Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “deep”

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 1,135 records · Page 63

Completion design improvement using a deep convolutional network

Maximizing stimulated natural and hydraulic fracture network is one of the primary hydraulic fracturing concerns for economic production from a horizontal shale gas well. Geomechanical facies and preexisting fractures in each stage are identified based on similarities in formation characteristics to optimize the locations of perforation clusters. This often requires analyzing large volumes of drilling, Logging While Drilling (LWD) and Measurement While Drilling (MWD) data. In this paper, we develop a methodology that calculates the mechanical specific energy (MSE) using real-time drill string acceleration signals directly from its definition. High resolution vibration signals have been collected using a tri-axial accerlometer, which was an auxiliary tool included in acoustic borehole imager. This technique provides a cost-efficient solution for engineered completion design. Furthermore, we adopt deep Convolutional Neural Network (CNN) with signal processing to build a data pipeline that effectively extracts patterns from dynamic acceleration signals for rock lateral MSE classification. First, we apply discrete wavelet transform and Short-Time Fourier Transform (STFT) for signal denoising and pattern recognition. Then we construct an image dataset using multi-scale image fusion at pixel level from 3 sensor channels, including axial, lateral acceleration spectrograms and zero-padded revolutions per minute (RPM). The resulted RGB image dataset includes 4,000 images of 5 MSE ranges with various rock strength conditions. Our results demonstrate that the proposed deep learning model can achieve more than 90% classification accuracy. The deep learning results, as a reference source, were applied in selected Marcellus Shale Energy and Environmental Lab (MSEEL) wells engineered completion located in the Marcellus shale gas site.

03 NATURAL GAS↗

Deep Learning for Pion Identification and Energy Calibration with the ATLAS Detector

Separating charged and neutral pions as well as calibrating the pion energy response is a core component of reconstruction in the ATLAS calorimeter. This note presents an investigation of deep learning techniques for these tasks, representing the signal in the ATLAS calorimeter layers as pixelated images. Deep learning approaches outperform the classification applied in the baseline local hadronic calibration and are able to improve the energy resolution for a wide range in particle momenta, especially for low energy pions. This work demonstrates the potential of deep-learning-based low-level hadronic calibrations to significantly improve the quality of particle reconstruction in the ATLAS calorimeter.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

A Search for Sterile-Neutrino-Based Muon Neutrino Disappearance Using the MicroBooNe Deep Learning Analysis

We describe a search for νµ disappearance using the MicroBooNE Deep Learning analysis 1µ1p selection. Presently, the unexplained MiniBooNE and LSND anomalies could be explained by a sterile neutrino impacting neutrino oscillations. Our analysis searches for the allowed parameter space that could describe such a sterile neutrino. We determine the allowed and excluded region of a 3+1 sterile-based muon neutrino disappearance model in MicroBooNE at 90% confidence. Our allowed region includes both the null model, and current global best fit model. Context for the underlying Deep Learning analysis is provided and several validation studies surrounding both the disappearance search, and originating 1µ1p selection are performed to strengthen confidence in the result. In addition, a next-generation deep learning tool for cosmic-ray-muon event discrimination is proposed and evaluated, demonstrating a removal of 70% of the remaining event background when added to current methods, under the cut criteria used.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

A Search for Sterile-Neutrino-Based Muon Neutrino Disappearance Using the MicroBooNe Deep Learning Analysis

We describe a search for νµ disappearance using the MicroBooNE Deep Learning analysis 1µ1p selection. Presently, the unexplained MiniBooNE and LSND anomalies could be explained by a sterile neutrino impacting neutrino oscillations. Our analysis searches for the allowed parameter space that could describe such a sterile neutrino. We determine the allowed and excluded region of a 3+1 sterile-based muon neutrino disappearance model in MicroBooNE at 90% confidence. Our allowed region includes both the null model, and current global best fit model. Context for the underlying Deep Learning analysis is provided and several validation studies surrounding both the disappearance search, and originating 1µ1p selection are performed to strengthen confidence in the result. In addition, a next-generation deep learning tool for cosmic-ray-muon event discrimination is proposed and evaluated, demonstrating a removal of 70% of the remaining event background when added to current methods, under the cut criteria used.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

