Search NASA⌕ Search

SEARCH · Search NASA

Results for “dimensional reduction”

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 253 records · Page 14

Emergence of charge density waves and a pseudogap in single-layer TiTe 2

Two-dimensional materials constitute a promising platform for developing nanoscale devices and systems. Their physical properties can be very different from those of the corresponding three-dimensional materials because of extreme quantum confinement and dimensional reduction. Here in this paper we report a study of TiTe 2 from the single-layer to the bulk limit. Using angle-resolved photoemission spectroscopy and scanning tunneling microscopy and spectroscopy, we observed the emergence of a (2 × 2) charge density wave order in single-layer TiTe 2 with a transition temperature of 92 ± 3 K. Also observed was a pseudogap of about 28 meV at the Fermi level at 4.2 K. Surprisingly, no charge density wave transitions were observed in two-layer and multi-layer TiTe 2 , despite the quasi-two-dimensional nature of the material in the bulk. The unique charge density wave phenomenon in the single layer raises intriguing questions that challenge the prevailing thinking about the mechanisms of charge density wave formation.

36 MATERIALS SCIENCE↗

Canonical analysis of electrodynamics in 5D

We review the canonical analysis of Maxwell's theory in five dimensions (5D). In the first part we summarize the canonical analysis of the free theory in Minkowski space-time. As is well known the theory has two first class constraints and therefore two gauge conditions are needed in order to fix the gauge completely. In particular we discuss the properties of the axial gauge. In the second part we discuss the theory in a different 5D background. We consider a space-time in which the fifth dimension is a circle (Kaluza-Klein dimensional reduction). We show that the axial gauge is not consistent with the periodic boundary conditions in the fifth coordinate. As a consequence it is necessary to introduce a different gauge, the 'almost axial gauge'.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Machine-learned closure of URANS for stably stratified turbulence: connecting physical timescales & data hyperparameters of deep time-series models

Stably stratified turbulence (SST), a model that is representative of the turbulence found in the oceans and atmosphere, is strongly affected by fine balances between forces and becomes more anisotropic in time for decaying scenarios. Moreover, there is a limited understanding of the physical phenomena described by some of the terms in the Unsteady Reynolds-Averaged Navier–Stokes (URANS) equations—used to numerically simulate approximate solutions for such turbulent flows. Rather than attempting to model each term in URANS separately, it is attractive to explore the capability of machine learning (ML) to model groups of terms, i.e. to directly model the force balances. We develop deep time-series ML for closure modeling of the URANS equations applied to SST. We consider decaying SST which are homogeneous and stably stratified by a uniform density gradient, enabling dimensionality reduction. We consider two time-series ML models: long short-term memory and neural ordinary differential equation. Both models perform accurately and are numerically stable in a posteriori (online) tests. Furthermore, we explore the data requirements of the time-series ML models by extracting physically relevant timescales of the complex system. We find that the ratio of the timescales of the minimum information required by the ML models to accurately capture the dynamics of the SST corresponds to the Reynolds number of the flow. The current framework provides the backbone to explore the capability of such models to capture the dynamics of high-dimensional complex dynamical system like SST flows.

97 MATHEMATICS AND COMPUTING↗

Geometry and energy of non-Abelian vortices

We study pure Yang-Mills theory on {Sigma}xS{sup 2}, where {Sigma} is a compact Riemann surface, and invariance is assumed under rotations of S{sup 2}. It is well known that the self-duality equations in this setup reduce to vortex equations on {Sigma}. If the Yang-Mills gauge group is SU(2), the Bogomolny vortex equations of the Abelian Higgs model are obtained. For larger gauge groups, one generally finds vortex equations involving several matrix-valued Higgs fields. Here we focus on Yang-Mills theory with gauge group SU(N)/Z{sub N} and a special reduction which yields only one non-Abelian Higgs field. One of the new features of this reduction is the fact that while the instanton number of the theory in four dimensions is generally fractional with denominator N, we still obtain an integral vortex number in the reduced theory. We clarify the relation between these two topological charges at a bundle geometric level. Another striking feature is the emergence of nontrivial lower and upper bounds for the energy of the reduced theory on {Sigma}. These bounds are proportional to the area of {Sigma}. We give special solutions of the theory on {Sigma} by embedding solutions of the Abelian Higgs model into the non-Abelian theory, and we relate our work to the language of quiver bundles, which has recently proved fruitful in the study of dimensional reduction of Yang-Mills theory.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Reduced-order modeling of advection-dominated systems with recurrent neural networks and convolutional autoencoders

