Low-order approximations for the normal probability integral and the error function
Rational fractions of the form 0.5/(a + bx +
Carta, D. G.↗
SEARCH · Search NASA
Results for “Function 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.
Approximating Data with Stochastic and Physical Dependence using the Functional Tensor Train Models.
Abstract not provided.
Safta, Cosmin↗
Static stability of seaplane floats and hulls
Values of lateral and longitudinal metacentric heights for various seaplanes were calculated by means of approximate formulae derived here. The data are given in tabular form. Upon plotting these metacentric heights against the corresponding gross weights, it appears that the metacentric height is approximately a straight line function of the gross weight. For the lateral metacentric height GM = 13 + .002 W and for longitudinal metacentric height GM = 15 + .002 W, GM is in feet and the gross weight (W) is in pounds. Although only approximate, it is thought that the values indicated here are a reliable guide to current practice. It is recommended that the longitudinal and lateral metacentric heights be made equal and of the value given by GM = 15 = .002 W. The proper length or spacing required to satisfy the indicated value may then be obtained from substitution in the approximate formulae for metacentric height.
Diehl, W S↗
Calculation of Dynamical Response Functions Using a Bound-State Method
Abstract We investigate a method to extract response functions (dynamical polarisabilities) directly from a bound-state approach applied to calculations of perturbation-induced reactions. The use of a square-integrable basis leads to a response in the form of a sum of $$\delta $$ δ functions. We integrate this over energy and fit a smooth function to the resulting stepwise-continuous one. Its derivative gives the final approximation to the physical response function. We show that the method reproduces analytical results where known, and analyse the details for a variety of models. We apply it to some simple models, using the stochastic variational method as the numerical method. Albeit we find that this approach, and other numerical techniques, have some difficulties with the threshold behavior in coupled-channel problems with multiple thresholds, its stochastic nature allows us to extract robust results even for such cases.
71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
The solution of the relaxation problem for the Boltzmann equation by the integral iteration method
The Boltzmann equation is considered in terms of the problem of relaxation of some initial distribution function which depends only on velocities, to Maxwell's distribution function. The Boltzmann equation is given for the relaxation problem in which the distribution function f(t, u, v) is time dependent and is also dependent on two other variables u and v (the velocities of rigid spherical molecules). An iteration process is discussed in which the velocity space u, v is subdivided into squares, the distribution function in each square being approximated by the second-order surface from the values of the distribution function at nine points. The set of all of these points forms a network of u, v values at the nodes of which the distribution function can be found.
Limar, Y. F.↗
The Helmholtz-Kirchhoff integral relation as a framework for developing algorithms for sound propagation through inhomogeneous moving media
Transient sound propagation in an inhomogeneous moving medium is considered. For circumstances in which the medium is slowly varying over distances of a wavelength but possibly varying substantially over the propagation distance, a derivation is given of a new wave equation, which implicitly allows for diffraction and scattering and which also is consistent with earlier geometrical acoustics formulations. This wave equation is used as a starting point to derive a version of the Helmholtz-Kirchhoff integral relation that applies to inhomogeneous moving medium. It is suggested that a good approximation to the Green's function that appears in this relation is that derived from geometrical acoustics, the approximation becoming progressively better the shorter the distance between surfaces in the computation. It should also be at least as good as conventional ray acoustics, but can account for diffraction effects, such as at caustics.
Pierce, Allan D.↗
A Patch Density Recommendation based on Convergence Studies for Vehicle Panel Vibration Response resulting from Excitation by a Diffuse Acoustic Field
Producing fluid structural interaction estimates of panel vibration from an applied pressure field excitation are quite dependent on the spatial correlation of the pressure field. There is a danger of either over estimating a low frequency response or under predicting broad band panel response in the more modally dense bands if the pressure field spatial correlation is not accounted for adequately. It is a useful practice to simulate the spatial correlation of the applied pressure field over a 2d surface using a matrix of small patch area regions on a finite element model (FEM). Use of a fitted function for the spatial correlation between patch centers can result in an error if the choice of patch density is not fine enough to represent the more continuous spatial correlation function throughout the intended frequency range of interest. Several patch density assumptions to approximate the fitted spatial correlation function are first evaluated using both qualitative and quantitative illustrations. The actual response of a typical vehicle panel system FEM is then examined in a convergence study where the patch density assumptions are varied over the same model. The convergence study results illustrate the impacts possible from a poor choice of patch density on the analytical response estimate. The fitted correlation function used in this study represents a diffuse acoustic field (DAF) excitation of the panel to produce vibration response.
