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 919 records · Page 51

How Well Can Quantum Embedding Method Predict the Reaction Profiles for Hydrogenation of Small Li Clusters?

Quantum computing leverages the principles of quantum mechanics in novel ways to tackle complex chemistry problems that cannot be accurately addressed using traditional quantum chemistry methods. However, the high computational cost and available number of physical qubits with high fidelity limit its application to small chemical systems. This work employed a quantum-classical framework which features a quantum active space-embedding approach to perform simulations of chemical reactions that require up to 14 qubits. This framework was applied to prototypical example metal hydrogenation reactions: the coupling between hydrogen and Li 2 , Li 3 , and Li 4 clusters. Particular attention was paid to the computation of barriers and reaction energies. The predicted reaction profiles compare well with advanced classical quantum chemistry methods, demonstrating the potential of the quantum embedding algorithm to map out reaction profiles of realistic gas-phase chemical reactions to ascertain qualitative energetic trends. Additionally, the predicted potential energy curves provide a benchmark to compare against both current and future quantum embedding approaches.

36 MATERIALS SCIENCE↗

Theoretical research program to predict the properties of molecules and clusters containing transition metal atoms

The primary focus of this research has been the theoretical study of transition metal (TM) chemistry. A major goal of this work is to provide reliable information about the interaction of H atoms with iron metal. This information is needed to understand the effect of H atoms on the processes of embrittlement and crack propagation in iron. The method in the iron hydrogen studies is the cluster method in which the bulk metal is modelled by a finite number of iron atoms. There are several difficulties in the application of this approach to the hydrogen iron system. First the nature of TM-TM and TM-H bonding for even diatomic molecules was not well understood when these studies were started. Secondly relatively large iron clusters are needed to provide reasonable results.

Walch, S.↗

Bergman-type medium range order in amorphous Zr77Rh23 alloy studied by ab initio molecular dynamics simulations

In recent years, some arguments about the existence of medium-range order (MRO) in the Zr-Rh system have been put forward. However, research on the structural features of the Zr-Rh binary alloy at the atomic level is still lacking. This study uses ab initio molecular dynamics simulations to systematically study the local structures of Zr77Rh23 from the liquid to the glassy states. Pair correlation function (PCF), coordination number (CN), Honeycutt–Anderson(HA) index, bond-angle distribution functions, and the Voronoi tessellation method are used to reveal a clear icosahedral-like configuration in the amorphous Zr77Rh23 alloy. It is noteworthy that the splitting in the second peak of the partial PDF implies the existence of a medium range order (MRO) in the Zr77Rh23 system. We obtain the local order in three-dimensional atomic density distributions by using a new atomistic cluster alignment (ACA) method. Interestingly, a Bergman-type MRO is observed in the glassy Zr77Rh23. Furthermore, the spatial distribution and connections of icosahedral-like clusters are shown to further demonstrate the MRO network. Our findings shed light on the nature of atomic local structures of amorphous Zr77Rh23 alloy and have important implications to understanding the formation of various MROs in metallic glasses.

Guo, Y. R.↗

Density functional investigation of intermolecular effects on {sup 13}C NMR chemical-shielding tensors modeled with molecular clusters

A quantum-chemical method for modeling solid-state nuclear magnetic resonance chemical-shift tensors by calculations on large symmetry-adapted clusters of molecules is demonstrated. Four hundred sixty five principal components of the {sup 13}C chemical-shielding tensors of 24 organic materials are analyzed. The comparison of calculations on isolated molecules with molecules in clusters demonstrates that intermolecular effects can be successfully modeled using a cluster that represents a local portion of the lattice structure, without the need to use periodic-boundary conditions (PBCs). The accuracy of calculations which model the solid state using a cluster rivals the accuracy of calculations which model the solid state using PBCs, provided the cluster preserves the symmetry properties of the crystalline space group. The size and symmetry conditions that the model cluster must satisfy to obtain significant agreement with experimental chemical-shift values are discussed. The symmetry constraints described in the paper provide a systematic approach for incorporating intermolecular effects into chemical-shielding calculations performed at a level of theory that is more advanced than the generalized gradient approximation. Specifically, NMR parameters are calculated using the hybrid exchange-correlation functional B3PW91, which is not available in periodic codes. Calculations on structures of four molecules refined with density plane waves yield chemical-shielding values that are essentially in agreement with calculations on clusters where only the hydrogen sites are optimized and are used to provide insight into the inherent sensitivity of chemical shielding to lattice structure, including the role of rovibrational effects.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Integrated Approach to Density-Based Spatial Clustering of Applications with Noise and Dynamic Time Warping for Breakout Prediction in Slab Continuous Casting

