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 577 records · Page 32

TNet: A Model-Constrained Tikhonov Network Approach for Inverse Problems

Deep learning (DL), in particular deep neural networks, by default is purely data-driven and in general does not require physics. This is the strength of DL but also one of its key limitations when applied to science and engineering problems in which underlying physical properties—such as stability, conservation, and positivity—and accuracy are required. DL methods in their original forms are often not capable of respecting the underlying mathematical models or achieving desired accuracy even in big-data regimes. On the other hand, many data-driven science and engineering problems, such as inverse problems, typically have limited experimental or observational data, and DL would overfit the data in this case. Leveraging information encoded in the underlying mathematical models, we argue, not only compensates for missing information in low data regimes but also provides opportunities to equip DL methods with the underlying physics, hence promoting better generalization. This paper develops a model-constrained DL approach and its variant TNet—a Tikhonov neural network—which are capable of learning not only information hidden in the training data but also in the underlying mathematical models to solve inverse problems governed by partial differential equations in low data regimes. We provide the constructions and some theoretical results for the proposed approaches for both linear and nonlinear inverse problems. Since TNet is designed to learn inverse solutions with Tikhonov regularization, it is interpretable: in fact it recovers Tikhonov solutions for linear cases while potentially approximating Tikhonov solutions for nonlinear inverse problems. We also prove that data randomization can enhance not only the smoothness of the networks but also their generalizations. Comprehensive numerical results confirm the theoretical findings and show that with even as little as 1 training data sample for one-dimensional (1D) deconvolution, 5 for an inverse 2D heat conductivity problem, 100 for inverse initial conditions for a time-dependent 2D Burgers’s equation, and 50 for inverse initial conditions for 2D Navier–Stokes equations, TNet solutions can be as accurate as Tikhonov solutions while being several orders of magnitude faster. Furthermore, this is possible owing to the model-constrained term, replications, and randomization.

97 MATHEMATICS AND COMPUTING↗

A robust hierarchical solver for ill-conditioned systems with applications to ice sheet modeling

A hierarchical solver is proposed for solving sparse ill-conditioned linear systems in parallel. The solver is based on a modification of the LoRaSp method, but employs a deferred-compression technique, which provably reduces the approximation error and significantly improves efficiency. Moreover, the deferred-compression technique introduces minimal overhead and does not affect parallelism. As a result, the new solver achieves linear computational complexity under mild assumptions and excellent parallel scalability. To demonstrate the performance of the new solver, we focus on applying it to solve sparse linear systems arising from ice sheet modeling. The strong anisotropic phenomena associated with the thin structure of ice sheets creates serious challenges for existing solvers. To address the anisotropy, we additionally developed a customized partitioning scheme for the solver, which captures the strong-coupling direction accurately. In general, the partitioning can be computed algebraically with existing software packages, and thus the new solver is generalizable for solving other sparse linear systems. Our results show that ice sheet problems of about 300 million degrees of freedom have been solved in just a few minutes using a thousand processors.

97 MATHEMATICS AND COMPUTING↗

A framework for testing soil carbon dynamics post land-use transition in a multisector dynamics model

Soil carbon plays a crucial role in the global carbon cycle. Changes in land use can determine whether carbon is stored or is emitted into the atmosphere as carbon dioxide, which has broad implications for the human and Earth systems. These feedbacks to the carbon cycle and their socio-economic drivers are modelled by many global multisector dynamics models to project future possibilities for the human-Earth system. One notable model of this class is the Global Change Analysis Model (GCAM), which uses a simplified process to model soil organic carbon (SOC) content after land-use transition across 384 land units. While the current GCAM soil carbon framework is based on scientific principles, it has not been tested against experimental data. This work examines rates of SOC change from GCAM input data. Specifically, first order rate constants derived from model inputs were compared to values from two syntheses to assess GCAM’s accuracy. Welch’s t-tests and linear models were used to determine if rate constants were consistent across all tested geographical areas and land-use transition types. While we found that there was general agreement on the direction and magnitude (i.e., rate) of SOC change, the rate constant derived from GCAM and empirical values differed strongly in a subset of specific instances. These results indicate that GCAM’s current SOC dynamics during land use transition successfully capture broad patterns of change in this critical carbon pool, but should be interpreted with caution at finer spatial scales. One potential cause of these discrepancies is our highly aggregated variable, soil timescale, which could be made more granular to improve accuracy. When using economically rooted multisector dynamics models, such as GCAM, it is critical to understand such model limitations for representing specific Earth system processes.

