SEARCH · Search NASA
Results for “Ensemble computing.”
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.
Enhanced analysis of experimental x-ray spectra through deep learning
X-ray spectroscopic data from high-energy-density laser-produced plasmas has long required thorough, time-consuming analysis to extract meaningful source conditions. There are often confounding factors due to rapidly evolving states and finite spatial gradients (e.g., the existence of multi-temperature, multi-density, multi-ionization states, etc.) that make spectral measurements and analysis difficult. Here, in this paper, we demonstrate how deep learning can be applied to enhance x-ray spectral data analysis in both speed and intricacy. Neural networks (NNs) are trained on ensemble atomic physics simulations so that they can subsequently construct a model capable of extracting plasma parameters directly from experimental spectra. Through deep learning, the models can extract temperature distributions as opposed to single or dual temperature/density fits from standard trial-and-error atomic modeling at a significantly reduced computational cost compared to traditional trial-and-error methods. These NNs are envisioned to be deployed with high repetition rate x-ray spectrometers in order to provide detailed real-time analysis of experimental spectra.
Charmed and light pseudoscalar meson decay constants from four-flavor lattice QCD with physical light quarks
We compute the leptonic decay constants f D+ , f Ds , and f K+ , and the quark-mass ratios m c =m s and m s =m l in unquenched lattice QCD using the experimentally determined value of f π+ for normalization. We use the MILC Highly Improved Staggered Quark (HISQ) ensembles with four dynamical quark flavors -- up, down, strange, and charm -- and with both physical and unphysical values of the light sea-quark masses. The use of physical pions removes the need for a chiral extrapolation, thereby eliminating a significant source of uncertainty in previous calculations. Four different lattice spacing ranging from a ≈ 0:06 fm to 0:15 fm are included in the analysis to control the extrapolation to the continuum limit. Our primary results are f D+ = 212:6(0:4)$(^{+1.0}_{-1.2})$ MeV, f Ds = 249:0(0:3)$(^{+1.1}_{-1.5})$ MeV, and f Ds /f D+ = 1:1712(10)$(^{+29}_{-32})$, where the errors are statistical and total systematic, respectively. The errors on our results for the charm decay constants and their ratio are approximately two to four times smaller than those of the most precise previous lattice calculations. We also obtain f K+ / f π+ = 1:1956(10)$(^{+26}_{-18})$, updating our previous result, and determine the quark-mass ratios m s /m l = 27:35(5)$(^{+10}_{-7})$ and m c /m s = 11:747(19)$(^{+59}_{-43})$. When combined with experimental measurements of the decay rates, our results lead to precise determinations of the CKM matrix elements !Vus! = 0:22487(51)(29)(20)(5), !Vcd! = 0:217(1)(5)(1) and !Vcs! = 1:010(5)(18)(6), where the errors are from this calculation of the decay constants, the uncertainty in the experimental decay rates, structure-dependent electromagnetic corrections, and, in the case of !Vus!, the uncertainty in |Vud|, respectively.
Improved lattice computation of proton decay matrix elements
In this paper, we present an improved result for the lattice computation of the proton decay matrix elements in N f = 2 + 1 QCD. In this study, by adopting the error reduction technique of all-mode-averaging, a significant improvement of the statistical accuracy is achieved for the relevant form factor of proton (and also neutron) decay on the gauge ensemble of N f = 2 + 1 domain-wall fermions with m π = 0.34 – 0.69 GeV on a 2.7 fm 3 lattice, as used in our previous work. We improve the total accuracy of matrix elements to 10–15% from 30–40% for p → πe + or from 20–40% for p → K$\bar{ν}$. The accuracy of the low-energy constants α and β in the leading-order baryon chiral perturbation theory (BChPT) of proton decay are also improved. The relevant form factors of p → π estimated through the “direct” lattice calculation from the three-point function appear to be 1.4 times smaller than those from the “indirect” method using BChPT with α and β . It turns out that the utilization of our result will provide a factor 2–3 larger proton partial lifetime than that obtained using BChPT. Lastly, we also discuss the use of these parameters in a dark matter model.