A common strategy for the dimensionality reduction of nonlinear partial differential equations (PDEs) relies on the use of the proper orthogonal decomposition (POD) to identify a reduced subspace and the Galerkin projection for evolving dynamics in this reduced space. However, advection-dominated PDEs are represented poorly by this methodology since the process of truncation discards important interactions between higher-order modes during time evolution. In this study, we demonstrate that encoding using convolutional autoencoders (CAEs) followed by a reduced-space time evolution by recurrent neural networks overcomes this limitation effectively. We demonstrate that a truncated system of only two latent space dimensions can reproduce a sharp advecting shock profile for the viscous Burgers equation with very low viscosities, and a six-dimensional latent space can recreate the evolution of the inviscid shallow water equations. Additionally, the proposed framework is extended to a parametric reduced-order model by directly embedding parametric information into the latent space to detect trends in system evolution. Furthermore, our results show that these advection-dominated systems are more amenable to low-dimensional encoding and time evolution by a CAE and recurrent neural network combination than the POD-Galerkin technique.

97 MATHEMATICS AND COMPUTING↗

Intrusive analysis for NEK5000: development of intrusive uncertainty quantification for high-dimensional, high-fidelity codes.

We study the use of lower-fidelity training data for uncertainty quantification of complex simulation models. In our approach, computationally expensive full-model outputs are approximated applying proper orthogonal-decomposition-based dimensionality reduction to the full model. A Gaussian-processes-based machine learning approach is then used to model the difference (or, rather, the correspondence) between the higher-fidelity and the lower-fidelity data. This stochastic model can be constructed by using few additional full code evaluations; it is used to calibrate arbitrarily many lower-quality outputs in order to create additional training data for sampling-based analysis. In effect, an adequate approximation of the simulation model’s response to uncertainty can be constructed at a modest computational cost. In fact, we aim at the number of full-model evaluations comparable to that required for a linear approximation (as opposed to numbers traditionally associated with sampling in high-dimensional spaces). In this report, we explain the basic algorithm, suggest some performance tuning options, and give a first characterization for the class of simulation models of nuclear engineering for which the suggested report is effective.

Roderick, O.↗

A novel coupling of noise reduction algorithms for particle flow simulations

Proper orthogonal decomposition (POD) and its extension based on time-windows have been shown to greatly improve the effectiveness of recovering smooth ensemble solutions from noisy particle data. However, to successfully de-noise any molecular system, a large number of measurements still need to be provided. In order to achieve a better efficiency in processing time-dependent fields, we have combined POD with a well-established signal processing technique, wavelet-based thresholding. In this novel hybrid procedure, the wavelet filtering is applied within the POD domain and referred to as WAVinPOD. The algorithm exhibits promising results when applied to both synthetically generated signals and particle data. In this work, the simulations compare the performance of our new approach with standard POD or wavelet analysis in extracting smooth profiles from noisy velocity and density fields. Numerical examples include molecular dynamics and dissipative particle dynamics simulations of unsteady force- and shear-driven liquid flows, as well as phase separation phenomenon. Simulation results confirm that WAVinPOD preserves the dimensionality reduction obtained using POD, while improving its filtering properties through the sparse representation of data in wavelet basis. This paper shows that WAVinPOD outperforms the other estimators for both synthetically generated signals and particle-based measurements, achieving a higher signal-to-noise ratio from a smaller number of samples. The new filtering methodology offers significant computational savings, particularly for multi-scale applications seeking to couple continuum informations with atomistic models. It is the first time that a rigorous analysis has compared de-noising techniques for particle-based fluid simulations.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