carbon↗

Potentials for soft-wall AdS/QCD

Soft-wall models in AdS/QCD generally have dilaton and scalar fields that vary with the fifth-dimension coordinate. These fields can be parametrized to yield hadron mass spectra with linear radial trajectories and to incorporate spontaneous breaking of chiral symmetry. We show how to construct scalar potentials which lead to such solutions.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Phase diagram of hot QCD in an external magnetic field: Possible splitting of deconfinement and chiral transitions

The structure of the phase diagram for strong interactions becomes richer in the presence of a magnetic background, which enters as a new control parameter for the thermodynamics. Motivated by the relevance of this physical setting for current and future high-energy heavy-ion collision experiments and for the cosmological QCD transitions, we use the linear sigma model coupled to quarks and to Polyakov loops as an effective theory to investigate how the chiral and the deconfining transitions are affected, and present a general picture for the temperature-magnetic field phase diagram. We compute and discuss each contribution to the effective potential for the approximate order parameters, and uncover new phenomena such as the paramagnetically induced breaking of global Z{sub 3} symmetry, and possible splitting of deconfinement and chiral transitions in a strong magnetic field.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Evaluating Lyman- α constraints for general dark-matter velocity distributions: Multiple scales and cautionary tales

One of the most important observational constraints on possible models of dark-matter physics exploits the Lyman-α absorption spectrum associated with photons traversing the intergalactic medium. Because these data allow us to probe the linear matter power spectrum with great accuracy down to relatively small distance scales, finding ways of accurately evaluating such Lyman-α constraints across large classes of candidate models of dark-matter physics is of paramount importance. While such Lyman-α constraints have been evaluated for dark-matter models that give rise to relatively simple dark-matter velocity distributions, more complex models—particularly those whose dark-matter velocity distributions stretch across multiple scales—have recently been receiving increasing attention. In this paper, we undertake a study of the Lyman-α constraints associated with general dark-matter velocity distributions. Although such Lyman-α constraints are difficult to evaluate in principle, in practice there currently exist two classes of methods in the literature through which such constraints can be recast into forms which are easier to evaluate and which therefore allow a more rapid determination of whether a given dark-matter model is ruled in or out. Accordingly, we utilize both of these recasts in order to determine the Lyman-α bounds on different classes of dark-matter velocity distributions. Here we also develop a general method by which the results of these different recasts can be compared. For relatively simple dark-matter velocity distributions, we demonstrate that these two classes of recasts tend to align and give similar results. However, we find that the situation is far more complex for distributions involving multiple velocity scales: while these two classes of recasts continue to yield similar results within certain regions of parameter space, they nevertheless yield dramatically different results within precisely those regions of parameter space which are likely to be phenomenologically relevant. This, then, serves as a cautionary tale regarding the use of such recasts for complex dark-matter velocity distributions.

79 ASTRONOMY AND ASTROPHYSICS↗

Dedicated beam position monitor pair for model-independent lattice characterization at NSLS-II

This paper reports recent lattice characterization results obtained at the National Synchrotron Light Source II (NSLS-II) storage ring, conducted without reliance on a lattice model. A pair of beam position monitors (BPMs) with bunch-by-bunch (B×B) resolution, were recently installed in a section of the storage ring free of magnetic fields. The new BPM pair measured the beam, or bunch’s transverse Poincar´e map precisely after the beam was excited. Linear one-turn-matrices (OTM) were then derived, and from these, the 4-dimensional coupled Twiss parameters were extracted at the locations of the BPM pair. By normalizing beam oscillation amplitudes with the Twiss parameters, the global action-variables were obtained. These action-variables facilitated the measurement of the local Twiss parameters observed by other BPMs independent on lattice model. This method is general, and particularly useful in certain scenarios such as a round beam mode in a diffraction-limited light source ring. We applied it to assess both weakly and strongly coupled lattices at the NSLS-II ring. Through analysis of the strongly coupled lattice, the quadrupole tilt errors were estimated to be less than 400 µrad. Utilizing the BPMs’ B×B resolution, for the first time we observed the variations of the linear lattice along a long bunch-train.