Oak Ridge Institutional Cluster Autotune Test Drive Report
The Oak Ridge Institutional Cluster (OIC) provides general purpose computational resources for the ORNL staff to run computation heavy jobs that are larger than desktop applications but do not quite require the scale and power of the Oak Ridge Leadership Computing Facility (OLCF). This report details the efforts made and conclusions derived in performing a short test drive of the cluster resources on Phase 5 of the OIC. EnergyPlus was used in the analysis as a candidate user program and the overall software environment was evaluated against anticipated challenges experienced with resources such as the shared memory-Nautilus (JICS) and Titan (OLCF). The OIC performed within reason and was found to be acceptable in the context of running EnergyPlus simulations. The number of cores per node and the availability of scratch space per node allow non-traditional desktop focused applications to leverage parallel ensemble execution. Although only individual runs of EnergyPlus were executed, the software environment on the OIC appeared suitable to run ensemble simulations with some modifications to the Autotune workflow. From a standpoint of general usability, the system supports common Linux libraries, compilers, standard job scheduling software (Torque/Moab), and the OpenMPI library (the only MPI library) for MPI communications. The file system is a Panasas file system which literature indicates to be an efficient file system.
Applications of the worldline Monte Carlo formalism in quantum mechanics
Highlights: Universal approach to the numerical computation of the Euclidean path integral. Inspired by recent work in relativistic quantum field theory. Here adapted to non-relativistic quantum mechanics. Worked out for the computation of propagators and ground-state energies. Special smoothing procedure for singular potentials. -- Abstract: In recent years efficient algorithms have been developed for the numerical computation of relativistic single-particle path integrals in quantum field theory. Here, we adapt this “worldline Monte Carlo” approach to the standard problem of the numerical approximation of the non-relativistic path integral, resulting in a formalism whose characteristic feature is the fast, non-recursive generation of an ensemble of trajectories that is independent of the potential, and is thus universally applicable. The numerical implementation discretises the trajectories with respect to their time parametrisation but maintains a continuous spatial domain. In the case of singular potentials, the discretised action gets adapted to the singularity through a “smoothing” procedure. We show for a variety of examples (the harmonic oscillator in various dimensions, the modified Pöschl–Teller potential, delta-function potentials, the Coulomb and Yukawa potentials) that the method allows one to obtain fast and reliable estimates for the Euclidean propagator and use them in a certain time window suitable for extracting the ground state energy. Usually Monte Carlo calculations in quantum mechanics are performed using algorithms that involve a more specific adaptation to the potential, and we briefly compare our method with the most widely used such algorithms, which are of Metropolis type. As an aside, we apply it for studying the classical limit where nearly classical trajectories are expected to dominate in the path integral. We expect the advances made here to be useful also in the relativistic case.
Stochastic generation of explicit pore structures by thresholding Gaussian random fields
We provide a description and computational investigation of an efficient method to stochastically generate realistic pore structures. Smolarkiewicz and Winter introduced this specific method in pores resolving simulation of Darcy flows (Smolarkiewicz and Winter, 2010 [1]) without giving a complete formal description or analysis of the method, or indicating how to control the parameterization of the ensemble. We address both issues in this paper. The method consists of two steps. First, a realization of a correlated Gaussian field, or topography, is produced by convolving a prescribed kernel with an initial field of independent, identically distributed random variables. The intrinsic length scales of the kernel determine the correlation structure of the topography. Next, a sample pore space is generated by applying a level threshold to the Gaussian field realization: points are assigned to the void phase or the solid phase depending on whether the topography over them is above or below the threshold. Hence, the topology and geometry of the pore space depend on the form of the kernel and the level threshold. Manipulating these two user prescribed quantities allows good control of pore space observables, in particular the Minkowski functionals. Extensions of the method to generate media with multiple pore structures and preferential flow directions are also discussed. To demonstrate its usefulness, the method is used to generate a pore space with physical and hydrological properties similar to a sample of Berea sandstone. -- Graphical abstract: -- Highlights: •An efficient method to stochastically generate realistic pore structures is provided. •Samples are generated by applying a level threshold to a Gaussian field realization. •Two user prescribed quantities determine the topology and geometry of the pore space. •Multiple pore structures and preferential flow directions can be produced. •A pore space based on Berea sandstone is generated.
