SEARCH · Search NASA

Results for “Estimation method”

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 307 records · Page 17

All sources NASA NTRS DOE OSTI

DOE OSTI2025

Source Characterization at Low Yields from Physics Experiment One A

The detonation of a well-recorded coupled explosion at the well-characterized Nevada National Security Site, as part of Physics Experiment One, has provided us with an opportunity to compare and contrast methods and results of yield estimation and other event characterization methods. Yield estimates are made using several methods, including local/regional magnitudes, waveform envelopes, acoustic amplitudes, joint seismoacoustic yield, and moment tensors. We include in our analysis a discussion and comparison of applicability, calibrations, assumptions, and uncertainties involved in each. With the exception of magnitude-based methods, all the techniques were more or less successful in recovering the known yield. It is, however, useful to be cognizant of requirements and calibrations, such as station coverage, emplacement conditions, propagation characterization, and uncertainties, to apply the methods to other areas with less complete information. We demonstrate that we can be successful at monitoring low-yield explosions using a variety of methods assuming that the signal-to-noise ratio is adequate at recording stations, and that the material properties are taken into account. Depth estimates are also important, particularly if one allows the possibility of a surface explosion.

Pasyanos, Michael E. [Lawrence Livermore National ↗

DOE OSTI2023

Quantifying spatiotemporal variability in occupant exposure to an indoor airborne contaminant with an uncertain source location

Well-mixed zone models are often employed to compute indoor air quality and occupant exposures. While effective, a potential downside to assuming instantaneous, perfect mixing is underpredicting exposures to high intermittent concentrations within a room. When such cases are of concern, more spatially resolved models, like computational-fluid dynamics methods, are used for some or all of the zones. But, these models have higher computational costs and require more input information. A preferred compromise would be to continue with a multi-zone modeling approach for all rooms, but with a better assessment of the spatial variability within a room. Here to do so, we present a quantitative method for estimating a room's spatiotemporal variability, based on influential room parameters. Our proposed method disaggregates variability into the variability in a room's average concentration, and the spatial variability within the room relative to that average. This enables a detailed assessment of how variability in particular room parameters impacts the uncertain occupant exposures. To demonstrate the utility of this method, we simulate contaminant dispersion for a variety of possible source locations. We compute breathing-zone exposure during the releasing (source is active) and decaying (source is removed) periods. Using CFD methods, we found after a 30 minutes release the average standard deviation in the spatial distribution of exposure was approximately 28% of the source average exposure, whereas variability in the different average exposures was lower, only 10% of the total average. We also find that although uncertainty in the source location leads to variability in the average magnitude of transient exposure, it does not have a particularly large influence on the spatial distribution during the decaying period, or on the average contaminant removal rate. By systematically characterizing a room's average concentration, its variability, and the spatial variability within the room important insights can be gained as to how much uncertainty is introduced into occupant exposure predictions by assuming a uniform in-room contaminant concentration. We discuss how the results of these characterizations can improve our understanding of the uncertainty in occupant exposures relative to well-mixed models.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

DOE OSTI2024

Optimal hybrid power plants for electric vehicle charging demand

Transmission constraints, increasing motivations to decarbonize, and concerns over peak electric vehicle (EV) load impacts on local grids have driven electric customers to consider behind-the-meter, hybrid power plant generation and storage at the distributed-grid level for EV charging. In this study, we develop capabilities to optimize hybrid power plant component capacities for EV charging. We then demonstrate these capabilities in a case study for Boulder, Colorado, using public EV charging data as well as wind and solar resource data. Our results show system designs that balance the cost of energy with load-meeting and peak shaving performance. Within the case study, systems designed for wind, solar photovoltaic (PV), and storage resulted in lower cost of energy than those optimized for PV and storage only. This indicates that in areas where wind resource exists, hybrid power plants that include wind, PV, and battery assets can better meet EV charging loads (including peak loads that are prone to overloading local grids) than PV and battery assets alone. Future work to address limitations in this paper include extending cost modeling to include performance losses (e.g., based on operations or weather) and charging station costs to estimate levelized cost of charging, and quantifying uncertainty and error in our aggregation methods for estimating EV charging loads at the hourly timescale.

