SEARCH · Engineering Papers

Results for “learning dynamics”

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 1,783 records · Page 99

All sources NASA NTRS DOE OSTI

DOE OSTI2022

Accurate and Timely Forecasts of Geologic Carbon Storage using Machine Learning Methods

Carbon capture and storage is one strategy to reduce greenhouse gas emissions. One approach to storing the captured CO2 is to inject it into deep saline aquifers. However, dynamics of the injected CO2 plume is uncertain and the potential for leakage back to the atmosphere must be assessed. Thus, accurate and timely forecasts of CO2 storage via real-time measurements integration becomes very crucial. This study proposes a learning-based, inverse-free prediction method that can accurately and rapidly forecast CO2 movement and distribution with uncertainty quantification based on limited simulation and observation data. The machine learning techniques include dimension reduction, multivariate data analysis, and Bayesian learning. The outcome is expected to provide CO2 storage site operators with an effective tool for real-time decision making.

Lu, Dan↗

DOE OSTI2023

Configuration of an optical switch fabric using machine learning

An optical switch fabric comprises two or more optical switch elements. The optical switch elements are configured in a topology. A switch control has a plurality of bias control signals. The switch control can address one or more of the optical switch elements and can apply one of the bias control signals to bias of the addressed optical switch element to establish a switch setting. The topology and switch settings determine how each of one of the inputs is connected to each of one of the outputs of the optical switch fabric. The switch settings are determined by a machine learning process which includes a model creation. The model can be made to adapt dynamically during optical switch fabric operation.

Dupuis, Nicolas↗

DOE OSTI2024

Digitalization mapping and assessment process supporting ION strategic transformation activities

The existing fleet of commercial nuclear power plants (NPPs) are an important asset in the nation’s portfolio of electrical generating resources. Their continued safe and reliable operation are critical to providing a large source of carbon-free electricity to power the nation’s economy. The United States Department of Energy’s (DOE) Light Water Reactor Sustainability (LWRS) Program develops the scientific bases, methods, and tools, for the continued safe and economical operation of the nation's commercial NPPs. The Plant Modernization Pathway within LWRS Program focuses on providing guidance to industry on the full-scale implementation of modernization solutions for NPPs that significantly reduce the technical and financial risks associated with modernization. This research is focused on helping the nuclear industry understand how to digitize and digitalize their NPPs so that they can design their modernization solutions to be scalable, sustainable, and integrated both laterally and horizontally within their organization. That is, this research creates a digital transformation in NPPs by reshaping cultural mindsets and by identifying business efficiencies. In partnership with industry, and using four previously established guiding principles for digitalization, this research supported NPP modernization through assessing readiness for digitalization as a means to achieve integrated operations for nuclear. Specifically, this research created an assessment to review an entire organization’s work processes to gather information about the digitalization health of the plant. The assessment tools were administered to plant employees, and the results were used to develop a digitalization plan. The survey assessment identified the optimal candidate processes that would most benefit from a digitalization initiative which were revealed through analytical frameworks. One analysis calculated mean digitalization health indicator scores for all endorsed activities which allowed the researchers to rank and color code the results for easy identification. Individual health indicator scores are also provided, should our industry partner wish to understand these findings according to their own organizational priorities, business considerations and desired end-state. The results were also analyzed from the perspective that organizations are comprised of different types of innovators (e.g., generators, optimizers, conceptualizers, and implementers), which differentially affects the organization’s ability to comprehend and adapt to change (i.e., opportunities to innovate). Understanding the relative composition of innovator types at an NPP allows them to gather insights into the strengths and weaknesses they have in innovating how work is performed. For the utility that partnered with this research team, the results showed that implementers make up the largest portion of respondents and conceptualizers the smallest portion. Knowing the proportion of innovator types gave this organization insights on how they can effectively implement their innovation solutions. Additionally, the results were analyzed from a technical, economic, and risk perspective to identify and quantify work reduction opportunities (WROs). Recognizing that not all cost-saving opportunities are the same, a Technical, Economic and Risk Assessment (TERA) was performed to evaluate WROs to identify areas of greatest potential and lowest risk. The key results from TERA included a digitalization opportunity score for each activity, and a calculation of potential cost savings. These two outputs formed the bases for calculating a priority index and rank for the activities/processes assessed. From the prioritization calculations, TERA can then help the utility 1) decide what digitalization priorities to invest money in implementing and then 2) calculates how much should be invested in the digitalization initiatives selected to achieve cost savings and/or an acceptable return on investment. Last, onsite interviews revealed several inefficiencies in the standard work processes that occur cross-departmentally that are due to the absence of digitized and digitalized processes. Examples of these include time spent scanning paper documents and then uploading the documents electronically, obtaining signatures, and searching for desired information. This represents a digital but not digitalized process. Over 15 opportunities to improve work processes were identified through this multi-method digitalization assessment. The various analytical assessments used (e.g., TERA, digitalization health indicator scores), as well as discussions with the utility partner, corroborated that all the opportunities identified had a strong potential to make work processes more efficient and to improve overall performance of the NPP.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

