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 577 records · Page 32

Shipper/receiver difference verification of spent fuel by use of PDET

Spent fuel storage pools in most countries are rapidly approaching their design limits with the discharge of over 10,000 metric tons of heavy metal from global reactors. Countries like UK, France or Japan have adopted a closed fuel cycle by reprocessing spent fuel and recycling MOX fuel while many other countries opted for above ground interim dry storage for their spent fuel management strategy. Some countries like Finland and Sweden are already well on the way to setting up a conditioning plant and a deep geological repository for spent fuel. For all these situations, shipments of spent fuel are needed and the number of these shipments is expected to increase significantly. Although shipper/receiver difference (SRD) verification measurements are needed by IAEA when the recipient facility receives spent fuel, these are not being practiced to the level that IAEA has desired due to lack of a credible measurement methodology and instrument that can reliably perform these measurements to verify non-diversion of spent fuel during shipment and confirm facility operator declarations on the spent fuel. In this paper, we describe a new safeguards method and an associated instrument, Partial Defect Tester (PDET), which can detect pin diversion from Pressurized Water Reactor (PWR) Spent Fuel Assemblies in an in-situ condition. The PDET uses multiple tiny neutron and gamma detectors in the form of a cluster and a simple, yet highly precise, gravity-driven system to obtain underwater radiation measurements inside a Pressurized Water Reactor (PWR) spent fuel assembly. The method takes advantage of the PWR fuel design which contains multiple guide tubes which can be accessed from the top. The data obtained in such a manner can provide spatial distribution of neutron and gamma flux within a spent fuel assembly. Our simulation study as well as validation measurements indicated that the ratio of the gamma signal to the thermal neutron signal at each detector location normalized to the peak ratio of all the detector locations gives a unique signature that is sensitive to missing pins. The signature is principally dependent on the geometry of the detector locations, and little sensitive to enrichment or burn-up variations. A small variation in the fuel bundle, such as a few missing pins, changes the shape of the signature to enable detection. After verification of the non-diversion of spent fuel pins, the neutron signal and gamma signal are subsequently used to verify the consistency of the operator declaration on the fuel burn-up and cooling time. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

The coupling of fluids, dynamics, and controls on advanced architecture computers

This grant provided for the demonstration of coupled controls, body dynamics, and fluids computations in a workstation cluster environment; and an investigation of the impact of peer-peer communication on flow solver performance and robustness. The findings of these investigations were documented in the conference articles.The attached publication, 'Towards Distributed Fluids/Controls Simulations', documents the solution and scaling of the coupled Navier-Stokes, Euler rigid-body dynamics, and state feedback control equations for a two-dimensional canard-wing. The poor scaling shown was due to serialized grid connectivity computation and Ethernet bandwidth limits. The scaling of a peer-to-peer communication flow code on an IBM SP-2 was also shown. The scaling of the code on the switched fabric-linked nodes was good, with a 2.4 percent loss due to communication of intergrid boundary point information. The code performance on 30 worker nodes was 1.7 (mu)s/point/iteration, or a factor of three over a Cray C-90 head. The attached paper, 'Nonlinear Fluid Computations in a Distributed Environment', documents the effect of several computational rate enhancing methods on convergence. For the cases shown, the highest throughput was achieved using boundary updates at each step, with the manager process performing communication tasks only. Constrained domain decomposition of the implicit fluid equations did not degrade the convergence rate or final solution. The scaling of a coupled body/fluid dynamics problem on an Ethernet-linked cluster was also shown.

Atwood, Christopher↗

Structure in the 3D Galaxy Distribution: III. Fourier Transforming the Universe: Phase and Power Spectra

We demonstrate the effectiveness of a relatively straightforward analysis of the complex 3D Fourier transform of galaxy coordinates derived from redshift surveys. Numerical demonstrations of this approach are carried out on a volume-limited sample of the Sloan Digital Sky Survey redshift survey. The direct unbinned transform yields a complex 3D data cube quite similar to that from the Fast Fourier Transform (FFT) of finely binned galaxy positions. In both cases deconvolution of the sampling window function yields estimates of the true transform. Simple power spectrum estimates from these transforms are roughly consistent with those using more elaborate methods. The complex Fourier transform characterizes spatial distributional properties beyond the power spectrum in a manner different from (and we argue is more easily interpreted than) the conventional multi-point hierarchy. We identify some threads of modern large scale inference methodology that will presumably yield detections in new wider and deeper surveys.

