Search NASA⌕ Search

SEARCH · Search NASA

Results for “general linear methods”

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 1,423 records · Page 79

Data Imbalance, Uncertainty Quantification, and Transfer Learning in Data‐Driven Parameterizations: Lessons From the Emulation of Gravity Wave Momentum Transport in WACCM

Abstract Neural networks (NNs) are increasingly used for data‐driven subgrid‐scale parameterizations in weather and climate models. While NNs are powerful tools for learning complex non‐linear relationships from data, there are several challenges in using them for parameterizations. Three of these challenges are (a) data imbalance related to learning rare, often large‐amplitude, samples; (b) uncertainty quantification (UQ) of the predictions to provide an accuracy indicator; and (c) generalization to other climates, for example, those with different radiative forcings. Here, we examine the performance of methods for addressing these challenges using NN‐based emulators of the Whole Atmosphere Community Climate Model (WACCM) physics‐based gravity wave (GW) parameterizations as a test case. WACCM has complex, state‐of‐the‐art parameterizations for orography‐, convection‐, and front‐driven GWs. Convection‐ and orography‐driven GWs have significant data imbalance due to the absence of convection or orography in most grid points. We address data imbalance using resampling and/or weighted loss functions, enabling the successful emulation of parameterizations for all three sources. We demonstrate that three UQ methods (Bayesian NNs, variational auto‐encoders, and dropouts) provide ensemble spreads that correspond to accuracy during testing, offering criteria for identifying when an NN gives inaccurate predictions. Finally, we show that the accuracy of these NNs decreases for a warmer climate (4 × CO 2 ). However, their performance is significantly improved by applying transfer learning, for example, re‐training only one layer using ∼1% new data from the warmer climate. The findings of this study offer insights for developing reliable and generalizable data‐driven parameterizations for various processes, including (but not limited to) GWs.

54 ENVIRONMENTAL SCIENCES↗

On the time growth of the error of the DG method for advective problems

Abstract In this paper we derive a priori $L^{\infty }(L^{2})$ and L2(L2) error estimates for a linear advection–reaction equation with inlet and outlet boundary conditions. The goal is to derive error estimates for the discontinuous Galerkin method that do not blow up exponentially with respect to time, unlike the usual case when Gronwall’s inequality is used. While this is possible in special cases, such as divergence-free advection fields, we take a more general approach using exponential scaling of the exact and discrete solutions. Here we use a special scaling function, which corresponds to time taken along individual pathlines of the flow. For advection fields, where the time that massless particles carried by the flow spend inside the spatial domain is uniformly bounded from above by some $\widehat{T}$, we derive $\mathcal{O}$(hp+1/2) error estimates where the constant factor depends only on $\widehat{T}$, but not on the final time T. This can be interpreted as applying Gronwall’s inequality in the error analysis along individual pathlines (Lagrangian setting), instead of physical time (Eulerian setting).

Mathematics↗

Load Mitigating Feedback Control of WECs

This report outlines the development of load-mitigating feedback control for wave energy converters. A simple, self-tuning multi-objective controller is demonstrated in simulation for a 3-DOF (surge, heave, pitch) point absorber. In previous work, the proposed control architecture has been shown to be effective in experiment for a variety of device archetypes for the single objective of the maximization of electrical power capture: here this architecture is extended to reduce device loading as well. In particular, PTO actuation forces and the minimization of fatigue damage (determined from the sum of wave-exerted and PTO forces) are considered as additional objectives for the self-tuning controller. This controller is demonstrated for two similar, but distinct systems: one described by the identified linear models from physical testing of the WaveBot device, and another based upon a WEC-Sim simulation that expands upon boundary element method data from the WaveBot device. In both cases, because the power surface is consistently fairly flat in the vicinity of control parameters that maximize power capture in contrasting sea-states, it is found to be generally possible to mitigate either fatigue damage or PTO load. However, PTO load is found to conflict with fatigue damage in some sea-states, limiting the efficacy of control objectives that attempt to mitigate both simultaneously. Additionally, coupling between the surge and pitch DOFs also limits the extent to which fatigue damage can be mitigated for both DOFs in some sea-states. Because control objectives can be considered a function of the sea-state (e.g., load mitigation may not be a concern until the sea is sufficiently large) a simple transition strategy is proposed and demonstrated. This transition strategy is found to be effective with some caveats: firstly, it cannot circumvent the aforementioned objective contradictions. Secondly, this objective transition is too slow to act as a system constraint, and objective thresholds must thus be considered quite conservatively. Improvement of the adjustment strategy is demonstrated through the addition of an integral term. Selection of well-performing transition parameters can be a function of sea-state. While a simple selection procedure is proposed, it is non-optimal, and a more robust selection procedure is suggested for future work.

