Search NASA⌕ Search

SEARCH · Search NASA

Results for “Variational approach”

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 307 records · Page 17

A micromorphic filter for determining stress and deformation measures from direct numerical simulations of lower length scale behavior

A micromorphic filter is presented for the extraction of stress and deformation measures from underlying direct numerical simulation (DNS) of microstructural mechanical response. The filter is consistent with the micromorphic continuum theory of Eringen and Suhubi (1964), and allows the interrogation of a DNS to generate higher order constitutive models. The filter adopts aspects of the overlap coupling techniques for atomistic-continuum scale-bridging in order to update the nodal degrees of freedom of the finite-element-based filter. Here, two methods for the extraction of quantities of interest from the underlying DNS are presented. The first is a least squares fitting technique which is appropriate for finite-element-style macro-scale domains, and the second is a variationally based approach which generalizes the method to any macro-scale domain which has a variational form. No restrictions on the form of the micro-scale simulation are inherent to the approach, provided that the required quantities (stress, position, etc.) can be provided. The filter is applied to a simple homogeneous medium undergoing homogeneous deformation to demonstrate the recovery of a classical continuum response, what we call the null case. We then apply the framework to the same homogeneous medium with a central and edge penny-shaped crack to show micromorphic effects. We demonstrate that varying the size of the filter causes the micromorphic behaviors to arise and then dissipate as the filter becomes smaller than the crack's length scale. Additionally we show the effect of filter size on a chiral element network along with a plate perforated with evenly spaced and perturbed holes through the thickness. The micromorphic effects are shown to be present for the chiral elements under any loading but become most pronounced when a single element is located within each averaging domain. Micromorphic effects become pronounced for the plate upon the onset of localization.

97 MATHEMATICS AND COMPUTING↗

Evidence for variations in magma production along oceanic spreading centers - A critical appraisal

Recent studies of the oceanic lithosphere near fracture zones have resulted in the proposal that the 'magma budget', defined as the amount of magma delivered to magma chambers or conduits beneath a spreading center for a given amount of spreading, decreases as fracture zones are approached. Geochemical variations in basaltic glasses collected near fracture zones are consistent with a decrease in partial melting as fracture zones are approached, but they could also be produced by variations in open-system magmatic processes with no change in the extent of partial melting. Although a decrease in the magma budget as fracture zones are approached is consistent with these data, so are alternative models that incorporate a constant magma budget.

Karson, J. A.↗

Optimizing training trajectories in variational autoencoders via latent Bayesian optimization approach *

Unsupervised and semi-supervised ML methods such as variational autoencoders (VAE) have become widely adopted across multiple areas of physics, chemistry, and materials sciences due to their capability in disentangling representations and ability to find latent manifolds for classification and/or regression of complex experimental data. Like other ML problems, VAEs require hyperparameter tuning, e.g. balancing the Kullback–Leibler and reconstruction terms. However, the training process and resulting manifold topology and connectivity depend not only on hyperparameters, but also their evolution during training. Because of the inefficiency of exhaustive search in a high-dimensional hyperparameter space for the expensive-to-train models, here we have explored a latent Bayesian optimization (zBO) approach for the hyperparameter trajectory optimization for the unsupervised and semi-supervised ML and demonstrated for joint-VAE with rotational invariances. We have demonstrated an application of this method for finding joint discrete and continuous rotationally invariant representations for modified national institute of standards and technology database (MNIST) and experimental data of a plasmonic nanoparticles material system. The performance of the proposed approach has been discussed extensively, where it allows for any high dimensional hyperparameter trajectory optimization of other ML models.