Mold breakout is a catastrophic accident that has serious impacts on smooth production, slab quality, and caster equipment. Accurate identification and prediction of an impending breakout are always top priorities in continuous casting operations. In view of crucial common features of mold copper plate temperatures during a breakout, such as time lag and space inversion, the concepts of density-based spatial clustering of applications with noise and dynamic time warping are introduced, and an integrated novel method for breakout prediction is developed. Through extracting and fusing the representative singularity and approximation of temperature variation, the typical temporal and spatial temperature characteristics during breakout can be distinguished and predicted accurately. Compared with traditional methods of logical judgment and artificial neural network, the method based on clustering does not need to modify forecast thresholds or parameters artificially, which overcomes the limitation of model dependence on human beings, and demonstrates excellent adaptability and robustness for online abnormality prevention.

36 MATERIALS SCIENCE↗

Cloud field classification based upon high spatial resolution textural features. II - Simplified vector approaches

This paper compares the results of cloud-field classification derived from two simplified vector approaches, the Sum and Difference Histogram (SADH) and the Gray Level Difference Vector (GLDV), with the results produced by the Gray Level Cooccurrence Matrix (GLCM) approach described by Welch et al. (1988). It is shown that the SADH method produces accuracies equivalent to those obtained using the GLCM method, while the GLDV method fails to resolve error clusters. Compared to the GLCM method, the SADH method leads to a 31 percent saving in run time and a 50 percent saving in storage requirements, while the GLVD approach leads to a 40 percent saving in run time and an 87 percent saving in storage requirements.

Chen, D. W.↗

Geochemical Signatures as a Tool for Vermiculite Provenance Determination

Thirty-eight samples of known origin (China, Libby MT, South Africa, South Carolina) and 6 vermiculite product samples of unknown origin were analyzed for major and trace elements, including rare earth elements to determine the feasibility of distinguishing the provenance of the samples based upon a geochemical signature. Probability plots suggest that two of the four groups (Libby, South Carolina) were comprised of two subgroups. Results of hierarchical cluster analysis are highly sensitive to the linkage method chosen. Ward’s method is the most useful for this data and suggests that there are five groups within the data set (South African samples, two subsets of the Libby samples, a subset of the South Carolina samples, and a second subset of the South Carolina samples combined with the China samples). Similar results were obtained using k-cluster analysis. Neither clustering method was able to distinguish samples from China from the South Carolina samples. Discriminant analysis was used on a four-category model comprised of the original four groups and on a six-category model comprised of the five categories identified from the cluster analysis but with the China samples grouped into a sixth category. The discriminant/classification model was able to distinguish all of the groups including the China samples from one another for both the four- and six-category models with 100% of the samples properly classified. The 6 unknown product samples were classified with a probability of consistency of 99%. Both discriminant models were also run with a subset of our analyte set to be consistent with the smaller Gunter et al., (2005) analyte set. Twenty vermiculite samples (nine of known origin and eleven of unknown origin) from their study were classified based on our discriminant models with the reduced set of analytes. Of the twenty samples, Gunter et al. (2005) was able to classify 16 with cluster analysis while our 4-category discriminant analysis model allowed us to classify all twenty. Of the 16 samples Gunter et al. (2005) classified using cluster analysis, all but one sample was assigned the same classification by our 4-category model. Of the nine samples with known origin, all were correctly classified. Similar results were obtained for the six-category model. Comparison of the plots of the canonical roots, the Wilks’ L, and the square Mahalanobis distances suggest the full analyte set provides better discrimination of the groups than the reduced analyte set. The six-category model is more consistent with the results of the probability plots and the cluster analysis. Discriminant analysis of geochemical data from vermiculite ore is a powerful technique for determining the ore’s provenance.

