Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Kernel”

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 145 records · Page 8
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Kernel learning backward SDE filter for data assimilation

In this paper, we develop a kernel learning backward SDE filter method to estimate the state of a stochastic dynamical system based on its partial noisy observations. A system of forward backward stochastic differential equations is used to propagate the state of the target dynamical model, and Bayesian inference is applied to incorporate the observational information. Further, to characterize the dynamical model in the entire state space, we introduce a kernel learning method to learn a continuous global approximation for the conditional probability density function of the target state by using discrete approximated density values as training data. Numerical experiments demonstrate that the kernel learning backward SDE is highly effective.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2018

A kernel machine method for detecting higher order interactions in multimodal datasets: Application to schizophrenia

Technological advances are enabling us to collect multimodal datasets at an increasing depth and resolution while with decreasing labors. Understanding complex interactions among multimodal datasets, however, is challenging. Here in this study, we tested the interaction effect of multimodal datasets using a novel method called the kernel machine for detecting higher order interactions among biologically relevant multimodal data. Using a semiparametric method on a reproducing kernel Hilbert space, we formulated the proposed method as a standard mixed-effects linear model and derived a score-based variance component statistic to test higher order interactions between multimodal datasets. The method was evaluated using extensive numerical simulation and real data from the Mind Clinical Imaging Consortium with both schizophrenia patients and healthy controls. Our method identified 13-triplets that included 6 gene-derived SNPs, 10 ROIs, and 6 gene-specific DNA methylations that are correlated with the changes in hippocampal volume, suggesting that these triplets may be important for explaining schizophrenia-related neurodegeneration. The performance of the proposed method is compared with the following methods: test based on only first and first few principal components followed by multiple regression, and full principal component analysis regression, and the sequence kernel association test. With strong evidence (p-value ≤0.000001), the triplet (MAGI2, CRBLCrus1.L, FBXO28) is a significant biomarker for schizophrenia patients. This novel method can be applicable to the study of other disease processes, where multimodal data analysis is a common task.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2014

Fission product release and survivability of UN-kernel LWR TRISO fuel

A thermomechanical assessment of the LWR application of TRISO fuel with UN kernels was performed. Fission product release under operational and transient temperature conditions was determined by extrapolation from fission product recoil calculations and limited data from irradiated UN pellets. Both fission recoil and diffusive release were considered and internal particle pressures computed for both 650 and 800 um diameter kernels as a function of buffer layer thickness. These pressures were used in conjunction with a finite element program to compute the radial and tangential stresses generated within a TRISO particle undergoing burnup. Creep and swelling of the inner and outer pyrolytic carbon layers were included in the analyses. A measure of reliability of the TRISO particle was obtained by computing the probability of survival of the SiC barrier layer and the maximum tensile stress generated in the pyrolytic carbon layers from internal pressure and thermomechanics of the layers. These reliability estimates were obtained as functions of the kernel diameter, buffer layer thickness, and pyrolytic carbon layer thickness. The value of the probability of survival at the end of irradiation was inversely proportional to the maximum pressure.

99 GENERAL AND MISCELLANEOUS↗
DOE OSTI2017

Measurement of kernel swelling and buffer densification in irradiated UCO-TRISO particles

Radiation-induced volume changes in the fuel kernels and buffer layers of UCO-TRISO particles irradiated to an average burnup of 16.1% FIMA have been determined. Measurements of particle dimensions were made on polished cross-sections of 56 irradiated particles at several different polish planes. The data were then analyzed to compute the equivalent spherical diameters of the kernels and the various coating layers, and these were compared to the average as-fabricated values to determine changes due to irradiation. Finally, the kernel volume was found to have increased by an average of 26 ± 6%. Buffer volume decreased by an average of 39 ± 2% due to densification.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2015

Relativistic four-component calculations of indirect nuclear spin-spin couplings with efficient evaluation of the exchange-correlation response kernel

In this work, we report on the development and implementation of a new scheme for efficient calculation of indirect nuclear spin-spin couplings in the framework of four-component matrix Dirac-Kohn-Sham approach termed matrix Dirac-Kohn-Sham restricted magnetic balance resolution of identity for J and K, which takes advantage of the previous restricted magnetic balance formalism and the density fitting approach for the rapid evaluation of density functional theory exchange-correlation response kernels. The new approach is aimed to speedup the bottleneck in the solution of the coupled perturbed equations: evaluation of the matrix elements of the kernel of the exchange-correlation potential. The performance of the new scheme has been tested on a representative set of indirect nuclear spin-spin couplings. The obtained results have been compared with the corresponding results of the reference method with traditional evaluation of the exchange-correlation kernel, i.e., without employing the fitted electron densities. Overall good agreement between both methods was observed, though the new approach tends to give values by about 4%-5% higher than the reference method. On the average, the solution of the coupled perturbed equations with the new scheme is about 8.5 times faster compared to the reference method.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2020

