Search NASA⌕ Search

SEARCH · Search NASA

Results for “Kernel methods”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
DOE OSTI2016

Identification of the Diffusion Parameter in Nonlocal Steady Diffusion Problems

The problem of identifying the diffusion parameter appearing in a nonlocal steady diffusion equation is considered. The identification problem is formulated as an optimal control problem having a matching functional as the objective of the control and the parameter function as the control variable. The analysis makes use of a nonlocal vector calculus that allows one to define a variational formulation of the nonlocal problem. In a manner analogous to the local partial differential equations counterpart, we demonstrate, for certain kernel functions, the existence of at least one optimal solution in the space of admissible parameters. We introduce a Galerkin finite element discretization of the optimal control problem and derive a priori error estimates for the approximate state and control variables. Using one-dimensional numerical experiments, we illustrate the theoretical results and show that by using nonlocal models it is possible to estimate non-smooth and discontinuous diffusion parameters.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2020

Method for measurement of TRISO kernel and layer volumes by X-ray computed tomography

Layer dimensions are key parameters for as-fabricated tristructural-isotropic (TRISO) particle fuel as well as for post-irradiation examination of particle performance. Layer thicknesses are typically measured by optical microscopy of the particle cross section near mid-plane, while layer volumes are estimated with serial sectioning and microscopy. This method for measuring layer volumes is limited due to the resolution limit imposed by slice thickness and the effect of the polishing process on delicate irradiated particle microstructures. In this study, image processing software has been developed to segregate the three-dimensional (3D) TRISO particle images provided by x-ray computed tomography (XCT) into high-resolution volumetric data for the kernel, all particle layers, and any internal voids introduced during irradiation. These data can be used to analyze key in-reactor behaviors of TRISO particles such as kernel swelling or buffer shrinkage, which are important inputs for fuel performance modeling.

36 MATERIALS SCIENCE↗
DOE OSTI2011

OPTIMAL LINEAR IMAGE COMBINATION

A simple, yet general, formalism for the optimized linear combination of astrophysical images is constructed and demonstrated. The formalism allows the user to combine multiple undersampled images to provide oversampled output at high precision. The proposed method is general and may be used for any configuration of input pixels and point spread function; it also provides the noise covariance in the output image along with a powerful metric for describing undesired distortion to the image convolution kernel. The method explicitly provides knowledge and control of the inevitable compromise between noise and fidelity in the output image. We present a first prototype implementation of the method, outlining the steps taken to generate an efficient algorithm. This implementation is then put to practical use in reconstructing fully sampled output images using simulated, undersampled input exposures that are designed to mimic the proposed Wide-field InfraRed Survey Telescope (WFIRST). We examine results using randomly rotated and dithered input images, while also assessing better-known 'ideal' dither patterns: comparing results, we illustrate the use of the method as a survey design tool. Finally, we use the method to test the robustness of linear image combination when subjected to practical realities such as missing input pixels and focal plane plate scale variations.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2023

An $\bar{F}$ Meshfree Treatment for Nearly Incompressible Materials

The computational modeling of nearly incompressible materials is a difficult task for many numerical methods, and even after several decades of investigation, it is still an active research area. This report seeks to address the treatment of incompressible materials in meshfree methods using a synergistic combination of two treatments. The first treatment is an $\bar{F}$ method, where the decomposed dilatational and deviatoric parts are calculated over different smoothing domains. The second treatment “activates” additional nodes throughout the domain to increase the flexibility of the model. We implement this synergistic combination in the context of the reproducing kernel particle method (RKPM) and present results for the Cook’s membrane benchmark problem. The results are compared with those using the composite tet10 finite element with a volume-averaged J formulation. We show that the combined treatment is an effective way to deal with nearly incompressible materials in a meshfree framework and compares well with other highly-effective treatments.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2015