43 PARTICLE ACCELERATORS↗

Towards optimal extraction of cosmological information from nonlinear data

One of the main unsolved problems of cosmology is how to maximize the extraction of information from nonlinear data. If the data are nonlinear the usual approach is to employ a sequence of statistics (N-point statistics, counting statistics of clusters, density peaks or voids etc.), along with the corresponding covariance matrices. However, this approach is computationally prohibitive and has not been shown to be exhaustive in terms of information content. Here we instead develop a hierarchical Bayesian approach, expanding the likelihood around the maximum posterior of linear modes, which we solve for using optimization methods. By integrating out the modes using perturbative expansion of the likelihood we construct an initial power spectrum estimator, which for a fixed forward model contains all the cosmological information if the initial modes are gaussian distributed. We develop a method to construct the window and covariance matrix such that the estimator is explicitly unbiased and nearly optimal. We then generalize the method to include the forward model parameters, including cosmological and nuisance parameters, and primordial non-gaussianity. We apply the method in the simplified context of nonlinear structure formation, using either simplified 2-LPT dynamics or N-body simulations as the nonlinear mapping between linear and nonlinear density, and 2-LPT dynamics in the optimization steps used to reconstruct the initial density modes. We demonstrate that the method gives an unbiased estimator of the initial power spectrum, providing among other a near optimal reconstruction of linear baryonic acoustic oscillations.

79 ASTRONOMY AND ASTROPHYSICS↗

Mitigation of two-plasmon decay in direct-drive inertial confinement fusion through the manipulation of ion-acoustic and Langmuir wave damping

The extended Zakharov model of the two-plasmon decay instability in an inhomogeneous plasma [D. F. DuBois et al., Phys. Rev. Lett. 74, 3983 (1995); D. A. Russell and D. F. DuBois, Phys. Rev. Lett. 86, 428 (2001)] is further generalized to include the evolution of the electron distribution function in the quasi-linear approximation [cf., e.g., K. Y. Sanbonmatsu et al. Phys. Plasmas 7, 2824 (2000); D. A. Russell et al., paper presented at the Workshop on SRS/SBS Saturation, Wente Vineyards, Livermore, CA, 2–5 April 2002]. This makes it possible to investigate anomalous absorption of laser light and hot electron production due to the two-plasmon decay instability of multiple overlapping electromagnetic waves. Scalings of hot-electron production in the (stationary) nonlinearly saturated regime relevant to recent experiments [B. Yaakobi et al., Phys. Plasmas 19, 012704 (2012); D. H. Froula et al., Phys. Rev. Lett. 108, 165003 (2012)] have been obtained. They indicate a sensitivity to ion-acoustic wave (IAW) damping and to the collisional absorption of Langmuir waves. Such a sensitivity might be exploited in inertial confinement fusion target design by the use of mid-Z ablators.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Instantaneous spatially local projective measurements are consistent in a relativistic quantum field

Suppose the postulate of measurement in quantum mechanics can be extended to quantum field theory; then a local projective measurement at some moment on an object locally coupled with a relativistic quantum field will result in a projection or collapse of the wavefunctional of the combined system defined on the whole time-slice associated with the very moment of the measurement, if the relevant degrees of freedom have nonzero correlations. This implies that the wavefunctionals in the same Hamiltonian system but defined in different reference frames would collapse on different time-slices passing through the same local event where the measurement was done. Are these post-measurement states consistent with each other? We illustrate that the quantum states of the Raine-Sciama-Grove detector-field system started with the same initial Gaussian state defined on the same initial time-slice, then collapsed by the measurements on the pointlike detectors on different time-slices in different frames, will evolve to the same state of the combined system up to a coordinate transformation when compared on the same final time-slice. Such consistency is guaranteed by the spatial locality of interactions and the general covariance in a relativistic system, together with the spatial locality of measurements and the linearity of quantum dynamics in its quantum theory. - Highlights: Black-Right-Pointing-Pointer Spatially local quantum measurements in detector-field models are studied. Black-Right-Pointing-Pointer Local quantum measurement collapses the wavefunctional on the whole time-slice. Black-Right-Pointing-Pointer In different frames wavefunctionals of a field would collapse on different time-slices. Black-Right-Pointing-Pointer States collapsed by the same measurement will be consistent on the same final slice.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Unified first law in inhomogeneous and anisotropic spacetime models: Some general prescription

In this paper we investigate the interrelationship between the unified first law of thermodynamics and the Friedmann equations for three anisotropic and inhomogeneous spacetime metrics namely Lemaître–Tolman–Bondi and Friedmann–Lemaître–Robertson–Walker type inhomogeneous spacetime model and Kantowski–Sachs type model. In all these models we find that Einstein field equations are obtained by projecting the unified first law along and perpendicular to the Kodama vector and also along two linearly independent directions.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Analytic marginalization of N(z) uncertainties in tomographic galaxy surveys

In this paper, we present a new method to marginalize over uncertainties in redshift distributions, N(z), within tomographic cosmological analyses applicable to current and upcoming photometric galaxy surveys. We allow for arbitrary deviations from the best-guess N(z) governed by a general covariance matrix describing the uncertainty in our knowledge of redshift distributions. In principle, this is marginalization over hundreds or thousands of new parameters describing potential deviations as a function of redshift and tomographic bin. However, by linearly expanding the theory predictions around a fiducial model, this marginalization can be performed analytically, resulting in a modified data covariance matrix that effectively downweights the modes of the data vector that are more sensitive to redshift distribution variations. We showcase this method by applying it to the galaxy clustering measurements from the Hyper Suprime-Cam first data release. We illustrate how to marginalize over sample variance of the calibration sample and a large general systematic uncertainty in photometric estimation methods, and explore the impact of priors imposing smoothness in the redshift distributions.

