Search NASA⌕ Search

SEARCH · Search NASA

Results for “Generalized Linear 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 613 records · Page 34

The compressed state Kalman filter for nonlinear state estimation: Application to large-scale reservoir monitoring

Reservoir monitoring aims to provide snapshots of reservoir conditions and their uncertainties to assist operation management and risk analysis. These snapshots may contain millions of state variables, e.g., pressures and saturations, which can be estimated by assimilating data in real time using the Kalman filter (KF). However, the KF has a computational cost that scales quadratically with the number of unknowns, m, due to the cost of computing and storing the covariance and Jacobian matrices, along with their prod-ucts. The compressed state Kalman filter (CSKF) adapts the KF for solving large-scale monitoring problems. The CSKF uses N preselected orthogonal bases to compute an accurate rank-N approximation of the covari-ance that is close to the optimal spectral approximation given by SVD. The CSKF has a computational cost that scales linearly in m and uses an efficient matrix-free approach that propagates uncertainties using N + 1 forward model evaluations, where N << m. Here in this paper we present a generalized CSKF algorithm for nonlin-ear state estimation problems such as CO 2 monitoring. For simultaneous estimation of multiple types of state variables, the algorithm allows selecting bases that represent the variability of each state type. Through synthetic numerical experiments of CO 2 monitoring, we show that the CSKF can reproduce the Kal-man gain accurately even for large compression ratios (m/N). For a given computational cost, the CSKF uses a robust and flexible compression scheme that gives more reliable uncertainty estimates than the ensemble Kalman filter, which may display loss of ensemble variability leading to suboptimal uncertainty estimates.

97 MATHEMATICS AND COMPUTING↗

The influence of seawater carbonate chemistry, mineralogy, and diagenesis on calcium isotope variations in Lower-Middle Triassic carbonate rocks

The geological calcium cycle is linked to the geological carbon cycle through the weathering and burial of carbonate rocks. As a result, calcium (Ca) isotope ratios ( 44 Ca/ 40 Ca, expressed as δ 44/40 Ca) can help to constrain ancient carbon cycle dynamics if Ca cycle behavior can be reconstructed. But, the δ 44/40 Ca of carbonate rocks is influenced not only by the δ 44/40 Ca of seawater but also by diagenetic processes and fractionation associated with carbonate precipitation. In this study, we investigate the dominant controls on carbonate δ 44/40 Ca in Upper Permian to Middle Triassic limestones (ca. 253 to 244 Ma) from south China and Turkey. This time interval is ideal for assessing controls on Ca isotope ratios in carbonate rocks because fluctuations in seawater δ 44/40 Ca may be expected based on several large carbon isotope (δ 13 C) excursions ranging from -2 to + 8‰. Parallel negative δ 13 C and δ 44/40 Ca excursions were previously identified across the end-Permian extinction horizon. Here, we find a second negative excursion in δ 44/40 Ca of ~ 0.2‰ within Lower Triassic strata in both south China and Turkey; however, this excursion is not synchronous between regions and thus cannot be interpreted to reflect secular change in the δ 44/40 Ca of global seawater. Additionally, δ 44/40 Ca values from Turkey are consistently 0.3‰ lower than contemporaneous samples from south China, providing further support for local or regional influences. By measuring δ 44/40 Ca and Sr concentrations ([Sr]) in two stratigraphic sections located at opposite margins of the Paleo-Tethys Ocean, we can determine whether the data represent global conditions (e.g., secular variations in the δ 44/40 Ca of seawater) versus local controls (e.g., original mineralogy or diagenetic alteration). The [Sr] and δ 44/40 Ca data from this study are best described statistically by a log-linear correlation that also exists in many previously published datasets of various geological ages. By using a model of early marine diagenetic water-rock interaction, we illustrate that this general correlation can be explained by the chemical evolution of bulk carbonate sediment samples with different initial mineralogical compositions that subsequently underwent recrystallization. Although early diagenetic resetting and carbonate mineralogy strongly influence the carbonate δ 44/40 Ca values, the relationship between [Sr] and δ 44/40 Ca holds potential for reconstructing first-order secular changes in seawater δ 44/40 Ca composition.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Damping of linear waves via ionization and recombination in homogeneous plasmas

An oscillation-center model is proposed that analytically describes transformation of an arbitrary homogeneous linear wave at gradual ionization and recombination in homogeneous plasma. For the case when either of the processes dominates, general adiabatic invariants are found, from which the wave energy is derived as a function of the frequency.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