Advanced Nodal P 3 /SP 3 Axial Transport Solvers for the MPACT 2D/1D Scheme

As part of its initiative to provide multiphysics simulations of nuclear reactor cores, the Consortium for Advanced Simulation of Light Water Reactors (CASL) is developing the Virtual Environment for Reactor Applications Core Simulator (VERA-CS). The MPACT code, which is the primary neutron transport solver of VERA-CS, employs the two-dimensional/one-dimensional (2D/1D) method to solve 3-dimensional neutron transport problems and provide sub-pin-level resolution of the power distribution. While 2D method of characteristics is used to solve for the transport effects within each plane, 1D-nodal methods are used axially. There have been extensive studies of the 2D/1D method with a variety nodal methods, and the P 3 /SP 3 solver has proved to be an effective method of providing higher-fidelity solutions while maintaining a low computational burden.The current implementation in MPACT wraps a one-node nodal expansion method (NEM) kernel for each moment, iterating between them and performing multiple sweeps to resolve flux distributions. However, it has been observed that this approach is more sensitive to convergence problems. This paper documents the theory and application two new nodal P 3 /SP 3 approaches to be used within the 2D/1D method in MPACT. These two approaches aim to provide enhanced stability compared with the pre-existing one-node approach. Results from the HY-NEM-SP 3 solver show that the accuracy is consistent with the one-node formulations and provides improved convergence for some problems; but the solver has issues with cases in thin planes. Although the 2N-SENM-SP 3 solver is still under development, it is intended to resolve the issues with HY-NEM-SP 3 but it will incur some additional computational burden by necessitating an additional 1D-CMFD-P 3 solver to generate the second moment cell-averaged scalar flux.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2016

High-performance implementation of Chebyshev filter diagonalization for interior eigenvalue computations

We study Chebyshev filter diagonalization as a tool for the computation of many interior eigenvalues of very large sparse symmetric matrices. In this technique the subspace projection onto the target space of wanted eigenvectors is approximated with filter polynomials obtained from Chebyshev expansions of window functions. After the discussion of the conceptual foundations of Chebyshev filter diagonalization we analyze the impact of the choice of the damping kernel, search space size, and filter polynomial degree on the computational accuracy and effort, before we describe the necessary steps towards a parallel high-performance implementation. Because Chebyshev filter diagonalization avoids the need for matrix inversion it can deal with matrices and problem sizes that are presently not accessible with rational function methods based on direct or iterative linear solvers. To demonstrate the potential of Chebyshev filter diagonalization for large-scale problems of this kind we include as an example the computation of the 10{sup 2} innermost eigenpairs of a topological insulator matrix with dimension 10{sup 9} derived from quantum physics applications.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2018

A meshless Galerkin method for non-local diffusion using localized kernel bases

Here, we introduce a meshless method for solving both continuous and discrete variational formulations of a volume constrained, nonlocal diffusion problem. We use the discrete solution to approximate the continuous solution. Our method is nonconforming and uses a localized Lagrange basis that is constructed out of radial basis functions. By verifying that certain inf-sup conditions hold, we demonstrate that both the continuous and discrete problems are well-posed, and also present numerical and theoretical results for the convergence behavior of the method. The stiffness matrix is assembled by a special quadrature routine unique to the localized basis. Combining the quadrature method with the localized basis produces a well-conditioned, symmetric matrix. This then is used to find the discretized solution.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Localization dynamics in a centrally coupled system

In systems in which interactions couple a central degree of freedom and a bath, one would expect signatures of the bath's phase to be reflected in the dynamics of the central degree of freedom. This has been recently explored in connection with many-body localized baths coupled with a central qubit or a single-cavity mode - systems with growing experimental relevance in various platforms. Such models also have an interesting connection with Floquet many-body localization via quantizing the external drive, although this has been relatively unexplored. Here we adapt the multilayer multiconfigurational time-dependent Hartree (ML-MCTDH) method, a well-known tree tensor network algorithm, to numerically simulate the dynamics of a central degree of freedom, represented by a d-level system (qudit), coupled to a disordered interacting one-dimensional spin bath. ML-MCTDH allows us to reach ≈10 2 lattice sites, a far larger system size than what is feasible with exact diagonalization or kernel polynomial methods. From the intermediate time dynamics, we find a well-defined thermodynamic limit for the qudit dynamics upon appropriate rescaling of the system-bath coupling. The spin system shows similar scaling collapse in the Edward-Anderson spin-glass order parameter or entanglement entropy at relatively short times. At longer timescales, we see slow growth of the entanglement, which may arise from dephasing mechanisms in the localized system or long-range interactions mediated by the central degree of freedom. Similar signs of localization are shown to appear as well with unscaled system-bath coupling.