Galactic clusters↗

Nucleon clustering at kinetic freezeout of heavy-ion collisions via path-integral Monte Carlo

Clustering of the four-nucleon system at kinetic freezeout conditions is studied using path-integral Monte Carlo techniques. This method seeks to improve upon previous calculations which relied on approximate semiclassical methods or few-body quantum mechanics. Estimates are given for the decay probabilities of the 4N system into various light nuclei decay channels and the strength of spatial correlations is characterized. Additionally, a simple model is presented to describe the impact of this clustering on nucleon multiplicity distributions. Additionally, the effects of a possible modification of the inter-nucleon interaction due to the close critical line (and hypothetical QCD critical point) on the clustering are also studied.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Dissociation of migrating particle from trap with long-range interaction field

Some lattice defects, such as dislocations, interact with migrating species, e.g. vacancies, interstitial atoms and their clusters, via long-range strain fields. In this paper, an equation for the mean dissociation time of a migrating particle from a trap is derived in terms of the potential well function for the interaction energy. The distribution of dissociation times is studied by the Monte Carlo method, and the problem of particle exchange between spatially separated traps is considered.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Scale-up Unlearnable Examples Learning with High-performance Computing

Recent advancements in AI models, like ChatGPT, are structured to retain user interactions, which could inadvertently include sensitive healthcare data. In the healthcare field, particularly when radiologists use AI-driven diagnostic tools hosted on online platforms, there is a risk that medical imaging data may be repurposed for future AI training without explicit consent, spotlighting critical privacy and intellectual property concerns around healthcare data usage. Addressing these privacy challenges, a novel approach known as Unlearnable Examples (UEs) has been introduced, aiming to make data unlearnable to deep learning models. A prominent method within this area, called Unlearnable Clustering (UC), has shown improved UE performance with larger batch sizes but was previously limited by computational resources (e.g., a single workstation). To push the boundaries of UE performance with theoretically unlimited resources, we scaled up UC learning across various datasets using Distributed Data Parallel (DDP) training on the Summit supercomputer. Our goal was to examine UE efficacy at high-performance computing (HPC) levels to prevent unauthorized learning and enhance data security, particularly exploring the impact of batch size on UE’s unlearnability. Utilizing the robust computational capabilities of the Summit, extensive experiments were conducted on diverse datasets such as Pets, MedMNist, Flowers, and Flowers102. Our findings reveal that both overly large and overly small batch sizes can lead to performance instability and affect accuracy. However, the relationship between batch size and unlearnability varied across datasets, highlighting the necessity for tailored batch size strategies to achieve optimal data protection. The use of Summit’s high-performance GPUs, along with the efficiency of the DDP framework, facilitated rapid updates of model parameters and consistent training across nodes. Our results underscore the critical role of selecting appropriate batch sizes based on the specific characteristics of each dataset to prevent learning and ensure data security in deep learning applications. The source code is publicly available at https: // github. com/ hrlblab/ UE_ HPC .

Zhu, Yanfan [Vanderbilt University, Nashville, TN,↗

A Deep Reinforcement Learning-based Reserve Optimization in Active Distribution Systems for Tertiary Frequency Regulation

Federal Energy Regulatory Commission (FERC)Orders 841 and 2222 have recommended that distributed energy resources (DERs) should participate in energy and reserve markets; therefore, a mechanism needs to be developed to facilitate DERs’ participation at the distribution level. Although the available reserve from a single distribution system may not be sufficient for tertiary frequency regulation, stacked and coordinated contributions from several distribution systems can enable them participate in tertiary frequency regulation at scale. This paper proposes a deep reinforcement learning (DRL)-based approach for optimization of requested aggregated reserves by system operators among the clusters of DERs. The co-optimization of cost of reserve, distribution network loss, and voltage regulation of the feeders are considered while optimizing the reserves among participating DERs. The proposed framework adopts deep deterministic policy gradient (DDPG), which is an algorithm based on an actor-critic method. The effectiveness of the proposed method for allocating reserves among DERs is demonstrated through case studies on a modified IEEE 34-node distribution system.

