Search NASA⌕ Search

SEARCH · Search NASA

Results for “parameter 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 19 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Parameter Reduction of Composite Load Model Using Active Subspace Method

Over the past decades, the increasing penetration of distributed energy resources (DERs) has dramatically changed the power load composition in the distribution networks. The traditional static and dynamic load models can hardly capture the dynamic behavior of modern loads especially for fault-induced delayed voltage recovery (FIDVR) events. Thus, a more comprehensive composite load model with combination of static load, different types of induction motors, single-phase A/C motor, electronic load and DERs has been proposed by Western Electricity Coordinating Council (WECC). However, due to the large number of parameters and model complexity, the WECC composite load model (WECC CMLD) raises new challenges to power system studies. To overcome these challenges, in this paper, a cutting-edge parameter reduction (PR) approach for WECC CMLD based on active subspace method (ASM) is proposed. Firstly, the WECC CMLD is parameterized in a discrete-time manner for the application of the proposed method. Then, parameter sensitivities are calculated by discovering the active subspace, which is a lower-dimensional linear subspace of the parameter space of WECC CMLD in which the dynamic response is most sensitive. The interdependency among parameters can be taken into consideration by our approach. Finally, the numerical experiments validate the effectiveness and advantages of the proposed approach for WECC CMLD model.

active subspace↗
DOE OSTI2022

Mapping the effects of physical and chemical reduction parameters on local atomic distributions within bimetallic nanoparticles

Bimetallic nanoparticles prove advantageous over their monometallic counterparts due to the tunable, hybrid properties that result from combining different atomic species in a controlled way. The favorable optical and catalytic properties resulting from AgAu nanoparticle formation have been widely attributed to the existence of Ag–Au bonds, the maximization of which assumes the formation of a homogeneous alloy. Despite the importance of atomic scale structure in these systems, synthetic studies are typically not paired with structural characterization at the atomic scale. Herein, a comprehensive synthetic exploration of physical and chemical reduction parameters of resulting nanoparticle products is complemented with thorough X-ray characterization to probe how these parameters affect atomic scale alloy distributions within AgAu nanoparticles. Presented evidence shows Ag is substantially underincorporated into nanoparticle constructs compared with solution Ag : Au ratios regardless of precursor :reductant ratio or volume of reductant added. Both Ag and Au exhibit significant local clustering, with Ag distributed preferentially towards the nanoparticle surface. Most significantly, the results of this investigation suggest that reduction parameters alone can affect the local alloy distributions and homogeneity within bimetallic nanoparticles, even when the ratio of metallic precursors remains constant. Overall, this investigation presents the ability to control alloy distributions using kinetics and provides new considerations for optimizing synthetic methods to produce functional bimetallic nanoparticles.

77 NANOSCIENCE AND NANOTECHNOLOGY↗
DOE OSTI2024

Flavor in SU(5)$SU(5)$ Finite Grand Unified Models

Abstract Four supersymmetric models which exhibit and/or symmetries are studied, that are finite to two or all loops, and their corresponding mass matrices. The first is an all‐loop finite model based on an flavor symmetry, which leads to phenomenologically nonviable mass matrices. The remaining models, based on cyclic symmetries, show various mass textures, some of which are phenomenologically promising. For the two‐loop finite models, the parametric solutions to the finiteness conditions determine completely some of the Yukawa couplings, and lead to a restricted range of values for other ones at the GUT scale, with a considerable reduction in the number of free parameters. One particular solution of the two‐loop models shows an enhanced symmetry, leading to an all‐loop finite model, which has a significant parameter reduction and could in principle reproduce the observed quark masses and mixing pattern. In this case the finiteness conditions determine the absolute value of all the Yukawa couplings at the unification scale. Finally, the minimum number of phases in the mass matrices and their position are determined, a task not previously done in Finite Unified Theories, which contributes towards the reduction of parameters and a better understanding of the Yukawa couplings.

