Search NASA⌕ Search

SEARCH · Search NASA

Results for “distributed clustering 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 505 records · Page 28

A study of the threshold method utilizing raingage data

The threshold method for estimation of area-average rain rate relies on determination of the fractional area where rain rate exceeds a preset level of intensity. Previous studies have shown that the optimal threshold level depends on the climatological rain-rate distribution (RRD). It has also been noted, however, that the climatological RRD may be composed of an aggregate of distributions, one for each of several distinctly different synoptic conditions, each having its own optimal threshold. In this study, the impact of RRD variations on the threshold method is shown in an analysis of 1-min rainrate data from a network of tipping-bucket gauges in Darwin, Australia. Data are analyzed for two distinct regimes: the premonsoon environment, having isolated intense thunderstorms, and the active monsoon rains, having organized convective cell clusters that generate large areas of stratiform rain. It is found that a threshold of 10 mm/h results in the same threshold coefficient for both regimes, suggesting an alternative definition of optimal threshold as that which is least sensitive to distribution variations. The observed behavior of the threshold coefficient is well simulated by assumption of lognormal distributions with different scale parameters and same shape parameters.

Short, David A.↗

Performance and power modeling and prediction using MuMMI and 10 machine learning methods

Energy-efficient scientific applications require insight into how high performance computing system features impact the applications' power and performance. This insight can result from the development of performance and power models. Here, in this article, we use the modeling and prediction tool MuMMI (Multiple Metrics Modeling Infrastructure) and 10 machine learning methods to model and predict performance and power consumption and compare their prediction error rates. We use an algorithm-based fault-tolerant linear algebra code and a multilevel checkpointing fault-tolerant heat distribution code to conduct our modeling and prediction study on the Cray XC40 Theta and IBM BG/Q Mira at Argonne National Laboratory and the Intel Haswell cluster Shepard at Sandia National Laboratories. Our experimental results show that the prediction error rates in performance and power using MuMMI are less than 10% for most cases. By utilizing the models for runtime, node power, CPU power, and memory power, we identify the most significant performance counters for potential application optimizations, and we predict theoretical outcomes of the optimizations. Based on two collected datasets, we analyze and compare the prediction accuracy in performance and power consumption using MuMMI and 10 machine learning methods.

97 MATHEMATICS AND COMPUTING↗

Fitting FFT-derived spectra: Theory, tool, and application to solar radio spike decomposition

Spectra derived from fast Fourier transform (FFT) analysis of time-domain data intrinsically contain statistical fluctuations whose distribution depends on the number of accumulated spectra contributing to a measurement. The tail of this distribution, which is essential for separating the true signal from the statistical fluctuations, deviates noticeably from the normal distribution for a finite number of accumulations. In this paper, we develop a theory to properly account for the statistical fluctuations when fitting a model to a given accumulated spectrum. The method is implemented in software for the purpose of automatically fitting a large body of such FFT-derived spectra. We apply this tool to analyze a portion of a dense cluster of spikes recorded by our FASR Subsystem Testbed instrument during a record-breaking event that occurred on 2006 December 6. The outcome of this analysis is briefly discussed.

79 ASTRONOMY AND ASTROPHYSICS↗

Necessary conditions for accurate computations of three-body partial decay widths

The partial width for decay of a resonance into three fragments is largely determined at distances where the energy is smaller than the effective potential producing the corresponding wave function. At short distances the many-body properties are accounted for by preformation or spectroscopic factors. We use the adiabatic expansion method combined with the WKB approximation to obtain the indispensable cluster model wave functions at intermediate and larger distances. We test the concept by deriving conditions for the minimal basis expressed in terms of partial waves and radial nodes. We compare results for different effective interactions and methods. Agreement is found with experimental values for a sufficiently large basis. We illustrate the ideas with realistic examples from {alpha} emission of {sup 12}C and two-proton emission of {sup 17}Ne. Basis requirements for accurate momentum distributions are briefly discussed.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