58 GEOSCIENCES↗

Improving Dark Energy Constraints Using Low-Redshift Large-Scale Structures

The primary goal of this project was to improve constraints on dark energy measurements by improving our ability to extract cosmological information from low redshift large-scale structures. PI Clowe's project's primary aim was to reduce the bias in measurements of the masses of clusters of galaxies to a level where the evolution of the cluster mass function can be used in the Vera Rubin Observatory's Legacy Survey of Space and Time Dark Energy Science Collaboration survey to improve the accuracy of the measurement of dark energy and other cosmological parameters. Co-PI Seo's project studied observational systematics affecting large-scale clustering of galaxies, which will be used to improve dark energy constraints from the Dark Energy Spectroscopic Instrument (DESI). The cluster lensing project employed a series of simulations and observations of clusters of galaxies to test numerous potential systematic errors in cluster mass measurements using weak gravitational lensing as the accuracy of current weak lensing measurements are more than order of magnitude worse than what is required to use clusters of galaxies for accurate determination of dark energy parameters. PI Clowe and group developed and analyzed simulations to test for, and correct biases introduced in, the weak lensing measurement process. Finally, PI Clowe and group developed a method of detecting clusters using galaxy overdensities and applied the method to the BLISS and DES surveys. The success of spectroscopic dark energy mission such as the extended Baryon Oscillation Spectroscopic Survey (eBOSS) and the Dark Energy Spectroscopic Instrument (DESI) will depend on a thorough understanding of various observational systematics in the target density fluctuations that would give rise to spurious, non-cosmological signals. PI Seo and group developed a deep learning, artificial neural network (ANN) technique that modeled and mitigated such effects, aimed at deriving more robust galaxy clustering signals not only for the baryon acoustic oscillation feature and redshift-space distortions but also for primordial non-Gaussianity constraint.

79 ASTRONOMY AND ASTROPHYSICS↗

DEDICATED FILTER FOR DEFECTS CLUSTERING IN RADIOGRAPHIC IMAGE

Defect clusters such as linear or clustered porosity are in some cases even more important than single flaws. This paper presents two methods of defect clustering and algorithm for calculation of distances between flaws in digital radiographic image. Dedicated lookup table based filter is used for calculation of distances between objects in the specified range. For defect clustering two functions were developed. First one is based on MMD (Minimum Mean Distance) algorithm. Second one uses hierarchical procedures for clustering defects of various types, shapes and size.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Construction of ground-state preserving sparse lattice models for predictive materials simulations

First-principles based cluster expansion models are the dominant approach in ab initio thermodynamics of crystalline mixtures enabling the prediction of phase diagrams and novel ground states. However, despite recent advances, the construction of accurate models still requires a careful and time-consuming manual parameter tuning process for ground-state preservation, since this property is not guaranteed by default. In this paper, we present a systematic and mathematically sound method to obtain cluster expansion models that are guaranteed to preserve the ground states of their reference data. The method builds on the recently introduced compressive sensing paradigm for cluster expansion and employs quadratic programming to impose constraints on the model parameters. The robustness of our methodology is illustrated for two lithium transition metal oxides with relevance for Li-ion battery cathodes, i.e., Li 2x Fe 2(1-x) O 2 and Li 2x Ti 2(1-x) O 2 , for which the construction of cluster expansion models with compressive sensing alone has proven to be challenging. We demonstrate that our method not only guarantees ground-state preservation on the set of reference structures used for the model construction, but also show that out-of-sample ground-state preservation up to relatively large supercell size is achievable through a rapidly converging iterative refinement. This method provides a general tool for building robust, compressed and constrained physical models with predictive power.

