Search NASA⌕ Search

SEARCH · Search NASA

Results for “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 415 records · Page 23

Solution to sign problems in models of interacting fermions and quantum spins

We show that solutions to fermion sign problems that are found in the formulation where the path integral is expanded in powers of the interaction in continuous time can be extended to systems involving fermions interacting with dynamical quantum spins. While these sign problems seem unsolvable in the auxiliary field approach, solutions emerge in the world-line representation of quantum spins. Combining this idea with meron-cluster methods, we are able to further extend the class of models that are solvable. We demonstrate these solutions to sign problems by considering several examples of strongly correlated systems that contain the physics of semimetals, insulators, superfluidity, and antiferromagnetism.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Precision Calculation of Blackbody Radiation Shifts for Optical Frequency Metrology

We show that three group IIIB divalent ions, B{sup +}, Al{sup +}, and In{sup +}, have anomalously small blackbody radiation (BBR) shifts of the ns{sup 2} {sup 1}S{sub 0}-nsnp {sup 3}P{sub 0}{sup o} clock transitions. The fractional BBR shifts for these ions are at least 10 times smaller than those of any other present or proposed optical frequency standards at the same temperature, and are less than 0.3% of the Sr clock shift. We have developed a hybrid configuration-interaction + coupled-cluster method that provides accurate treatment of correlation corrections in such ions and yields a rigorous upper bound on the uncertainty of the final results. We reduce the BBR contribution to the fractional frequency uncertainty of the Al{sup +} clock to 4x10{sup -19} at T=300 K. We also reduce the uncertainties due to this effect at room temperature to 10{sup -18} level for B{sup +} and In{sup +} to facilitate further development of these systems for metrology and quantum sensing.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

First Principles Description of the Giant Dipole Resonance in O 16

We present a calculation of the giant dipole resonance in 16 O based on a nucleon-nucleon (NN) interaction from chiral e ective eld theory that reproduces NN scattering data with high accuracy. By merging the Lorentz integral transform and the coupled-cluster methods, we extend the previous theoretical limits for break-up observables in light nuclei with mass numbers (A ≤ 7), and address the collective giant dipole resonance of 16 O. We successfully benchmark the new approach against virtually exact results from the hyperspherical harmonics method in 4 He. Our results for 16 O reproduce the position and the total strength (bremsstrahlung sum rule) of the dipole response very well. When compared to the cross section from photo-absorption experiments the theoretical curve exhibits a smeared form of the peak. The tail region between 40 and 100 MeV is reproduced within uncertainties.

74 ATOMIC AND MOLECULAR PHYSICS↗

Dynamics of Density Fluctuations in Atomic Nuclei

We study the spatiotemporal patterns of density fluctuations in 16,24O and 48Ca using nuclear interactions from chiral effective field theory and the time-dependent coupled-cluster method. We find that two-particle–two-hole excitations generate small-amplitude fluctuations that are fast, short-ranged, and of stochastic character.

Bonaiti, Francesca [ORNL] (ORCID:0000000239261609)↗

Combining Tensor Decompositions and Graph Analytics to Provide Cyber Situational Awareness at HPC Scale

This paper describes MADHAT (Multidimensional Anomaly Detection fusing HPC, Analytics, and Tensors), an integrated workflow that demonstrates the applicability of HPC resources to the problem of maintaining cyber situational awareness. MADHAT combines two high-performance packages: ENSIGN for large-scale sparse tensor decompositions and HAGGLE for graph analytics. Tensor decompositions isolate coherent patterns of network behavior in ways that common clustering methods based on distance metrics cannot. Parallelized graph analysis then uses directed queries on a representation that combines the elements of identified patterns with other available information (such as additional log fields, domain knowledge, network topology, whitelists and blacklists, prior feedback, and published alerts) to confirm or reject a threat hypothesis, collect context, and raise alerts. MADHAT was developed using the collaborative HPC Architecture for Cyber Situational Awareness (HACSAW) research environment and evaluated on structured network sensor logs collected from Defense Research and Engineering Network (DREN) sites using HPC resources at the U.S. Army Engineer Research and Development Center DoD Supercomputing Resource Center (ERDC DSRC). To date, MADHAT has analyzed logs with over 650 million entries.