NWChem: scalable parallel computational chemistry

NWChem is a general purpose computational chemistry code specifically designed to run on distributed memory parallel computers. The core functionality of the code focuses on molecular dynamics, Hartree-Fock and density functional theory methods for both plane-wave basis sets as well as Gaussian basis sets, tensor contraction engine based coupled cluster capabilities and combined quantum mechanics/molecular mechanics descriptions. It was realized from the beginning that scalable implementations of these methods required a programming paradigm inherently different from what message passing approaches could offer. In response a global address space library, the Global Array Toolkit, was developed. The programming model it offers is based on using predominantly one-sided communication. This model underpins most of the functionality in NWChem and the power of it is exemplified by the fact that the code scales to tens of thousands of processors. In this paper the core capabilities of NWChem are described as well as their implementation to achieve an efficient computational chemistry code with high parallel scalability. NWChem is a modern, open source, computational chemistry code1 specifically designed for large scale parallel applications2. To meet the challenges of developing efficient, scalable and portable programs of this nature a particular code design was adopted. This code design involved two main features. First of all, the code is build up in a modular fashion so that a large variety of functionality can be integrated easily. Secondly, to facilitate writing complex parallel algorithms the Global Array toolkit was developed. This toolkit allows one to write parallel applications in a shared memory like approach, but offers additional mechanisms to exploit data locality to lower communication overheads. This framework has proven to be very successful in computational chemistry but is applicable to any engineering domain. Within the context created by the features above NWChem has grown into a general purpose computational chemistry code that supports a wide variety of energy expressions and capabilities to calculate properties based there upon. The main energy expressions are classical mechanics force fields, Hartree-Fock and DFT both for finite systems and condensed phase systems, coupled cluster, as well as QM/MM. For most energy expressions single point calculations, geometry optimizations, excited states, and other properties are available. Below we briefly discuss each of the main energy expressions and the critical points involved in scalable implementations thereof.

99 GENERAL AND MISCELLANEOUS↗

Predicting U 3 O 8 powder processing conditions: An AI/ML approach analyzing deep learning embeddings of SEM micrographs

High-resolution SEM images of uranium-oxide powders encode micro- and nanoscale clues to their synthesis route and calcination temperature. We trained a ResNet-50 model on 11 commercial-scale U₃O₈ classes, ammonium diuranate (ADU) or uranyl peroxide (H₂O₂) precursors calcined at temperatures ranging from 400 to 750 °C and added a 256-D projection head before the classifier to analyze the learned representation. The best of eight seeds reached 92.4 % accuracy on reserved testing data, but our focus is the structure of the embedding space rather than the accuracy and labels. We quantify class relatedness in the original 256-D space using centroid similarity and distributional distances, and we use Uniform Manifold Approximation Projection (UMAP) for visualization. ‘Unknown’ images from different preparation methods, SEM operators, and from the literature localized near the expected classes under a nearest-centroid analysis without retraining, as well as clustered in similar UMAP space. In conclusion, this embedding-centered workflow complements black-box classification by providing quantitative, similarity-based comparisons of U₃O₈ morphologies and reduces storage space by up to 98 % for image data used in millisecond vector search comparisons.

36 MATERIALS SCIENCE↗

Microscopic approach to the kinetics of pattern formation of charged molecules on surfaces

A microscopic formalism based on computing many-particle densities is applied to the analysis of the diffusion-controlled kinetics of pattern formation in oppositely charged molecules on surfaces or adsorbed at interfaces with competing long-range Coulomb and short-range Lennard-Jones interactions. Particular attention is paid to the proper molecular treatment of energetic interactions driving pattern formation in inhomogeneous systems. The reverse Monte Carlo method is used to visualize the spatial molecular distribution based on the calculated radial distribution functions (joint correlation functions). We show the formation of charge domains for certain combinations of temperature and dynamical interaction parameters. The charge segregation evolves into quasicrystalline clusters of charges, due to the competing long- and short-range interactions. The clusters initially co-exist with a gas phase of charges that eventually add to the clusters, generating ”fingers” or line of charges of the same sign, very different than the nanopatterns expected by molecular dynamics in systems with competing interactions in two dimensions, such as strain or dipolar versus van der Waals interactions.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