42 ENGINEERING↗

Resource-Efficient Chemistry on Quantum Computers with the Variational Quantum Eigensolver and The Double Unitary Coupled-Cluster approach

Applications of quantum simulation algorithms to obtain electronic energies of molecules on noisy intermediate-scale quantum (NISQ) devices require careful consideration of resources describing the complex electron correlation effects. In modeling second-quantized problems, the biggest challenge confronted is that the number of qubits scales linearly with the size of molecular basis. This poses a significant limitation on the size of the basis sets and the number of correlated electrons included in quantum simulations of chemical processes. To address this issue and to enable more realistic simulations on NISQ computers, we employ the double unitary coupled-cluster (DUCC) method to effectively downfold correlation effects into the reduced-size orbital space, commonly referred to as the active space. Using downfolding techniques, we demonstrate that properly constructed effective Hamiltonians can capture the effect of the whole orbital space in small-size active spaces. Combining the downfolding pre-processing technique with the Variational Quantum Eigensolver, we solve for the ground-state energy of H2 and Li2 in the cc-pVTZ basis using the DUCC-reduced active spaces. We compare these results to full configuration-interaction and high-level coupled-cluster reference calculations.

quantum computing, variational quantum solver, cou↗

Flight-test measurement of the noise reduction of a jet transport delayed flap approach procedure

A delayed flap approach procedure was flight tested using the NASA CV-990 airplane to measure and analyze the noise produced beneath the flight path. Three other types of landing approaches were also flight tested to provide a comparison of the noise reduction benefits to the delayed flap approach. The conventional type of approach was used as a baseline to compare the effectiveness of the other approaches. The decelerating approach is a variation of the delayed flap approach. A detailed comparison of the ground perceived noise generated during the approaches is presented. For this comparison, the measured noise data were normalized to compensate for variations in aircraft weight and winds that occurred during the flight tests. The data show that the reduced flap approach offers some noise reduction, while the delayed flap and decelerating approaches offer significant noise reductions over the conventional approach.

Foster, J. D.↗

Next generation initiation techniques

Four-dimensional data assimilation strategies can generally be classified as either current or next generation, depending upon whether they are used operationally or not. Current-generation data-assimilation techniques are those that are presently used routinely in operational-forecasting or research applications. They can be classified into the following categories: intermittent assimilation, Newtonian relaxation, and physical initialization. It should be noted that these techniques are the subject of continued research, and their improvement will parallel the development of next generation techniques described by the other speakers. Next generation assimilation techniques are those that are under development but are not yet used operationally. Most of these procedures are derived from control theory or variational methods and primarily represent continuous assimilation approaches, in which the data and model dynamics are 'fitted' to each other in an optimal way. Another 'next generation' category is the initialization of convective-scale models. Intermittent assimilation systems use an objective analysis to combine all observations within a time window that is centered on the analysis time. Continuous first-generation assimilation systems are usually based on the Newtonian-relaxation or 'nudging' techniques. Physical initialization procedures generally involve the use of standard or nonstandard data to force some physical process in the model during an assimilation period. Under the topic of next-generation assimilation techniques, variational approaches are currently being actively developed. Variational approaches seek to minimize a cost or penalty function which measures a model's fit to observations, background fields and other imposed constraints. Alternatively, the Kalman filter technique, which is also under investigation as a data assimilation procedure for numerical weather prediction, can yield acceptable initial conditions for mesoscale models. The third kind of next-generation technique involves strategies to initialize convective scale (non-hydrostatic) models.

Warner, Tom↗

The initiation, propagation, and effect of matrix microcracks in cross-ply and related laminates

Recently, a variational mechanics approach was used to determine the thermoelastic stress state in cracked laminates. Described here is a generalization of the variational mechanics techniques to handle other cross-ply laminates, related laminates, and to account for delaminations emanating from microcrack tips. Microcracking experiments on Hercules 3501-6/AS4 carbon fiber/epoxy laminates show a staggered cracking pattern. These results can be explained by the variational mechanics analysis. The analysis of delaminations emanating from microcrack tips has resulted in predictions about the structural and material variables controlling competition between microcracking and delamination failure modes.