On the Stochastic Stability of Deep Markov Models

Deep Markov models (DMM) are generative models which are scalable and expressive generalization of Markov models for representation, learning, and inference problems. DMMs using deep neural networks to parametrize the transition of Markov probability distributions have recently been shown to provide more expressiveness in modeling sequential data and dynamical system responses. However, the fundamental stochastic stability guarantees of such models have not been thoroughly investigated. In this paper, we present a rigorous analytical method to prove the necessary and sufficient conditions of DMM's stochastic stability. This task is achieved by spectral analysis of the efficiently computed Jacobians of probabilistic maps modeled by deep neural networks. We make theoretical connections between the eigenvalues of neural network's weights and the different activation function types used on the stability and overall dynamic behavior of DMMs with Gaussian distributions. We empirically substantiate our theoretical results on stochastic stability and eigenvalue spectra via several numerical experiments. Formal stability guarantees of DMMs can substantially improve their robustness and trustworthiness, necessary for reliable use in safety-critical real-world applications.

Drgona, Jan↗

Diffusion Dominant Solute Transport Modelling in Fractured Media Under Deep Geological Environment - 12211

Deep geologic disposal of high activity and long-lived radioactive waste is gaining increasing support in many countries, where suitable low permeability geological formation in combination with engineered barriers are used to provide long term waste contaminant and minimise the impacts to the environment and risk to the biosphere. This modelling study examines the solute transport in fractured media under low flow velocities that are relevant to a deep geological environment. In particular, reactive solute transport through fractured media is studied using a 2-D model, that considers advection and diffusion, to explore the coupled effects of kinetic and equilibrium chemical processes. The effects of water velocity in the fracture, matrix porosity and diffusion on solute transport are investigated and discussed. Some illustrative modelled results are presented to demonstrate the use of the model to examine the effects of media degradation on solute transport, under the influences of hydrogeological (diffusion dominant) and microbially mediated chemical processes. The challenges facing the prediction of long term degradation such as cracks evolution, interaction and coalescence are highlighted. The potential of a novel microstructure informed modelling approach to account for these effects is discussed, particularly with respect to investigating multiple phenomena impact on material performance. The GRM code is used to examine the effects of media degradation for a geological waste disposal package, under the combined hydrogeological (diffusion dominant) and chemical effects in low groundwater flow conditions that are typical of deep geological disposal systems. An illustrative reactive transport modelling application demonstrates the use of the code to examine the interplay of kinetic controlled biogeochemical reactive processes with advective and diffusive transport, under the influence of media degradation. The initial model results are encouraging which show the disposal system to evolve in a physically realistic manner. In the example presented the reactive-transport coupling develops chemically reducing zones, which limit the transport of uranium. This illustrates the potential significance of media degradation and chemical effect on the transport of radionuclides which would need to be taken into account when examining the long-term behaviour and containment properties of the geological disposal system. Microstructure-informed modelling and its potential linkage with continuum flow modelling is a subject of ongoing studies. The approach of microstructure-informed modelling is discussed to provide insight and a mechanistic understanding of macroscopic parameters and their evolution. The proposed theoretical and methodological basis for microstructure-informed modelling of porous quasi-brittle media has the potential to develop into an explanatory and predictive tool for deriving mechanism-based, as opposed to phenomenological, evolution laws for macroscopic properties. These concepts in micro-scale modelling are likely to be applicable to the diffusion process, in addition to advective transport illustrated here for porous media. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Terminal Sinking Velocity for Waste Packages Falling in a Deep Borehole - 17063

