperature gradients. DISORT has been designed to be an exemplar of good scientific software as well as a program of intrinsic utility. An extraordinary effort has been made to make it numerically well-conditioned, error-resistant, and user-friendly, and to take advantage of robust existing software tools. A thorough test suite is provided to verify the program both against published results, and for consistency where there are no published results. This careful attention to software design has been just as important in DISORT's popularity as its powerful algorithmic content.

Tsay, Si-Chee↗

Validation of Aura Microwave Limb Sounder HCl Measurements

The Earth Observing System (EOS) Microwave Limb Sounder (MLS) aboard the Aura satellite has provided daily global HCl profiles since August 2004. We provide a characterization of the resolution, random and systematic uncertainties, and known issues for the version 2.2 MLS HCl data. The MLS sampling allows for comparisons with many (~1500 to more than 3000) closely matched profiles from the Halogen Occultation Experiment (HALOE) and Atmospheric Chemistry Experiment Fourier Transform Spectrometer (ACE-FTS). These data sets provide HCl latitudinal distributions that are, overall, very similar to those from (coincident) MLS profiles, although there are some discrepancies in the upper stratosphere between the MLS and HALOE gradients. As found in previous work, MLS and ACE HCl profiles agree very well (within approximately 5%, on average), but the MLS HCl abundances are generally larger (by 10-20%) than HALOE HCl. The bias versus HALOE is unlikely to arise mostly from MLS, as a similar systematic bias (of order 15%) is not observed between average MLS and balloon-borne measurements of HCl, obtained over Fort Sumner, New Mexico, in 2004 and 2005. At the largest pressure (147 hPa) for MLS HCl, a high bias (approximately 0.2 ppbv) is apparent in analyses of low to midlatitude data versus in situ aircraft chemical ionization mass spectrometry (CIMS) HCl measurements from the Aura Validation Experiment (AVE) campaigns in 2004, 2005, and 2006; this bias is also observed in comparisons of MLS and aircraftHCl/O3 correlations. Good agreement between MLS and CIMS HCl is obtained at 100 to 68 hPa. The recommended pressure range for MLS HCl is from 100 to 0.15 hPa.

Earth Observing System (EOS)↗

Light Scattering by Small Multilayer Particles: A Generalized Separation of Variables Method

A Rayleigh approximation is constructed for light scattering by small multilayer axisymmetric particles, in which their polarizability is determined by the generalized separation of variables method (SVM). In this method, scalar potentials, the gradients of which yield the electric-field strengths, are represented as expansions in spherical harmonics of the Laplace equation. Unknown coefficients of expansions are determined from the boundary conditions, which are reduced to infinite systems of linear algebraic equations (ISLAEs), since the separation of variables is incomplete. The T matrix of the electrostatic problem, principal element T{sub 11} of which is proportional to the particle polarizability, is determined. The necessary condition for the ISLAEs solvability for the SVM coincides with the condition of the correct application of the extended boundary conditions method (EBCM). However, numerical calculations in which finite-dimensional (i.e., reduced) systems are solved, yield different results in ranges of variation of parameters that are close to the boundary of the range of applicability. An analysis of the numerical calculations of the scattering and absorption cross sections for two-layer confocal spheroids, an exact solution for which can be obtained using spheroidal harmonics, shows that the SVM is preferable to the EBCM. It turned out that the proposed method yields workable results in a wider range of variation of parameters. Even outside the range of applicability, in which it should be regarded as a certain approximate solution, its use in a number of cases is quite acceptable. Additional calculations for three-layer nonconfocal spheroids, as well as for three-layer similar pseudospheroids and Pascal’s snails, which can be obtained from spheroids as a result of the inversion with respect to the coordinate origin and one of the foci, respectively, confirm these inferences. We note that, for certain values of the parameters, the shapes of the latter particles are nonconvex.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

A memory-distributed quasi-Newton solver for nonlinear programming problems with a small number of general constraints

We approach the problem of parallelizing state-of-the-art nonlinear programming optimization algorithms.Specifically, we focus on parallelizing quasi-Newton interior-point methods that use limited-memory secant Hessian approximations. Such interior-point methods are known to have better convergence properties and to be more effective on large-scale problems than gradient-based and derivative-free optimization algorithms. We target nonlinear and potentially nonconvex optimization problems with an arbitrary number of bound constraints and a small number of general equality and inequality constraints on the optimization variables. These problems occur for example in the form of optimal control, optimal design, and inverse problems governed by ordinary or partial differential equations, whenever they are expressed in a “reduced-space” optimization approach. We introduce and analyze the time and space complexity of a decomposition method for solving the quasi-Newton linear systems that leverages the fact that the quasi-Newton Hessian matrix has a small number of dense blocks that border a low-rank update of a diagonal matrix. This enables an efficient parallelization on memory-distributed computers of the iterations of the optimization algorithm, a state-of-the-art filter line-search interior-point algorithm by Wächter et. al. We illustrate the efficiency of the proposed method by solving structural topology optimization problems on up to 4608 cores on a parallel machine.

97 MATHEMATICS AND COMPUTING↗

Alfvén eigenmode stability and critical gradient energetic particle transport using the Trapped-Gyro-Landau-Fluid model

The Trapped-Gyro-Landau-Fluid (TGLF) transport model is a physically realistic and comprehensive theory based on a local quasilinear transport model fitted to linear and nonlinear GYRO gyrokinetic simulations [Staebler et al., Phys. Plasmas 14, 55909 (2007)]. This work presents the first use of the TGLF model to treat low-n Alfvén eigenmode (AE) stability and energetic particle (EP) transport. TGLF accurately recovers the local GYRO toroidicity-induced AE (TAE) and energetic particle mode (EPM) linear growth and frequency rates for a fusion alpha case. With a very high grid resolution, TGLF can quickly find the critical EP pressure gradient profile for stiff EP transport based on an AE linear threshold given the background thermal plasma profiles in DIII-D. The TGLF critical gradient profile using the recipe γ AE =0, that is the linear AE growth rate without additional driving rates from the background plasma gradients, matches the more expensive linear GYRO results with a single worst toroidal mode number n. TGLF can easily find the minimum critical gradient profile with testing multiple ns. From a database of runs using a newly developed TGLFEP code, a rough but insightful parametric “power law” scaling for critical EP beta is demonstrated. An important toroidal stabilization condition on the EP pressure gradient p EP /L$EP\atop{p}$ drive is isolated: R/L$EP\atop{p}$ > C R ~ 3, where L$EP\atop{p}$ is the EP pressure gradient length and R is the tokamak major radius. This paper also demonstrates that relaxation of the fixed slowing down EP profile shape approximation often used to find the critical EP density profile has little effect on the resulting EP transport. The single EP species critical gradient model is generalized to handle two EP species.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