SEARCH · Engineering Papers

Results for “Functional priors”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 73 records · Page 4

All sources NASA NTRS DOE OSTI

NASA NTRS2017

Reconfigurable Drive Current System

A reconfigurable drive current system includes drive stages, each of which includes a high-side transistor and a low-side transistor in a totem pole configuration. A current monitor is coupled to an output of each drive stage. Input channels are provided to receive input signals. A processor is coupled to the input channels and to each current monitor for generating at least one drive signal using at least one of the input signals and current measured by at least one of the current monitors. A pulse width modulation generator is coupled to the processor and each drive stage for varying the drive signals as a function of time prior to being supplied to at least one of the drive stages.

Alhorn, Dean C.↗

DOE OSTI2022

Performance Evaluation of Distributed Energy Resource Management Systems Using Software- and Hardware-in-the-Loop

Distributed energy resource management systems (DERMS) are an important emerging technology for coordinating distributed energy resources (DERs) integrated in power distribution networks. These solutions implement a software platform to monitor, control, and coordinate DERs in real time to provide aggregate grid services to the electric utility or distribution system operator. To move DERMS technology forward, research and performance validation in a laboratory environment are needed to ensure proper function of DERMS prior to field deployment. This presentation presents a generic hardware-in-the-loop (HIL) platform to evaluate performance of DERMS using software-in-the-loop (SIL) and power HIL. The platform provides realistic testing in a laboratory environment, including accurate modeling of full-scale distribution systems, DERMS software controllers, and at-power distributed energy resources (DERs), such as solar and battery inverters. Two case studies are presented to demonstrate how the platform can be used independently as an integrated development and validation approach for different projects. Laboratory testing results are included to showcase the voltage regulation performance of each study and the benefits of implementing such a platform for DERMS evaluation.

DERMS↗

DOE OSTI2025

Physics-Informed Gaussian Process Inference of Liquid Structure from Scattering Data

We present a nonparametric Bayesian framework to infer radial distribution functions from experimental scattering measurements with uncertainty quantification using nonstationary Gaussian processes. The Gaussian process prior mean and kernel functions are designed to mitigate well-known numerical challenges with the Fourier transform, including discrete measurement binning and detector windowing, while encoding fundamental yet minimal physical knowledge of the liquid structure. We demonstrate uncertainty propagation of the Gaussian process posterior to unmeasured quantities of interest. Experimental radial distribution functions of liquid argon and water with uncertainty quantification are provided as both a proof of principle for the method and a benchmark for molecular models.

Chemical structure↗

NASA NTRS2010

Coaxial Propellant Injectors With Faceplate Annulus Control

An improved design concept for coaxial propellant injectors for a rocket engine (or perhaps for a non-rocket combustion chamber) offers advantages of greater robustness, less complexity, fewer parts, lower cost, and less bulk, relative to prior injectors of equivalent functionality. This design concept is particularly well suited to small, tight-tolerance injectors, for which prior designs are not suitable because the practical implementation of those designs entails very high costs and difficulty in adhering to the tolerances.

Horn, Mark D.↗

DOE OSTI2023

Bayesian Monte Carlo Evaluation of Imperfect (n, 233 U) Data and Model

Conventional nuclear data evaluation methods using generalized linear least squares make the following assumptions: prior and posterior probability distribution functions (PDFs) of all model parameters and data are normal (Gaussian); the linear approximation is sufficiently accurate to minimize the cost function (even for nonlinear models); the model (e.g., of neutron cross section) and experimental data (including covariance data) are without defect and prior PDFs of parameters and measured data are known perfectly. Neglect of covariance between model parameters and measured data in conventional evaluations contributes to imperfections. These assumptions are inherent to the generalized linear least squares minimization method commonly used for resolved resonance region neutron cross section evaluations but are often not justified due to the presence of non-normal PDFs, nonlinear models (e.g., R-matrix formalism), and inherent imperfections in data and models (e.g., imperfect covariance data). Here, these assumptions are removed in a mathematical framework of Bayes’ theorem, which is implemented using the Metropolis-Hastings Monte Carlo method. Most importantly, new parameters are introduced to parameterize discrepancies between the theoretical model and measured data to quantify judgement about discrepancies or imperfections in a reproducible manner. An evaluation of 233U in the eV region using the ENDF-B/VIII.0 library and transmission data (Guber et al.) is presented, and posterior parameters are compared to those obtained by conventional evaluation methods. This example illustrates the effects of removing the most harmful assumption: that of model-data perfection.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