Smith, Andrew↗
Study of the zinc-silver oxide battery system
Theoretical and experimental models for the evaluation of current distribution in flooded, porous electrodes are discussed. An approximation for the local current distribution function was derived for conditions of a linear overpotential, a uniform concentration, and a very conductive matrix. By considering the porous electrode to be an analog of chemical catalyst structures, a dimensionless performance parameter was derived from the approximated current distribution function. In this manner the electrode behavior was characterized in terms of an electrochemical Thiele parameter and an effectiveness factor. It was shown that the electrochemical engineering approach makes possible the organizations of theoretical descriptions and of practical experience in the form of dimensionless parameters, such as the electrochemical Thiele parameters, and hence provides useful information for the design of new electrochemical systems.
Nanis, L.↗
Incorporating spatial context into statistical classification of multidimensional image data
Compound decision theory is employed to develop a general statistical model for classifying image data using spatial context. The classification algorithm developed from this model exploits the tendency of certain ground-cover classes to occur more frequently in some spatial contexts than in others. A key input to this contextural classifier is a quantitative characterization of this tendency: the context function. Several methods for estimating the context function are explored, and two complementary methods are recommended. The contextural classifier is shown to produce substantial improvements in classification accuracy compared to the accuracy produced by a non-contextural uniform-priors maximum likelihood classifier when these methods of estimating the context function are used. An approximate algorithm, which cuts computational requirements by over one-half, is presented. The search for an optimal implementation is furthered by an exploration of the relative merits of using spectral classes or information classes for classification and/or context function estimation.
Bauer, M. E.↗
Fast two-stream method for computing diurnal-mean actinic flux in vertically inhomogeneous atmospheres
Based on a derivation of the two-stream daytime-mean equations of radiative flux transfer, a method for computing the daytime-mean actinic fluxes in the absorbing and scattering vertically inhomogeneous atmosphere is suggested. The method applies direct daytime integration of the particular solutions of the two-stream approximations or the source functions. It is valid for any duration of period of averaging. The merit of the method is that the multiple scattering computation is carried out only once for the whole averaging period. It can be implemented with a number of widely used two-stream approximations. The method agrees with the results obtained with 200-point multiple scattering calculations. The method was also tested in runs with a 1-km cloud layer with optical depth of 10, as well as with aerosol background. Comparison of the results obtained for a cloud subdivided into 20 layers with those obtained for a one-layer cloud with the same optical parameters showed that direct integration of particular solutions possesses an 'analytical' accuracy. In the case of the source function interpolation, the actinic fluxes calculated above the one-layer and 20-layer clouds agreed within 1%-1.5%, while below the cloud they may differ up to 5% (in the worst case). The ways of enhancing the accuracy (in a 'two-stream sense') and computational efficiency of the method are discussed.
Filyushkin, V. V.↗
Unveiling the Role of Surface Ir-Oxo Species in O 2 Evolution at IrO 2 Electrocatalysts via Embedded Cluster Multireference Calculations
Understanding the mechanisms driving the oxygen evolution reaction (OER) on iridium oxide (IrO 2 )-based catalysts is essential to improving their performance and enabling an actual scale-up of water-splitting photoelectrochemical cells. The mechanistic pathways at IrO 2 interfaces have been extensively investigated computationally using density functional theory (DFT), which predicts a high-energy barrier for the last step of the OER of molecular oxygen detachment and release from the catalyst surface. Nevertheless, surface O 2 over- and under-binding results by standard generalized gradient approximation and hybrid density functionals, respectively, call for further analysis of this crucial step via multireference methods. Aiming at unveiling the nature of such a barrier, we hereby address the formation of O 2 from the most-stable IrO 2 (110) surface with both periodic DFT and an electrostatic embedded cluster approach at the n-electron valence-state perturbation theory. With this multireference approach, we find a value for the aforementioned energy barrier that is much closer to experimental indications than DFT ones. Here, an in-depth analysis of the involved molecular orbitals suggests that the origin of this barrier is related to the breaking of a π interaction between O 2 and Ir surface atom and to a significant additional O 2 interaction with adjacent electrophilic Ir-oxo species, which is present under experimental operating conditions. Besides shedding light on the mechanism of the OER on IrO 2 , these findings point out the importance of multireference methods for dissecting complex reactions at electrocatalytic interfaces and pave the route for further investigations with effective embedding approaches.