Estrada Ramos, Luis Odín↗
DOE OSTI2025

CONCURRENT, CONDENSED STEIN VARIATIONAL GRADIENT DESCENT FOR UNCERTAINTY QUANTIFICATION OF NEURAL NETWORKS

In this work, we propose a Stein variational gradient descent (SVGD) method to concurrently sparsify, train, and provide uncertainty quantification (UQ) of a complexly parameterized model, such as a neural network (NN). It employs a graph reconciliation and condensation process to reduce complexity and increase similarity in the Stein ensemble of parameterizations. Therefore, the proposed concurrent, condensed SVGD (ccSVGD) method can provide UQ on parameters, not just outputs. Furthermore, the parameter reduction speeds up the convergence of the Stein gradient descent as it reduces the combinatorial complexity by aligning and differentiating the sensitivity to parameters. These properties are demonstrated with an illustrative example and an application to a mechanical response representation problem in solid mechanics.

42 ENGINEERING↗
DOE OSTI2022

Towards Compact Neural Networks via End-to-End Training: A Bayesian Tensor Approach with Automatic Rank Determination

Post-training model compression can reduce the inference costs of deep neural networks, but uncompressed training still consumes enormous hardware resources and energy. To enable low-energy training on edge devices, it is highly desirable to directly train a compact neural network from scratch with a low memory cost. Low-rank tensor decomposition is an effective approach to reduce the memory and computing costs of large neural networks. However, directly training low-rank tensorized neural networks is a very challenging task because it is hard to determine a proper tensor rank a priori, and the tensor rank controls both model complexity and accuracy. Here, this paper presents a novel end-to-end framework for low-rank tensorized training. We first develop a Bayesian model that supports various low-rank tensor formats (e.g., CANDECOMP/PARAFAC, Tucker, tensor-train, and tensor-train matrix) and reduces neural network parameters with automatic rank determination during training. Then we develop a customized Bayesian solver to train large-scale tensorized neural networks. Our training methods shows orders-of-magnitude parameter reduction and little accuracy loss (or even better accuracy) in the experiments. On a very large deep learning recommendation system with over 4.2 ×10 9 model parameters, our method can reduce the parameter number to 1.6 ×10 5 automatically in the training process (i.e., by 2.6 ×10 4 times) while achieving almost the same accuracy. Code is available at https://github.com/colehawkins/bayesian-tensor-rank-determination.

compact neural networks↗
DOE OSTI2022

Gamma-ray burst data strongly favour the three-parameter fundamental plane (Dainotti) correlation over the two-parameter one

ABSTRACT Gamma-ray bursts (GRBs), observed to redshift z = 9.4, are potential probes of the largely unexplored z ∼ 2.7–9.4 part of the early Universe. Thus, finding relevant relations among GRB physical properties is crucial. We find that the Platinum GRB data compilation, with 50 long GRBs (with relatively flat plateaus and no flares) in the redshift range 0.553 ≤ z ≤ 5.0, and the LGRB95 data compilation, with 95 long GRBs in 0.297 ≤ z ≤ 9.4, as well as the 145 GRB combination of the two, strongly favour the 3D Fundamental Plane (Dainotti) correlation (between the peak prompt luminosity, the luminosity at the end of the plateau emission, and its rest-frame duration) over the 2D one (between the luminosity at the end of the plateau emission and its duration). The 3D Dainotti correlations in the three data sets are standardizable. We find that while LGRB95 data have ∼50 per cent larger intrinsic scatter parameter values than the better-quality Platinum data, they provide somewhat tighter constraints on cosmological-model and GRB-correlation parameters, perhaps solely due to the larger number of data points, 95 versus 50. This suggests that when compiling GRB data for the purpose of constraining cosmological parameters, given the quality of current GRB data, intrinsic scatter parameter reduction must be balanced against reduced sample size.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