1-dimensional spin chains↗
NASA NTRS2004

A Kernel-Free Particle-Finite Element Method for Hypervelocity Impact Simulation

An improved hybrid particle-finite element method has been developed for the simulation of hypervelocity impact problems. Unlike alternative methods, the revised formulation computes the density without reference to any kernel or interpolation functions, for either the density or the rate of dilatation. This simplifies the state space model and leads to a significant reduction in computational cost. The improved method introduces internal energy variables as generalized coordinates in a new formulation of the thermomechanical Lagrange equations. Example problems show good agreement with exact solutions in one dimension and good agreement with experimental data in a three dimensional simulation.

Park, Young-Keun↗
DOE OSTI2019

Conforming window functions for meshfree methods

Window functions provide a base for the construction of approximation functions in many meshfree methods. They control the smoothness and extent of the approximation functions and are commonly defined using Euclidean distances which helps eliminate the need for a meshed discretization, simplifying model development for some classes of problems. However, for problems with complicated geometries such as nonconvex or multi-body domains, poor solution accuracy and convergence can occur unless the extents of the window functions, and thus approximation functions, are carefully controlled, often a time consuming or intractable task. Here in this paper, we present a method to provide more control in window function design, allowing efficient and systematic handling of complex geometries. “Conforming” window functions are constructed using Bernstein–Bézier splines defined on local triangulations with constraints imposed to control smoothness. Graph distances are used in conjunction with Euclidean metrics to provide adequate information for shaping the window functions. The conforming window functions are demonstrated using the Reproducing Kernel Particle Method showing improved accuracy and convergence rates for problems with challenging geometries. Conforming window functions are also demonstrated as a means to simplify the imposition of essential boundary conditions.

, rapid design-to-analysis↗
DOE OSTI2017

Vehicle Classification and Identification Using Multi-Modal Sensing and Signal Learning

Vehicle counting, time-of-travel analysis, and other traffic studies frequently require the classification and identification of vehicles in a roadway. Unfortunately, many current technologies for identifying vehicles, such as image-based methods that use cameras and machine vision, are not appropriate for studies that require low power consumption and low cost. In addition, privacy issues are becoming a larger concern with the increasing controversy surrounding the public collection of imagery. Some monitoring solutions, then, will require modalities beyond images. In this work we evaluate a multi-modal sensing approach to vehicle classification and identification. An ensemble of sensors measures electromagnetic emanations and acoustic signatures, and the signals are evaluated with a kernel regression method that is optimized with a genetic algorithm. We show that these signals are sufficient to classify categories of vehicles and identify individual vehicles that have already been seen, all without the benefit of images. The strength of our approach is demonstrated on a new publicly available dataset

Kerekes, Ryan↗
DOE OSTI2020

Normal modes in thermal AdS via the Selberg zeta function

The heat kernel and quasinormal mode methods of computing 1-loop partition functions of spin s s fields on hyperbolic quotient spacetimes \mathbb{H}^{3}/\mathbb{Z} ℍ 3 / ℤ are related via the Selberg zeta function. We extend that analysis to thermal \text{AdS}_{2n+1} AdS 2 n + 1 backgrounds, with quotient structure \mathbb{H}^{2n+1}/\mathbb{Z} ℍ 2 n + 1 / ℤ . Specifically, we demonstrate the zeros of the Selberg function encode the normal mode frequencies of spin fields upon removal of non-square-integrable modes. With this information we construct the 1-loop partition functions for symmetric transverse traceless tensors in terms of the Selberg zeta function and find exact agreement with the heat kernel method.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2019

Kernel Density Estimators for SAMPLER Analyses [Slides]

This presentation discusses Sampler, specifically in terms of uncertainty quantification and states that for many simulation tasks nominal input parameters have complicated influence on the uncertainty in the final result. Sampler provides a convenient means to perform stochastic uncertainty quantification, brute force sampling parameter space and tallying changes in outputs, and it can be coupled with any SCALE modules. Perturbed responses can be collected and analyzed with standard statistical measures; however, two problems have been identified. The first problem is that for Monte Carlo simulations the individual response uncertainties are ignored, and the second problem is that features of distributions are easily hidden in histograms. They found that the solution is to add Kernel Density Estimator (KDE) analysis to the Sampler output. In conclusion, Sampler is a unique tool for stochastic uncertainty propagation, and it can be coupled with any module in SCALE. Current functionality is limited to histograms, and it may not adequately account for sample uncertainties and/or not adequately display underlying distribution features. Work is in progress to implement KDE in the Sampler module and no additional flagging is needed.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2023