PhILMs: Collaboratory on Mathematics and Physics-Informed Learning Machines for Multiscale and Multiphysics Problems

The landscape of computational science and engineering is continually evolving, with the challenge of high-dimensional regression problems standing as a significant hurdle in numerous scientific endeavors. Addressing this challenge, our research, funded by this award, has led to the development of an innovative computational framework known as Probabilistic Partition of Unity Networks (PPOU-Nets). This initiative represents a collaborative effort to harness the potential of mathematics and physics-informed machine learning in tackling multiscale and multiphysics problems prevalent in high-dimensional spaces. Through this work, we have proposed a novel methodology that seamlessly integrates adaptive dimensionality reduction and a mixture of experts model, thereby facilitating a more efficient and accurate approximation of complex functions. This research effort has not only advanced the state of computational science but also opened new avenues for exploration in quantum computing and beyond. This report outlines the motivation, methodology, key findings, and implications of our work, underscoring our contributions to the broader scientific community and the potential pathways for future research.

97 MATHEMATICS AND COMPUTING↗

Understanding latent timescales in neural ordinary differential equation models of advection-dominated dynamical systems

The neural ordinary differential equation (ODE) framework has shown considerable promise in recent years in developing highly accelerated surrogate models for complex physical systems characterized by partial differential equations (PDEs). For PDE-based systems, state-of-the-art neural ODE strategies leverage a two-step procedure to achieve this acceleration: a nonlinear dimensionality reduction step provided by an autoencoder, and a time integration step provided by a neural-network based model for the resultant latent space dynamics (the neural ODE). This work explores the applicability of such autoencoder-based neural ODE strategies for PDEs in which advection terms play a critical role. More specifically, alongside predictive demonstrations, physical insight into the sources of model acceleration (i.e., how the neural ODE achieves its acceleration) is the scope of the current study. Such investigations are performed by quantifying the effects of both autoencoder and neural ODE components on latent system time-scales using eigenvalue analysis of dynamical system Jacobians. To this end, the sensitivity of various critical training parameters – de-coupled versus end-to-end training, latent space dimensionality, and the role of training trajectory length, for example – to both model accuracy and the discovered latent system timescales is quantified. Furthermore, this work specifically uncovers the key role played by the training trajectory length (the number of rollout steps in the loss function during training) on the latent system timescales: larger trajectory lengths correlate with an increase in limiting neural ODE time-scales, and optimal neural ODEs are found to recover the largest time-scales of the full-order (ground-truth) system. Demonstrations are performed across fundamentally different unsteady fluid dynamics configurations influenced by advection: (1) the Kuramoto–Sivashinsky equations (2) Hydrogen-Air channel detonations (the compressible reacting Navier–Stokes equations with detailed chemistry), and (3) 2D Atmospheric flow.

Advection-dominated dynamical systems↗

Black hole in the expanding universe from intersecting branes

We study physical properties and global structures of a time-dependent, spherically symmetric solution obtained via the dimensional reduction of intersecting M-branes. We find that the spacetime describes a maximally charged black hole which asymptotically tends to the Friedmann-Lemaitre-Robertson-Walker universe filled by a stiff matter. The metric solves the field equations of the Einstein-Maxwell-dilaton system, in which four Abelian gauge fields couple to the dilation with different coupling constants. The spacetime satisfies the dominant energy condition and is characterized by two parameters, Q and {tau}, related to the Maxwell charge and the relative ratio of black-hole horizon radii, respectively. In spite of the nontrivial time dependence of the metric, it turns out that the black-hole event horizon is a Killing horizon. This unexpected symmetry may be ascribed to the fact that the 11-dimensional brane configurations are supersymmetric in the static limit. Finally, combining with laws of the trapping horizon, we discuss the thermodynamic properties of the black hole. It is shown that the horizon possesses a nonvanishing temperature, contrary to the extremal Reissner-Nordstroem solution.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