NASA NTRS2007

Computational Support for Technology- Investment Decisions

Strategic Assessment of Risk and Technology (START) is a user-friendly computer program that assists human managers in making decisions regarding research-and-development investment portfolios in the presence of uncertainties and of non-technological constraints that include budgetary and time limits, restrictions related to infrastructure, and programmatic and institutional priorities. START facilitates quantitative analysis of technologies, capabilities, missions, scenarios and programs, and thereby enables the selection and scheduling of value-optimal development efforts. START incorporates features that, variously, perform or support a unique combination of functions, most of which are not systematically performed or supported by prior decision- support software. These functions include the following: Optimal portfolio selection using an expected-utility-based assessment of capabilities and technologies; Temporal investment recommendations; Distinctions between enhancing and enabling capabilities; Analysis of partial funding for enhancing capabilities; and Sensitivity and uncertainty analysis. START can run on almost any computing hardware, within Linux and related operating systems that include Mac OS X versions 10.3 and later, and can run in Windows under the Cygwin environment. START can be distributed in binary code form. START calls, as external libraries, several open-source software packages. Output is in Excel (.xls) file format.

Adumitroaie, Virgil↗

NASA NTRS1987

Comparing hard and soft prior bounds in geophysical inverse problems

In linear inversion of a finite-dimensional data vector y to estimate a finite-dimensional prediction vector z, prior information about X sub E is essential if y is to supply useful limits for z. The one exception occurs when all the prediction functionals are linear combinations of the data functionals. Two forms of prior information are compared: a soft bound on X sub E is a probability distribution p sub x on X which describeds the observer's opinion about where X sub E is likely to be in X; a hard bound on X sub E is an inequality Q sub x(X sub E, X sub E) is equal to or less than 1, where Q sub x is a positive definite quadratic form on X. A hard bound Q sub x can be softened to many different probability distributions p sub x, but all these p sub x's carry much new information about X sub E which is absent from Q sub x, and some information which contradicts Q sub x. Both stochastic inversion (SI) and Bayesian inference (BI) estimate z from y and a soft prior bound p sub x. If that probability distribution was obtained by softening a hard prior bound Q sub x, rather than by objective statistical inference independent of y, then p sub x contains so much unsupported new information absent from Q sub x that conclusions about z obtained with SI or BI would seen to be suspect.

Backus, George E.↗

NASA NTRS1988

Comparing hard and soft prior bounds in geophysical inverse problems

In linear inversion of a finite-dimensional data vector y to estimate a finite-dimensional prediction vector z, prior information about X sub E is essential if y is to supply useful limits for z. The one exception occurs when all the prediction functionals are linear combinations of the data functionals. Two forms of prior information are compared: a soft bound on X sub E is a probability distribution p sub x on X which describes the observer's opinion about where X sub E is likely to be in X; a hard bound on X sub E is an inequality Q sub x(X sub E, X sub E) is equal to or less than 1, where Q sub x is a positive definite quadratic form on X. A hard bound Q sub x can be softened to many different probability distributions p sub x, but all these p sub x's carry much new information about X sub E which is absent from Q sub x, and some information which contradicts Q sub x. Both stochastic inversion (SI) and Bayesian inference (BI) estimate z from y and a soft prior bound p sub x. If that probability distribution was obtained by softening a hard prior bound Q sub x, rather than by objective statistical inference independent of y, then p sub x contains so much unsupported new information absent from Q sub x that conclusions about z obtained with SI or BI would seen to be suspect.

Backus, George E.↗

NASA NTRS1980

The multicategory case of the sequential Bayesian pixel selection and estimation procedure

A Bayesian technique for stratified proportion estimation and a sampling based on minimizing the mean squared error of this estimator were developed and tested on LANDSAT multispectral scanner data using the beta density function to model the prior distribution in the two-class case. An extention of this procedure to the k-class case is considered. A generalization of the beta function is shown to be a density function for the general case which allows the procedure to be extended.

Pore, M. D.↗

NASA NTRS1966

Human transfer functions used to predict system performance parameters

Automatic, parameter-tracking, model-matching technique compares the responses of a human operator with those of an analog computer model of a human operator to predict and analyze the performance of mechanical or electromechanical systems prior to construction. Transfer functions represent the input-output relation of an operator controlling a closed-loop system.