This evaluation of terminal velocity for a waste package sinking in a deep borehole filled with emplacement fluid was done to support concept development for the Deep Borehole Field Test (DBFT). The DBFT objective is to evaluate the safety and feasibility of the deep borehole disposal (DBD) concept for certain radioactive wastes. The DBD concept calls for emplacing waste packages in the lower 2 km of a borehole drilled into crystalline basement rock to a depth of 5 km. Previous studies identified that accidentally dropping one or more waste packages during emplacement could be an important risk, particularly if a waste package is breached. This study evaluated the range of terminal sinking velocity, which can be used to design waste packages and impact limiters to ensure that breach does not occur. Based on an analytical model of a sinking container, a highly turbulent flow is expected to develop in the emplacement fluid. Turbulence is a complex three-dimensional phenomenon and the applicability of analytical models is limited. To backstop the analytical model and to extend predictions to other package dimensions and fluid properties, numerical simulation was performed using the ANSYS Fluent computational fluid dynamics code. Terminal velocity was calculated for downhole temperatures up to 1200 C, and three emplacement fluids (water, NaCl brine, and NaBr brine). Simulation results confirmed that flow around the sinking container is highly turbulent (Reynolds number 5.6 x 10{sup 4} to 7.3 x 10{sup 5}). Turbulence increases with temperature and decreases with density. Calculated terminal velocity ranges from 1.95 (200 C) to 2.13 m/s (1200 C) in water; 1.61 (200 C) to 1.79 m/s (1200 C) in NaCl; and 1.30 (200 C) to 1.46 m/s (1200 C) in NaBr. Terminal velocity is inversely related to fluid density, and the temperature effect on density appears to have a stronger effect on terminal velocity than the temperature effect on viscosity. The foregoing results are for non-perforated casing, but in practice the casing may be perforated to prepare for cementing, and to allow relief of fluid pressure from thermal expansion after emplacement of heat generating waste, and after final borehole closure. To represent the effect from perforations, the analytical solution was modified to include bypass flow through perforations and in the annulus. A workable perforation scheme is identified, consistent with borehole construction and closure requirements, using 2-cm and 5-cm diameter perforations. This scheme is estimated to limit the terminal velocity to 3 m/s, for reference waste packages falling in water at 400 C. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Cleaning up Hanford: A Deep Borehole Disposal Concept for the Cs/Sr Capsules - 16249

The Hanford site in Washington State is possibly the most radioactive-contaminated site in the USA presenting a major environmental problem requiring urgent clean-up. Sealed capsules containing halides of strontium (Sr) and cesium (Cs) account for over a third of the total radioactivity. Interim storage for a period of 10-half-lives is an unappealing option since many of the capsules contain Cs-135, which has a half-life of 2.3 million years, and the dose received from the shorter-lived isotopes would remain dangerously high for several generations. Deep geological disposal of the capsules would be a safer solution. We discuss a new and revised conceptual model for disposal of the entire Hanford capsule inventory in a single deep geological borehole with a diameter that could be as little as 31 cm. The concept involves placing thirty capsules inside a steel-disposal container. Holes up to this size are regularly drilled to depths much greater than 5 km by the hydrocarbon industry and are well within current drilling capabilities. Numerical modeling work demonstrates the feasibility of the conceptual model. In particular we examine the thermal envelope generated as a function of time enabling decisions to be made on possible fillings for the borehole annulus. Variations on the disposal concept include disposing different combinations of capsules to control the heat output and thus the temperature rise. The results lend support to a comment attributed to Secretary Moniz that the Hanford capsules could be the world's first HLW to be disposed of in a deep borehole. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Advances in Thermal-Hydrologic Modeling of Nuclear Waste Disposal in Deep Boreholes - 16303

Deep borehole disposal is one of the potential options for safely isolating high-level radioactive waste. It is expected that existing drilling technology can provide reliable and cost-effective construction of suitable deep boreholes. In addition, favorable disposal conditions such as low permeability host rock, high salinity, and geochemically reducing conditions, exist at depth in many locations. Coupled thermal-hydrologic processes induced by heat from the radioactive waste may impact fluid flow and the associated migration of radionuclides. This work looks at those processes as was also done with recent studies. Simulations of thermal-hydrology for the emplacement of cesium and strontium capsules in a deep borehole are presented. The simulations looked at disposal options such as different disposal configurations and aging of waste to reduce thermal effects. The simulations studied temperature and fluid flux in the vicinity of the borehole. Simulation results include temperature and vertical flux profiles around the borehole at selected depths. Of particular importance are peak temperatures, and fluxes above the disposal zone. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Development of a Universal Canister for Disposal of High-Level Waste in Deep Boreholes - 16482