Cyber security, Tensor decomposition, graph analyt↗

Joint Management and Optimization of Residential Natural Gas and Electricity Distribution Networks Coupled via Fuel Cells

The attractive features of natural gas as well as the growing electric power demand worldwide have created increasing interest in natural-gas-based distributed generation applications for electric distribution networks. Here, this paper investigates the interdependency between a residential natural gas network and an electric distribution network that are linked together via fuel cells. The modeling of the natural gas network is introduced first, and then the algorithm for gas flow study is presented. The optimal placement and sizing of fuel cell based distributed generation systems are formulated to minimize the losses in both the natural gas network and the electric distribution grid, subject to the constraints imposed by both networks. In addition, a probabilistic model for both gas and electricity demands is developed based on historical electricity and natural gas demand data. A K-means clustering method is used to determine the hourly load states to solve the joint probabilistic optimization problem. Simulation studies are carried out on an integrated system consisting of the IEEE 69-bus distribution network and a radial 27-node natural gas network to verify the developed optimization model and the proposed method.

24 POWER TRANSMISSION AND DISTRIBUTION↗

A Zonal Volt/VAR Control Mechanism for High PV Penetration Distribution Systems

This paper presents a zonal Volt/VAR control scheme for regulating voltage in unbalanced 3-phase distribution systems using inverter-based resources (IBR). First, the dependency between nodal voltage changes and IBR reactive power injections is derived via voltage sensitivity studies. Then, a fast-incremental clustering method is used to divide the distribution circuit into weakly-coupled zones based on correlations between nodal voltage sensitivities. The weak-coupling allows the voltage to be regulated independently within each zone using a rule-based voltage controller to dispatch IBR for voltage corrections. Simulation results on actual distribution feeder show that the proposed zone-based Volt/VAR control method maintains system voltages within their operational limits while reducing the runtime of the Volt/VAR controller from tens of second to a couple of milliseconds compared with centralized, optimization-based Volt/VAR control methods.

Alrushoud, Asmaa↗

Robust System Separation Strategy Considering Online Wide-Area Coherency Identification and Uncertainties of Renewable Energy Sources

With the fast growth of renewable energy sources (RES), more and more uncertainties are involved and influencing the stable operation of power systems. Controlled islanding is the last measure to prevent power system blackouts, thus this paper aims to propose a novel model of system separation based on Online Coherency Identification and Adjustable Robust Optimization Programming (OCI-AROP) for minimizing load shedding considering the uncertainties of RES. First, Fuzzy C-Means (FCM) clustering method with F-statistics is utilized to identify the coherent generator groups with the frequency data measured by Phasor Measurement Units (PMUs). Then, the OCI-AROP model considering coherent group constraints, connectivity constraints and robustness constraints about RES are presented. Finally, the case studies on IEEE-39 bus system and WECC-179 bus system are employed to demonstrate the effectiveness of the proposed OCI-AROP model, and comparisons among the OCI-AROP model and the other models are also given to show its superiority.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Magnetic hyperfine interactions of {sup 119}Sn probe atoms in the binary perovskite CaCu{sub 3}Mn{sub 4}O{sub 12}

The manganite CaCu{sub 3}Mn{sub 4}O{sub 12} doped by {sup 119}Sn atoms (about 1 at % with respect to Mn atoms) is studied by Moessbauer spectroscopy. The introduction of diamagnetic tin atoms is found not to affect the structure of the manganite. Tetravalent tin atoms are shown to substitute for the isovalent manganese atoms that are located in an octahedral oxygen surrounding. The cluster method of molecular orbitals is used to calculate the contributions of Mn{sup 4+} and Cu{sup 2+} cations that belong to different structural sublattices to the hyperfine magnetic field at {sup 119}Sn nuclei (H{sub Sn} = 105 kOe at T = 77 K). These partial contributions are analyzed, and the intrasublattice Mn{sup 4+}-O-Mn{sup 4+} exchange interactions are found to play a significant role in the formation of the magnetic structure of the manganite.

36 MATERIALS SCIENCE↗

Dynamics of the order parameter and the potential of the hydrogen bond in a ferroelectric DKDP crystal

Within a pseudospin model of a KD{sub 2}PO{sub 4} crystal, the relaxation time of the average pseudospin (the order parameter) in the mean-field approximation depends on the activation energy of the jump of a deuteron between the minima of the hydrogen bond potential, the average equilibrium value of the pseudospin, and the dielectric susceptibility. Calculating the average equilibrium spin and the susceptibility within the four-particle cluster approximation and fitting the data for the width of the central peak from a Raman scattering experiment, we numerically estimate the activation energy. For the same values of the parameters of the cluster method, the calculated inverse relaxation time is in good agreement with the data of ultrasonic measurements.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Ab initio study of Hg-Hg and E112-E112 van der Waals interactions

The ground electronic state of the eka-mercury dimer (E112{sub 2}) was studied within the model of generalized relativistic effective core potentials. A combined procedure based on describing correlation effects by the scalar relativistic coupled-cluster method and on taking into account spin-dependent interactions by means of density-functional theory was used in the calculations. A high accuracy of this approach was confirmed by the results of similar calculations for the mercury dimer. It was found that the bond length is nearly 0.4 A shorter in E112{sub 2} than in Hg{sub 2} and that the dissociation energy of the former is approximately twice as high as that of the latter dimer.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Nonperturbative model of hot SU(2) gluodynamics