N=1 Wess-Zumino model in d=3 at zero and finite temperature

Supersymmetric renormalization group flow equations for the effective superpotential of the three-dimensional Wess-Zumino model are derived at zero and non-zero temperature. This model with fermions and bosons interacting via a Yukawa term possesses a supersymmetric analogue of the Wilson-Fisher fixed-point. At zero temperature we determine the phase-transition line in coupling-constant space separating the supersymmetric from the nonsupersymmetric phase. At finite temperature we encounter dimensional reduction from 3 to 2 dimensions in the infrared regime. We determine the finite-temperature phase diagram for the restoration of the global Z{sub 2}-symmetry and show that for temperatures above the Z{sub 2} phase transition the pressure obeys the Stefan-Boltzmann law of a gas of massless bosons in 2+1 dimensions.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

150 Shades of Green: Using the Full Spectrum of Remote Sensing Reflectance to Elucidate Color Shifts in the Ocean

This article proposes a simple and intuitive classification system by which to define full spectral remote sensing reflectance (Rrs(λ)) data with a quantitative output that enables a more manageable handling of spectral information for aquatic science applications. The weighted harmonic mean of the Rrs(λ) wavelengths outputs an Apparent Visible Wavelength (in units of nanometers), representing a one-dimensional geophysical metric of color that is inherently correlated to spectral shape. This dimensionality reduction of spectral information combined with the output along a continuum of wavelength values offers a robust and user-friendly means to describe and analyze spectral Rrs(λ) in terms of spatial and temporal trends and variability. The uncertainty in the algorithm's estimation of spectral shape is demonstrated on a global scale, in addition to the utility of the algorithm to discern spectral-spatial-temporal trends in the ocean, on a per-pixel basis for the entire 22 year continuous ocean color (SeaWiFS and MODIS-Aqua) time-series. This technique can be applied to datasets of varying multi- and hyper-spectral resolutions, providing continuity between heritage and future satellite sensors, and further enabling an effective means of elucidating similarities or differences in complex spectral signatures within the constraints of two dimensions. This straightforward means of conceptualizing multi-dimensional variability can help maximize the potential of the spectral information embedded in remote sensing data.

ocean color↗

Four-dimensional N=1 super Yang-Mills theory from a matrix model

We consider a supersymmetric matrix quantum mechanics. This is obtained by adding Myers and mass terms to the dimensional reduction of 4D N=1 super Yang-Mills theory to one dimension. Using this model we construct 4D N=1 super Yang-Mills theory in the planar limit by using the Eguchi-Kawai equivalence. This regularization turns out to be free from the sign problem at the regularized level. The same matrix quantum mechanics is also used to provide a nonperturbative formulation of 4D N=1 super Yang-Mills theory on a noncommutative space.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Kaluza-Klein theory in the limit of large number of extra dimensions

The Kaluza-Klein compactification in the limit of a large number of extra dimensions is studied. The starting point is the Einstein-Hilbert action plus cosmological constant in 4+D dimensions. It is shown that in the large D limit the effective four-dimensional cosmological constant is of order 1/D, whereas the size of the extra dimensions remains finite. A 't Hooft-like large D expansion of the effective Lagrangian for the Kaluza-Klein scalar and gauge fields arising from the dimensional reduction is considered. It is shown that the propagator of the scalar field associated to the determinant of the metric of the extra dimensions is strongly suppressed. This is an interesting result as in standard Kaluza-Klein theory this scalar degree of freedom is responsible for the constraint on the gauge fields which makes it impossible to recover the usual Yang-Mills equations. Moreover in the large D limit it turns out that the ultraviolet divergences due to the interactions between gauge and scalar fields are softened.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Long Baseline Neutrino Experiment (LBNE) Target Material Radiation Damage from Energetic Protons of the Brookhaven Linear Isotope Production (BLIP) Facility