Inelastic Neutron Scattering Data for Ba₃Zn(Ru₀.₁Sb₁.₉)O₉

This dataset contains inelastic neutron scattering measurements on the magnetically dilute compound Ba₃Zn(Ru₀.₁Sb₁.₉)O₉, collected at the Spallation Neutron Source using the Cold Neutron Chopper Spectrometer (CNCS) and the SEQUOIA Fine-Resolution Fermi Chopper Spectrometer. The sample consists of nominally 5% magnetic Ru⁵⁺ ions (S = 3/2) randomly substituted into a diamagnetic Sb⁵⁺ matrix. Measurements were performed at base temperature (2–5 K) and 300 K using multiple incident energies (1.00, 3.32, 12.0, and 100 meV) to resolve magnetic excitations from isolated monomers and small Ru clusters (dimers, trimers, etc.). The dataset includes raw event-mode files, momentum- and energy-resolved spectra, and supporting metadata for instrument configuration, sample environment, and reduction parameters.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Derivation of the Future Time Equation for Analog, Non-Multiplying Monte Carlo Simulation

The expected computational time required to simulate a particle from a point in phase space through a Monte Carlo history, termed the expected future time, is found by solving the Future Time Equation (FTE). The expected future time may be useful when generating variance reduction parameters for a Monte Carlo simulation with a method such as Consistent Adjoint Driven Importance Sampling (CADIS). This report presents a detailed derivation of the Future Time Probability Density Function (FTPDF) and FTE for neutral particle Monte Carlo transport to aid future researchers. For simplicity, this derivation only considers analog transport in non-multiplying media.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2025

Compressing Vision Transformers in Geospatial Transfer Learning with Manifold-Constrained Optimization

Deploying geospatial foundation models on resource-constrained edge devices demands compact architectures that maintain high downstream performance. However, their large parameter counts and the accuracy loss often induced by compression limit practical adoption.In this work, we leverage manifold-constrained optimization framework DLRT to compress large vision transformer–based geospatial foundation models during transfer learning. By enforcing structured low-dimensional parameterizations aligned with downstream objectives, this approach achieves strong compression while preserving task-specific accuracy. We show that the method outperforms of-the-shelf low-rank methods as LoRA. Experiments on diverse geospatial benchmarks confirm substantial parameter reduction with minimal accuracy loss, enabling high-performing, on-device geospatial models.

Snyder, Thomas [Yale University]↗
DOE OSTI2022

Plasma parameters and the reduction potential at a plasma-liquid interface

We report nonthermal plasmas in contact with liquids have been shown to generate a variety of reactive species capable of initiating reduction–oxidation (redox) reactions at the electrochemically active plasma–liquid interface. In conventional electrochemical cells, selective redox chemistry is achieved by controlling the reduction potential at the solid electrode–electrolyte interface by applying a bias via an external circuit. In the case of plasma–liquid systems, an analogous means of tuning the reduction potential near the interface has not clearly been identified. When treated as a floating surface, the liquid is expected to adopt a net negative charge to balance the flux of hot electrons and relatively cold positive ions. The reduction potential near the plasma–liquid interface is hypothesized to be proportional to the floating potential, which can be approximated using an analytical model provided the plasma parameters are known. Herein, we present a framework for correlating the electron density and electron temperature of a noble gas plasma jet to the reduction potential near the plasma–liquid interface. The plasma parameters were acquired for an argon atmospheric plasma jet in contact with an aqueous solution by means of laser Thomson scattering. The reduction potential was determined using identical reference electrodes to measure the potential difference between the plasma–liquid interface and bulk solution. Interestingly, the measured reduction potentials near the plasma–liquid interface were found to be in good agreement with the model-predicted values determined using the plasma parameters obtained from the Thomson scattering experiments.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2021

Feasibility of DEIM for retrieving the initial field via dimensionality reduction

When parameter estimation is solved in a high-dimensional space, the dimensionality reduction strategy becomes the primary consideration for alleviating the tremendous computational cost. Here, the discrete empirical interpolation method (DEIM) is explored to retrieve the initial condition (IC) by combining the polynomial chaos (PC) based ensemble Kalman filter (i.e. PC-EnKF), where a non-intrusive PC expansion is considered as a surrogate model in place of the forward model in the prediction step of the ensemble Kalman filter, resulting in fewer forward model integrations but with a comparable accuracy as Monte Carlo-based approaches. The DEIM acts as a hyper-reduction tool to provide the low-dimensional input for the high-dimensional initial field, which can be reconstructed using the information on the sparse interpolation grid points that is adaptively obtained through PC-EnKF data assimilation method. Thus an innovative framework to reconstruct the IC is developed. The detailed procedure at each assimilation iteration includes: the determination of the spatial interpolation points, the estimation of the initial values on the interpolation locations using the optimal observations, and the reconstruction of IC in the full space. The current study uses the reconstruction field of initial conditions of the Navier-Stokes equations as an example to illustrate the efficacy of our method. The experimental results demonstrate the proposed algorithm achieves a satisfactory reconstruction for the initial field. The proposed method helps to extend the applicable area of DEIM in solving inverse problems.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2023