Volume of the set of unistochastic matrices of order 3 and the mean Jarlskog invariant
A bistochastic matrix B of size N is called unistochastic if there exists a unitary U such that B{sub ij}=|U{sub ij}|{sup 2} for i,j=1,...,N. The set U{sub 3} of all unistochastic matrices of order N=3 forms a proper subset of the Birkhoff polytope, which contains all bistochastic (doubly stochastic) matrices. We compute the volume of the set U{sub 3} with respect to the flat (Lebesgue) measure and analytically evaluate the mean entropy of an unistochastic matrix of this order. We also analyze the Jarlskog invariant J, defined for any unitary matrix of order three, and derive its probability distribution for the ensemble of matrices distributed with respect to the Haar measure on U(3) and for the ensemble which generates the flat measure on the set of unistochastic matrices. For both measures the probability of finding |J| smaller than the value observed for the Cabbibo-Kobayashi-Maskawa matrix, which describes the violation of the CP parity, is shown to be small. Similar statistical reasoning may also be applied to the Maki-Nakagawa-Sakata matrix, which plays role in describing the neutrino oscillations. Some conjectures are made concerning analogous probability measures in the space of unitary matrices in higher dimensions.
Importance sampling large deviations in nonequilibrium steady states. I
Large deviation functions hold information on the stability and response of systems driven into nonequilibrium steady states and in such a way are similar to free energies for systems at equilibrium. As with equilibrium free energies, evaluating large deviation functions numerically for all but the simplest systems is difficult because by construction they depend on exponentially rare events. In this first paper of a series, we assess different trajectory-based sampling methods capable of computing large deviation functions of time integrated observables within nonequilibrium steady states. We highlight some convergence criteria and best practices using a number of different models, including a biased Brownian walker, a driven lattice gas, and a model of self-assembly. We show how two popular methods for sampling trajectory ensembles, transition path sampling and diffusion Monte Carlo, suffer from exponentially diverging correlations in trajectory space as a function of the bias parameter when estimating large deviation functions. Improving the efficiencies of these algorithms requires introducing guiding functions for the trajectories.
Classical spin glass system in external field with taking into account relaxation effects
We study statistical properties of disordered spin systems under the influence of an external field with taking into account relaxation effects. For description of system the spatial 1D Heisenberg spin-glass Hamiltonian is used. In addition, we suppose that interactions occur between nearest-neighboring spins and they are random. Exact solutions which define angular configuration of the spin in nodes were obtained from the equations of stationary points of Hamiltonian and the corresponding conditions for the energy local minimum. On the basis of these recurrent solutions an effective parallel algorithm is developed for simulation of stabile spin-chains of an arbitrary length. It is shown that by way of an independent order of N{sup 2} numerical simulations (where N is number of spin in each chain) it is possible to generate ensemble of spin-chains, which is completely ergodic which is equivalent to full self-averaging of spin-chains' vector polarization. Distributions of different parameters (energy, average polarization by coordinates, and spin-spin interaction constant) of unperturbed system are calculated. In particular, analytically is proved and numerically is shown, that for the Heisenberg nearest-neighboring Hamiltonian model, the distribution of spin-spin interaction constants as opposed to widely used Gauss-Edwards-Anderson distribution satisfies Levy alpha-stable distribution law. This distribution is nonanalytic function and does not have variance. In the work we have in detail studied critical properties of an ensemble depending on value of external field parameters (from amplitude and frequency) and have shown that even at weak external fields the spin-glass systemis strongly frustrated. It is shown that frustrations have fractal behavior, they are selfsimilar and do not disappear at scale decreasing of area. By the numerical computation is shown that the average polarization of spin-glass on a different coordinates can have values which can lead to catastrophes in the equation ofClausius-Mossotti for dielectric constant. In other words, for some values of external field parameter, a critical phenomenon occurs in the system which is impossible to describe by the real-valued Heisenberg spin-glass Hamiltonian. For the solution of this problem at first the complex-valued disordered Hamiltonian is used. Physically this type of extension of Hamiltonian allows to consider relaxation effects which occur in the system under the influence of an external field. On the basis of developed approach an effective parallel algorithm is developed for simulation of statistic parameters of spin-glass system under the influence of an external field.