A High-Throughput Solver for Marginalized Graph Kernels on GPU

Here, we present the design and optimization of a solver for efficient and high-throughput computation of the marginalized graph kernel on General Purpose GPUs. The graph kernel is computed using the conjugate gradient method to solve a generalized Laplacian of the tensor product between a pair of graphs. To cope with the large gap between the instruction throughput and the memory bandwidth of the GPUs, our solver forms the graph tensor product on-the-fly without storing it in memory. This is achieved by using threads in a warp cooperatively to stream the adjacency and edge label matrices of individual graphs by small square matrix blocks called tiles, which are then staged in registers and the shared memory for later reuse. Warps across a thread block can further share tiles via the shared memory to increase data reuse. We exploit the sparsity of the graphs hierarchically by storing only non-empty tiles using a coordinate format and nonzero elements within each tile using bitmaps. We propose a new partition-based reordering algorithm for aggregating nonzero elements of the graphs into fewer but denser tiles to further exploit sparsity. We carry out extensive theoretical analyses on the graph tensor product primitives for tiles of various density and evaluate their performance on synthetic and real-world datasets. Our solver delivers three to four orders of magnitude speedup over existing CPU-based solvers such as GraKeL and GraphKernels. The capability of the solver enables kernel-based learning tasks at unprecedented scales.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Fault Detection via Occupation Kernel Principal Component Analysis

Reliable operation of automatic systems is heavily dependent on the ability to detect faults in the underlying dynamics. While traditional model-based methods have been widely used for fault detection, data-driven approaches have garnered increasing attention due to their ease of deployment and minimal need for expert knowledge. In this letter, we present a novel principal component analysis (PCA) method that uses occupation kernels. Occupation kernels result in feature maps that are tailored to the measured data, have inherent noise-robustness due to the use of integration, and can utilize irregularly sampled system trajectories of variable lengths for PCA. The occupation kernel PCA method is used to develop a reconstruction error approach to fault detection and its efficacy is validated using numerical simulations.

42 ENGINEERING↗
DOE OSTI2015

Fast generation of sparse random kernel graphs

The development of kernel-based inhomogeneous random graphs has provided models that are flexible enough to capture many observed characteristics of real networks, and that are also mathematically tractable. We specify a class of inhomogeneous random graph models, called random kernel graphs, that produces sparse graphs with tunable graph properties, and we develop an efficient generation algorithm to sample random instances from this model. As real-world networks are usually large, it is essential that the run-time of generation algorithms scales better than quadratically in the number of vertices n. We show that for many practical kernels our algorithm runs in time at most ο(n(logn)²). As an example, we show how to generate samples of power-law degree distribution graphs with tunable assortativity.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2016

Connectivity from a different perspective: comparing seed dispersal kernels in connected vs. unfragmented landscapes

Habitat fragmentation can create significant impediments to dispersal. A technique to increase dispersal between otherwise isolated fragments is the use of corridors. Although previous studies have compared dispersal between connected fragments to dispersal between unconnected fragments, it remains unknown how dispersal between fragments connected by a corridor compares to dispersal in unfragmented landscapes. To assess the extent to which corridors can restore dispersal in fragmented landscapes to levels observed in unfragmented landscapes, we employed a stable‐isotope marking technique to track seeds within four unfragmented landscapes and eight experimental landscapes with fragments connected by corridors. We studied two wind‐ and two bird‐dispersed plant species, because previous community‐based research showed that dispersal mode explains how connectivity effects vary among species. We constructed dispersal kernels for these species in unfragmented landscapes and connected fragments by marking seeds in the center of each landscape with 15 N and then recovering marked seeds in seed traps at distances up to 200 m. For the two wind‐dispersed plants, seed dispersal kernels were similar in unfragmented landscapes and connected fragments. In contrast, dispersal kernels of bird‐dispersed seeds were both affected by fragmentation and differed in the direction of the impact: Morella cerifera experienced more and Rhus copallina experienced less long‐distance dispersal in unfragmented than in connected landscapes. These results show that corridors can facilitate dispersal probabilities comparable to those observed in unfragmented landscapes. Although dispersal mode may provide useful broad predictions, we acknowledge that similar species may respond uniquely due to factors such as seasonality and disperser behavior. Furthermore, our results further indicate that prior work has likely underestimated dispersal distances of wind‐dispersed plants and that factors altering long‐distance dispersal may have a greater impact on the spread of species than previously thought.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2016