14 SOLAR ENERGY↗

DOE OSTI2022

Quantitative Ceilometer-Radar Studies of Clouds Field Campaign Report

The Long Island Solar Farm (LISF) is a 32-megawatt solar photovoltaic power plant built through a collaboration including BP Solar, the Long Island Power Authority (LIPA), and the U.S. Department of Energy (DOE). The LISF, located on the Brookhaven National Laboratory site, began delivering power to the LIPA grid in November 2011, and is currently one of the largest solar photovoltaic power plants in the eastern United States. It is generating enough renewable energy to power approximately 4,500 homes and is helping New York State meet its clean energy and carbon reduction goals. Brookhaven National Laboratory (BN)L has the LISF instrumented and is capturing solar insolation and power data that can be used for research purposes. Additional information on LISF is available here: https://www.bnl.gov/lisf/. A network of nine high-definition (HD cameras) and 32 pyranometers are deployed within the LISF for the purpose of monitoring the location and characteristics of clouds and available global horizontal irradiance at the surface across the region. Information collected by this network every 30 s is currently used in BNL’s solar NowCasting algorithm, which forecasts near-term solar energy availability accounting for the behavior of clouds. The most important parameters for forecasting the near-term solar energy availability are accurate estimations of their horizontal extend, horizontal motion, and cloud base height. Currently, images from nearly located cameras are used to estimate the cloud base height. In order to evaluate the potential of this method for estimating the cloud base height, we decided to deploy a Vaisala ceilometer at the LISF.

14 SOLAR ENERGY↗

DOE OSTI2020

Variable-Speed Pump Efficiency Calculation For Fluid Flow Systems with and without Static Head

To accurately calculate pump energy savings gained from implementing variable frequency drive (VFD) controls, the variation of pump efficiency must be considered when operating conditions transition from the design operating point to new operating points. Many software tools require users to specify the new pump efficiency, or it is assumed to be unchanged. Unfortunately, many users have challenges of estimating the pump efficiency at new operating points. This paper presents a simplified method of estimating centrifugal pump efficiency at new operating speeds when the pump is controlled by a VFD. This methodology applies to systems with and without static head when the system curve is not affected by the change, and also systems where the change in operation changes the system curve.A hypothetical fluid flow system and centrifugal pump were used to demonstrate the calculation process for these scenarios. For this hypothetical system, the pump’s efficiency at new operating points was up to 5.4% lower than the design operating point.

42 ENGINEERING↗

DOE OSTI2021

Toward Robust and Routine Determination of Mw for Small Earthquakes: Application to the 2020 Mw 5.7 Magna, Utah, Seismic Sequence

To better characterize seismic hazard, particularly, for induced seismicity, there is an increasing interest in methods to estimate moment magnitude (M w ) for small earthquakes. M w is generally preferred over other magnitude types, but, it is difficult to estimate M w for earthquakes with local magnitude (M L ) <3-3.5, using conventional moment tensor (MT) inversion. The 2020 M ww 5.7 Magna, Utah, seismic sequence provides an opportunity to illustrate and evaluate the value of spectral methods for this purpose. Starting with a high-quality seismic catalog of 2103 earthquakes (M L <5.6), we estimate M w using two independent spectral methods—one based on direct waves, yielding M w,direct , and the other based on coda waves, yielding M w,coda . For the direct-wave method, we present a non-parametric (NP) inversion scheme that solves for apparent geometrical spreading, G(R), and site effects (S), similar to other NP procedures that have been used to calibrate regional M L scales. The NP inversion is constrained using M ws derived from MTs for nine events in the Magna sequence. We recover statistically robust and physically reasonable G(R) and S and compute M w,direct for 635 Magna earthquakes down to M L 0.7. For the coda-wave method, we consider two separate calibration schemes involving previous MT solutions and compute M w,coda for 311 earthquakes down to M L 1.0. For 280 of the events that were processed with both methods—M w,direct and M w,coda —are strongly correlated (r = 0.98), with a mean difference of only 0.05. We compare M w,direct and M w,coda with M L and find reasonably good agreement for M L <3.6 with the theoretically predicted relationship of M w =(2/3)M L +C, in which C is a regional constant. Our results imply that seismic network operators can use spectral-based M w estimates to replace M L estimates for events with M L ≥1.0, and possibly smaller. The main requirement is the existence of a small number of MT solutions for calibration purposes.