An Algorithm for Characterization of Fiber Aggregation in Composite Microstructures

Composite structures are susceptible to localized flaws, or variability, that drive global failure. To capture this variance, a multiscale model needs to be introduced that not only accurately represents the statistical nature of the composite microstructure but is also efficient. At the microscale, fiber aggregation creates local stress concentrations, where failure is likely to occur sooner than expected. A method of rapid microstructure generation, in conjunction with cluster characterization, is examined to develop accurate reproductions of 2D composite cross-sections. Parameterization of shape and size of fiber clusters is used to characterize representative volume elements. The discrete element method is used to generate pseudo-microstructures. The clustering parameters from the pseudo- and actual microstructure arrangements (obtained from micrographs) can be compared to determine the validity of the representative volume element generated with the discrete element method. Results show a promising approach to evaluating randomness of fiber distributions and how to accurately recreate microstructures for strength analysis.

Carbon fiber↗

MO-FG-202-06: Improving the Performance of Gamma Analysis QA with Radiomics- Based Image Analysis

Purpose: The use of gamma analysis for IMRT quality assurance has well-known limitations. Traditionally, a simple thresholding technique is used to evaluated passing criteria. However, like any image the gamma distribution is rich in information which thresholding mostly discards. We therefore propose a novel method of analyzing gamma images that uses quantitative image features borrowed from radiomics, with the goal of improving error detection. Methods: 368 gamma images were generated from 184 clinical IMRT beams. For each beam the dose to a phantom was measured with EPID dosimetry and compared to the TPS dose calculated with and without normally distributed (2mm sigma) errors in MLC positions. The magnitude of 17 intensity histogram and size-zone radiomic features were derived from each image. The features that differed most significantly between image sets were determined with ROC analysis. A linear machine-learning model was trained on these features to classify images as with or without errors on 180 gamma images.The model was then applied to an independent validation set of 188 additional gamma distributions, half with and half without errors. Results: The most significant features for detecting errors were histogram kurtosis (p=0.007) and three size-zone metrics (p<1e-6 for each). The sizezone metrics detected clusters of high gamma-value pixels under mispositioned MLCs. The model applied to the validation set had an AUC of 0.8, compared to 0.56 for traditional gamma analysis with the decision threshold restricted to 98% or less. Conclusion: A radiomics-based image analysis method was developed that is more effective in detecting error than traditional gamma analysis. Though the pilot study here considers only MLC position errors, radiomics-based methods for other error types are being developed, which may provide better error detection and useful information on the source of detected errors. This work was partially supported by a grant from the Agency for Healthcare Research and Quality, grant number R18 HS022244-01.

60 APPLIED LIFE SCIENCES↗

SLUG-STOCHASTICALLY LIGHTING UP GALAXIES. I. METHODS AND VALIDATING TESTS

The effects of stochasticity on the luminosities of stellar populations are an often neglected but crucial element for understanding populations in the low-mass or the low star formation rate regime. To address this issue, we present SLUG, a new code to 'Stochastically Light Up Galaxies'. SLUG synthesizes stellar populations using a Monte Carlo technique that properly treats stochastic sampling including the effects of clustering, the stellar initial mass function, star formation history, stellar evolution, and cluster disruption. This code produces many useful outputs, including (1) catalogs of star clusters and their properties such as their stellar initial mass distributions and their photometric properties in a variety of filters, (2) two dimensional histograms of color-magnitude diagrams of every star in the simulation, and (3) the photometric properties of field stars and the integrated photometry of the entire simulated galaxy. After presenting the SLUG algorithm in detail, we validate the code through comparisons with STARBURST99 in the well-sampled regime, and with observed photometry of Milky Way clusters. Finally, we demonstrate SLUG's capabilities by presenting outputs in the stochastic regime. SLUG is publicly distributed through the Web site http://sites.google.com/site/runslug/.