deep reinforcement learning, distributed energy re↗

NUMERICAL MODELING OF THE COAGULATION AND POROSITY EVOLUTION OF DUST AGGREGATES

Porosity evolution of dust aggregates is crucial in understanding dust evolution in protoplanetary disks. In this study, we present useful tools to study the coagulation and porosity evolution of dust aggregates. First, we present a new numerical method for simulating dust coagulation and porosity evolution as an extension of the conventional Smoluchowski equation. This method follows the evolution of the mean porosity for each aggregate mass simultaneously with the evolution of the mass distribution function. This method reproduces the results of previous Monte Carlo simulations with much less computational expense. Second, we propose a new collision model for porous dust aggregates on the basis of our N-body experiments on aggregate collisions. As the first step, we focus on 'hit-and-stick' collisions, which involve neither compression nor fragmentation of aggregates. We first obtain empirical data on porosity changes between the classical limits of ballistic cluster-cluster and particle-cluster aggregation. Using the data, we construct a recipe for the porosity change due to general hit-and-stick collisions as well as formulae for the aerodynamical and collisional cross sections. Our collision model is thus more realistic than a previous model of Ormel et al. based on the classical aggregation limits only. Simple coagulation simulations using the extended Smoluchowski method show that our collision model explains the fractal dimensions of porous aggregates observed in a full N-body simulation and a laboratory experiment. By contrast, similar simulations using the collision model of Ormel et al. result in much less porous aggregates, meaning that this model underestimates the porosity increase upon unequal-sized collisions. Besides, we discover that aggregates at the high-mass end of the distribution can have a considerably small aerodynamical cross section per unit mass compared with aggregates of lower masses. This occurs when aggregates drift under uniform acceleration (e.g., gravity) and their collision is induced by the difference in their terminal velocities. We point out an important implication of this discovery for dust growth in protoplanetary disks.

79 ASTRONOMY AND ASTROPHYSICS↗

HIGH ANGULAR RESOLUTION INTEGRAL-FIELD SPECTROSCOPY OF THE GALAXY'S NUCLEAR CLUSTER: A MISSING STELLAR CUSP?

We report on the structure of the nuclear star cluster in the innermost 0.16 pc of the Galaxy as measured by the number density profile of late-type giants. Using laser guide star adaptive optics in conjunction with the integral field spectrograph, OSIRIS, at the Keck II telescope, we are able to differentiate between the older, late-type (approx 1 Gyr) stars, which are presumed to be dynamically relaxed, and the unrelaxed young (approx 6 Myr) population. This distinction is crucial for testing models of stellar cusp formation in the vicinity of a black hole, as the models assume that the cusp stars are in dynamical equilibrium in the black hole potential. In the survey region, we classified 60 stars as early-type (22 newly identified) and 74 stars as late-type (61 newly identified). We find that contamination from young stars is significant, with more than twice as many young stars as old stars in our sensitivity range (K' < 15.5) within the central arcsecond. Based on the late-type stars alone, the surface stellar number density profile, SIGMA(R) propor to R {sup -G}AMMA, is flat, with GAMMA = -0.27 +- 0.19. Monte Carlo simulations of the possible de-projected volume density profile, n(r) propor tor {sup -g}amma, show that gamma is less than 1.0 at the 99.7% confidence level. These results are consistent with the nuclear star cluster having no cusp, with a core profile that is significantly flatter than that predicted by most cusp formation theories, and even allows for the presence of a central hole in the stellar distribution. Of the possible dynamical interactions that can lead to the depletion of the red giants observable in this survey-stellar collisions, mass segregation from stellar remnants, or a recent merger event-mass segregation is the only one that can be ruled out as the dominant depletion mechanism. The lack of a stellar cusp around a supermassive black hole would have important implications for black hole growth models and inferences on the presence of a black hole based upon stellar distributions.