Source record↗

DOE OSTI2021

A variational phase-field model For ductile fracture with coalescence dissipation

A novel phase-field model for ductile fracture is presented. The model is developed within a consistent variational framework in the context of finite-deformation kinematics. A novel coalescence dissipation introduces a new coupling mechanism between plasticity and fracture by degrading the fracture toughness as the equivalent plastic strain increases. The proposed model is compared with a recent alternative where plasticity and fracture are strongly coupled. Several representative numerical examples motivate specific modeling choices. In particular, a linear crack geometric function provides an “unperturbed” ductile response prior to crack initiation, and Lorentz-type degradation functions ensure that the critical fracture strength remains independent of the phase-field regularization length. In addition, the response of the model is demonstrated to converge with a vanishing phase-field regularization length. Additionally, the model is then applied to calibrate and simulate a three-point bending experiment of an aluminum alloy specimen with a complex geometry. The effect of the proposed coalescence dissipation coupling on simulations of the experiment is first investigated in a two-dimensional plane strain setting. The calibrated model is then applied to a three-dimensional calculation, where the calculated load-deflection curves and the crack trajectory show excellent agreement with experimental observations. Finally, the model is applied to simulate crack nucleation and growth in a specimen from a recent Sandia Fracture Challenge.

42 ENGINEERING↗

DOE OSTI2021

A Multi-Objective Bayesian Optimization Approach Using the Weighted Tchebycheff Method

Abstract Bayesian optimization (BO) is a low-cost global optimization tool for expensive black-box objective functions, where we learn from prior evaluated designs, update a posterior surrogate Gaussian process model, and select new designs for future evaluation using an acquisition function. This research focuses upon developing a BO model with multiple black-box objective functions. In the standard multi-objective (MO) optimization problem, the weighted Tchebycheff method is efficiently used to find both convex and non-convex Pareto frontiers. This approach requires knowledge of utopia values before we start optimization. However, in the BO framework, since the functions are expensive to evaluate, it is very expensive to obtain the utopia values as a prior knowledge. Therefore, in this paper, we develop a MO-BO framework where we calibrate with multiple linear regression (MLR) models to estimate the utopia value for each objective as a function of design input variables; the models are updated iteratively with sampled training data from the proposed MO-BO. These iteratively estimated mean utopia values are used to formulate the weighted Tchebycheff MO acquisition function. The proposed approach is implemented in two numerical test examples and one engineering design problem of optimizing thin tube geometries under constant loading of temperature and pressure, with minimizing the risk of creep-fatigue failure and design cost, along with risk-based and manufacturing constraints. Finally, the model accuracy with frequentist, Bayesian and without MLR-based calibration are compared to true Pareto solutions.

Engineering↗

DOE OSTI2022

Comparing correlation components and approximations in Hartree–Fock and Kohn–Sham theories via an analytical test case study

The asymmetric Hubbard dimer is a model that allows for explicit expressions of the Hartree–Fock (HF) and Kohn–Sham (KS) states as analytical functions of the external potential, Δv, and of the interaction strength, U. We use this unique circumstance to establish a rigorous comparison between the individual contributions to the correlation energies stemming from the two theories in the {U, Δv} parameter space. Within this analysis of the Hubbard dimer, we observe a change in the sign of the HF kinetic correlation energy, compare the indirect repulsion energies, and derive an expression for the “traditional” correlation energy, i.e., the one that corrects the HF estimate, in a pure site-occupation function theory spirit [Eq. (45)]. Next, we test the performances of the Liu–Burke and the Seidl–Perdew–Levy functionals, which model the correlation energy based on its weak- and strong-interaction limit expansions and can be used for both the traditional and the KS correlation energies. Further, our results show that, in the Hubbard dimer setting, they typically work better for the HF reference, despite having been originally devised for KS. These conclusions are somewhat in line with prior assessments of these functionals on various chemical datasets. However, the Hubbard dimer model allows us to show the extent of the error that may occur in using the strong-interaction ingredient for the KS reference in place of the one for the HF reference, as has been carried out in most of the prior assessments.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

NASA NTRS2003

Solution-Adaptive Program for Computing 2D/Axi Viscous Flow

A computer program solves the Navier- Stokes equations governing the flow of a viscous, compressible fluid in an axisymmetric or two-dimensional (2D) setting. To obtain solutions more accurate than those generated by prior such programs that utilize regular and/or fixed computational meshes, this program utilizes unstructured (that is, irregular triangular) computational meshes that are automatically adapted to solutions. The adaptation can refine to regions of high change in gradient or can be driven by a novel residual minimization technique. Starting from an initial mesh and a corresponding data structure, the adaptation of the mesh is controlled by use of minimization functional. Other improvements over prior such programs include the following: (1) Boundary conditions are imposed weakly; that is, following initial specification of solution values at boundary nodes, these values are relaxed in time by means of the same formulations as those used for interior nodes. (2) Eigenvalues are limited in order to suppress expansion shocks. (3) An upwind fluctuation-splitting distribution scheme applied to inviscid flux requires fewer operations and produces less artificial dissipation than does a finite-volume scheme, leading to greater accuracy of solutions.