79 ASTRONOMY AND ASTROPHYSICS↗

A domain decomposition method of stochastic PDEs: An iterative solution techniques using a two-level scalable preconditioner

Recent advances in high performance computing systems and sensing technologies motivate computational simulations with extremely high resolution models with capabilities to quantify uncertainties for credible numerical predictions. A two-level domain decomposition method is reported in this investigation to devise a linear solver for the large-scale system in the Galerkin spectral stochastic finite element method (SSFEM). In particular, a two-level scalable preconditioner is introduced in order to iteratively solve the large-scale linear system in the intrusive SSFEM using an iterative substructuring based domain decomposition solver. The implementation of the algorithm involves solving a local problem on each subdomain that constructs the local part of the preconditioner and a coarse problem that propagates information globally among the subdomains. The numerical and parallel scalabilities of the two-level preconditioner are contrasted with the previously developed one-level preconditioner for two-dimensional flow through porous media and elasticity problems with spatially varying non-Gaussian material properties. A distributed implementation of the parallel algorithm is carried out using MPI and PETSc parallel libraries. The scalabilities of the algorithm are investigated in a Linux cluster.

42 ENGINEERING↗

Adaptive Hierarchical Cyber Attack Detection and Localization in Active Distribution Systems

Development of a cyber security strategy for the active distribution systems is challenging due to the inclusion of distributed renewable energy generations. Here this paper proposes an adaptive hierarchical cyber attack detection and localization framework for distributed active distribution systems via analyzing electrical waveforms. Cyber attack detection is based on a sequential deep learning model, via which even minor cyber attacks can be identified. The two-stage cyber attack localization algorithm first estimates the cyber attack sub-region, and then localize the specified cyber attack within the estimated subregion. We propose a modified spectral clustering-based network partitioning method for the hierarchical cyber attack ‘coarse’ localization. Next, to further narrow down the cyber attack location, a normalized impact score based on waveform statistical metrics is proposed to obtain a ‘fine’ cyber attack location by characterizing different waveform properties. Finally, compared with classical and state-of-art methods, a comprehensive quantitative evaluation with two case studies shows promising estimation results of the proposed framework.

42 ENGINEERING↗

THE LACK OF DIFFUSE, NON-THERMAL HARD X-RAY EMISSION IN THE COMA CLUSTER: THE SWIFT BURST ALERT TELESCOPE'S EYE VIEW

The Coma Cluster of galaxies hosts the brightest radio halo known and has therefore been the target of numerous searches for associated inverse Compton (IC) emission, particularly at hard X-ray energies where the IC signal must eventually dominate over thermal emission. The most recent search with the Suzaku Hard X-ray Detector failed to confirm previous IC detections with RXTE and BeppoSAX, instead setting an upper limit 2.5 times below their non-thermal flux. However, this discrepancy can be resolved if the IC emission is very extended, beyond the scale of the cluster radio halo. Using reconstructed sky images from the 58-month Swift Burst Alert Telescope (BAT) all-sky survey, the feasibility of such a solution is investigated. Building on Renaud et al., we test and implement a method for extracting the fluxes of extended sources, assuming specified spatial distributions. BAT spectra are jointly fit with an XMM-Newton EPIC-pn spectrum derived from mosaic observations. We find no evidence for large-scale IC emission at the level expected from the previously detected non-thermal fluxes. For all non-thermal spatial distributions considered, which span the gamut of physically reasonable IC models, we determine upper limits for which the largest (most conservative) limit is {approx}<4.2 x 10{sup -12} erg s{sup -1} cm{sup -2} (20-80 keV), which corresponds to a lower limit on the magnetic field B > 0.2 {mu} G. A nominal flux upper limit of <2.7 x 10{sup -12} erg s{sup -1} cm{sup -2}, with corresponding B > 0.25 {mu} G, is derived for the most probable IC distribution given the size of the radio halo and likely magnetic field radial profile.