79 ASTRONOMY AND ASTROPHYSICS↗

Exciton-driven highly hyperthermal O-atom desorption from nanostructured CaO

We report qualitatively new highly-hyperthermal (HHT) oxygen atom emission from nanostructured CaO excited by 6.4 eV nanosecond laser pulses. The kinetic energy distribution of emitted O-atoms peaks at 0.7 eV, which is over four times greater than previously observed. Excitation of MgO and CaO nanostructures with UV laser pulses is known to result in thermal and hyperthermal emission of oxygen atoms when photons with energies below and above the band gap, respectively, are used. The highly energetic atomic desorption we observe, following bulk excitation, challenges the conventional view that bulk excitation can only induce thermal desorption. Using density functional theory and an embedded cluster method, we propose a mechanism for this HHT feature based on the interaction of surface holes with bulk excitons. These experimental and theoretical results suggest that specific atomic desorption mechanisms in wide-bandgap materials can be controlled by selective electronic excitation of not only the surface but also the bulk of these materials

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

ERTS-1 imagery and native plant distributions

A method is developed for using ERTS spectral signature data to determine plant community distribution and phenology without resolving individual plants. An Exotech ERTS radiometer was used near ground level to obtain spectral signatures for a desert plant community, including two shrub species, ground covered with live annuals in April and dead ones in June, and bare ground. It is shown that comparisons of scene types can be made when spectral signatures are expressed as a ratio of red reflectivity to IR reflectivity or when they are plotted as red reflectivity vs. IR reflectivity, in which case the signature clusters of each component are more distinct. A method for correcting and converting the ERTS radiance values to reflectivity values for comparison with ground truth data is appended.

Musick, H. B.↗

Systematic calculation of {alpha} decay within a generalized density-dependent cluster model

We perform an extensive investigation on {alpha} decays in both even-A and odd-A nuclei within the new version of the generalized density-dependent cluster model. The microscopic deformed potential is numerically constructed in the double-folding model by the multipole expansion method. The coupled-channel effect resulting from nuclear deformation is included by using the coupled-channel Schroedinger equation with outgoing wave boundary conditions. The fine structure observed in {alpha} decay is well described by taking into account the angular momentum of the emitted {alpha} particle and the Boltzmann distribution of excitation spectrum in daughter nuclei. A good agreement between experiment and theory is achieved, and the results of our calculations are discussed in detail, together with the sensitivity of the calculated half-lives and branching ratios to some physical quantities used in the calculations.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Beyond the single-file fluid limit using transfer matrix method: Exact results for confined parallel hard squares

We extend the transfer matrix method of one-dimensional hard core fluids placed between confining walls for that case where the particles can pass each other and at most two layers can form. We derive an eigenvalue equation for a quasi-one-dimensional system of hard squares confined between two parallel walls, where the pore width is between σ and 3σ (σ is the side length of the square). The exact equation of state and the nearest neighbor distribution functions show three different structures: a fluid phase with one layer, a fluid phase with two layers, and a solid-like structure where the fluid layers are strongly correlated. The structural transition between differently ordered fluids develops continuously with increasing density, i.e., no thermodynamic phase transition occurs. The high density structure of the system consists of clusters with two layers which are broken with particles staying in the middle of the pore.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Clustering of nonvalently bonded NO{sub 2}…O{sub 2}N fragments at C(sp{sup 3}) atoms

Based on the 'invariom' approach and the 'Atoms in Molecules' theory, a simplified method is proposed to reveal and analyze bonding interatomic interactions using large arrays of structural data. The technique was tested taking nonvalent interactions between nitro groups bonded to C(sp{sup 3}) atoms as examples. Geometric clustering of structural fragments was performed and the interaction strength and covalent contribution distributions were analyzed. Relationships between the fragment clustering results and nonvalent bonding pattern features are discussed.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

pGraph: Efficient Parallel Construction of Large-Scale Protein Sequence Homology Graphs