37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
An operator calculus for surface and volume modeling
The mathematical techniques which form the foundation for most of the surface and volume modeling techniques used in practice are briefly described. An outline of what may be termed an operator calculus for the approximation and interpolation of functions of more than one independent variable is presented. By considering the linear operators associated with bivariate and multivariate interpolation/approximation schemes, it is shown how they can be compounded by operator multiplication and Boolean addition to obtain a distributive lattice of approximation operators. It is then demonstrated via specific examples how this operator calculus leads to practical techniques for sculptured surface and volume modeling.
Gordon, W. J.↗
Saturn guidance concepts
Four approximations to closed loop steering function in application to spacecraft guidance
APPROXIMATION METHOD↗
Data-driven predictive probability density function control of fiber length stochastic distribution shaping in refining process
Pulp is the most important raw material in paper in paper industries, whose Fiber Length Stochastic Distribution (FLSD) shaping directly determines the energy consumption and paper quality of the subsequent papermaking processes. However, the mean and variance are insufficient to describe the output FLSD shaping, which display non-Gaussian distribution properties. Therefore, the traditional control, method based on the mean and variance of fiber length is difficult to control the FLSD shaping effectively. In this paper, a novel data-driven predictive probability density function (PDF) control method is proposed for output FLSD shaping in refining process. Primarily, in order to improve the approximation accuracy of the PDF employing RBF neural network, the parameters (center value and width) of each RBF basis functions are tuned via utilizing iterative learning control (ILC) law, and the corresponding estimations of weights law can be obtained. Secondly, considering that the conventional linear model of weights vector has drawbacks of low accuracy and weak generalization ability, random vector functional link networks (RVFLNs) based data driven nonlinear modelling method is employed to characterize the prediction model between the input variables and weights vector. Finally, in order to reduce the randomness of the output FLSD, a minimum entropy control method under mean constraint is employed to design the predictive PDF controller for the output FLSD shaping with the help of the established nonlinear stochastic distribution model. Both simulations and experiments are given to illustrate the effectiveness and practicability of the proposed method.
Probability debsity function, Fiber length distrib↗
Investigation of Ethane Dehydrogenation and Hydrogenolysis on Pt(111), Pt(211), and Pt(100): Bayesian Quantification and Correction of DFT-Based Enthalpic and Entropic Uncertainties
Computational investigations of heterogeneously catalyzed reactions using density functional theory (DFT) are often inaccurate, largely due to uncertainties in the choice of DFT functional (enthalpic uncertainty) and approximations for modeling adsorbate movement along the catalyst surface (entropic uncertainty). This work illustrates that both uncertainties are significant in the investigation of ethane dehydrogenation (EDH) and hydrogenolysis on Pt catalysts by considering the complete deconstruction of ethane on Pt(111), Pt(211), and Pt(100) using microkinetic modeling (MKM). Hence, this work uses both noncalibrated and Bayesian-calibrated MKMs to quantify and correct inaccuracies in macroscopic properties due to both uncertainties. A Bayesian approach to the correction of entropic errors was introduced using a “Modified Fermi Function (MFF)” to calibrate between the two bounds of entropy represented by the harmonic oscillator (HO) and free translator (FT) approximations. Regardless of enthalpic and entropic uncertainties, all three surfaces are capable of ethane activation; however, Pt(211) was found to be the most active and is largely responsible for methane production. Next, Pt(111) is largely responsible for acetylene production, and Pt(100) has the highest ethylene selectivity but is most susceptible to coking. By comparison of different calibrated models, the FT entropy approximation was found to better describe EDH under typical experimental conditions. Statistical evidence was found to support Pt(111) as the active site for EDH, assuming that one single site is responsible for the chemistry. On the three surfaces, competing second dehydrogenations to CH 2 CH 2 and CH 3 CH were observed as well as isomerization of CH 3 CH back to CH 2 CH 2 and deeper dehydrogenation of CH 3 CH. In conclusion, C–C cleavage was found to largely proceed via the CH 3 C intermediate on Pt(100) and Pt(111), while on Pt(211), it was via both CHC and CH 3 C.