58 GEOSCIENCES↗

DOE OSTI2024

Functional Data Analysis in Wearable Body Sensor Networks

Improving response time of indirect room-size calorimeters is still an outstanding problem in metabolic research. Accurate estimates of instantaneous rates of gaseous exchange require numerical differentiation of measured gaseousgas concentrations. We propose a new method to estimate the instantaneous gaseousgas exchange rates in indirect calorimetry. In contrast to the previously developed techniques, the method addresses the problem of differentiation of gaseous concentrations as an ill-posed problem. By applying the method of regularization, the problem of differentiation is converted into a well-posed problem resulting in smooth and consistent gaseous exchange rates. The validity of the method is tested on a large dataset of calorimeter experiments which included 313 human experiments along with 231 alcohol combustion experiments. It is demonstrated that the method is able to reliably differentiate between the “unphysiological” process of alcohol combustion and physiological variations produced by human metabolism. The method also allowed unraveling the previously unreported relative kinetics of O2 consumption and Respiratory Quotient (RQ) in humans. It was found that the kinetics of oxidative fuel selection lags behind the energy expenditure in humans exhibiting some sort of oxidative inertia. The time lag varies from 2-3 min up to 30 min, depending on particular individual. No such lag was found in alcohol combustion experiments. In addition to the relative kinetics of substrate oxidation, two statistical indexes reflecting variability of minute-by-minute RQ were estimated. The indexes were the RQ’s standard deviation and RQ’s first-order derivative. Both indexes showed statistically significant difference between human experiments and alcohol combustion experiments. We conclude that the proposed method can consistently extract physiologically-relevant information from noisy calorimetry data and the aforesaid information can provide additional insights into the mechanism of metabolic fuel selection in humans.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2024

Functional Data Analysis in Wearable Body Sensor Networks

60 - APPLIED LIFE SCIENCES↗

DOE OSTI2020

An Inductor Current Estimation Approach for DC/DC Converters Based on Bisection Method

The trends toward higher power density and efficiency for the DC/DC converters has posed more strict requirements for inductors. The inductor current in turn affect the operating characteristics of the converters significantly and thus needs to be accurately estimated. In this paper, a new bisection-method-based inductor current estimation approach considering the variation of inductance is proposed. The proposed approach can be adapted and applied for various DC/DC converters, including the widely used buck and boost converters and etc. Comparing to the conventional estimation method using constant inductance, the proposed method can obtain more precise estimations of both the inductor current waveform and the power loss. Experimental results on a buck converter with a ferrite-core inductor are presented to validate the effectiveness of the proposed approach.

Zhao, Zhe↗

DOE OSTI2020

Estimating region-specific fuel economy in the United States from real-world driving cycles

Here we describe a method for estimating region-specific real-world light-duty vehicle fuel economy in the United States that is unique in both the size and representativeness of real-world driving that was considered, and for its ability to model regional variations in driving patterns. Over one million miles of national driving data were used to select real-world cycles representative of observed trip categories. The six cycles were compared to U.S. legislative cycles, revealing some key differences. Finally, a set of cycle weighting factors for 533 separate U.S. regions was derived from annual traffic statistics. Applying this method, it was found that regional fuel economy varies due to differences in driving patterns alone and that rural driving patterns lead to improved fuel economy (for conventional vehicles). The driving cycles and regional weighting factors described here are useful for testing and simulation studies, specifically those sensitive to regional variations in driving patterns.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