Finite Element Analysis System Workflow Tools
A collection of MATLAB functions and class definitions called System Workflow Tools (SWFT) are available to semi-automate steps in the simulation process. Some of these steps are often simple and routine for smaller finite element models, but if done directly by an analyst can quickly become labor intensive, cumbersome, and error prone for larger, system level models. Some of SWFT’s capabilities demonstrated in this report includes writing Sierra input decks and processing Quantities of Interest (QOI) from results files. SWFT also writes scripts in order to utilize other software programs such as Cubit (separating system level CAD into subassemblies and components, creating nodesets and sidesets), DAKOTA (ensemble management), and ParaView (contour plots and animations). Detailed commands and workflows from mesh generation to report generation are provided as examples for analysts to utilize SWFT capabilities.
Temporal Coarse Graining for Classical Stochastic Noise in Quantum Systems
Simulations of quantum systems with Hamiltonian classical stochastic noise can be challenging when the noise exhibits temporal correlations over a multitude of time scales, such as for 1/f noise in solid-state quantum information processors. Here we present an approach for simulating Hamiltonian classical stochastic noise that performs temporal coarse-graining by effectively integrating out the high-frequency components of the noise. We focus on the case where the stochastic noise can be expressed as a sum of Ornstein-Uhlenbeck processes. Temporal coarse-graining is then achieved by conditioning the stochastic process on a coarse realization of the noise, expressing the conditioned stochastic process in terms of a sum of smooth, deterministic functions and bridge processes with boundaries fixed at zero, and performing the ensemble average over the bridge processes. For Ornstein-Uhlenbeck processes, the deterministic components capture all dependence on the coarse realization, and the stochastic bridge processes are not only independent but taken from the same distribution with correlators that can be expressed analytically, allowing the associated noise propagators to be precomputed once for all simulations. This combination of noise trajectories on a coarse time grid and ensemble averaging over bridge processes has practical advantages, such as a simple concatenation rule, that we highlight with numerical examples.
Fault location in High Voltage Multi-terminal dc Networks Using Ensemble Learning
Precise location of faults for large distance power transmission networks is essential for faster repair and restoration process. High Voltage direct current (HVdc) networks using modular multi-level converter (MMC) technology has found its prominence for interconnected multi-terminal networks. This allows for large distance bulk power transmission at lower costs. However, they cope with the challenge of dc faults. Fast and efficient methods to isolate the network under dc faults have been widely studied and investigated. After successful isolation, it is essential to precisely locate the fault. The post-fault voltage and current signatures are a function of multiple factors and thus accurately locating faults on a multi-terminal network is challenging. In this paper, we discuss a novel data-driven ensemble learning based approach for accurate fault location. Here we utilize the eXtreme Gradient Boosting (XGB) method for accurate fault location. The sensitivity of the proposed algorithm to measurement noise, fault location, resistance and current limiting inductance are performed on a radial three-terminal MTdc network designed in Power System Computer Aided Design (PSCAD)/Electromagnetic Transients including dc (EMTdc).
Diagnosing isopycnal diffusivity in an eddying, idealized midlatitude ocean basin via Lagrangian, in Situ, Global, High-Performance Particle Tracking (LIGHT)
Isopycnal diffusivity due to stirring by mesoscale eddies in an idealized, wind-forced, eddying, midlatitude ocean basin is computed using Lagrangian, in Situ, Global, High-Performance Particle Tracking (LIGHT). Simulation is performed via LIGHT within the Model for Prediction across Scales Ocean (MPAS-O). Simulations are performed at 4-, 8-, 16-, and 32-km resolution, where the first Rossby radius of deformation (RRD) is approximately 30 km. Scalar and tensor diffusivities are estimated at each resolution based on 30 ensemble members using particle cluster statistics. Each ensemble member is composed of 303 665 particles distributed across five potential density surfaces. Diffusivity dependence upon model resolution, velocity spatial scale, and buoyancy surface is quantified and compared with mixing length theory. The spatial structure of diffusivity ranges over approximately two orders of magnitude with values of O(10 5 ) m 2 s –1 in the region of western boundary current separation to O(10 3 ) m 2 s –1 in the eastern region of the basin. Dominant mixing occurs at scales twice the size of the first RRD. Model resolution at scales finer than the RRD is necessary to obtain sufficient model fidelity at scales between one and four RRD to accurately represent mixing. Mixing length scaling with eddy kinetic energy and the Lagrangian time scale yield mixing efficiencies that typically range between 0.4 and 0.8. In conclusion, a reduced mixing length in the eastern region of the domain relative to the west suggests there are different mixing regimes outside the baroclinic jet region.