THE NONLINEAR BIASING OF THE zCOSMOS GALAXIES UP TO z {approx} 1 FROM THE 10k SAMPLE

We use the zCOSMOS galaxy overdensity field to study the biasing of galaxies in the COSMOS field. By comparing the probability distribution function of the galaxy density contrast {delta}{sub g} to the lognormal approximation of the mass density contrast {delta}, we obtain the mean biasing function b({delta}, z, R) between the galaxy and matter overdensity fields and its second moments b-hat and b-tilde. Over the redshift interval 0.4 < z < 1 the conditional mean function ({delta}{sub g}|{delta}) = b({delta}, z, R){delta} is of a characteristic shape, requiring nonlinear biasing in the most overdense and underdense regions. Taking into account the uncertainties due to cosmic variance, we do not detect any significant evolution in the ({delta}{sub g}|{delta}) function, but we do detect a significant redshift evolution in the linear biasing parameter b-hat from 1.23 {+-} 0.11 at z {approx} 0.55 to 1.62 {+-} 0.14 at z {approx} 0.75, for a luminosity-complete sample of M{sub B} < -20 - z galaxies. The b-hat parameter does not change significantly with smoothing scale between 8 and 12 h{sup -1} Mpc, but increases systematically with luminosity (at 2{sigma}-3{sigma} significance between the M{sub B} < -20.5 - z and M{sub B} < -20 - z samples). The nonlinearity parameter b-tilde/b-hat is offset from unity by at most 2%, with an uncertainty of the same order. The b-tilde / b-hat parameter does not show any significant redshift evolution, dependence on the smoothing scale or on the luminosity. By matching the linear bias of galaxies to the halo bias, we infer that the M{sub B} < -20 - z galaxies reside in dark matter halos with a characteristic mass of about (2.6 - 5.6) x 10{sup 12} M{sub sun} with a small dependence on the adopted bias-mass relation. Our detailed error analysis and comparison with previous studies lead us to conclude that cosmic variance is the main contributor to the differences in the linear bias measured from different surveys. While our results support the general picture of biased galaxy formation up to z {approx} 1, the fine-tuning of the galaxy formation models is still limited by the restrictions of the current spectroscopic surveys at these redshifts.

79 ASTRONOMY AND ASTROPHYSICS↗

Inflation with a constant rate of roll

We consider an inflationary scenario where the rate of inflaton roll defined by {sup ··}φ/H φ-dot remains constant. The rate of roll is small for slow-roll inflation, while a generic rate of roll leads to the interesting case of 'constant-roll' inflation. We find a general exact solution for the inflaton potential required for such inflaton behaviour. In this model, due to non-slow evolution of background, the would-be decaying mode of linear scalar (curvature) perturbations may not be neglected. It can even grow for some values of the model parameter, while the other mode always remains constant. However, this always occurs for unstable solutions which are not attractors for the given potential. The most interesting particular cases of constant-roll inflation remaining viable with the most recent observational data are quadratic hilltop inflation (with cutoff) and natural inflation (with an additional negative cosmological constant). In these cases even-order slow-roll parameters approach non-negligible constants while the odd ones are asymptotically vanishing in the quasi-de Sitter regime.

79 ASTRONOMY AND ASTROPHYSICS↗

Impact of the equation of state on f - and p - mode oscillations of neutron stars

We investigate the impact of the neutron-star matter equation of state on the f- and p1-mode oscillations of neutron stars obtained within the Cowling approximation and linearized general relativity. The f- and p 1 -mode oscillation frequencies, and their damping times are calculated using representative sets of Skyrme Hartree-Fock and relativistic mean-field models, all of which reproduce nuclear systematics and support 2M ⊙ neutron stars. Our study shows strong correlations between the frequencies of f- and p 1 -modes and their damping times with the pressure of β-equilibrated matter at densities equal to or slightly higher than the nuclear saturation density ρ 0 . Such correlations are found to be almost independent of the composition of the stars. The frequency of the p 1 -mode of 1.4M ⊙ star is strongly correlated with the slope of the symmetry energy L 0 and β-equilibrated pressure at density ρ 0 . Compared to GR calculations, the error in the Cowling approximation for the f-mode is about 30% for neutron stars of low mass, whereas it decreases with increasing mass. Here, the accuracy of the p 1 -mode is better than 15% for neutron stars of maximum mass, and improves for lower masses and higher number of radial nodes. I.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Effect of the size of charged spherical macroparticles on their electrostatic interaction in an equilibrium plasma