DOE OSTI2025

AI Benchmark Democratization and Carpentry

Benchmarks are a cornerstone of modern machine learning, enabling reproducibility, comparison, and scientific progress. However, AI benchmarks are increasingly complex, requiring dynamic, AI-focused workflows. Rapid evolution in model architectures, scale, datasets, and deployment contexts makes evaluation a moving target. Large language models often memorize static benchmarks, causing a gap between benchmark results and real-world performance. Beyond traditional static benchmarks, continuous adaptive benchmarking frameworks are needed to align scientific assessment with deployment risks. This calls for skills and education in AI Benchmark Carpentry. From our experience with MLCommons, educational initiatives, and programs like the DOE's Trillion Parameter Consortium, key barriers include high resource demands, limited access to specialized hardware, lack of benchmark design expertise, and uncertainty in relating results to application domains. Current benchmarks often emphasize peak performance on top-tier hardware, offering limited guidance for diverse, real-world scenarios. Benchmarking must become dynamic, incorporating evolving models, updated data, and heterogeneous platforms while maintaining transparency, reproducibility, and interpretability. Democratization requires both technical innovation and systematic education across levels, building sustained expertise in benchmark design and use. Benchmarks should support application-relevant comparisons, enabling informed, context-sensitive decisions. Dynamic, inclusive benchmarking will ensure evaluation keeps pace with AI evolution and supports responsible, reproducible, and accessible AI deployment. Community efforts can provide a foundation for AI Benchmark Carpentry.

von Laszewski, Gregor [Virginia U.]↗

DOE OSTI2012

AVESTAR Center for Operational Excellence of Electricity Generation Plants

To address industry challenges in attaining operational excellence for electricity generation plants, the U.S. Department of Energy’s (DOE) National Energy Technology Laboratory (NETL) has launched a world-class facility for Advanced Virtual Energy Simulation Training and Research (AVESTARTM). This presentation will highlight the AVESTARTM Center simulators, facilities, and comprehensive training, education, and research programs focused on the operation and control of high-efficiency, near-zero-emission electricity generation plants. The AVESTAR Center brings together state-of-the-art, real-time, high-fidelity dynamic simulators with full-scope operator training systems (OTSs) and 3D virtual immersive training systems (ITSs) into an integrated energy plant and control room environment. AVESTAR’s initial offering combines--for the first time--a “gasification with CO2 capture” process simulator with a “combined-cycle” power simulator together in a single OTS/ITS solution for an integrated gasification combined cycle (IGCC) power plant with carbon dioxide (CO2) capture. IGCC systems are an attractive technology option for power generation, especially when capturing and storing CO2 is necessary to satisfy emission targets. The AVESTAR training program offers a variety of courses that merge classroom learning, simulator-based OTS learning in a control-room operations environment, and immersive learning in the interactive 3D virtual plant environment or ITS. All of the courses introduce trainees to base-load plant operation, control, startups, and shutdowns. Advanced courses require participants to become familiar with coordinated control, fuel switching, power-demand load shedding, and load following, as well as to problem solve equipment and process malfunctions. Designed to ensure work force development, training is offered for control room and plant field operators, as well as engineers and managers. Such comprehensive simulator-based instruction allows for realistic training without compromising worker, equipment, and environmental safety. It also better prepares operators and engineers to manage the plant closer to economic constraints while minimizing or avoiding the impact of any potentially harmful, wasteful, or inefficient events. The AVESTAR Center is also used to augment graduate and undergraduate engineering education in the areas of process simulation, dynamics, control, and safety. Students and researchers gain hands-on simulator-based training experience and learn how the commercial-scale power plants respond dynamically to changes in manipulated inputs, such as coal feed flow rate and power demand. Students also analyze how the regulatory control system impacts power plant performance and stability. In addition, students practice start-up, shutdown, and malfunction scenarios. The 3D virtual ITSs are used for plant familiarization, walk-through, equipment animations, and safety scenarios. To further leverage the AVESTAR facilities and simulators, NETL and its university partners are pursuing an innovative and collaborative R&D program. In the area of process control, AVESTAR researchers are developing enhanced strategies for regulatory control and coordinated plant-wide control, including gasifier and gas turbine lead, as well as advanced process control using model predictive control (MPC) techniques. Other AVESTAR R&D focus areas include high-fidelity equipment modeling using partial differential equations, dynamic reduced order modeling, optimal sensor placement, 3D virtual plant simulation, and modern grid. NETL and its partners plan to continue building the AVESTAR portfolio of dynamic simulators, immersive training systems, and advanced research capabilities to satisfy industry’s growing need for training and experience with the operation and control of clean energy plants. Future dynamic simulators under development include natural gas combined cycle (NGCC) and supercritical pulverized coal (SCPC) plants with post-combustion CO2 capture. These dynamic simulators are targeted for use in establishing a Virtual Carbon Capture Center (VCCC), similar in concept to the DOE’s National Carbon Capture Center for slipstream testing. The VCCC will enable developers of CO2 capture technologies to integrate, test, and optimize the operation of their dynamic capture models within the context of baseline power plant dynamic models. The objective is to provide hands-on, simulator-based “learn-by-operating” test platforms to accelerate the scale-up and deployment of CO2 capture technologies. Future AVESTAR plans also include pursuing R&D on the dynamics, operation, and control of integrated electricity generation and storage systems for the modern grid era. Special emphasis will be given to combining load-following energy plants with renewable and distributed generating supplies and fast-ramping energy storage systems to provide near constant baseload power.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