79 ASTRONOMY AND ASTROPHYSICS↗

CLASH: MASS DISTRIBUTION IN AND AROUND MACS J1206.2-0847 FROM A FULL CLUSTER LENSING ANALYSIS

We derive an accurate mass distribution of the galaxy cluster MACS J1206.2-0847 (z = 0.439) from a combined weak-lensing distortion, magnification, and strong-lensing analysis of wide-field Subaru BVR{sub c} I{sub c} z' imaging and our recent 16-band Hubble Space Telescope observations taken as part of the Cluster Lensing And Supernova survey with Hubble program. We find good agreement in the regions of overlap between several weak- and strong-lensing mass reconstructions using a wide variety of modeling methods, ensuring consistency. The Subaru data reveal the presence of a surrounding large-scale structure with the major axis running approximately northwest-southeast (NW-SE), aligned with the cluster and its brightest galaxy shapes, showing elongation with a {approx}2: 1 axis ratio in the plane of the sky. Our full-lensing mass profile exhibits a shallow profile slope dln {Sigma}/dln R {approx} -1 at cluster outskirts (R {approx}> 1 Mpc h{sup -1}), whereas the mass distribution excluding the NW-SE excess regions steepens farther out, well described by the Navarro-Frenk-White form. Assuming a spherical halo, we obtain a virial mass M{sub vir} = (1.1 {+-} 0.2 {+-} 0.1) Multiplication-Sign 10{sup 15} M{sub Sun} h{sup -1} and a halo concentration c{sub vir} = 6.9 {+-} 1.0 {+-} 1.2 (c{sub vir} {approx} 5.7 when the central 50 kpc h{sup -1} is excluded), which falls in the range 4 {approx}< (c) {approx}< 7 of average c(M, z) predictions for relaxed clusters from recent {Lambda} cold dark matter simulations. Our full-lensing results are found to be in agreement with X-ray mass measurements where the data overlap, and when combined with Chandra gas mass measurements, they yield a cumulative gas mass fraction of 13.7{sup +4.5}{sub -3.0}% at 0.7 Mpc h{sup -1}( Almost-Equal-To 1.7 r{sub 2500}), a typical value observed for high-mass clusters.

79 ASTRONOMY AND ASTROPHYSICS↗

The panchromatic Hubble Andromeda Treasury. V. Ages and masses of the year 1 stellar clusters

We present ages and masses for 601 star clusters in M31 from the analysis of the six filter integrated light measurements from near-ultraviolet to near-infrared wavelengths, made as part of the Panchromatic Hubble Andromeda Treasury (PHAT). We derive the ages and masses using a probabilistic technique, which accounts for the effects of stochastic sampling of the stellar initial mass function. Tests on synthetic data show that this method, in conjunction with the exquisite sensitivity of the PHAT observations and their broad wavelength baseline, provides robust age and mass recovery for clusters ranging from ∼10{sup 2} to 2 × 10{sup 6} M {sub ☉}. We find that the cluster age distribution is consistent with being uniform over the past 100 Myr, which suggests a weak effect of cluster disruption within M31. The age distribution of older (>100 Myr) clusters falls toward old ages, consistent with a power-law decline of index –1, likely from a combination of fading and disruption of the clusters. We find that the mass distribution of the whole sample can be well described by a single power law with a spectral index of –1.9 ± 0.1 over the range of 10{sup 3}-3 × 10{sup 5} M {sub ☉}. However, if we subdivide the sample by galactocentric radius, we find that the age distributions remain unchanged. However, the mass spectral index varies significantly, showing best-fit values between –2.2 and –1.8, with the shallower slope in the highest star formation intensity regions. We explore the robustness of our study to potential systematics and conclude that the cluster mass function may vary with respect to environment.

79 ASTRONOMY AND ASTROPHYSICS↗

Molecular level structure in water-alcohol mixtures; from shots to pints: L’Chaim!