Nairn, John A.↗

Extended self-energy functional approach for strongly correlated lattice bosons in the superfluid phase

Among the various numerical techniques to study the physics of strongly correlated quantum many-body systems, the self-energy functional approach (SFA) has become increasingly important. In its previous form, however, SFA is not applicable to Bose-Einstein condensation or superfluidity. In this paper, we show how to overcome this shortcoming. To this end, we identify an appropriate quantity, which we term D, that represents the correlation correction of the condensate order parameter, as it does the self-energy for Green's function. An appropriate functional is derived, which is stationary at the exact physical realization of D and of the self-energy. Its derivation is based on a functional-integral representation of the grand potential followed by an appropriate sequence of Legendre transformations. The approach is not perturbative and, therefore, applicable to a wide range of models with local interactions. We show that the variational cluster approach based on the extended self-energy functional is equivalent to the ''pseudoparticle'' approach proposed in Phys. Rev. B 83, 134507 (2011). We present results for the superfluid density in the two-dimensional Bose-Hubbard model, which shows a remarkable agreement with those of quantum-Monte-Carlo calculations.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

A comprehensive modeling procedure for estimating statistical properties of forced ignition

In this work, a comprehensive modeling procedure for estimating the probability of ignition with application to high altitude relights of aircraft combustors is developed. In these configurations, an ignitor is used to introduce high-enthalpy discharge into a fuel-laden but stratified flow. Because of the inherent variabilities inflow conditions, kernel discharge process, and the chaotic turbulent flow, ignition cannot be described deterministically, but only as a probabilistic measure. The proposed modeling framework consists of three components. The turbulent flow is represented using the large eddy simulation (LES) framework. The ignition process is modeled using a manifold approach, where the initial kernel evolution is represented using a homogeneous reactor while the latter part of the evolution is represented as a competition between diffusion and chemical reactions using a flamelet-type mapping. A combined lookup table that can track the evolution of the kernel through these two distinct reaction stages is developed. The table lookup variables are solved within the LES framework. The resulting simulation tool is then embedded within an uncertainty quantification approach, where variations in the turbulent flow, as well as operating and kernel properties, are simulated using a Monte-Carlo-based sampling approach. Techniques to reduce computational cost are used to obtain a robust, numerically accurate, and physically representative model for engine relight. The method is validated using experimental data for ignition of methane/air mixtures. Due to the comprehensive nature of the modeling procedure, it is found that the simulation tool reproduces experimentally observed ignition probabilities over a wide range of operating conditions.

42 ENGINEERING↗

Variational quantum solver employing the PDS energy functional

In our previous work (J. Chem. Phys. 2020, 153, 201102),we reported a new class of quantum algorithms that are based on the quantum computation of the connected moment expansion to find the ground and excited state energies. In particular, the Peeters-Devreese-Soldatov (PDS) formulation is found variational and bearing the potential for further combining with the existing variational quantum infrastructure. Following this direction, here we propose a variational quantum solver employing the PDS energy gradient. In comparison with the usual variational quantum eigensolver (VQE) and the original static PDS approach, the proposed variational quantum solver offers an effective approach to achieve high ac-curacy at finding the ground state and its energy through the rotation of the trial wave function of modest quality guided by the low order PDS energy gradients, thus improves the ac-curacy and efficiency of the quantum simulation. We demonstrate the performance of the proposed variational quantum solver for toy models, H2molecule, and strongly correlated planar H4system in some challenging situations. In all the case studies, the proposed variational quantum approach out-performs the usual VQE and static PDS calculations even at the lowest order.

Peng, Bo↗

ESTIMATING BLACK HOLE MASSES IN HUNDREDS OF QUASARS

We explore the practical feasibility of active galactic nucleus (AGN) broadband reverberation mapping and present first results. We lay out and apply a rigorous approach for the stochastic reverberation mapping of unevenly sampled multi-broadband flux measurements, assuming that the broad-line region (BLR) line flux is contributing up to 15% in some bands, and is directly constrained by one spectroscopical epoch. The approach describes variations of the observed flux as the continuum, modeled as a stochastic Gaussian process, and emission line contribution, modeled as a scaled, smoothed, and delayed version of the continuum. This approach can be used not only to interpolate in time between measurements, but also to determine confidence limits on continuum—line emission delays. This approach is applied to Sloan Digital Sky Survey observations in Stripe 82 (S82), providing flux measurements that are precise to 2% at ∼60 epochs over ∼10 yr. The strong annual variations in the epoch sampling prove a serious limitation in practice. In addition, suitable redshift ranges must be identified where strong, broad emission lines contribute to one filter, but not to another. By generating and evaluating problem-specific mock data, we verify that S82-like data can constrain τ{sub delay} for a simple transfer function model. In application to real data, we estimate τ{sub delay} for 323 AGNs with 0.225 < z < 0.846, combining information for different objects through the ensemble-scaling relationships for BLR size and black hole mass. Our analysis tentatively indicates a 1.7 times larger BLR size of Hα and Mg II compared to Kaspi et al. and Vestergaard, but the seasonal data sampling casts doubt on the robustness of the inference.