16 TIDAL AND WAVE POWER↗

Effect of composition on the density of multi-component molten nitrate salts.

The density of molten nitrate salts was measured to determine the effects of the constituents on the density of multi-component mixtures. The molten salts consisted of various proportions of the nitrates of potassium, sodium, lithium and calcium. Density measurements ere performed using an Archimedean method and the results were compared to data reported in the literature for the individual constituent salts or simple combinations, such as the binary Solar Salt mixture of NaNO3 and KNO3. The addition of calcium nitrate generally ncreased density, relative to potassium nitrate or sodium nitrate, while lithium nitrate decreased density. The temperature dependence of density is described by a linear equation regardless of composition. The molar volume, and thereby, density of multi-component mixtures an be calculated as a function of temperature using a linear additivity rule based on the properties of the individual constituents.

25 ENERGY STORAGE↗

A Kalman filter for a two-dimensional shallow-water model

A two-dimensional Kalman filter is described for data assimilation for making weather forecasts. The filter is regarded as superior to the optimal interpolation method because the filter determines the forecast error covariance matrix exactly instead of using an approximation. A generalized time step is defined which includes expressions for one time step of the forecast model, the error covariance matrix, the gain matrix, and the evolution of the covariance matrix. Subsequent time steps are achieved by quantifying the forecast variables or employing a linear extrapolation from a current variable set, assuming the forecast dynamics are linear. Calculations for the evolution of the error covariance matrix are banded, i.e., are performed only with the elements significantly different from zero. Experimental results are provided from an application of the filter to a shallow-water simulation covering a 6000 x 6000 km grid.

Parrish, D. F.↗

Finding curvilinear features in speckled images

A method for finding curves in digital images with speckle noise is described. The solution method differs from standard linear convolutions followed by thresholds in that it explicitly allows curvature in the features. Maximum a posteriori (MAP) estimation is used, together with statistical models for the speckle noise and for the curve-generation process, to find the most probable estimate of the feature, given the image data. The estimation process is first described in general terms. Then, incorporation of the specific neighborhood system and a multiplicative noise model for speckle allows derivation of the solution, using dynamic programming, of the estimation problem. The detection of curvilinear features is considered separately. The detection results allow the determination of the minimal size of detectable feature. Finally, the estimation of linear features, followed by a detection step, is shown for computer-simulated images and for a SAR image of sea ice.

Samadani, Ramin↗

Effect of Strain Rate on Mechanical Properties of Wrought Sintered Tungsten at Temperatures above 2500 F

Specimens of wrought sintered commercially pure tungsten were made from 1/8-inch swaged rods. All the specimens were recrystallized at 4050 F for 1 hour prior to testing at temperatures from 2500 to 4000 F at various strain rates from 0.002 to 20 inches per inch per minute. Results showed that, at a constant temperature, increasing the strain rate increased the ultimate tensile strength significantly. The effects of both strain rate and temperature on the ultimate tensile strength of tungsten may be correlated by the linear parameter method of Manson and Haferd and may be used to predict the ultimate tensile strength at higher temperatures, 4500 and 5000 F. As previously reported, ductility, as measured by reduction of area in a tensile test, decreases with increasing temperature above about 3000 F. Increasing the strain rate at temperatures above 3000 F increases the ductility. Fractures are generally transgranular at the higher strain rates and intergranular at the lower strain rates.