79 ASTRONOMY AND ASTROPHYSICS↗

Modeling direct band-to-band tunneling: From bulk to quantum-confined semiconductor devices

A rigorous framework to study direct band-to-band tunneling (BTBT) in homo- and hetero-junction semiconductor nanodevices is introduced. An interaction Hamiltonian coupling conduction and valence bands (CVBs) is derived using a multiband envelope method. A general form of the BTBT probability is then obtained from the linear response to the “CVBs interaction” that drives the system out of equilibrium. Simple expressions in terms of the one-electron spectral function are developed to compute the BTBT current in two- and three-dimensional semiconductor structures. Additionally, a two-band envelope equation based on the Flietner model of imaginary dispersion is proposed for the same purpose. In order to characterize their accuracy and differences, both approaches are compared with full-band, atomistic quantum transport simulations of Ge, InAs, and InAs-Si Esaki diodes. As another numerical application, the BTBT current in InAs-Si nanowire tunnel field-effect transistors is computed. It is found that both approaches agree with high accuracy. The first one is considerably easier to conceive and could be implemented straightforwardly in existing quantum transport tools based on the effective mass approximation to account for BTBT in nanodevices.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Fouling modeling and prediction approach for heat exchangers using deep learning

In this article, we develop a generalized and scalable statistical model for accurate prediction of fouling resistance using commonly measured parameters of industrial heat exchangers. This prediction model is based on deep learning where a scalable algorithmic architecture learns non-linear functional relationships between a set of target and predictor variables from large number of training samples. Here, the efficacy of this modeling approach is demonstrated for predicting fouling in an analytically modeled cross-flow heat exchanger, designed for waste heat recovery from flue-gas using room temperature water. The performance results of the trained models demonstrate that the mean absolute prediction errors are under 10 –4 KW –1 for flue-gas side, water side and overall fouling resistances. The coefficients of determination (R 2 ), which characterize the goodness of fit between the predictions and observed data, are over 99%. Even under varying levels of measurement noise in the inputs, we demonstrate that predictions over an ensemble of multiple neural networks achieves better accuracy and robustness to noise. We find that the proposed deep-learning fouling prediction framework learns to follow heat exchanger flow and heat transfer physics, which we confirm using locally interpretable model agnostic explanations around randomly selected operating points. Overall, we provide a robust algorithmic framework for fouling prediction that can be generalized and scaled to various types of industrial heat exchangers.

42 ENGINEERING↗

Modeling non-steady state radioisotope transport in the vadose zone--A case study using uranium isotopes at Pena Blanca, Mexico

Current models using U- and Th-series disequilibria to study radioisotope transport in groundwater systems mostly consider a steady-state situation. These models have limited applicability to the vadose zone (UZ) where the concentration and migratory behavior of radioisotopes in fluid are often transitory. We present here, as a first attempt of its kind, a model simulating the non-steady state, intermittent fluid transport in vadose layers. It provides quantitative constraints on in-situ migration of dissolved and colloidal radioisotopes in terms of retardation factor and rock-water interaction (or water transit) time. For uranium, the simulation predicts that intermittent flushing in the UZ leads to a linear relationship between reciprocal U concentration and {sup 234}U/{sup 238}U ratio in percolating waters, with the intercept and slope bearing information on the rates of dissolution and {alpha}-recoil of U isotopes, respectively. The general validity of the model appears to be borne out by the measurement of uranium isotopes in UZ waters collected at various times over a period during 1995-2006 from a site in the Pena Blanca mining district, Mexico, where the Nopal I uranium deposit is located. Enhanced {sup 234}U/{sup 238}U ratios in vadose-zone waters resulting from lengthened non-flushing time as prescribed by the model provide an interpretative basis for using {sup 234}U/{sup 238}U in cave calcites to reconstruct the regional changes in hydrology and climate. We also provide a theoretical account of the model's potential applications using radium isotopes.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Nonlinear Time Domain Seismic Soil-Structure Interaction (SSI) Deep Soil Site Methodology Development