The mission of the United States Department of Energy's (DOE's) Office of Environmental Management is to complete the safe cleanup of the environmental legacy brought about from five decades of nuclear weapons development and government-sponsored nuclear energy research. Some of the wastes that must be managed have been identified as good candidates for disposal in a deep borehole in crystalline rock. In particular, wastes that can be disposed of in a small package are good candidates for this disposal concept. A canister-based system that can be used for handling these wastes during the disposition process (i.e., storage, transfer, transportation, and disposal) could facilitate the eventual disposal of these wastes. Development of specifications for the universal canister system will consider the regulatory requirements that apply to storage, transportation, and disposal of the capsules, as well as operational requirements and limits that could affect the design of the canister (e.g., deep borehole diameter). In addition, there are risks and technical challenges that need to be recognized and addressed as Universal Canister system specifications are developed. This paper provides an approach to developing specifications for such a canister system that is integrated with the overall efforts of the DOE's Used Fuel Disposition Campaign's Deep Borehole Field Test and compatible with planned storage of potential borehole candidate wastes. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Hydrologic Modeling of Subsurface Brine Density Stratification for Deep Borehole Disposal - 20213

Density-dependent and topographically-driven subsurface brine flow was modeled in regional-scale and Deep Borehole Disposal-scale domains under varying conditions. These conditions included variable basement permeability, brine density, and hydraulic head. Findings revealed the possibility of hydraulic isolation of deep basement for at least 1 million years under appropriate conditions, which supports the viability of the Deep Borehole Disposal concept. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

A Study of Physics-Informed Deep Learning for System Fluid Dynamics Closures

The sub-grid-scale (SGS) physics models, or so-called closure relations (CR), are essential in thermal-fluid modeling and simulation codes, ranging from large-eddy simulation to system-level thermo-fluid dynamic problems. While details vary, all thermo-fluid codes solve coupled mass-momentum-energy conservation partial differential equations (PDE). Traditionally, the analytical form of closure models has to be known, and then we calibrate the model using data from a set of applicable experiments. Characteristically, it takes long effort to gain insights and mechanistic understanding needed for developing closure models. Such a model is a compact form of data. The long time required for new model development limits the pace of model application while dealing with new geometries, new coolants or system conditions, particularly in newly designed systems. Alternative to mechanistic and semi-analytical models, it is possible to discover the pattern of closure relations by applying the modern machine learning methods to the body of relevant experimental data. Deep learning (DL) is a relatively new branch in the machine learning family, and a major breakthrough has been achieved since Hinton introduces a fast algorithm to efficiently train a feed-forward neural network (NN) with multilayer perceptrons. Generally, any NN with more than two layers is considered as deep learning, and it contains a large number of hyper-parameters to achieve pattern recognition or regression for complex datasets. Hornik shows that multilayer NN is compatible with universal approximation theorem. There is no constraint for a multilayer NN to capture the properties of any measurable information. This capability makes DL attractive for closure models in a system thermal-hydraulic code. As system geometry and/or operating conditions change, new physical experiments (increasingly sophisticated) are conducted to support evaluation and revision of closure models. Such a so-called data-driven modeling relaxes the structural form of system codes to directly assimilate data instead of hard-wiring the model to pre-calibrated correlations. The objective of this research is to enable cost-effective development of a DL-based closure that reflects the data. The previous work on applying machine learning techniques to fluid dynamics focused on statistical learning with basic NN from direct numerical simulations (DNS) dataset. In the present work, we show that DL-based closure models can cause the fluid calculation to diverge if multilayer NN is used. To address the issue, we investigate conditions by which the DL-based closure models work compatibly, stably, and effectively with PDE-constrained forward prediction problems. Optimally meeting the compatibility conditions constitutes a physics-informed deep learning (PIDL) strategy. We refer to this process as PDE-constrained machine learning or PDE-ML. To develop the technical basis for PDE-ML using DL-based fluid dynamics closure, numerical experiments are performed for manufactured problems of increasing complexity. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