Sikora, Paul F.↗

Designing Agent Collectives For Systems With Markovian Dynamics

The "Collective Intelligence" (COIN) framework concerns the design of collectives of agents so that as those agents strive to maximize their individual utility functions, their interaction causes a provided "world" utility function concerning the entire collective to be also maximized. Here we show how to extend that framework to scenarios having Markovian dynamics when no re-evolution of the system from counter-factual initial conditions (an often expensive calculation) is permitted. Our approach transforms the (time-extended) argument of each agent's utility function before evaluating that function. This transformation has benefits in scenarios not involving Markovian dynamics, in particular scenarios where not all of the arguments of an agent's utility function are observable. We investigate this transformation in simulations involving both linear and quadratic (nonlinear) dynamics. In addition, we find that a certain subset of these transformations, which result in utilities that have low "opacity (analogous to having high signal to noise) but are not "factored" (analogous to not being incentive compatible), reliably improve performance over that arising with factored utilities. We also present a Taylor Series method for the fully general nonlinear case.

Wolpert, David H.↗

Problem-free time-dependent variational principle for open quantum systems

Methods of quantum nuclear wave-function dynamics have become very efficient in simulating large isolated systems using the time-dependent variational principle (TDVP). However, a straightforward extension of the TDVP to the density matrix framework gives rise to methods that do not conserve the energy in the isolated system limit and the total system population for open systems where only energy exchange with environment is allowed. These problems arise when the system density is in a mixed state and is simulated using an incomplete basis. Thus, the basis set incompleteness, which is inevitable in practical calculations, creates artificial channels for energy and population dissipation. To overcome this unphysical behavior, we have introduced a constrained Lagrangian formulation of TDVP applied to a non-stochastic open system Schrödinger equation [L. Joubert-Doriol, I. G. Ryabinkin, and A. F. Izmaylov, J. Chem. Phys. 141, 234112 (2014)]. While our formulation can be applied to any variational ansatz for the system density matrix, derivation of working equations and numerical assessment is done within the variational multiconfiguration Gaussian approach for a two-dimensional linear vibronic coupling model system interacting with a harmonic bath.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Linearized Pair-Density Functional Theory for Vertical Excitation Energies

Multiconfiguration pair-density functional theory (MC-PDFT) is a computationally efficient method that computes the energies of electronic states in a state specific or state average framework via an on-top functional. However, MC-PDFT does not include state interaction among these states since the final energies do not come from the diagonalization of an effective model-space Hamiltonian. Recently, multistate extensions such as linearized PDFT (L-PDFT) have been developed to accurately model the potentials near conical intersections and avoided crossings. However, there has not been any systematic study evaluating their performance for predicting vertical excitations at the equilibrium geometry of a molecule, when the excited states are generally well separated. In this paper, we report the performance of L-PDFT on the extensive QUESTDB data set of vertical excitations using a database of automatically selected active spaces. We show that L-PDFT performs well on all these excitations and successfully reproduces the performance of MC-PDFT. These results further demonstrate the potential of L-PDFT, as its scaling is constant with the number of states included in the state-average manifold, whereas MC-PDFT scales linearly in this regard.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Long-time uncertainty propagation using generalized polynomial chaos and flow map composition

We present an efficient and accurate method for long-time uncertainty propagation in dynamical systems. Uncertain initial conditions and parameters are both addressed. The method approximates the intermediate short-time flow maps by spectral polynomial bases, as in the generalized polynomial chaos (gPC) method, and uses flow map composition to construct the long-time flow map. In contrast to the gPC method, this approach has spectral error convergence for both short and long integration times. The short-time flow map is characterized by small stretching and folding of the associated trajectories and hence can be well represented by a relatively low-degree basis. The composition of these low-degree polynomial bases then accurately describes the uncertainty behavior for long integration times. The key to the method is that the degree of the resulting polynomial approximation increases exponentially in the number of time intervals, while the number of polynomial coefficients either remains constant (for an autonomous system) or increases linearly in the number of time intervals (for a non-autonomous system). The findings are illustrated on several numerical examples including a nonlinear ordinary differential equation (ODE) with an uncertain initial condition, a linear ODE with an uncertain model parameter, and a two-dimensional, non-autonomous double gyre flow.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