A hot-gluodynamics model based on taking consistently into account nonperturbative fluctuations of the gluon field in the temperature-deconfining phase is proposed. In order to perform averaging over nonperturbative fields, the gluon system (partition function) is broken down into a fast and a slow subsystem. In one of these subsystems, nonperturbative fields play the role of rapidly changing (stochastic) degrees of freedom, whose effect on soft gluons is considered within the cluster-expansion method for Green's functions in the nonperturbative background field. The effect of long-wavelength chromomagnetic fluctuations on hard gluons is studied within adiabatic perturbation theory in the strong-field approximation. The pressure and the energy density of SU(2) gluon matter are calculated for T > T{sub c}.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

NuLattice: Ab initio computations of atomic nuclei on lattices

Here, we introduce NuLattice, a Python software package for ab initio computations of atomic nuclei on lattices. The computational tools consist of Hartree Fock, the coupled cluster method, the in-medium similarity renormalization group, and full configuration interaction. At present, the employed interactions are from pion-less effective field theory at leading order and consist of two-body and three-body contacts. We present results for light nuclei 2 H, 3,4 He, 8 Be, 12 C, and 16 O. NuLattice algorithms exploit the sparsity and locality of lattice interactions, and as a result computations can be run on laptops.

Rothman, Maxwell [Univ. of Tennessee, Knoxville, T↗

A Framework for Load Balancing of Tensor Contraction Expressions via Dynamic Task Partitioning

In this paper, we introduce the Dynamic Load-balanced Tensor Contractions (DLTC), a domain-specific library for efficient task parallel execution of tensor contraction expressions, a class of computation encountered in quantum chemistry and physics. Our framework decomposes each contraction into smaller unit of tasks, represented by an abstraction referred to as iterators. We exploit an extra level of parallelism by having tasks across independent contractions executed concurrently through a dynamic load balancing run- time. We demonstrate the improved performance, scalability, and flexibility for the computation of tensor contraction expressions on parallel computers using examples from coupled cluster methods.

Tensor contraction, domain-specific language, dyna↗

Autotuning algorithmic choice for input sensitivity

A daunting challenge faced by program performance autotuning is input sensitivity, where the best autotuned configuration may vary with different input sets. This paper presents a novel two-level input learning algorithm to tackle the challenge for an important class of autotuning problems, algorithmic autotuning. Here, the new approach uses a two-level input clustering method to automatically refine input grouping, feature selection, and classifier construction. Its design solves a series of open issues that are particularly essential to algorithmic autotuning, including the enormous optimization space, complex influence by deep input features, high cost in feature extraction, and variable accuracy of algorithmic choices. Experimental results show that the new solution yields up to a 3x speedup over using a single configuration for all inputs, and a 34x speedup over a traditional one-level method for addressing input sensitivity in program optimizations.

97 MATHEMATICS AND COMPUTING↗

Insights on sources and formation mechanisms of liquid-bearing clouds over MOSAiC examined from a Lagrangian framework

Understanding Arctic stratiform liquid-bearing cloud life cycles and properly representing these life cycles in models is crucial for evaluations of cloud feedbacks as well as the faithfulness of climate projections for this rapidly warming region. Examination of cloud life cycles typically requires analyses of cloud evolution and origins on short time scales, on the order of hours to several days. Measurements from the recent Multidisciplinary drifting Observatory for the Study of Arctic Climate (MOSAiC) expedition provide a unique view of the current state of the central Arctic over an annual cycle. Here, we use the MOSAiC radiosonde measurements to detect liquid-bearing cloud layers over full atmospheric columns and to examine the cloud-generating air masses’ properties. We perform 5-day (120 h) back-trajectory calculations for every detected cloud and cluster them using a unique set of variables extracted from these trajectories informed by ERA5 reanalysis data. This clustering method enables us to separate between the air mass source regions such as ice-covered Arctic and midlatitude open water. We find that moisture intrusions into the central Arctic typically result in multilayer liquid-bearing cloud structures and that more than half of multilayer profiles include overlying liquid-bearing clouds originating in different types of air masses. Finally, we conclude that Arctic cloud formation via prolonged radiative cooling of elevated stable subsaturated air masses circulating over the Arctic can occur frequently (up to 20% of detected clouds in the sounding data set) and may lead to a significant impact of ensuing clouds on the surface energy budget, including net surface warming in some cases.

54 ENVIRONMENTAL SCIENCES↗