Improving the accuracy of freight mode choice models: A case study using the 2017 CFS PUF data set and ensemble learning techniques
Here, the US Census Bureau has collected two rounds of experimental data from the Commodity Flow Survey, providing shipment-level characteristics of nationwide commodity movements, published in 2012 (i.e., Public Use Microdata) and in 2017 (i.e., Public Use File). With this information, data-driven methods have become increasingly valuable for understanding detailed patterns in freight logistics. In this study, we used the 2017 Commodity Flow Survey Public Use File data set to explore building a high-performance freight mode choice model, considering three main improvements: (1) constructing local models for each separate commodity/industry category; (2) extracting useful geographical features, particularly the derived distance of each freight mode between origin/destination zones; and (3) applying additional ensemble learning methods such as stacking or voting to combine results from local and unified models for improved performance. The proposed method achieved over 92% accuracy without incorporating external information, an over 19% increase compared to directly fitting Random Forests models over 10,000 samples. Furthermore, SHAP (Shapely Additive Explanations) values were computed to explain the outputs and major patterns obtained from the proposed model. The model framework could enhance the performance and interpretability of existing freight mode choice models.
Engineering Molecular Transformations for Sustainable Energy Conversion
Future strategies for sustainable energy production will undoubtedly require processes and materials that can efficiently convert renewable resources into fuels. Nature’s enzymes can exquisitely integrate highly active catalytic centers within flexible environments that can adaptively guide reactants to products with very high activities and selectivities. They are limited, however, by their stability and ability to integrate into large scale production processes. The design of more robust heterogeneous catalytic materials that mimic the performance of enzymes, however, has been hindered by our limited understanding of how such transformations proceed. The tremendous advances in ab initio quantum mechanical methods, atomistic simulations, and high performance computing that have occurred over the past two decades, however, provide unprecedented ability to track molecular transformations and how they proceed at specific sites and within particular environments. This information together with the advances in in situ spectroscopic methods that follow such transformations can begin to enable the design of atomic surface ensembles and nanoscale reaction environments. This paper provides the author’s perspective on how theory and simulation can be used to move from current onedimensional design efforts based on catalytic descriptors to the design of two-dimensional surfaces, threedimensional reaction environments, and proton-coupled electron transfer systems that mimic enzymes in the transformation of molecules.
Nonadiabatic dynamics with classical trajectories: The problem of an initial coherent superposition of electronic states
Advances in coherent light sources and development of pump–probe techniques in recent decades have opened the way to study electronic motion in its natural time scale. When an ultrashort laser pulse interacts with a molecular target, a coherent superposition of electronic states is created and the triggered electron dynamics is coupled to the nuclear motion. A natural and computationally efficient choice to simulate this correlated dynamics is a trajectory-based method where the quantum-mechanical electronic evolution is coupled to a classical-like nuclear dynamics. These methods must approximate the initial correlated electron–nuclear state by associating an initial electronic wavefunction to each classical trajectory in the ensemble. Different possibilities exist that reproduce the initial populations of the exact molecular wavefunction when represented in a basis. We show that different choices yield different dynamics and explore the effect of this choice in Ehrenfest, surface hopping, and exact-factorization-based coupled-trajectory schemes in a one-dimensional two-electronic-state model system that can be solved numerically exactly. Furthermore, this work aims to clarify the problems that standard trajectory-based techniques might have when a coherent superposition of electronic states is created to initialize the dynamics, to discuss what properties and observables are affected by different choices of electronic initial conditions and to point out the importance of quantum-momentum-induced electronic transitions in coupled-trajectory schemes.