The effect of the size of two charged spherical macroparticles on their electrostatic interaction in an equilibrium plasma is analyzed within the linearized Poisson–Botzmann model. It is established that, under the interaction of two charged dielectric macroparticles in an equilibrium plasma, the forces acting on each particle turn out to be generally unequal. The forces become equal only in the case of conducting macroparticles or in the case of dielectric macroparticles of the same size and charge. They also turn out to be equal when the surface potentials of the macroparticles remain constant under the variation of interparticle distances. Formulas are proposed that allow one to calculate the interaction force with a high degree of accuracy under the condition that the radii of macroparticles are much less than the screening length, which is usually satisfied in experiments with dusty plasmas.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Evaluation of data driven low-rank matrix factorization for accelerated solutions of the Vlasov equation

Low-rank methods have shown success in accelerating simulations of a collisionless plasma described by the Vlasov equation, but still rely on computationally costly linear algebra every time step. We propose a data-driven factorization method using artificial neural networks, specifically with convolutional layer architecture, that trains on existing simulation data. At inference time, the model outputs a low-rank decomposition of the distribution field of the charged particles, and we demonstrate that this step is faster than the standard linear algebra technique. Numerical experiments show that the method achieves comparable reconstruction accuracy for interpolation tasks, generalizing to unseen test data in a manner beyond just memorizing training data; patterns in factorization also inherently followed the same numerical trend as those within algebraic methods (e.g., truncated singular-value decomposition). However, when training on the first 70% of a time-series data and testing on the remaining 30%, the method fails to meaningfully extrapolate. Despite this limiting result, the technique may have benefits for simulations in a statistical steady-state or otherwise showing temporal stability. These results suggest that while the model offers a computationally efficient alternative for datasets with temporal stability, its current formulation is best suited for interpolation rather than for predicting future states in time-evolving systems. This study thus lays the groundwork for further refinement of neural network-based approaches to low-rank matrix factorization in high-dimensional plasma simulations.

97 MATHEMATICS AND COMPUTING↗

Quantization of the Maxwell fish-eye problem and the quantum-classical correspondence

The so-called fish-eye model, originally investigated by Maxwell in geometrical optics, is studied both in the classical as well as in the quantum formulations. The best agreement between the two approaches is achieved by using a suitably constructed coherent state, which is of the SU(2) type. The perfect quantum-classical correspondence is obtained in the sense that classical rays go exactly over maxima of the corresponding quantum probability distributions. The distributions are made of linear combinations of the E=0 bound states of the considered model.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Toward particle-resolved accuracy in Euler–Lagrange simulations of multiphase flow using machine learning and pairwise interaction extended point-particle (PIEP) approximation

This study presents two different machine learning approaches for the modeling of hydrodynamic force on particles in a particle-laden multiphase flow. Results from particle-resolved direct numerical simulations (PR-DNS) of flow over a random array of stationary particles for eight combinations of particle Reynolds number (Re) and volume fraction (φ) are used in the development of the models. The first approach follows a two-step process. In the first flow prediction step, the perturbation flow due to a particle is obtained as an axisymmetric superposable wake using linear regression. In the second force prediction step, the force on a particle is evaluated in terms of the perturbation flow induced by all its neighbors using the generalized Faxén form of the force expression. In the second approach, the force data on all the particles from the PR-DNS simulations are used to develop an artificial neural network (ANN) model for direct prediction of force on a particle. Due to the unavoidable limitation on the number of fully resolved particles in the PR-DNS simulations, direct force prediction with the ANN model tends to over-fit the data and performs poorly in the prediction of test data. In contrast, due to the millions of grid points used in the PR-DNS simulations, accurate flow prediction is possible, which then allows accurate prediction of particle force. This hybridization of multiphase physics and machine learning is particularly important, since it blends the strength of each, and the resulting pairwise interaction extended point-particle model cannot be developed by either physics or machine learning alone.

42 ENGINEERING↗

Massively parallel and linear-scaling algorithm for second-order Moller–Plesset perturbation theory applied to the study of supramolecular wires

Here, we present a scalable cross-platform hybrid MPI/OpenMP/OpenACC implementation of the Divide–Expand–Consolidate (DEC) formalism with portable performance on heterogeneous HPC architectures. The Divide–Expand–Consolidate formalism is designed to reduce the steep computational scaling of conventional many-body methods employed in electronic structure theory to linear scaling, while providing a simple mechanism for controlling the error introduced by this approximation. Our massively parallel implementation of this general scheme has three levels of parallelism, being a hybrid of the loosely coupled task-based parallelization approach and the conventional MPI +X programming model, where X is either OpenMP or OpenACC. We demonstrate strong and weak scalability of this implementation on heterogeneous HPC systems, namely on the GPU-based Cray XK7 Titan supercomputer at the Oak Ridge National Laboratory. Using the “resolution of the identity second-order Moller–Plesset perturbation theory” (RI-MP2) as the physical model for simulating correlated electron motion, the linear-scaling DEC implementation is applied to 1-aza-adamantane-trione (AAT) supramolecular wires containing up to 40 monomers (2440 atoms, 6800 correlated electrons, 24 440 basis functions and 91 280 auxiliary functions). This represents the largest molecular system treated at the MP2 level of theory, demonstrating an efficient removal of the scaling wall pertinent to conventional quantum many-body methods.