Wood, William A.↗

NASA NTRS2006

Studying Functions of All Yeast Genes Simultaneously

A method of studying the functions of all the genes of a given species of microorganism simultaneously has been developed in experiments on Saccharomyces cerevisiae (commonly known as baker's or brewer's yeast). It is already known that many yeast genes perform functions similar to those of corresponding human genes; therefore, by facilitating understanding of yeast genes, the method may ultimately also contribute to the knowledge needed to treat some diseases in humans. Because of the complexity of the method and the highly specialized nature of the underlying knowledge, it is possible to give only a brief and sketchy summary here. The method involves the use of unique synthetic deoxyribonucleic acid (DNA) sequences that are denoted as DNA bar codes because of their utility as molecular labels. The method also involves the disruption of gene functions through deletion of genes. Saccharomyces cerevisiae is a particularly powerful experimental system in that multiple deletion strains easily can be pooled for parallel growth assays. Individual deletion strains recently have been created for 5,918 open reading frames, representing nearly all of the estimated 6,000 genetic loci of Saccharomyces cerevisiae. Tagging of each deletion strain with one or two unique 20-nucleotide sequences enables identification of genes affected by specific growth conditions, without prior knowledge of gene functions. Hybridization of bar-code DNA to oligonucleotide arrays can be used to measure the growth rate of each strain over several cell-division generations. The growth rate thus measured serves as an index of the fitness of the strain.

Stolc, Viktor↗

DOE OSTI2022

Psb27, a photosystem II assembly protein, enables quenching of excess light energy during its participation in the PSII lifecycle

Photosystem II (PSII), the enzyme responsible for oxidizing water into molecular oxygen, undergoes a complex lifecycle during which multiple assembly proteins transiently bind to and depart from PSII assembly intermediate complexes. Psb27 is one such protein. It associates with the CP43 chlorophyll-binding subunit of PSII to form a Psb27-PSII sub-complex that constitutes 7–10% of the total PSII pool. Psb27 remains bound to PSII assembly intermediates and dissociates prior to the formation of fully functional PSII. In this study, we compared a series of Psb27 mutant strains in the cyanobacterium Synechocystis sp. PCC 6803 with varied expression levels of Psb27: wild type (WT); psb27 genetic deletion (Del27), genetically complemented psb27 (Com27); and over-expressed Psb27 (OE27). The Del27 strain demonstrated decreased non-photochemical fluorescence quenching, while the OE27 strain showed increased non-photochemical quenching and tolerance to fluctuating light conditions. Multiple flashes and fluorescence decay analysis indicated that OE27 has the least affected maximum PSII quantum yield of the mutants. OE27 also displayed a minimal impact on the half-life of the fast component of QA – reoxidation over multiple flashes, indicating robust PSII function. Further, we propose that the close association between Psb27 and CP43, and the absence of a fully functional manganese cluster in the Psb27-PSII complex create a PSII sub-population that dissipates excitation energy prior to its recruitment into the functional PSII pool. Efficient energy dissipation prevents damage to this pre-PSII pool and allows for efficient PSII repair and maturation. Participation of Psb27 in the PSII life cycle ensures high-quality PSII assembly.

59 BASIC BIOLOGICAL SCIENCES↗

NASA NTRS1997

Nickel-Hydrogen Battery Fault Clearing at Low State of Charge

Fault clearing currents were achieved and maintained at discharge rates from C/2 to C/3 at high and low states of charge. The fault clearing plateau voltage is strong function of: discharge current, and voltage-prior-to-the-fault-clearing-event and a weak function of state of charge. Voltage performance, for the range of conditions reported, is summarized.

Lurie, C.↗

DOE OSTI2026

The Sensitivity of Variational Bayesian Neural Network Performance to Hyperparameters

In scientific applications, predictive modeling is often of limited use without accurate uncertainty quantification (UQ) to indicate when a model may be extrapolating or when more data needs to be collected. Bayesian Neural Networks (BNNs) produce predictive uncertainty by propagating uncertainty in neural network (NN) weights and offer the promise of obtaining not only an accurate predictive model but also accurate UQ. However, in practice, obtaining accurate UQ with BNNs is difficult due in part to the approximations used for model training (such as those made in variational inference) and in part to the need to choose a suitable set of hyperparameters; these hyperparameters outnumber those needed for traditional NNs and often have opaque effects on the results. We aim to shed light on the effects of hyperparameter choices for variational BNNs by performing a global sensitivity analysis of variational BNN performance under varying hyperparameter settings. Our results indicate that many of the hyperparameters interact with each other to affect both predictive accuracy and UQ. For improved usage of variational BNNs in real-world applications, we suggest that thorough hyperparameter tuning, including tuning of prior hyperparameters and loss function parameters, is essential for accurate UQ in variational BNNs.

97 MATHEMATICS AND COMPUTING↗