In search of a low-Z pion production target for the Long Baseline Neutrino Facility (LBNF) of the Deep Underground Neutrino Experiment (DUNE) various grades of graphite were irradiated with 140-180 MeV protons to fluence up to ~6 10 20 cm -2 and irradiation temperatures in the range of 100-200oC. The objective of the irradiation experiment was to assess the radiation induced changes in the graphite onset by its interaction with energetic protons and deduce some measure of performance and survivability under the design conditions of LBNF. A multi-faceted post-irradiation evaluation was conducted to both qualify and quantify the induced material property changes. These studies included dimensional stability and thermal expansion, damage annealing via thermal cycling, high temperature post-irradiation annealing to mimic irradiation at elevated temperatures, stress-strain behavior, oxidation contribution to the postirradiation analysis, potential influence of the argon atmosphere under which irradiation took place, and lattice evolution. Dilatometry-based techniques were coupled with ultrasound-based annealing methodologies as well as DSC and TGA analyses. X-ray diffraction experiments assessed the lattice parameter changes and SEM analyses were used to study the different graphite grade structures. The multi-faceted post-irradiation evaluation, in conjunction with off-beam analyses, revealed some important facts regarding the choice of a graphite grade. Specifically, it was observed that the microstructure of a grade plays an important role in its dimensional stability. X-ray diffraction data indicated variability in basal plane spacing prior to irradiation with POCO d 002 =3.3833 Å, IG-430 d 002 =3.36814 Å and R7650 d 002 =3.37944 Å. Isotropic graphite grades such as POCO and IG-430 are confirmed to undergo microstructural shrinking under irradiation. All four irradiated graphite grades experience irradiation-induced creep leading to increased Young’s modulus with changes as high as 50% that varies with microstructure. X-ray diffraction on irradiated graphite confirmed previously made observations where under both energetic protons as well as fast neutrons of fluence > 5 10 20 cm -2 and at temperatures around 200°C acceleration of the effects on both the micro-and macrostructure is observed. Amongst the irradiated grades POCO graphite showed tendency to increase crystallinity order but also shows turbostratic state resulting from the interaction of its lattice with protons. This report presents and discusses the details of the irradiation experiment and the results of the postirradiation evaluation with main objective the assessment of the applicability and survivability of the selected four graphite grades as high-power accelerators pion production targets where the structural integrity, density reduction, dimensional and physical property stability are of paramount importance.

43 PARTICLE ACCELERATORS↗

Generative learning for slow manifolds and bifurcation diagrams

In dynamical systems characterized by separation of time scales, the approximation of so called “slow manifolds”, on which the long term dynamics lie, is a useful step for model reduction. Initializing on such slow manifolds is a useful step in modeling, since it circumvents fast transients, and is crucial in multiscale algorithms (like the equation-free approach) alternating between fine scale (fast) and coarser scale (slow) simulations. In a similar spirit, when one studies the infinite time dynamics of systems depending on parameters, the system attractors (e.g., its steady states) lie on bifurcation diagrams (curves for one-parameter continuation, and more generally, on manifolds in state parameter space. Sampling these manifolds gives us representative attractors (here, steady states of ODEs or PDEs) at different parameter values. Algorithms for the systematic construction of these manifolds (slow manifolds, bifurcation diagrams) are required parts of the “traditional” numerical nonlinear dynamics toolkit. In more recent years, as the field of Machine Learning develops, conditional score-based generative models (cSGMs) have been demonstrated to exhibit remarkable capabilities in generating plausible data from target distributions that are conditioned on some given label. It is tempting to exploit such generative models to produce samples of data distributions (points on a slow manifold, steady states on a bifurcation surface) conditioned on (consistent with) some quantity of interest (QoI, observable). In this work, we present a framework for using cSGMs to quickly (a) initialize on a low-dimensional (reduced-order) slow manifold of a multi-time-scale system consistent with desired value(s) of a QoI (a “label”) on the manifold, and (b) approximate steady states in a bifurcation diagram consistent with a (new, out-of-sample) parameter value. This conditional sampling can help uncover the geometry of the reduced slow-manifold and/or approximately “fill in” missing segments of steady states in a bifurcation diagram. Finally, the quantity of interest, which determines how the sampling is conditioned, is either known a priori or identified using manifold learning-based dimensionality reduction techniques applied to the training data.