97 MATHEMATICS AND COMPUTING↗

Nonlinear modelling of polymer electrolyte membrane fuel cell stack using nonlinear cancellation technique

Fuel cells are promising new energy conversion devices that are friendly to the environment. A set of control systems are required in order to operate a fuel cell based power plant system optimally. For the purpose of control system design, an accurate fuel cell stack model in describing the dynamics of the real system is needed. Currently, linear model are widely used for fuel cell stack control purposes, but it has limitations in narrow operation range. While nonlinear models lead to nonlinear control implemnetation whos more complex and hard computing. In this research, nonlinear cancellation technique will be used to transform a nonlinear model into a linear form while maintaining the nonlinear characteristics. The transformation is done by replacing the input of the original model by a certain virtual input that has nonlinear relationship with the original input. Then the equality of the two models is tested by running a series of simulation. Input variation of H2, O2 and H2O as well as disturbance input I (current load) are studied by simulation. The error of comparison between the proposed model and the original nonlinear model are less than 1 %. Thus we can conclude that nonlinear cancellation technique can be used to represent fuel cell nonlinear model in a simple linear form while maintaining the nonlinear characteristics and therefore retain the wide operation range.

30 DIRECT ENERGY CONVERSION↗

Hydrodynamic limit of Wigner-Poisson kinetic theory: Revisited

In this paper, we revisit the hydrodynamic limit of the Langmuir wave dispersion relation based on the Wigner-Poisson model in connection with that obtained directly from the original Lindhard dielectric function based on the random-phase-approximation. It is observed that the (fourth-order) expansion of the exact Lindhard dielectric constant correctly reduces to the hydrodynamic dispersion relation with an additional term of fourth-order, beside that caused by the quantum diffraction effect. It is also revealed that the generalized Lindhard dielectric theory accounts for the recently discovered Shukla-Eliasson attractive potential (SEAP). However, the expansion of the exact Lindhard static dielectric function leads to a k{sup 4} term of different magnitude than that obtained from the linearized quantum hydrodynamics model. It is shown that a correction factor of 1/9 should be included in the term arising from the quantum Bohm potential of the momentum balance equation in fluid model in order for a correct plasma dielectric response treatment. Finally, it is observed that the long-range oscillatory screening potential (Friedel oscillations) of type cos(2k{sub F}r)/r{sup 3}, which is a consequence of the divergence of the dielectric function at point k = 2k{sub F} in a quantum plasma, arises due to the finiteness of the Fermi-wavenumber and is smeared out in the limit of very high electron number-densities, typical of white dwarfs and neutron stars. In the very low electron number-density regime, typical of semiconductors and metals, where the Friedel oscillation wavelength becomes much larger compared to the interparticle distances, the SEAP appears with a much deeper potential valley. It is remarked that the fourth-order approximate Lindhard dielectric constant approaches that of the linearized quantum hydrodynamic in the limit if very high electron number-density. By evaluation of the imaginary part of the Lindhard dielectric function, it is shown that the Landau-damping region in ω-k plane increases dramatically by increase of the electron number-density.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Aspect ratio effects on limited scrape-off layer plasma turbulence

The drift-reduced Braginskii model describing turbulence in the tokamak scrape-off layer is written for a general magnetic configuration with a limiter. The equilibrium is then specified for a circular concentric magnetic geometry retaining aspect ratio effects. Simulations are then carried out with the help of the global, flux-driven fluid three-dimensional code GBS [Ricci et al., Plasma Phys. Controlled Fusion 54, 124047 (2012)]. Linearly, both simulations and simplified analytical models reveal a stabilization of ballooning modes. Nonlinearly, flux-driven nonlinear simulations give a pressure characteristic length whose trends are correctly captured by the gradient removal theory [Ricci and Rogers, Phys. Plasmas 20, 010702 (2013)], that assumes the profile flattening from the linear modes as the saturation mechanism. More specifically, the linear stabilization of ballooning modes is reflected by a 15% increase in the steady-state pressure gradient obtained from GBS nonlinear simulations when going from an infinite to a realistic aspect ratio.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Consistency check of {Lambda}CDM phenomenology