The two objectives of this work are to look at the molecular level mixing in water-alcohol liquid mixtures and to determine how well the Effective Fragment Potential (EFP) method performs at describing small, heterogenous water-alcohol mixtures. In this work we look both at small clusters, size 2 – 24 molecules, as well as bulk mixtures, 64 or more molecules. For small clusters (2 – 24 molecules), the molecular structure was studied using the EFP method and MP2. Extensive Monte Carlo/simulated annealing (MC/SA) global optimizations were used to locate low energy structures for all water, alcohol, and water-alcohol mixtures. The alcohols studied include the small, straight-chain, primary alcohols: methanol, ethanol, propanol, and 1-butanol as well as the secondary alcohol 2-butanol, and tertiary butanol. It is shown that a number of low energy mixed clusters retain the structural characteristics of the pure clusters. It is also shown that the EFP method performs well at describing small, heterogenous water-alcohol mixtures. The molecular structure in bulk water, alcohol, and water-alcohol mixtures is studied with the EFP method and molecular dynamics simulations. Radial distribution functions are used to analyze structure and to benchmark the EFP method against experimental results. The effect of alcohol concentration and alcohol chain length on the local structure of water is investigated. Preliminary results indicate that the EFP method can potentially model bulk water-alcohol systems well as compared to experiment. Results on the effects of alcohol chain length are forthcoming.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Overlapping clusters for distributed computation.

Scalable, distributed algorithms must address communication problems. We investigate overlapping clusters, or vertex partitions that intersect, for graph computations. This setup stores more of the graph than required but then affords the ease of implementation of vertex partitioned algorithms. Our hope is that this technique allows us to reduce communication in a computation on a distributed graph. The motivation above draws on recent work in communication avoiding algorithms. Mohiyuddin et al. (SC09) design a matrix-powers kernel that gives rise to an overlapping partition. Fritzsche et al. (CSC2009) develop an overlapping clustering for a Schwarz method. Both techniques extend an initial partitioning with overlap. Our procedure generates overlap directly. Indeed, Schwarz methods are commonly used to capitalize on overlap. Elsewhere, overlapping communities (Ahn et al, Nature 2009; Mishra et al. WAW2007) are now a popular model of structure in social networks. These have long been studied in statistics (Cole and Wishart, CompJ 1970). We present two types of results: (i) an estimated swapping probability {rho}{infinity}; and (ii) the communication volume of a parallel PageRank solution (link-following {alpha} = 0.85) using an additive Schwarz method. The volume ratio is the amount of extra storage for the overlap (2 means we store the graph twice). Below, as the ratio increases, the swapping probability and PageRank communication volume decreases.

99 GENERAL AND MISCELLANEOUS↗

Probing Heterogeneity in Bovine Enamel Composition through Nanoscale Chemical Imaging using Atom Probe Tomography

Objective: The aim of this study was to determine the heterogeneity in chemical composition of bovine enamel using atom probe tomography, and thereby evaluate the suitability of bovine enamel as a substitute for human enamel in in vitro dental research. Design: Enamel samples from extracted bovine incisor teeth were first sectioned using a diamond saw and then milled into needle-like samples (<100 nm diameter) by focused ion beam (FIB) coupled with a scanning electron microscope (SEM). The samples were then analyzed in the atom probe to acquire three-dimensional (3D) images and quantify the atomic chemistry and distribution in bovine enamel. Results: For the first time, the atomic-level composition and clustering of major constituents and impurities within bovine enamel were determined and imaged. We discovered that the chemical composition of bovine enamel is spatially inhomogeneous at the atomic scale. The average bulk Ca/P ratio, ~1.4, was in agreement with previously reported literature values from alternative conventional methods. When assessed locally at the atomic scale, the Ca/P ratio varied between 1.1 and 2.03. We also discovered that the Mg impurities were significantly segregated throughout the enamel, and such clustering influenced the variation of Ca/P ratios. The increase in Mg concentrations, near the Mg clusters, correlated with increased Ca and decreased P concentrations. Conclusion: In conclusion, the presented findings of variability in local composition should be taken into account when interpreting dental research results from bovine enamel.

59 BASIC BIOLOGICAL SCIENCES↗