EVs@Scale High-Power Charging (HPC) Pillar Deep-Dive Technical Meeting

EVs@Scale Lab Consortium is addressing challenges, developing solutions, and enabling technologies for transportation electrification ecosystem. The consortium has five research pillars one of which focuses on high power charging (HPC). HPC pillar brings together hardware and software expertise, capabilities, and facilities related to high power EV charging, charge management, and grid integration. Deep-dive technical meetings are organized twice a year and provides an opportunity for more industry engagement and technical feedback for the national labs throughout the project lifecycle. High-Power Charging pillar has two active projects: (i) Next-Gen Profiles (NGP) and (ii) High-Power Electric Vehicle Charging Hub Integration Platform (eCHIP). This presentation summarizes the progress made in both projects during the past six months focusing on specific topics. There will be two deep-dive technical meetings twice a year. Every deep-dive meeting will focus on different aspects of the project progress.

ADVANCED PROPULSION SYSTEMS↗

DeepUQ: Assessing the Aleatoric Uncertainties from two Deep Learning Methods

Assessing the quality of aleatoric uncertainty estimates from uncertainty quantification (UQ) deep learning methods is important in scientific contexts, where uncertainty is physically meaningful and important to characterize and interpret exactly. We systematically compare aleatoric uncertainty measured by two UQ techniques, Deep Ensembles (DE) and Deep Evidential Regression (DER). Our method focuses on both zero-dimensional (0D) and two-dimensional (2D) data, to explore how the UQ methods function for different data dimensionalities. We investigate uncertainty injected on the input and output variables and include a method to propagate uncertainty in the case of input uncertainty so that we can compare the predicted aleatoric uncertainty to the known values. We experiment with three levels of noise. The aleatoric uncertainty predicted across all models and experiments scales with the injected noise level. However, the predicted uncertainty is miscalibrated to $\rm{std}(\sigma_{\rm al})$ with the true uncertainty for half of the DE experiments and almost all of the DER experiments. The predicted uncertainty is the least accurate for both UQ methods for the 2D input uncertainty experiment and the high-noise level. While these results do not apply to more complex data, they highlight that further research on post-facto calibration for these methods would be beneficial, particularly for high-noise and high-dimensional settings.

Nevin, Rebecca↗

Nov. 2024 EVs@Scale High-Power Charging Deep Dive Technical Meeting

EVs@Scale Lab Consortium is addressing challenges, developing solutions, and enabling technologies for transportation electrification ecosystem. The consortium has five research pillars one of which focuses on high power charging (HPC). HPC pillar brings together hardware and software expertise, capabilities, and facilities related to high power EV charging, charge management, and grid integration. Deep-dive technical meetings are organized twice a year and provides an opportunity for more industry engagement and technical feedback for the national labs throughout the project lifecycle. High-Power Charging pillar has two active projects: (i) Next-Gen Profiles (NGP) and (ii) High-Power Electric Vehicle Charging Hub Integration Platform (eCHIP). This presentation summarizes the progress made in both projects during the past six months focusing on specific topics. There will be two deep-dive technical meetings twice a year. Every deep-dive meeting will focus on different aspects of the project progress.

ADVANCED PROPULSION SYSTEMS,DIRECT ENERGY CONVERSI↗

Controlling distributed energy resources via deep reinforcement learning for load flexibility and energy efficiency