36 MATERIALS SCIENCE↗

A simulation of the intracluster medium with feedback from cluster galaxies

We detail method and report first results from a three-dimensional hydrodynamical and N-body simulation of the formation and evolution of a Coma-sized cluster of galaxies, with the intent of studying the history of the hot, X-ray emitting intracluster medium. Cluster gas, galaxies, and dark matter are included in the model. The galaxies and dark matter fell gravitational forces; the cluster gas also undergoes hydrodynamical effects such as shock heating and PdV work. For the first time in three dimensions, we include modeling of ejection of processed gas from the simulated galaxies by winds, including heating and heavy element enrichment. For comparison, we employ a `pure infall' simulation using the same initial conditions but with no galaxies or winds. We employ an extreme ejection history for galactic feedback in order to define the boundary of likely models. As expected, feedback raises the entropy of the intracluster gas, preventing it from collapsing to densities as high as those attained in the infall model. The effect is more pronounced in subclusters formed at high redshift. The cluster with feedback is always less X-ray luminous, but experiences more rapid luminosity evolution, than the pure infall cluster. Even employing an extreme ejection model, the final gas temperature is only approximately 15% larger than in the infall model. The radial temperature profile is very nearly isothermal within 1.5 Mpc. The cluster galaxies in the feedback model have a velocity dispersion approximately 15% lower than the dark matter. This results in the true ratio of specific energies in galaxies to gas being less than one, beta(sub spec) approximately 0.7. The infall model predicts beta(sub spec) approximately 1.2. Large excursions in these values occur over time, following the complex dynamical history of the cluster. The morphology of the X-ray emission is little affected by feedback. The emission profiles of both clusters are well described by the standard beta-model with beta(sub fit) approximately equal to 0.7 - 0.9. X-ray mass estimates based on the assumptions of hydrostatic equilibrium and the applicability of the beta-model are quite accurate in both cases. A strong, radial iron abundance gradient is present, which develops as a consequence of the steepening of the galaxy density profile over time. Spectroscopic observations using nonimaging detectors with wide (approximately 45 min) fields of view dramatically smear the gradient. Observations with arcminute resolution, made available with the ASCA satellite, would readily resolve the gradient.

Metzler, Christopher A.↗

Polynomial Scaling Localized Active Space Unitary Selective Coupled Cluster Singles and Doubles

We present a polynomial-scaling algorithm for the localized active space unitary selective coupled cluster singles and doubles (LAS-USCCSD) method. In this approach, cluster excitations are selected based on a threshold ϵ determined by the absolute gradients of the LAS-UCCSD energy with respect to cluster amplitudes. Using the generalized Wick’s theorem for multireference wave functions, we derive the gradient expression as a polynomial function of one-, two-, and three-body reduced density matrices and 1- and 2-electron integrals, valid for any multireference wave function. The resulting gradient implementation exhibits a memory scaling of 𝒪(N 6 ), with N spin orbitals in the combined active space of all fragments. The variational quantum eigensolver is used to optimize the selected cluster excitations on a quantum simulator. Furthermore, by plotting the energy error, defined as the difference between the LAS-USCCSD and corresponding CASCI energies, against the inverse cluster amplitude selection threshold (ϵ –1 ) for polyene chains containing 2 to 5 π-bond units, we establish a relationship between the energy error and the threshold. To further validate the accuracy of LAS-USCCSD, we computed the cis–trans isomerization energy of stilbene (a 20-qubit system) and the magnetic coupling constant of the tris-hydroxo-bridged chromium dimer [Cr 2 (OH) 3 (NH 3 ) 6 ] 3+ (evaluated as both 12- and 20-qubit systems) using the Qiskit-Qulacs simulator. Assessing such examples is important to determine the practical feasibility of quantum simulations for chemically realistic systems. Toward this goal, with the LAS-USCCSD algorithm we estimated the quantum resources required for simulating an active space of (30e,22o) in [Cr 2 (OH) 3 (NH 3 ) 6 ] 3+ , a size that remains beyond the reach of current quantum simulators for accurate treatment.