Dynamical systems↗

Goal-Oriented Optimal Approximations of Bayesian Linear Inverse Problems

We propose optimal dimensionality reduction techniques for the solution of goal-oriented linear-Gaussian inverse problems, where the quantity of interest (QoI) is a function of the inversion parameters. These approximations are suitable for large-scale applications. In particular, we study the approximation of the posterior covariance of the QoI as a low-rank negative update of its prior covariance and prove optimality of this update with respect to the natural geodesic distance on the manifold of symmetric positive definite matrices. Assuming exact knowledge of the posterior mean of the QoI, the optimality results extend to optimality in distribution with respect to the Kullback--Leibler divergence and the Hellinger distance between the associated distributions. We also propose the approximation of the posterior mean of the QoI as a low-rank linear function of the data and prove optimality of this approximation with respect to a weighted Bayes risk. Both of these optimal approximations avoid the explicit computation of the full posterior distribution of the parameters and instead focus on directions that are well informed by the data and relevant to the QoI. These directions stem from a balance among all the components of the goal-oriented inverse problem: prior information, forward model, measurement noise, and ultimate goals. We illustrate the theory using a high-dimensional inverse problem in heat transfer.

97 MATHEMATICS AND COMPUTING↗

Modeling and simulation challenges pursued by the Consortium for Advanced Simulation of Light Water Reactors (CASL)

The Consortium for the Advanced Simulation of Light Water Reactors (CASL), the first Energy Innovation Hub of the Department of Energy, was established in 2010 with the goal of providing modeling and simulation (M&S) capabilities that support and accelerate the improvement of nuclear energy's economic competitiveness and the reduction of spent nuclear fuel volume per unit energy, and all while assuring nuclear safety. To accomplish this requires advances in M&S capabilities in radiation transport, thermal-hydraulics, fuel performance and corrosion chemistry. To focus CASL's R&D, industry challenge problems have been defined, which equate with long standing issues of the nuclear power industry that M&S can assist in addressing. To date CASL has developed a multi-physics “core simulator” based upon pin-resolved radiation transport and subchannel (within fuel assembly) thermal-hydraulics, capitalizing on the capabilities of high performance computing. CASL's fuel performance M&S capability can also be optionally integrated into the core simulator, yielding a coupled multi-physics capability with untapped predictive potential. Material models have been developed to enhance predictive capabilities of fuel clad creep and growth, along with deeper understanding of zirconium alloy clad oxidation and hydrogen pickup. Understanding of corrosion chemistry (e.g., CRUD formation) has evolved at all scales: micro, meso and macro. CFD R&D has focused on improvement in closure models for subcooled boiling and bubbly flow, and the formulation of robust numerical solution algorithms. For multiphysics integration, several iterative acceleration methods have been assessed, illuminating areas where further research is needed. Finally, uncertainty quantification and data assimilation techniques, based upon sampling approaches, have been made more feasible for practicing nuclear engineers via R&D on dimensional reduction and biased sampling. Industry adoption of CASL's evolving M&S capabilities, which is in progress, will assist in addressing long-standing and future operational and safety challenges of the nuclear industry. - Highlights: • Complexity of physics based modeling of light water reactor cores being addressed. • Capability developed to help address problems that have challenged the nuclear power industry. • Simulation capabilities that take advantage of high performance computing developed.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