Currently the Department of Energy (DOE) and the nuclear industry perform seismic soil-structure interaction (SSI) analysis using equivalent linear numerical analysis tools. For lower levels of ground motion, these tools should produce reasonable in-structure response values for evaluation of existing and new facilities. For larger levels of ground motion these tools likely overestimate the in-structure response (and therefore structural demand) since they do not consider geometric nonlinearities (such as gaping and sliding between the soil and structure) and are limited in the ability to model nonlinear soil behavior. The current equivalent linear SSI (SASSI) analysis approach either joins the soil and structure together in both tension and compression or releases the soil from the structure for both tension and compression. It also makes linear approximations for material nonlinearities and generalizes energy absorption with viscous damping. This produces the potential for inaccurately establishing where the structural concerns exist and/or inaccurately establishing the amplitude of the in-structure responses. Seismic hazard curves at nuclear facilities have continued to increase over the years as more information has been developed on seismic sources (i.e. faults), additional information gathered on seismic events, and additional research performed to determine local site effects. Seismic hazard curves are used to develop design basis earthquakes (DBE) that are used to evaluate nuclear facility response. As the seismic hazard curves increase, the input ground motions (DBE’s) used to numerically evaluation nuclear facility response increase causing larger in-structure response. As ground motions increase so does the importance of including nonlinear effects in numerical SSI models. To include material nonlinearity in the soil and geometric nonlinearity using contact (gaping and sliding) it is necessary to develop a nonlinear time domain methodology. This methodology will be known as, NonLinear Soil-Structure Interaction (NLSSI). In general NLSSI analysis should provide a more accurate representation of the seismic demands on nuclear facilities their systems and components. INL, in collaboration with a Nuclear Power Plant Vender (NPP-V), will develop a generic Nuclear Power Plant (NPP) structural design to be used in development of the methodology and for comparison with SASSI. This generic NPP design has been evaluated for the INL soil site because of the ease of access and quality of the site specific data. It is now being evaluated for a second site at Vogtle which is located approximately 15 miles East-Northeast of Waynesboro, Georgia and adjacent to Savanna River. The Vogtle site consists of many soil layers spanning down to a depth of 1058 feet. The reason that two soil sites are chosen is to demonstrate the methodology across multiple soil sites. The project will drive the models (soil and structure) using successively increasing acceleration time histories with amplitudes. The models will be run in time domain codes such as ABAQUS, LS-DYNA, and/or ESSI and compared with the same models run in SASSI. The project is focused on developing and documenting a method for performing time domain, non-linear seismic soil structure interaction (SSI) analysis. Development of this method will provide the Department of Energy (DOE) and industry with another tool to perform seismic SSI analysis.

58 GEOSCIENCES↗

An efficient numerical method for computing dynamics of spin F = 2 Bose-Einstein condensates

In this paper, we extend the efficient time-splitting Fourier pseudospectral method to solve the generalized Gross-Pitaevskii (GP) equations, which model the dynamics of spin F = 2 Bose-Einstein condensates at extremely low temperature. Using the time-splitting technique, we split the generalized GP equations into one linear part and two nonlinear parts: the linear part is solved with the Fourier pseudospectral method; one of nonlinear parts is solved analytically while the other one is reformulated into a matrix formulation and solved by diagonalization. We show that the method keeps well the conservation laws related to generalized GP equations in 1D and 2D. We also show that the method is of second-order in time and spectrally accurate in space through a one-dimensional numerical test. We apply the method to investigate the dynamics of spin F = 2 Bose-Einstein condensates confined in a uniform/nonuniform magnetic field.

97 MATHEMATICS AND COMPUTING↗

Theory on excitations of drift Alfvén waves by energetic particles. II. The general fishbone-like dispersion relation

The theoretical framework of the general fishbone-like dispersion relation (GFLDR), presented and discussed in the Companion Paper [Phys. Plasmas 21, 072120 (2014)], is applied to cases of practical interest of shear/drift Alfvén waves (SAWs/DAWs) excited by energetic particles (EPs) in toroidal fusion plasmas. These applications demonstrate that the GFLDR provides a unified approach that allows analytical and numerical calculations of stability properties, as well as mode structures and, in general, nonlinear evolutions, based on different models and with different levels of approximation. They also show the crucial importance of kinetic descriptions, accurate geometries and boundary conditions for predicting linear as well as nonlinear SAW/DAW and EP behaviors in burning plasmas. Thus, the GFLDR unified theoretical framework elevates the interpretative capability for both experimental and numerical simulation results.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