Viral Dynamic Models During COVID‐19: Are We Ready for the Next Pandemic?

Mathematical models have been used for about 30 years to improve our understanding of virus-host interaction, in particular during chronic infections. During the COVID-19 pandemic, these models have been used to provide insights into the natural history of acute SARS-CoV-2 infection, optimize antiviral treatment strategies, understand factors associated with transmission, and optimize surveillance systems. The impact of modeling has been accelerated by the availability of unprecedented multidimensional immune data from animal and human systems, which enhanced partnerships between experimentalists and theorists and led to exciting new modeling and statistical developments. In this mini review, we examine the lessons learned from the COVID-19 pandemic and discuss the main insights provided by mathematical models of viral dynamics at the different stages of the outbreak. Although we focus on respiratory infection, we also consider the new areas for development in anticipation of future acute infections from new or reemerging pathogens.

59 BASIC BIOLOGICAL SCIENCES↗

NASA NTRSDate not supplied

Reinforcement Learning for Spacecraft Navigation & Environment Characterization in the Planar-Restricted Two-Body Problem

As science, exploration, and commercial space missions become increasingly complex, so does the need for efficient, autonomous, and integrated spacecraft navigation and operations techniques. Key operational functions, including data collection and transmission, environment characterization, systems constraints, human factors, and navigation, often are intertwined and conflicted. Deep Reinforcement Learning (DRL) offers a framework for addressing integrated spacecraft navigation and planning in an uncertain dynamical environment. The goal of this study is to evaluate the utility of DRL for integrated spacecraft navigation and planning. This is achieved by developing a simple environmental characterization training environment in the Planar-Restricted 2-Body Problem (PR2BP), establishing benchmarks and heuristic baselines, and designing a previously unstudied Markov Decision Process (MDP) formulation. This MDP formulation enables the spacecraft DRL agents to appropriately balance navigation and actuation capabilities. The resulting DRL-derived policy exceeds a random or untrained policy and meets or exceeds the level of performance of a heuristic without actuation. In the process, valuable intuition is gained about the problem with insight into how DRL methods could scale to increasingly more realistic scenarios, including net-work design and training architectures, efficient state space representations, and methods for encouraging exploration in a parametric action space, among others.

navigation↗

NASA NTRS2008

Orbit Determination During Spacecraft Emergencies with Sparse Tracking Data - THEMIS and TDRS-3 Lessons Learned