Fabrication of UN microsphere kernels for Fully Ceramic Microencapsulated fuel

Fully Ceramic Microencapsulated (FCM) fuel is of interest to the U.S. Department of Energy's Fuel Cycle Research and Development program as an Accident Tolerant Fuel (ATF). The FCM design uses a TRISO (Tri-Structural Isotropic) particle with a ZrO{sub 2} surrogate kernel, embedded in a dense SiC matrix. TRISO technology, originally developed for High Temperature Gas Reactors, has an established high level of performance retaining fission products and the SiC matrix of an FCM compact shows good steam oxidation behavior. This work reports on a fabrication route for producing dense UN microspheres used as the kernel of the TRISO particles in FCM fuel. The UN microspheres are of high purity and density with enough porosity to accommodate volatile fission products. Increasing the density is nevertheless of interest potentially allowing for an additional adjustable property for optimizing the fuel kernel and TRISO particle characteristics. Current work is focused on investigating the effect of a feedstock microstructure on final product density. (authors)

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2022

A neural network‐enhanced reproducing kernel particle method for modeling strain localization

Abstract Modeling the localized intensive deformation in a damaged solid requires highly refined discretization for accurate prediction, which significantly increases the computational cost. Although adaptive model refinement can be employed for enhanced effectiveness, it is cumbersome for the traditional mesh‐based methods to perform while modeling the evolving localizations. In this work, neural network‐enhanced reproducing kernel particle method (NN‐RKPM) is proposed, where the location, orientation, and shape of the solution transition near a localization is automatically captured by the NN approximation via a block‐level neural network (NN) optimization. The weights and biases in the blocked parameterization network control the location and orientation of the localization. The designed basic four‐kernel NN block is capable of capturing a triple junction or a quadruple junction topological pattern, while more complicated localization topological patters are captured by the superposition of multiple four‐kernel NN blocks. The standard RK approximation is then utilized to approximate the smooth part of the solution, which permits a much coarser discretization than the high‐resolution discretization needed to capture sharp solution transitions with the conventional methods. A regularization of the NN approximation is additionally introduced for discretization‐independent material responses. The effectiveness of the proposed NN‐RKPM is verified by a series of numerical verifications.

Baek, Jonghyuk↗
DOE OSTI2020

Renormalization and matching for the Collins-Soper kernel from lattice QCD

The Collins-Soper kernel, which governs the energy evolution of transverse- momentum dependent parton distribution functions (TMDPDFs), is required to accurately predict Drell-Yan like processes at small transverse momentum, and is a key ingredient for extracting TMDPDFs from experiment. Earlier we proposed a method to calculate this kernel from ratios of the so-called quasi-TMDPDFs determined with lattice QCD, which are defined as hadronic matrix elements of staple-shaped Euclidean Wilson line operators. Here we provide the one-loop renormalization of these operators in a regularization-independent momentum subtraction (RI'/MOM) scheme, as well as the conversion factor from the RI'/MOM-renormalized quasi-TMDPDF to the \( \overline{\mathrm{MS}} \) scheme. We also propose a procedure for calculating the Collins-Soper kernel directly from position space correlators, which simplifies the lattice determination.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2021

Determination of the Collins-Soper Kernel from Lattice QCD

We present lattice results for the non-perturbative Collins-Soper (CS) kernel, which describes the energy-dependence of transverse momentum-dependent parton distributions (TMDs). The CS kernel is extracted from the ratios of first Mellin moments of quasi-TMDs evaluated at different nucleon momenta. The analysis is done with dynamical N f = 2 + 1 clover fermions for the CLS ensemble H101 (a = 0. 0854 fm, m π = m K = 422 MeV). The computed CS kernel is in good agreement with experimental extractions and previous lattice studies.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2017

Determining the minimum required uranium carbide content for HTGR UCO fuel kernels

There are three important failure mechanisms that must be controlled in high-temperature gas-cooled reactor (HTGR) fuel for certain higher burnup applications are SiC layer rupture, SiC corrosion by CO, and coating compromise from kernel migration. All are related to high CO pressures stemming from free O generated when uranium present as UO 2 fissions and the O is not subsequently bound by other elements. Furthermore, in the HTGR UCO kernel design, CO buildup from excess O is controlled by the inclusion of additional uranium in the form of a carbide, UC x . An approach for determining the minimum UC x content to ensure negligible CO formation was developed and demonstrated using CALPHAD models and the Serpent 2 reactor physics and depletion analysis tool. Our results are intended to be more accurate than previous estimates by including more nuclear and chemical factors, in particular the effect of transmutation products on the oxygen distribution as the fuel kernel composition evolves with burnup.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2021