Behind-the-meter distributed energy resources (DERs), including building solar photovoltaic (PV) technology and electric battery storage, are increasingly being considered as solutions to support carbon reduction goals and increase grid reliability and resiliency. However, dynamic control of these resources in concert with traditional building loads, to effect efficiency and demand flexibility, is not yet commonplace in commercial control products. Traditional rule-based control algorithms do not offer integrated closed-loop control to optimize across systems, and most often, PV and battery systems are operated for energy arbitrage and demand charge management, and not for the provision of grid services. More advanced control approaches, such as MPC control have not been widely adopted in industry because they require significant expertise to develop and deploy. Recent advances in deep reinforcement learning (DRL) offer a promising option to optimize the operation of DER systems and building loads with reduced setup effort. However, there are limited studies that evaluate the efficacy of these methods to control multiple building subsystems simultaneously. Additionally, most of the research has been conducted in simulated environments as opposed to real buildings. This paper proposes a DRL approach that uses a deep deterministic policy gradient algorithm for integrated control of HVAC and electric battery storage systems in the presence of on-site PV generation. The DRL algorithm, trained on synthetic data, was deployed in a physical test building and evaluated against a baseline that uses the current best-in-class rule-based control strategies. Performance in delivering energy efficiency, load shift, and load shed was tested using price-based signals. The results showed that the DRL-based controller can produce cost savings of up to 39.6% as compared to the baseline controller, while maintaining similar thermal comfort in the building. The project team has also integrated the simulation components developed during this work as an OpenAIGym environment and made it publicly available so that prospective DRL researchers can leverage this environment to evaluate alternate DRL algorithms.

24 POWER TRANSMISSION AND DISTRIBUTION↗

The diffractive contribution to deep inelastic lepton-proton scattering: Implications for QCD momentum sum rules and parton distributions

The cross section for deep inelastic lepton-proton scattering (DIS) ℓp→ℓ'p'X includes a diffractive deep inelastic (DDIS) contribution ℓp→ℓ'p'X, in which the proton remains intact with a large longitudinal momentum fraction x F greater than 0.9 and small transverse momentum. The DDIS events, which can be identified with Pomeron exchange in the t-channel, account for approximately 10% of all of the DIS events. Thus, when one measures DIS, one automatically includes the leading-twist Bjorken-scaling DDIS events as a contribution to the DIS cross section, whether or not the final-state proton p' is detected. In such events, the missing momentum fraction x p ' ~0.9 carried by the final-state proton p' in the DDIS events could be misidentified with the light-front momentum fraction carried by sea quarks or gluons in the protons' Fock structure. As we shall show in this article, the underlying QCD Pomeron-exchange amplitude which produces the DDIS events does not obey the operator product expansion nor satisfy momentum sum rules. Thus we conclude that the quark and gluon distributions measured in DIS experiments will be misidentified, unless the measurements explicitly exclude the DDIS events and that a correct determination of the parton distribution functions (PDFs) derived from the DIS data requires the explicit subtraction of the DDIS contribution from the full DIS cross section.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Deep level defects in low-pressure chemical vapor deposition grown (010) β-Ga 2 O 3

This study provides the full-bandgap evaluation of defect state distributions in beta phase gallium oxide (β-Ga 2 O 3 ) grown by low-pressure chemical vapor deposition (LPCVD) on (010) β-Ga 2 O 3 substrates at high growth of up to 20 µm/h. Deep-level optical spectroscopy and deep-level transient spectroscopy measurements applied to Ni/β-Ga 2 O 3 Schottky diodes revealed the presence of a previously unreported defect state at E C -3.6 eV, which dominated the overall trap distribution in LPCVD grown material. However, states at E C -0.8 eV, E C -2.0 eV, and E C -4.4. eV were also detected, similar to prior studies on β-Ga 2 O 3 grown by other methods, with similar or lower concentrations for the LPCVD samples. The E C -0.8 eV and E C -2.0 eV states were previously connected to residual Fe impurities and gallium vacancies, respectively. The total concentration of traps in the LPCVD material was on par with or lower than the state-of-the-art metal–organic chemical vapor deposition-grown materials despite the much higher growth rate, and the distribution of states showed negligible dependence on SiCl 4 flow rate and doping concentration. These results demonstrate that the high growth rate of LPCVD-grown β-Ga 2 O 3 is very promising for achieving thick, low defect density, and high-quality layers needed for multi-kV device applications.

36 MATERIALS SCIENCE↗