The standard model of cosmology {Lambda}CDM assumes general relativity, flat space, and the presence of a positive cosmological constant. We relax these assumptions allowing spatial curvature, a time-dependent effective dark energy equation of state, as well as modifications of the Poisson equation for the lensing potential, and modifications of the growth of linear matter density perturbations in alternate combinations. Using six parameters characterizing these relations, we check {Lambda}CDM for consistency utilizing cosmic microwave background anisotropies, cross correlations thereof with high-redshift galaxies through the integrated Sachs-Wolfe effect, the Hubble constant, supernovae, and baryon acoustic oscillation distances, as well as the relation between weak gravitational lensing and galaxy flows. In all scenarios, we find consistency of the concordance model at the 95% confidence level. However, we emphasize that constraining supplementary background parameters and parametrizations of the growth of large-scale structure separately may lead to a priori exclusion of viable departures from the concordance model.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Pyomo.GDP: an ecosystem for logic based modeling and optimization development

We present three core principles for engineering-oriented integrated modeling and optimization tool sets—intuitive modeling contexts, systematic computer-aided reformulations, and flexible solution strategies—and describe how new developments in Pyomo.GDP for Generalized Disjunctive Programming (GDP) advance this vision. We describe a new logical expression system implementation for Pyomo.GDP allowing for a more intuitive description of logical propositions. The logical expression system supports automated reformulation of these logical constraints to linear constraints. We also describe two new logic-based global optimization solver implementations built on Pyomo.GDP that exploit logical structure to avoid “zero-flow” numerical difficulties that arise in nonlinear network design problems when nodes or streams disappear. These new solvers also demonstrate the capability to link to external libraries for expanded functionality within an integrated implementation. We present these new solvers in the context of a flexible array of solution paths available to GDP models. Finally, we present results on a new library of GDP models demonstrating the value of multiple solution approaches.

42 ENGINEERING↗

Survival probabilities in the Sisyphus random walk model with absorbing traps

We analyze the dynamics of the Sisyphus random walk model, a discrete Markov chain in which the walkers may randomly return to their initial position x{sub 0}. In particular, we present a remarkably compact derivation of the time-dependent survival probability function S(t;x{sub 0}) which characterizes the random walkers in the presence of an absorbing trap at the origin. The survival probabilities are expressed in a compact mathematical form in terms of the x{sub 0}-generalized Fibonacci-like numbers G{sub t}{sup (x{sub 0})}. Interestingly, it is proved that, as opposed to the standard random walk model in which the survival probabilities depend linearly on the initial distance x{sub 0} of the walkers from the trap and decay asymptotically as an inverse power of the time, in the Sisyphus random walk model the asymptotic survival probabilities decay exponentially in time and are characterized by a non-trivial (non-linear) dependence on the initial gap x{sub 0} from the absorbing trap. We use the analytically derived results in order to analyze the underlying dynamics of the ‘survival-game’, a highly risky investment strategy in which non-absorbed agents receive a reward P(t) which gradually increases in time.

97 MATHEMATICS AND COMPUTING↗

The Role of Subtropical Irreversible PV Mixing in the Zonal Mean Circulation Response to Global Warming-like Thermal Forcing

The atmospheric circulation response to the global warming-like tropical upper tropospheric heating is revisited using a dry atmospheric general circulation model (AGCM) in light of a new diagnostics based on the concept of finite-amplitude wave activity (FAWA) on equivalent latitude. For a given tropical heating profile, the linear Wentzel-Kramers-Brillouin (WKB) wave refraction analysis sometimes gives a very different and even opposite prediction of the eddy momentum flux response to that of the actual full model simulation, exposing the limitation of the traditional linear approach in understanding the full dynamics of the atmospheric response under global warming. The implementation of the FAWA diagnostics reveals that in response to the upper tropospheric heating, effective diffusivity, a measure of the mixing efficiency, increases and advances upward and poleward in the subtropics and the resultant enhancement and the poleward encroachment of eddy potential vorticity mixing leads to a poleward displaced potential vorticity (PV) gradient peak in the upper troposphere. The anomalous eddy PV flux, in balance with the PV dissipation, gives rise to a poleward shift in the eddy-driven jet and eddy-driven mean meridional circulation. Sensitivity experiments show that these irreversible dissipation processes in the upper troposphere are robust, regardless of the width of the tropical heating.

54 ENVIRONMENTAL SCIENCES↗