Algorithms↗

Robust Group Subspace Recovery: A New Approach for Multi-Modality Data Fusion

Robust Subspace Recovery (RoSuRe) algorithm was recently introduced as a principled and numerically efficient algorithm that unfolds underlying Unions of Subspaces (UoS) structure, present in the data. The union of Subspaces (UoS) is capable of identifying more complex trends in data sets than simple linear models. In this work, we build on and extend RoSuRe to prospect the structure of different data modalities individually. We propose a novel multi-modal data fusion approach based on group sparsity which we refer to as Robust Group Subspace Recovery (RoGSuRe). Relying on a bi-sparsity pursuit paradigm and non-smooth optimization techniques, the introduced framework learns a new joint representation of the time series from different data modalities, respecting an underlying UoS model. We subsequently integrate the obtained structures to form a unified subspace structure. The proposed approach exploits the structural dependencies between the different modalities data to cluster the associated target objects. The resulting fusion of the unlabeled sensors’ data from experiments on audio and magnetic data has shown that our method is competitive with other state of the art subspace clustering methods. The resulting UoS structure is employed to classify newly observed data points, highlighting the abstraction capacity of the proposed method.

47 OTHER INSTRUMENTATION↗

Permutation-adapted complete and independent basis for atomic cluster expansion descriptors

Atomic cluster expansion (ACE) methods provide a systematic way to describe particle local environments of arbitrary body order. For practical applications it is often required that the basis of cluster functions be symmetrized with respect to rotations and permutations. Existing methodologies yield sets of symmetrized functions that are over-complete. These methodologies thus require an additional numerical procedure, such as singular value decomposition (SVD), to eliminate redundant functions. In this work, it is shown that analytical linear relationships for subsets of cluster functions may be derived using recursion and permutation properties of generalized Wigner symbols. From these relationships, subsets (blocks) of cluster functions can be selected such that, within each block, functions are guaranteed to be linearly independent. It is conjectured that this block-wise independent set of permutation-adapted rotation and permutation invariant (PA-RPI) functions forms a complete, independent basis for ACE. Along with the first analytical proofs of block-wise linear dependence of ACE cluster functions and other theoretical arguments, numerical results are offered to demonstrate this. The utility of the method is demonstrated in the development of an ACE interatomic potential for tantalum. Using the new basis functions in combination with Bayesian compressive sensing sparse regression, some high degree descriptors are observed to persist and help achieve high-accuracy models.

Angular momentum↗

Peeking Network States with Clustered Patterns

Network traffic monitoring has long been a core element for effec- tive network management and security. However, it is still a chal- lenging task with a high degree of complexity for comprehensive analysis when considering multiple variables and ever-increasing traffic volumes to monitor. For example, one of the widely con- sidered approaches is to scrutinize probabilistic distributions, but it poses a scalability concern and multivariate analysis is not gen- erally supported due to the exponential increase of the complexity. In this work, we propose a novel method for network traffic moni- toring based on clustering, one of the powerful deep-learning tech- niques. We show that the new approach enables us to recognize clustered results as patterns representing the network states, which can then be utilized to evaluate “similarity” of network states over time. In addition, we define a new quantitative measure for the similarity between two compared network states observed in dif- ferent time windows, as a supportive means for intuitive analysis. Finally, we demonstrate the clustering-based network monitoring with public traffic traces, and show that the proposed approach us- ing the clustering method has a great opportunity for feasible, cost- effective network monitoring.

97 MATHEMATICS AND COMPUTING↗

Interpreting experimental measurements of helium bubbles using stochastic cluster dynamics models of heterogeneous nucleation and growth in irradiated ferritic alloys

Among a number of other advantageous features, ferritic/martensitic steels are being considered as fusion reactor structural materials due to their low intrinsic swelling under irradiation. However, under high-energy neutron irradiation, He produced through (n, α) reactions stabilizes vacancy clusters, which then act as seeds for bubble formation and growth, negating the intrinsic swelling resistance of these alloys. Standard models of irradiation damage accumulation and microstructural evolution consider homogeneous nucleation as the basis for bubble formation and growth. However, this generally does not explain the large bubble densities and sizes observed experimentally under a number of different conditions. Here, we propose a more realistic physical model of bubble nucleation, formation, and growth designed to capture recent experimental measurements of He-bubble formation and evolution during co-implantation of He and Fe ions in model ferritic alloys. We find that experimental results are explained only when the following three features are considered simultaneously: (i) heterogeneous nucleation of He-vacancy bubbles at defect sinks (e.g., dislocations, grain boundaries, and second-phase precipitates), (ii) nucleation and growth of bubbles via the ‘trap mutation’ mechanism (i.e., spontaneous production of Frenkel pairs due to absorption of He atoms), and (iii) transition from a growth-limited, He-stabilized bubble regime to a ‘runaway’ void-type growth scenario. The model is implemented into a stochastic cluster dynamics framework that takes advantage of cluster size grouping methods to accelerate the simulations, allowing us to reach 10 dpa of simulated irradiated dose, and to capture cluster sizes in excess of 20 nm. Further, a careful extrapolation exercise conducted assuming classical nucleation theory leads to excellent agreement with the experimental measurements at 50 dpa of irradiation.