Detecting sequence homology between protein sequences is a fundamental problem in computational molecular biology, with a pervasive application in nearly all analyses that aim to structurally and functionally characterize protein molecules. While detecting the homology between two protein sequences is relatively inexpensive, detecting pairwise homology for a large number of protein sequences can become computationally prohibitive for modern inputs, often requiring millions of CPU hours. Yet, there is currently no robust support to parallelize this kernel. In this paper, we identify the key characteristics that make this problemparticularly hard to parallelize, and then propose a new parallel algorithm that is suited for detecting homology on large data sets using distributed memory parallel computers. Our method, called pGraph, is a novel hybrid between the hierarchical multiple-master/worker model and producer-consumer model, and is designed to break the irregularities imposed by alignment computation and work generation. Experimental results show that pGraph achieves linear scaling on a 2,048 processor distributed memory cluster for a wide range of inputs ranging from as small as 20,000 sequences to 2,560,000 sequences. In addition to demonstrating strong scaling, we present an extensive report on the performance of the various system components and related parametric studies.

Environmental Molecular Sciences Laboratory↗

Non-equilibrium phase distribution in an Al-SiC composite

The phase distribution in an Al-SiC composite has been investigated using high resolution analytical electron microscopy. Particular attention was focused on Al-SiC interfaces, matrix boundaries and impurity phases which would impede the easy glide of dislocations. Small crystallites of MgO were distributed singly and in clusters along Al-SiC interfaces in all specimens. Interfacial segregation and precipitation involving alloy species apparently affected precipitation in the matrix, where the distribution of phases was found to be very heterogeneous. Matrix phases also included unusually large constituent particles and dispersoids, a consequence of the composite processing methods. The relationship between the observed microstructure and the composite mechanical behavior reported by others is discussed. The heterogeneous distribution of matrix phases is expected to result in a wide variaiton in local yield stress and local work-hardening rate within the composite.

Nutt, S. R.↗

Implementing Molecular Dynamics on Hybrid High Performance Computers - Particle-Particle Particle-Mesh

The use of accelerators such as graphics processing units (GPUs) has become popular in scientific computing applications due to their low cost, impressive floating-point capabilities, high memory bandwidth, and low electrical power requirements. Hybrid high-performance computers, machines with nodes containing more than one type of floating-point processor (e.g. CPU and GPU), are now becoming more prevalent due to these advantages. In this paper, we present a continuation of previous work implementing algorithms for using accelerators into the LAMMPS molecular dynamics software for distributed memory parallel hybrid machines. In our previous work, we focused on acceleration for short-range models with an approach intended to harness the processing power of both the accelerator and (multi-core) CPUs. To augment the existing implementations, we present an efficient implementation of long-range electrostatic force calculation for molecular dynamics. Specifically, we present an implementation of the particle-particle particle-mesh method based on the work by Harvey and De Fabritiis. We present benchmark results on the Keeneland InfiniBand GPU cluster. We provide a performance comparison of the same kernels compiled with both CUDA and OpenCL. We discuss limitations to parallel efficiency and future directions for improving performance on hybrid or heterogeneous computers.

97 MATHEMATICS AND COMPUTING↗

Influence of Asymmetric Cyclic Loading on Structural Evolution and Deformation Behavior of Cu-5 at.% Zr Alloy: An Atomistic Simulation-Based Study

Molecular dynamics (MD) simulation-based studies of tensile test and structural evolution of Cu-5 at.% Zr alloy under asymmetric cyclic loading (i.e., ratcheting behavior) considering various stress ratios such as − 0.2, − 0.4 and − 0.6 for different temperatures, viz.≈ 100, 300 and 500 K have been performed using embedded atom model Finnis–Sinclair potential. According to obtained stress–strain response from MD calculation, Cu-5 at.% Zr alloy specimen is pristine in nature as sudden drop in stress just after yield stress and subsequent elastic type deformation are observed for this alloy. Predicted ratcheting strain by MD simulation for Cu-5 at.% Zr alloy varies from 4.5 to 5%. Significant increase in ratcheting strain has been observed with the increase in temperature. Slight reduction in crystallinity is identified at the middle of the each loading cycle from the performed radial distribution function analysis and cluster analysis.

36 MATERIALS SCIENCE↗