79 ASTRONOMY AND ASTROPHYSICS↗

Dimensionality Reduction with Variational Encoders Based on Subsystem Purification

Efficient methods for encoding and compression are likely to pave the way toward the problem of efficient trainability on higher-dimensional Hilbert spaces, overcoming issues of barren plateaus. Here, we propose an alternative approach to variational autoencoders to reduce the dimensionality of states represented in higher dimensional Hilbert spaces. To this end, we build a variational algorithm-based autoencoder circuit that takes as input a dataset and optimizes the parameters of a Parameterized Quantum Circuit (PQC) ansatz to produce an output state that can be represented as a tensor product of two subsystems by minimizing $Tr(ρ^2)$. The output of this circuit is passed through a series of controlled swap gates and measurements to output a state with half the number of qubits while retaining the features of the starting state in the same spirit as any dimension-reduction technique used in classical algorithms. The output obtained is used for supervised learning to guarantee the working of the encoding procedure thus developed. We make use of the Bars and Stripes (BAS) dataset for an 8 × 8 grid to create efficient encoding states and report a classification accuracy of 95% on the same. Thus, the demonstrated example provides proof for the working of the method in reducing states represented in large Hilbert spaces while maintaining the features required for any further machine learning algorithm that follows.

97 MATHEMATICS AND COMPUTING↗

Physical discovery in representation learning via conditioning on prior knowledge

Recent advances in electron, scanning probe, optical, and chemical imaging and spectroscopy yield bespoke data sets containing the information of structure and functionality of complex systems. In many cases, the resulting data sets are underpinned by low-dimensional simple representations encoding the factors of variability within the data. The representation learning methods seek to discover these factors of variability, ideally further connecting them with relevant physical mechanisms. However, generally, the task of identifying the latent variables corresponding to actual physical mechanisms is extremely complex. Here, we present an empirical study of an approach based on conditioning the data on the known (continuous) physical parameters and systematically compare it with the previously introduced approach based on the invariant variational autoencoders. The conditional variational autoencoder (cVAE) approach does not rely on the existence of the invariant transforms and hence allows for much greater flexibility and applicability. Interestingly, cVAE allows for limited extrapolation outside of the original domain of the conditional variable. However, this extrapolation is limited compared to the cases when true physical mechanisms are known, and the physical factor of variability can be disentangled in full. We further show that introducing the known conditioning results in the simplification of the latent distribution if the conditioning vector is correlated with the factor of variability in the data, thus allowing us to separate relevant physical factors. We initially demonstrate this approach using 1D and 2D examples on a synthetic data set and then extend it to the analysis of experimental data on ferroelectric domain dynamics visualized via piezoresponse force microscopy.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Initial values for the integration scheme to compute the eigenvalues for propagation in ducts

A scheme for the calculation of eigenvalues in the problem of acoustic propagation in a two-dimensional duct is described. The computation method involves changing the coupled transcendental nonlinear algebraic equations into an initial value problem involving a nonlinear ordinary differential equation. The simplest approach is to use as initial values the hardwall eigenvalues and to integrate away from these values as the admittance varies from zero to its actual value with a linear variation. The approach leads to a powerful root finding routine capable of computing the transverse and axial wave numbers for two-dimensional ducts for any frequency, lining, admittance and Mach number without requiring initial guesses or starting points.