36 MATERIALS SCIENCE↗

The redshift evolution of the mean temperature, pressure, and entropy profiles in 80 SPT-selected galaxy clusters

We present the results of an X-ray analysis of 80 galaxy clusters selected in the 2500 deg{sup 2} South Pole Telescope survey and observed with the Chandra X-ray Observatory. We divide the full sample into subsamples of ∼20 clusters based on redshift and central density, performing a joint X-ray spectral fit to all clusters in a subsample simultaneously, assuming self-similarity of the temperature profile. This approach allows us to constrain the shape of the temperature profile over 0 < r < 1.5R {sub 500}, which would be impossible on a per-cluster basis, since the observations of individual clusters have, on average, 2000 X-ray counts. The results presented here represent the first constraints on the evolution of the average temperature profile from z = 0 to z = 1.2. We find that high-z (0.6 < z < 1.2) clusters are slightly (∼30%) cooler both in the inner (r < 0.1R {sub 500}) and outer (r > R {sub 500}) regions than their low-z (0.3 < z < 0.6) counterparts. Combining the average temperature profile with measured gas density profiles from our earlier work, we infer the average pressure and entropy profiles for each subsample. Confirming earlier results from this data set, we find an absence of strong cool cores at high z, manifested in this analysis as a significantly lower observed pressure in the central 0.1R {sub 500} of the high-z cool-core subset of clusters compared to the low-z cool-core subset. Overall, our observed pressure profiles agree well with earlier lower-redshift measurements, suggesting minimal redshift evolution in the pressure profile outside of the core. We find no measurable redshift evolution in the entropy profile at r ≲ 0.7R {sub 500}—this may reflect a long-standing balance between cooling and feedback over long timescales and large physical scales. We observe a slight flattening of the entropy profile at r ≳ R {sub 500} in our high-z subsample. This flattening is consistent with a temperature bias due to the enhanced (∼3×) rate at which group-mass (∼2 keV) halos, which would go undetected at our survey depth, are accreting onto the cluster at z ∼ 1. This work demonstrates a powerful method for inferring spatially resolved cluster properties in the case where individual cluster signal-to-noise is low, but the number of observed clusters is high.

79 ASTRONOMY AND ASTROPHYSICS↗

X-Ray Emission from a Simulated Cluster of Galaxies

Using the 1993 cluster simulation of Katz & White, we analyze the intracluster medium and investigate the accuracy of the standard hydrostatic method for determining cluster masses. We show that the simulated cluster gas is in hydrostatic equilibrium with a subsonic flow toward the center. Inside a radius of (approx.) 100 kpc, this flow is in a steady state. The cooling time is shorter than a Hubble time within the central 50 kpc. The flow rate is regulated by the gas sink in the middle of the cluster and the PdV work done as the gas flows in, verifying the standard cooling flow scenario. We simulate observations of the cluster using the instrument parameters of the EXOSAT ME detector and the Einstein IPC detector. Even though the intracluster gas is not isothermal, isothermal models of the cluster, excluding regions within 100 kpc of galaxies, fit the EXOSAT X-ray spectra as well as they fit real clusters. The X- ray surface brightness distribution is similar to that of real clusters, again excluding the galaxies. We simulate the procedure used to determine the masses of real clusters. We use the equation of hydrostatic equilibrium together with the temperature derived from an isothermal fit to the simulated EXOSAT spectrum and the density profile derived from a fit to the simulated IPC surface brightness profile to determine the mass. A comparison of the derived mass profile to the actual mass profile shows that errors of a factor of 2 are possible. If the actual temperature profile is used, the cluster mass is found to an accuracy of better than 25% within the virial radius.

Tsai, John C.↗