This paper provides an overview of the lessons learned from the National Aeronautics and Space Administration (NASA) Goddard Space Flight Center s (GSFC) Flight Dynamics Facility s (FDF) support of the Time History of Events and Macroscale Interactions during Substorms (THEMIS) spacecraft emergency in February 2007, and the Tracking and Data Relay Satellite-3 (TDRS-3) spacecraft emergency in March 2006. A successful and timely recovery from both of these spacecraft emergencies depended on accurate knowledge of the orbit. Unfortunately, the combination of each spacecraft emergency with very little tracking data contributed to difficulties in estimating and predicting the orbit and delayed recovery efforts in both cases. In both the THEMIS and TDRS-3 spacecraft emergencies, numerous factors contributed to problems with obtaining nominal tracking data measurements. This paper details the various causative factors and challenges. This paper further enumerates lessons learned from FDF s recovery efforts involving the THEMIS and TDRS-3 spacecraft emergencies and scant tracking data, as well as recommendations for improvements and corrective actions. In addition, this paper describes the broad range of resources and complex navigation methods employed within the FDF for supporting critical navigation activities during all mission phases, including launch, early orbit, and on-orbit operations.

Morinelli, Patrick J.↗

DOE OSTI2021

BETO 2021 Peer Review - WBS 4.1.2.32: Bioeconomy Scenario Analysis

The Bioeconomy Scenario Analysis project uses systems thinking and analysis to assess current and/or prospective techno-economics, research and development, deployment strategies, policy, and market conditions and their impact on the potential development trajectories of the bioenergy industry over time. Results from this project include identification of opportunities and constraints to industrial development, quantification of multiple metrics (energy, economic, environmental) and informing researchers, decision makers, and industry of the steps needed for a sustainable, nationwide biofuels industry. Analyses from this project enable the creation of a bioenergy industry by (1) inciting policy-makers to explore scenarios for nationwide biofuels production, identifying policy actions consistent with pathways for growth; (2) improving industry’s understanding of industry growth potential under different technology and investment conditions, better targeting their development efforts; and (3) providing universities and other interested stakeholders with tools and analyses that can be adapted to meet research and teaching objectives, connecting students with careers that build the industry. One of the many modeling tools used in this project, the Biomass Scenario Model (BSM) is a publicly-available, unique, validated, state-of-the-art, award-winning, fourth-generation model of the domestic biofuels supply chain which explicitly focuses on how and under what conditions biofuel technologies might be deployed to contribute to the U.S. transportation energy sector. We use models like the BSM to examine the implications of policies and incentives as well as their potential side-effects. The BSM uses a system-dynamics simulation to model dynamic interactions and transitions across the supply chain; it tracks the deployment of biofuels given industrial learning and the reaction of the investment community in the context of land availability, projected oil markets, consumer demand for biofuels, and government policies over time. Under expected market conditions, analyses using the BSM suggest that the biofuels industry may require significant external actions in the early years to thrive. Interventions that accelerate the industrial learning process (e.g. operation of pre-commercial and commercial facilities) have been identified as having strong influence in starting the growth of a commercial biofuel industry. Policies which are coordinated across the whole supply chain in BSM foster the growth of the biofuels industry and production of tens of billions of gallons of biofuels may occur under sufficiently favorable conditions.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

DOE OSTI2023

Reinforcement Learning of Structured Stabilizing Control for Linear Systems With Unknown State Matrix

This paper delves into designing feedback control gains for a continuous-time linear quadratic regulator (LQR) problem that is constrained to certain predefined structure with unknown state matrix. We bring forth the ideas from reinforcement learning (RL) in conjunction with sufficient stability and performance guarantees in order to design these structured gains using the trajectory measurements of states and controls. Here we first formulate a model-based framework using dynamic programming (DP) to embed the structural constraint to the LQR gain computation in the continuous-time setting, and then subsequently, formulate a policy iteration RL algorithm that can alleviate the requirement of known state matrix in conjunction with maintaining the feedback gain structure. The design enables a distributed learning control design which is necessary for many large-scale cyber-physical systems. Theoretical guarantees are provided for stability and convergence of the structured reinforcement learning (SRL) algorithm. We validate our theoretical results with numerical simulations on a multi-agent networked linear time-invariant (LTI) dynamic system.