DOE OSTI2023

Protection Against Graph-Based False Data Injection Attacks on Power Systems

Graph signal processing (GSP) has emerged as a powerful tool for practical network applications, including power system monitoring. By representing power system voltages as smooth graph signals, recent research has focused on developing GSP-based methods for state estimation, attack detection, and topology identification. Included, efficient methods have been developed for detecting false data injection (FDI) attacks, which until now were perceived as non-smooth with respect to the graph Laplacian matrix. Consequently, these methods may not be effective against smooth FDI attacks. In this paper, we propose a graph FDI (GFDI) attack that minimizes the Laplacian-based graph total variation (TV) under practical constraints. In addition, we develop a low-complexity algorithm that solves the non-convex GDFI attack optimization problem using ell_1-norm relaxation, the projected gradient descent (PGD) algorithm, and the alternating direction method of multipliers (ADMM). We then propose a protection scheme that identifies the minimal set of measurements necessary to constrain the GFDI output to high graph TV, thereby enabling its detection by existing GSP-based detectors. Our numerical simulations on the IEEE-57 bus test case reveal the potential threat posed by well-designed GSP-based FDI attacks. Moreover, we demonstrate that integrating the proposed protection design with GSP-based detection can lead to significant hardware cost savings compared to previous designs of protection methods against FDI attacks.

Morgenstern, Gal↗

DOE OSTI2026

Modeling the contributions to acoustic nonlinearity from complex dislocation networks using 3D dislocation dynamics

Nonlinear ultrasonic parameters are highly sensitive to microstructural features that affect macroscale material behavior, providing a nondestructive means to characterize their evolution. Although dislocations are known to be a strong source of acoustic nonlinearity, establishing quantitative links between the acoustic nonlinearity parameter (β), measured via Second Harmonic Generation, and dislocation morphology—such as dislocation length and density—remains an open challenge. This work advances the numerical modeling of dislocation–β relationships using 3D dislocation dynamics (DD) simulations in two approaches: a “static” method computing strain and stress fields from dislocation configurations in the absence of external loading, and a “quasi-static” method to estimate β from the curvature of dislocation lines under applied load. First, the static method is combined with finite element analysis to investigate a recent assertion that heterogeneous initial strain fields can induce higher harmonic generation in a linear elastic medium; the present results do not corroborate this outcome. Then, the quasi-static method is applied to multiple-dislocation scenarios through parametric studies, revealing behaviors not predicted by analytical models, such as the competing interactions of edge and screw dislocations and the significant influence of applied stress on β. Finally, the simulations are used to model SHG experimental results and validate the hypothesis that β can decrease during plastic deformation, despite increasing dislocation density. As the DD code used here is open-source, it provides a practical platform for future investigation into microstructure–β relationships important to the interpretation of SHG results.

Materials science↗

DOE OSTI2022

Experimental Investigation of Low-Frequency Distributed Acoustic Strain-Rate Responses to Propagating Fractures