Eversman, W.↗

Sensing Water Vapon via Spacecraft Radio Occultation Observations

The radio occultation technique has been used to characterize planetary atmospheres since the 1960's spanning atmospheric pressures from 16 microbars to several bars. In 1988, the use of GPS signals to make occultation observations of Earth's atmosphere was realized by Tom Yunck and Gunnar Lindal at JPL. In the GPS to low-Earth-orbiter limb- viewing occultation geometry, Fresnel diffraction yield a unique combination of high vertical resolution of 100 m to 1 km at long wavelengths (approx. 20 cm) insensitive to particulate scattering which allows routine limb sounding from the lower mesosphere through the troposphere. A single orbiting GPS/GLONASS receiver can observe - 1000 to 1400 daily occultations providing as many daily, high vertical resolution soundings as the present global radiosonde network, but with far more evenly distributed, global coverage. The occultations yield profiles of refractivity as a function of height. In the cold, dry conditions of the upper troposphere and above (T less than 240 K), profiles of density, pressure (geopotential), and temperature can be derived. Given additional temperature information, water vapor can be derived in the midddle and lower troposphere with a unique combination of vertical resolution, global distribution and insensitivity to clouds and precipitation to an accuracy of approx. 0.2 g/kg. At low latitudes, moisture profiles will be accurate to 1-5% within the convective boundary layer and better than 20% below 6 to 7 km. Accuracies of climatological averages should be approx. 0. 1 g/kg limited by the biases in the temperature estimates. To use refractivity to constrain water vapor, knowledge of temperature is required. The simplest approach is to use the temperature field from an analysis such as the 6 hour ECMWF global analysis interpolated to the locations of each occultation. A better approach is to combine the temperature and moisture fields from such an analysis with the occultation refractivity in a weighting scheme based on the errors in each data field. A ID variational combinational approach has been developed at the UKMO. We win present results from both approaches from GPS/MET data taken in June and July 1995 and compare them with the ECMWF global 6 hour moisture analyses which are derived largely from TOVS and radiosonde data. Overall, the atmosphere below the 500 mb level appears somewhat drier in general than the ECNIWF humidity field. A 2-D (latitude vs. height) climatological snapshot derived from a 2-week span of GPS/MET data will be compared to the humidity climatology of Peixoto and Oort derived from radiosonde data from 1963-1973. Differences between the GPS results and Peixoto and Oort may be the signature of a climate trend over the past 30 years.

Kursinski, E. Robert↗

DNN-based Signal Processing for Liquid Argon Time Projection Chambers

We investigate a deep learning-based signal processing for liquid argon time projection chambers (LArTPCs), a leading detector technology in neutrino physics. Identifying regions of interest (ROIs) in LArTPCs is challenging due to signal cancellation from bipolar responses and various detector effects observed in real data. We approach ROI identification as an image segmentation task, and employ a U-ResNet architecture. The network is trained on samples that incorporate detector geometry information and include a range of detector variations. Our approach significantly outperforms traditional methods while maintaining robustness across diverse detector conditions. This method has been adopted for signal processing in the Short-Baseline Neutrino program and provides a valuable foundation for future experiments such as the Deep Underground Neutrino Experiment.

Bhat, Avinay [Chicago U.]↗

Effects of heat treatments on microstructure and properties of Ti-6Al-4V ELI alloy fabricated by electron beam melting (EBM)

Electron beam melting (EBM) is a metal powder bed fusion additive manufacturing (AM) technology that is used to fabricate three-dimensional near-net-shaped parts directly from computer models. Ti-6Al-4V is the most widely used and studied alloy for this technology and is the focus of this work in its ELI (Extra Low Interstitial) variation. The mechanisms of microstructure formation, evolution, and its subsequent influence on mechanical properties of the alloy in as-fabricated condition have been documented by various researchers. In the present work, the thermal history resulting in the formation of the as-fabricated microstructure was analyzed and studied by a thermal simulation. Subsequently different heat treatments were performed based on three approaches in order to study the effects of heat treatments on the singular and exclusive microstructure formed during the EBM fabrication process. In the first approach, the effect of cooling rate after the solutionizing process was studied. In the second approach, the variation of α lath thickness during annealing treatment and correlation with mechanical properties was established. In the last approach, several solutionizing and aging experiments were conducted.

36 MATERIALS SCIENCE↗