A method of alternating characteristics with application to advection-dominated environmental systems

We present a numerical method for solving a class of systems of partial differential equations (PDEs) that arises in modeling environmental processes undergoing advection and biogeochemical reactions. The salient feature of these PDEs is that all partial derivatives appear in linear expressions. As a result, the system can be viewed as a set of ordinary differential equations (ODEs), albeit each one along a different characteristic. The method then consists of alternating between equations and integrating each one step-wise along its own characteristic, thus creating a customized grid on which solutions are computed. Since the solutions of such PDEs are generally smoother along their characteristics, the method offers the potential of using larger time steps while maintaining accuracy and reducing numerical dispersion. The advantages in efficiency and accuracy of the proposed method are demonstrated in two illustrative examples that simulate depth-resolved reactive transport and soil carbon cycling.

54 ENVIRONMENTAL SCIENCES↗

Development of Late Toxicity and International Prostate Symptom Score Resolution After External-Beam Radiotherapy Combined With Pulsed Dose Rate Brachytherapy for Prostate Cancer

Purpose: To investigate the development of gastrointestinal (GI) toxicity, genitourinary (GU) toxicity, erectile dysfunction, and International Prostate Symptom Score (IPSS) resolution in a cohort of patients treated with external-beam radiotherapy (EBRT) followed by a brachytherapy pulsed dose rate (PDR) boost. Methods and Materials: Between 2002 and 2008, 110 patients were treated with 46-Gy EBRT followed by PDR brachytherapy (24.96-28.80 Gy). The investigated outcome variables, GI toxicity, GU toxicity, erectile dysfunction, and IPSS were prospectively scored at several time points during follow-up. Association between time (as continuous and categorical variable) and the outcome variables was assessed using generalized linear models. Results: No statistically significant association was found between time (continuous) and GI toxicity (odds ratio [OR], 0.97; 95% confidence interval [CI], 0.89-1.06), GU toxicity (OR, 0.97; 95% CI, 0.91-1.03), erectile dysfunction (OR, 1.06; 95% CI, 0.99-1.11), and IPSS (-0.11; 95% CI, -0.41-0.20). Also, no statistically significant association was found between these variables and time as a categorical variable. GU toxicity was associated with IPSS resolution (OR, 1.16; 95% CI, 1.09-1.24). Posttreatment IPSS was associated with pretreatment IPSS (0.52; 95% CI, 0.25-0.79). Conclusions: No accumulation of high-grade toxicity over time could be established for a group of patients treated with EBRT and PDR brachytherapy for prostate cancer, probably because high-grade late toxicity resolves with time. Also, differences in IPSS values among patients are smaller after treatment than before treatment.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Thin film deposition of diamond using normal paraffins as source of diamond nucleation centers

Highlights: ► Paraffin compounds are diamond nucleation sources. ► Thermoconductivity of Cu–DTF device is higher than such conductivity of Cu. ► DTF growth in HFCVD reactor is not linear function of time. -- Abstract: We propose a process for diamond thin film (DTF) deposition using normal paraffins (nP) as source of diamond nucleation centers. We deposited micro-crystalline diamond thin films (MCDTF) on a Cu substrate using Hot Filament CVD (HFCVD) and Passive Pt/Pd Surface Catalysis (PPt/PdSC) methods. Beeswax and a 1:1 mixture of normal paraffins of the general formula CH{sub 3}(CH{sub 2}){sub n}CH{sub 3} with n = 22 and 26 were tested as nP starting material. The films obtained were characterized by scanning electronic microscopy (SEM), Raman scattering temperature dependent spectroscopy and X-ray diffraction (XRD) methods, all of which confirmed that the deposited material is MCDTF.