Bayesian model selection↗
Comparing first-principles density functionals plus corrections for the lattice dynamics of YBa 2 Cu 3 O 6
The enigmatic mechanism underlying unconventional high-temperature superconductivity, especially the role of lattice dynamics, has remained a subject of debate. Theoretical insights have long been hindered due to the lack of an accurate first-principles description of the lattice dynamics of cuprates. Recently, using the r2SCAN meta-generalized gradient approximation (meta-GGA) functional, we have been able to achieve accurate phonon spectra of an insulating cuprate YBa 2 Cu 3 O 6 and discover significant magnetoelastic coupling in experimentally interesting Cu–O bond stretching optical modes [Ning et al., Phys. Rev. B 107, 045126 (2023)]. We extend this work by comparing Perdew–Burke–Ernzerhof and r2SCAN performances with corrections from the on-site Hubbard U and the D4 van der Waals (vdW) methods, aiming at further understanding on both the materials science side and the density functional side. We demonstrate the importance of vdW and self-interaction corrections for accurate first-principles YBa 2 Cu 3 O 6 lattice dynamics. Since r2SCAN by itself partially accounts for these effects, the good performance of r2SCAN is now more fully explained. Finally, in addition, the performances of the Tao–Mo series of meta-GGAs, which are constructed in a different way from the strongly constrained and appropriately normed (SCAN) meta-GGA and its revised version r2SCAN, are also compared and discussed.
37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
Finite state modelling of unsteady aerodynamics and its application to a rotor dynamic problem
The paper presents a method for formulating finite state unsteady aerodynamic models in the time domain from frequency domain unsteady aerodynamics. The method is based on recognizing that the lift deficiency function represents an aerodynamic transfer function and utilizes the Bode plot technique, used in control systems engineering, to construct approximation to the lift deficiency function. Indicial response functions for both fixed wing and rotary wing applications are obtained, using these finite state unsteady aerodynamic models. It is shown that the rotary wing indicial response function is oscillatory and thus it is fundamentally different when compared to the fixed wing indicial response function which is nonoscillatory. Certain aspects of the finite state aerodynamic model are demonstrated by applying it to the flapping dynamics of an articulated helicopter rotor blade. The influence of unsteady aerodynamics on the damping characteristics of the rotor is examined. The same problem is also treated by using a different unsteady aerodynamic model, namely dynamic inflow. Based on a comparison of the results obtained with these two unsteady aerodynamic models, useful conclusions are drawn regarding some fundamental features of these theories.
Friedmann, P. P.↗
Train Like a (Var)Pro: Efficient Training of Neural Networks with Variable Projection
Deep neural networks (DNNs) have achieved state-of-the-art performance across a variety of traditional machine learning tasks, e.g., speech recognition, image classification, and segmentation. The ability of DNNs to efficiently approximate high-dimensional functions has also motivated their use in scientific applications, e.g., to solve partial differential equations and to generate surrogate models. In this paper, we consider the supervised training of DNNs, which arises in many of the above applications. We focus on the central problem of optimizing the weights of the given DNN such that it accurately approximates the relation between observed input and target data. Devising effective solvers for this optimization problem is notoriously challenging due to the large number of weights, nonconvexity, data sparsity, and nontrivial choice of hyperparameters. To solve the optimization problem more efficiently, we propose the use of variable projection (VarPro), a method originally designed for separable nonlinear least-squares problems. Our main contribution is the Gauss--Newton VarPro method (GNvpro) that extends the reach of the VarPro idea to nonquadratic objective functions, most notably cross-entropy loss functions arising in classification. These extensions make GNvpro applicable to all training problems that involve a DNN whose last layer is an affine mapping, which is common in many state-of-the-art architectures. In our four numerical experiments from surrogate modeling, segmentation, and classification, GNvpro solves the optimization problem more efficiently than commonly used stochastic gradient descent (SGD) schemes. Finally, GNvpro finds solutions that generalize well, and in all but one example better than well-tuned SGD methods, to unseen data points.
97 MATHEMATICS AND COMPUTING↗