Summary Low-frequency distributed acoustic sensing (LF-DAS) responds to strain changes due to far-field fracture propagation. To better understand the LF-DAS response to fracture propagation, we performed laboratory-scale hydraulic fracture experiments with embedded optical strain sensors. The objectives of this research are to generate hydraulic fractures of known geometry, measure the strain response along the embedded fiber-optic cable comparable to LF-DAS measurements, and use the results to inform the interpretation of field-derived LF-DAS data. The experiments were conducted in unconfined transparent cubic blocks with a dimension of 8 in. on each side. The block was made of transparent epoxy to visualize the fracture propagation. Fiber-optic sensing cables were embedded in the block at different distances to the source of injection. We injected dyed water through an injection tubing to generate a transverse, radial fracture along an initial flaw. An optical interrogator recorded the response of offset fiber Bragg grating (FBG) strain sensors normal to the plane of the fracture. The strain data were visualized on a waterfall plot, akin to visualizations of field-derived LF-DAS data. Dimensional analysis was used to scale the laboratory results to field conditions. We compared the evolution of the strain response at the fiber-optic cable, injection pressure, and rate with known fracture geometry. The measured strains were compared with Sneddon’s (1946) linear elastic solution for a penny-shaped crack and found to follow this behavior. The generated radial fractures in transparent media can be modeled with Sneddon’s linear elastic radial fracture model and a Mode I critical stress intensity factor. The LF-DAS characteristic response of a narrowing region of extension surrounded by compression was exhibited as a fracture that approached and intersected the fiber-optic cable. The experimentally derived strain and strain-rate waterfall plots with known fracture geometry, injection rate, and pressure response provide insight in understanding LF-DAS responses in the field. Furthermore, we developed a method to estimate fracture geometry evolution from the fiber-optic strain data and validated the method against the experimental data.

Engineering↗

DOE OSTI2022

A discussion on various experimental methods of impact ionization coefficient measurement in GaN

Impact ionization coefficients play a critical role in semiconductors. In addition to silicon, silicon carbide and gallium nitride are important semiconductors that are being seen more as mainstream semiconductor technologies. As a reflection of the maturity of these semiconductors, predictive modeling has become essential to device and circuit designers, and impact ionization coefficients play a key role here. Recently, several studies have measured impact ionization coefficients. We dedicated the first part of our study to comparing three experimental methods to estimate impact ionization coefficients in GaN, which are all based on photomultiplication but feature characteristic differences. The first method inserts an InGaN hole-injection layer, the accuracy of which is challenged by the dominance of ionization in InGaN, leading to possible overestimation of the coefficients. The second method utilizes the Franz–Keldysh effect for hole injection but not for electrons, where the mixed injection of induced carriers would require a margin of error. The third method uses complementary p–n and n–p structures that have been at the basis of this estimation in Si and SiC and leans on the assumption of a constant electric field, and any deviation would require a margin of error. In the second part of our study, we evaluated the models using recent experimental data from diodes demonstrating avalanche breakdown.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

DOE OSTI2026

Validation of an Erythema-Weighted UV Model Using Broadband Solar Irradiance Measurements From Eleven U.S. Sites: Preprint

Erythema-weighted UV solar irradiance (UV-E) has a potential impact on human health if the recommended maximum exposure times are exceeded. In spite of this, it is not measured at most sites that measure Global Horizontal Irradiance (GHI). However, since UV-E is highly correlated with GHI and total ozone content it can be estimated from this information with sufficient accuracy to assisst in public health recommendations. The Power Model (PM) provides a simple method to estimate the erythema-weighted UV irradiance (UV-E) from measured GHI, total ozone column and air mass. In this work, the performance of the PM method is assessed using high-quality data from 11 sites in the continental U.S. (part of SURFRAD and SOLRAD networks) and total ozone estimates publicly available from the MERRA-2 re-analysis database. A three year period (2021- 2023) at 1-minute frequency is considered. The results (for time aggregations of 5 and 60 minutes) show a high Pearson's correlations (> 0.99), consistently positive mean bias deviations (below 13%) and dispersions in the 8-19% range at all sites. Relative values are expressed in terms of the corresponding measurement mean. The performance indicators remain consistent across time resolutions (5 or 60 minutes), suggesting that the model's performance is robust and not significantly affected by short-term variability (which is captured by GHI). Spatial patterns reveal higher biases and RMSD in northern and eastern locations. These values represent a significant improvement over widely used satellite-based global UV-E estimates and open the possibility of using the PM with satellite-based GHI estimates for operational UV-E mapping over the contiguous U.S. territory.