When and why PINNs fail to train: A neural tangent kernel perspective

Physics-informed neural networks (PINNs) have lately received great attention thanks to their flexibility in tackling a wide range of forward and inverse problems involving partial differential equations. However, despite their noticeable empirical success, little is known about how such constrained neural networks behave during their training via gradient descent. More importantly, even less is known about why such models sometimes fail to train at all. Here in this work, we aim to investigate these questions through the lens of the Neural Tangent Kernel (NTK); a kernel that captures the behavior of fully-connected neural networks in the infinite width limit during training via gradient descent. Specifically, we derive the NTK of PINNs and prove that, under appropriate conditions, it converges to a deterministic kernel that stays constant during training in the infinite-width limit. This allows us to analyze the training dynamics of PINNs through the lens of their limiting NTK and find a remarkable discrepancy in the convergence rate of the different loss components contributing to the total training error. To address this fundamental pathology, we propose a novel gradient descent algorithm that utilizes the eigenvalues of the NTK to adaptively calibrate the convergence rate of the total training error. Finally, we perform a series of numerical experiments to verify the correctness of our theory and the practical effectiveness of the proposed algorithms.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2018

Gaussian kernel density functions for compositional quantification in atom probe tomography

Highlights: • A methodology for compositional quantification from atom probe tomography is proposed. • A probability density function p(x) given discrete data points is constructed. • Each atom approximated by a Gaussian function with a particular smoothing parameter. • Density Kernel functions are used to find the optimum smoothing parameter. • The method is validated by using simulated data of semiconductor materials. - Abstract: Atom probe tomography (APT) has the ability to identify the nature and position of single atoms in a material with an almost 3D atomic resolution. However, the quantification of the material composition requires an appropriate treatment of the discrete APT data. When the amount of atoms selected to quantify the composition is relatively small, the spatial resolution is enhanced but the statistical error worsens. Conversely, the increase in the number of atoms considered reduces the spatial resolution, but improving the statistics. We have developed a methodology to reach an optimum equilibrium between the positional and the statistical error in an unsupervised form using Gaussian Kernel density functions. The validity of the method has been tested using APT simulated data of semiconductor materials. It has been proved that the chemical quantification in these materials requires the appropriate selection of the smoothing parameter, obtained without user intervention. In addition, the results have been compared to the usual techniques for composition measurement from APT data (voxelization and proximity histograms), showing better precision for high spatial resolution. This work supplies a data driven methodology based in Gaussian Kernel density functions for the accurate quantification of the composition from APT data.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2021

Gaussian process analysis of electron energy loss spectroscopy data: multivariate reconstruction and kernel control

Abstract Advances in hyperspectral imaging including electron energy loss spectroscopy bring forth the challenges of exploratory and physics-based analysis of multidimensional data sets. The multivariate linear unmixing methods generally explore similarities in the energy dimension, but ignore correlations in the spatial domain. At the same time, Gaussian process (GP) explicitly incorporate spatial correlations in the form of kernel functions but is computationally intensive. Here, we implement a GP method operating on the full spatial domain and reduced representations in the energy domain. In this multivariate GP, the information between the components is shared via a common spatial kernel structure, while allowing for variability in the relative noise magnitude or image morphology. We explore the role of kernel constraints on the quality of the reconstruction, and suggest an approach for estimating them from the experimental data. We further show that spatial information contained in higher-order components can be reconstructed and spatially localized.

36 MATERIALS SCIENCE↗
DOE OSTI2017

Control of Early Flame Kernel Growth by Multi-Wavelength Laser Pulses for Enhanced Ignition

The present contribution examines the impact of plasma dynamics and plasma-driven fluid dynamics on the flame growth of laser ignited mixtures and shows that a new dual-pulse scheme can be used to control the kernel formation process in ways that extend the lean ignition limit. We do this by performing a comparative study between (conventional) single-pulse laser ignition (λ = 1064 nm) and a novel dual-pulse method based on combining an ultraviolet (UV) pre-ionization pulse (λ = 266 nm) with an overlapped near-infrared (NIR) energy addition pulse (λ = 1064 nm). We employ OH* chemiluminescence to visualize the evolution of the early flame kernel. For single-pulse laser ignition at lean conditions, the flame kernel separates through third lobe detachment, corresponding to high strain rates that extinguish the flame. In this work, we investigate the capabilities of the dual-pulse to control the plasma-driven fluid dynamics by adjusting the axial offset of the two focal points. In particular, we find there exists a beam waist offset whereby the resulting vorticity suppresses formation of the third lobe, consequently reducing flame stretch. With this approach, we demonstrate that the dual-pulse method enables reduced flame speeds (at early times), an extended lean limit, increased combustion efficiency, and decreased laser energy requirements.

42 ENGINEERING↗