DRIPS: A framework for dimension reduction and interpolation in parameter space

Reduced-order models are often used to describe the behavior of complex systems, whose simulation with a full model is too expensive, or to extract salient features from the full model’s output. We introduce a new model-reduction framework DRIPS (dimension reduction and interpolation in parameter space) that combines the offline local model reduction with the online parameter interpolation of reduced-order bases (ROBs). The offline step of this framework relies on dynamic mode decomposition (DMD) to build a low-rank linear surrogate model, equipped with a local ROB, for quantities of interest derived from the training data generated by repeatedly solving the (nonlinear) high-fidelity model for multiple parameter points. The online step consists of the construction of a parametric reduced-order model for each target/test point in the parameter space, with the interpolation of ROBs done on a Grassman manifold and the interpolation of reduced-order operators done on a matrix manifold. The DMD component enables DRIPS to model (typically low-dimensional) quantities of interest directly, without having to access the (typically high-dimensional and possibly nonlinear) operators in a high-fidelity model that governs the dynamics of the underlying high-dimensional state variables, as required in projection-based reduced-order modeling. A series of numerical experiments suggests that DRIPS yields a model reduction, which is computationally more efficient than the commonly used projection-based proper orthogonal decomposition; it does so without requiring a prior knowledge of the governing equation for quantities of interest. Furthermore, for the nonlinear systems considered, DRIPS is more accurate than Gaussian-process interpolation (Kriging).

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2022

Covalent Organic Framework (COF) Derived Ni–N–C Catalysts for Electrochemical CO 2 Reduction: Unraveling Fundamental Kinetic and Structural Parameters of the Active Sites

Electrochemical CO 2 reduction is a potential approach to convert CO 2 into valuable chemicals using electricity as feedstock. Abundant and affordable catalyst materials are needed to upscale this process in a sustainable manner. Nickel-nitrogen-doped carbon (Ni-N-C) is an efficient catalyst for CO 2 reduction to CO, and the single-site Ni-N x motif is believed to be the active site. However, critical metrics for its catalytic activity, such as active site density and intrinsic turnover frequency, so far lack systematic discussion. In this work, we prepared a set of covalent organic framework (COF)-derived Ni-N-C catalysts, for which the Ni-N x content could be adjusted by the pyrolysis temperature. The combination of high-angle annular dark-field scanning transmission electron microscopy and extended X-ray absorption fine structure evidenced the presence of Ni single-sites, and quantitative X-ray photoemission addressed the relation between active site density and turnover frequency.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

The Effect of Acceleration on the Separation Force in Constrained-Surface Stereolithography

Constrained-surface-based stereolithography has recently attracted much attention from both academic and industrial communities. Despite numerous experimental, numerical and theoretical efforts, the fundamental need to reduce the separation force between the newly cured part and constrained surface has not yet been completely solved. In this paper, we develop a fluid dynamics approach, proposed in our previous work, to theoretically model the separation force in 3D printing of a cylindrical part for flat and patterned windows. We demonstrate the possibility of separation force reduction with an accelerated movement of the printing platform. In particular, we investigate behaviors of transient parameter, its reduction rate, and separation force reduction with respect to elevation speed and time. The proposed approach involves deceleration and acceleration stages and allows to achieve the force reduction for the entire printing process. Finally, we provide implicit analytical solutions for time moments when switching between the stages can be done without noticeable increase of separation force and explicit expression for separation force in case of patterned window.

Gritsenko, Dmitry (ORCID:0000000248181512)↗