14 SOLAR ENERGY↗

DOE OSTI2023

System and method for simulation-assisted additive manufacturing involving estimating anisotropic deformation

A simulation-assisted additive manufacturing system and method for generating an anisotropic compensation to account for non-uniform deformation due to additive manufacturing and service loading. The predicted deformation may not be fully defined. The present disclosure provides a system and method for estimating the missing deformation data through regression analysis. The present disclosure also provides an integrated framework where the various simulated-assisted design modules are configured for two-way communication and sharing access to changes to the model.

Kim, Seokpum↗

DOE OSTI2024

NORTH DAKOTA CARBONSAFE PHASE III: SITE CHARACTERIZATION AND PERMITTING OF GEOLOGIC STORAGE OF CARBON DIOXIDE

The Energy & Environmental Research Center (EERC), in partnership with Minnkota Power Cooperative Inc. (Minnkota), SLB, and Computer Modelling Group Ltd. (CMG), supported wide-scale deployment of carbon capture and storage (CCS) as part of the U.S. Department of Energy (DOE) National Energy Technology Laboratory Carbon Storage Assurance Facility Enterprise (CarbonSAFE) Initiative Phase III. This phase included the acquisition, analysis, and development of information to fully characterize two storage complexes to demonstrate viable storage resources for commercial volumes of CO2 (defined by DOE as a minimum of 50 million tonnes [MMt] of CO2 within a 30-year period) (National Energy Technology Laboratory, 2024). Phase III also involved the preparation, submission, and approval of North Dakota underground injection control (UIC) Class VI storage facility permits (SFPs)—required precursors to applications for Class VI injection well permits. The presumed viability of commercial-scale CCS, situated adjacent to Minnkota’s Milton R. Young Station (MRYS), is validated by Minnkota’s continued pursuit of Project Tundra—an initiative to build the world’s largest lignite-based CCS project in central North Dakota (www.projecttundrand.com). Project Tundra comprises two scopes of work, Tundra Capture (installation of postcombustion CO2 capture at MRYS) and Tundra SGS (secure geologic storage). The efforts of North Dakota CarbonSAFE Phase III, Site Characterization and Permitting, supported Tundra SGS. Extensive site-specific characterization activities included a successful multimeasurement geophysical approach and drilling a stratigraphic test well (J-ROC 1, subsequently renamed Liberty-1) adjacent to MRYS. Core collection and analyses, downhole testing and fluid sampling, and geophysical logging were performed on J-ROC 1 and on a nearby stratigraphic test well (known as J-LOC 1), which was drilled, cored, and tested under a complementary project funded by the North Dakota Lignite Research Program. The injection tests performed on J-LOC 1 positively impacted the CarbonSAFE project, resulting in fewer proposed injection wells and significant construction, operations, and monitoring cost savings. The characterization data collected and analyses performed were integrated into geologic models, and successive numerical simulations were run to determine CO2 plume extent and subsurface pressure buildup associated with the planned CO2 injection rate of nearly 4 MMt per year. The latter doubles the CarbonSAFE Initiative goal with an estimated 100 MMt of CO2 stored in 20 years. Application of the U.S. Environmental Protection Agency’s (EPA’s) method for estimating the Class VI Rule area of review (AOR) to the overpressurized Broom Creek Formation inspired an alternative method of calculation, called risk-based AOR delineation. This peer-reviewed method was applied for the first time during the storage facility-permitting process. The two SFP applications submitted in 2021 successfully resulted in North Dakota Industrial Commission (NDIC) orders in 2022 authorizing the creation of the storage facility areas and amalgamation of pore space as well as establishing financial responsibility requirements. After approval of the SFPs, Minnkota filed in 2022 applications for permits to reenter the J-ROC 1 well and to drill two new wells—all with the intended purpose to become Class VI injection wells. To establish eligibility under the Internal Revenue Code for Section 45Q tax incentives, a monitoring, reporting, and verification (MRV) plan was prepared and submitted by Minnkota to EPA in 2022, resulting in the first such plan approved in North Dakota. Also in 2022, under the National Environmental Policy Act (NEPA), Minnkota prepared and submitted an environmental information volume (EIV) describing the proposed CCS project and associated potential environmental impacts. Based on the EIV, DOE determined that the proposed construction project required an environmental assessment, and Minnkota submitted the first draft in 2023 and a revised draft in 2024. Both submissions were followed by a public comment period. Subsequently, DOE issued a finding of no significant impact (FONSI) on September 13, 2024. A successful outreach program, strongly based in the production, presentation, and dissemination of informational material, fostered an environment to aid stakeholders in making informed decisions regarding the planned project. Opportunities for public input were provided at various steps along the way, including at county planning and zoning meetings, before and during the SFP administrative hearing, and during environmental assessment public comment periods. In addition, land/pore space owners and mineral owners had various points of contact, including granting access rights, securing pore space leasing, and mineral owner notifications. Based upon the successful storage facility permitting issued by NDIC, approval of the MRV plan by EPA, and receipt of a FONSI under the NEPA, Minnkota is continuing its pursuit of Project Tundra. In December 2023, the Office of Clean Energy Demonstrations under its Carbon Capture Demonstrations Projects Program announced funding for the capture system (Office of Clean Energy Demonstrations, 2023) and a proposal for CarbonSAFE Phase IV: Construction funding was submitted in March 2024 for the storage project. A go/no-go decision to proceed with construction and operations in the Broom Creek Formation is anticipated in 2024. References National Energy Technology Laboratory, CarbonSafe Initiative, https://netl.doe.gov/carbon-management/carbon-storage/carbonsafe (accessed August 2024). Office of Clean Energy Demonstrations, 2023, OCED selects three projects in CA, ND, and TX to reduce harmful carbon pollution, create new economic opportunities, and advance carbon reducing technologies, December, www.energy.gov/oced/articles/oced-selects-three-projects-ca-nd-and-tx-reduce-harmful-carbon-pollution-create-new (accessed August 2024).