Explaining machine-learning models for gamma-ray detection and identification

As more complex predictive models are used for gamma-ray spectral analysis, methods are needed to probe and understand their predictions and behavior. Recent work has begun to bring the latest techniques from the field of Explainable Artificial Intelligence (XAI) into the applications of gamma-ray spectroscopy, including the introduction of gradient-based methods like saliency mapping and Gradient-weighted Class Activation Mapping (Grad-CAM), and black box methods like Local Interpretable Model-agnostic Explanations (LIME) and SHapley Additive exPlanations (SHAP). In addition, new sources of synthetic radiological data are becoming available, and these new data sets present opportunities to train models using more data than ever before. In this work, we use a neural network model trained on synthetic NaI(Tl) urban search data to compare some of these explanation methods and identify modifications that need to be applied to adapt the methods to gamma-ray spectral data. We find that the black box methods LIME and SHAP are especially accurate in their results, and recommend SHAP since it requires little hyperparameter tuning. We also propose and demonstrate a technique for generating counterfactual explanations using orthogonal projections of LIME and SHAP explanations.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2018

A Data-Driven Framework for Residential Electric Vehicle Charging Load Profile Generation

Residential electric vehicle charging load profile is indispensable to achieve reliable control strategies for mitigating negative effects on power distribution system due to emerging electrified transportation. This paper introduces a data-driven framework of charging load profile generation for residential plug-in electric vehicles. Real world historical residential charging behavior data is utilized to construct empirical charging decision making model by using machine learning algorithm. A multiple channels method with kernel density estimation is proposed to construct probability density functions for estimating charging duration based on parking duration. A generation algorithm considering parking time and travel demand dependency is introduced to generate residential charging behaviors. This framework is extensible to generate various charging load profiles and simulate varied residential charging scenarios under different number of households and charging rates. This will be crucial for designing and validating residential charging control strategies.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2014

TH-E-BRE-01: A 3D Solver of Linear Boltzmann Transport Equation Based On a New Angular Discretization Method with Positivity for Photon Dose Calculation Benchmarked with Geant4

Purpose: The Linear Boltzmann Transport Equation (LBTE) solved through statistical Monte Carlo (MC) method provides the accurate dose calculation in radiotherapy. This work is to investigate the alternative way for accurately solving LBTE using deterministic numerical method due to its possible advantage in computational speed from MC. Methods: Instead of using traditional spherical harmonics to approximate angular scattering kernel, our deterministic numerical method directly computes angular scattering weights, based on a new angular discretization method that utilizes linear finite element method on the local triangulation of unit angular sphere. As a Result, our angular discretization method has the unique advantage in positivity, i.e., to maintain all scattering weights nonnegative all the time, which is physically correct. Moreover, our method is local in angular space, and therefore handles the anisotropic scattering well, such as the forward-peaking scattering. To be compatible with image-guided radiotherapy, the spatial variables are discretized on the structured grid with the standard diamond scheme. After discretization, the improved sourceiteration method is utilized for solving the linear system without saving the linear system to memory. The accuracy of our 3D solver is validated using analytic solutions and benchmarked with Geant4, a popular MC solver. Results: The differences between Geant4 solutions and our solutions were less than 1.5% for various testing cases that mimic the practical cases. More details are available in the supporting document. Conclusion: We have developed a 3D LBTE solver based on a new angular discretization method that guarantees the positivity of scattering weights for physical correctness, and it has been benchmarked with Geant4 for photon dose calculation.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2017

Coupled Bison/Thermochimica Simulation of Hydride Formation in Zircalloy Cladding

The report describes implementation of the thermodynamics models for simulation of zirconium hydride formation in LWR fuel cladding. The current models for ZrH x formation are based on empirical rate models that must calibrated to materials and regimes of interest. The thermodynamic approach enables hydride formation calculations based on fundamental thermodynamics principles and material composition. Thermodynamic formulations and models were implemented in the Idaho National Laboratory fuel performance code Bison using Material, Kernel and AuxKernel methods. The results for the Material and Kernel implementation are shown in this report. The Material method implementation naturally fits the concept of modeling evolution of material state in the material models. Given the relatively high cost of thermodynamic computation, the advantage of the AuxKernel compared to Material implementation is the ability to control the execution and computational cost. The implemented formulations are illustrated on example problems which fit the experimental data and satisfy the computational requirements of conservation equations.

36 MATERIALS SCIENCE↗