36 MATERIALS SCIENCE↗

Dosimetry of Microdistributed Dose-Enhancing Agents in X-ray Synchrotron Binary Therapy

Monte Carlo based modelling of the dose distribution in the vicinity of concentrates of iodine (I) and gold (Au) binary radiotherapy agents has been performed for monochromatised synchrotron X-rays. While the KERMA approximation, which ignores electron transport, is often acceptable for kilovoltage X-ray dosimetry in X-ray binary therapy, the range of photoelectrons and Auger electrons may be significant when compared to the microdostributed structure of the binary compound in which case corrections to the approximation may be necessary. Dose is calculated using EGSnrc for microdistributions associated with X-ray radiation synovecotomy, where iodine is taken up in the synovial lining. Dose as a function of the volume of aggregation for an Au-based contrast agent such as Au nanoparticles, ranging in diameter from 5 micron to 100 micron, were calculated using EGSnrc and Penelope, showing that the dose varies slowly for 90 keV X-rays, where much of the dose delivered by short range photoelectrons while 80 keV X-rays, just below the K-edge of Au (80.729 keV) increases linearly with diameter. In general the dose varies slowly as a function of volume suggesting that only small corrections will be needed to account for effects due to the failure of electronic equilibrium.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Multicomponent Transport in Polyatomic Reactive Gas Mixtures

We investigate multicomponent reactive flow models derived from the kinetic theory of gases. We discuss in particular the conservation equations, the transport fluxes and the transport coefficients in weak and strong magnetic fields. The mathematical properties of the resulting hyperbolic-parabolic systems of partial differential equations modeling multicomponent flows are deduced from the underlying kinetic framework. The structure and solution of the transport linear systems associated with the evaluation of transport coefficients are also addressed. In particular, the convergence of iterative techniques is deduced from the properties of the linearized Boltzmann collision operator. The impact of multicomponent transport is also discussed, notably the importance of Soret effects in various flows and the impact of volume viscosity.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Measuring Joule heating and strain induced by electrical current with Moire interferometry

This study proposes a new method to locate and measure the temperature of the hot spots caused by Joule Heating by measuring the free thermal expansion in-plane strain. It is demonstrated that the hotspot caused by the Joule heating in a thin metal film/plate structure can be measured by Phase shifting Moire interferometry with continuous wavelet transform (PSMI/CWT) at the microscopic scale. A demonstration on a copper film is conducted to verify the theory under different current densities. A correlation between the current density and strain in two orthogonal directions (one in the direction of the current flow) is proposed. The method can also be used for the measurement of the Joule heating in the microscopic solid structures in the electronic packaging devices. It is shown that a linear relationship exists between current density squared and normal strains.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Features of temperature control of fuel element cladding for pressurized water nuclear reactor “WWER-1000” while simulating reactor accidents

During the experiments simulating NPR (nuclear power reactor) accidents with a coolant loss fuel elements behavior in a steam-hydrogen medium was studied at the temperature changed with the rate from 1 to 100K/s within the range of 300÷1500 °C. Indications of the thermocouples fixed on the cladding notably differ from real values of the cladding temperatures in the area of measuring junction due to thermal resistance influence of the transition zones “cladding-junction” and “junction-coolant”. The estimating method of a measurement error was considered which can provide adequate accounting of the influence factors. The method is based on thermal probing of a thermocouple by electric current flashing through thermoelements under the coolant presence or absence, a response time registration and processing, calculation of thermal inertia value for a thermocouple junction. A formula was derived for calculation of methodical error under stationary mode and within the stage of linear increase in temperature, which will determine the conditions for the cladding depressurization. Some variants of the formula application were considered, and the values of methodical errors were established which reached ∼5% of maximum value by the final moment of the stage of linear increase in the temperature.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