Peck, Wesley↗

DOE OSTI2021

Multi-fidelity Bayesian neural networks: Algorithms and applications

Here we propose a new class of Bayesian neural networks (BNNs) that can be trained using noisy data of variable fidelity, and we apply them to learn function approximations as well as to solve inverse problems based on partial differential equations (PDEs). These multi-fidelity BNNs consist of three neural networks: The first is a fully connected neural network, which is trained following the maximum a posteriori probability (MAP) method to fit the low-fidelity data; the second is a Bayesian neural network employed to capture the cross-correlation with uncertainty quantification between the low- and high-fidelity data; and the last one is the physics-informed neural network, which encodes the physical laws described by PDEs. For the training of the last two neural networks, we first employ the mean-field variational inference (VI) to maximize the evidence lower bound (ELBO) to obtain informative prior distributions for the hyperparameters in the BNNs, and subsequently we use the Hamiltonian Monte Carlo (HMC) method to estimate accurately the posterior distributions for the corresponding hyperparameters. We demonstrate the accuracy of the present method using synthetic data as well as real measurements. Specifically, we first approximate a one- and four-dimensional function, and then infer the reaction rates in one- and two-dimensional diffusion-reaction systems. Moreover, we infer the sea surface temperature (SST) in the Massachusetts and Cape Cod Bays using satellite images and in-situ measurements. Taken together, our results demonstrate that the present method can capture both linear and nonlinear correlation between the low- and high-fidelity data adaptively, identify unknown parameters in PDEs, and quantify uncertainties in predictions, given a few scattered noisy high-fidelity data. Finally, we demonstrate that we can effectively and efficiently reduce the uncertainties and hence enhance the prediction accuracy with an active learning approach, using as examples a specific one-dimensional function approximation and an inverse PDE problem.

97 MATHEMATICS AND COMPUTING↗