42 ENGINEERING↗

DOE OSTI2023

Modeling the 4D discharge of lithium-ion batteries with a multiscale time-dependent deep learning framework

The lithium-ion battery (LIB) field is moving towards the direction of investigating spatially resolved physical phenomena in the 3D porous microstructure of electrodes. These pore-scale simulations give new insights into the local dynamics of lithiation/de-lithiation and charge transport, Nevertheless, the computational time of these simulations limits the integration of these models in optimization workflows of cycling conditions or electrode manufacturing processes. Machine learning models present a way of assessing in real-time the performance of materials. While several successful techniques for replicating simulations with machine learning have been proposed, this case study presents a more demanding problem, due to the necessity of understanding the behavior of heterogeneous 3D local data, as it evolves in time: this poses both a scientific and a technical challenge. To this end, we propose an autoregressive multiscale convolutional neural network model to predict relevant quantities at the pore-scale in the solid phase: the lithium concentration (in the active material) and potential (in the active material and carbon binder). Here, these are ultimately used to reconstruct the battery discharge curve. 3D images of the electrode microstructures are the input to the network, trained with a dataset of finite element method simulations to predict the discharge behavior of the cathode side in lithium ion batteries. We propose this machine learning model as a proof-of-concept of the applicability of multiscale networks for time-dependent physics problems. The trained model exhibits very high accuracy (with errors lower than 2 %) in forecasting the discharge behavior of new unseen cathodes.

25 ENERGY STORAGE↗

NASA NTRSDate not supplied

Deep Reinforcement Learning based Routing in an Air-to-Air Ad-hoc Network

This paper studies the Multiple Sources and Multiple Destinations (MSMD) routing problem in a dynamic Air-to-Air Ad-hoc Network (AAAN). We consider a spectrum limited scenario where multiple links have to share the same frequency channel so that co-channel interference becomes inevitable. As a result, routing decisions and spectrum access are coupled and must be jointly considered. This paper proposes a deep Q-learning based algorithm to find an optimal routing and channel selection strategy that minimizes the end-to-end communication delay. Specifically, under the assumption that only local information is available to every node, the Deep Q-Network (DQN) is trained offline to learn the optimal routing and channel selection strategy. After the trained DQN is implemented in every node, multiple relay nodes can simultaneously determine their next-hop relay and channel selections in real-time. Simulation results demonstrate the efficacy of our proposed algorithm.

AAAN↗

NASA NTRSDate not supplied

Deep Reinforcement Learning based Routing in an Air-to-Air Ad-hoc Network

AAAN↗

DOE OSTI2024

Experimental Setup and Learning-Based AI Model for Developing Accurate PV Inverter Models

The integration of power electronics-based interfaces presents challenges due to the absence of detailed models and the high computational complexity. Generic models used in system studies lack accuracy in capturing converter dynamics. This paper proposes a data-driven approach developed from experimental setup data. This approach enhances accuracy in photovoltaic inverter modeling. We used two types of PV inverters in the experiment. The recorded experimental data undergo processing through a machine learning model. Results from the model trained through machine learning is also presented.

artificial intelligence↗

DOE OSTI2022

Online Learning and Pricing for Network Revenue Management with Reusable Resources

We consider a price-based network revenue management problem with multiple products and multiple reusable resources. Each randomly arriving customer requests a product (service) that needs to occupy a sequence of reusable resources (servers). We adopt an incomplete information setting where the firm does not know the price-demand function for each product and the goal is to dynamically set prices of all products to maximize the total expected revenue of serving customers. We propose novel batched bandit learning algorithms for finding near-optimal pricing policies, and show that they admit a near-optimal cumulative regret bound of $O(J\sqrt{XT})$, where J, X, and T, are the numbers of products, candidate prices, and service periods, respectively. As part of our regret analysis, we develop the first finite-time mixing time analysis of an open network queueing system (i.e., the celebrated Jackson Network), which could be of independent interest. Our numerical studies show that the proposed approaches perform consistently well.

42 ENGINEERING↗

DOE OSTI2021

Parsimonious Potential Energy Surface Expansions Using Dictionary Learning with Multipass Greedy Selection

Potential energy surfaces fit with basis set expansions have been shown to provide accurate representations of electronic energies and have enabled a variety of high-accuracy dynamics, kinetics, and spectroscopy applications. The number of terms in these expansions scales poorly with system size, a drawback that challenges their use for systems with more than similar to 10 atoms. A solution is presented here using dictionary learning. Subsets of the full set of conventional basis functions are optimized using a newly developed multipass greedy regression method inspired by forward and backward selection methods from the statistics, signal processing, and machine learning literatures. Here, the optimized representations have accuracies comparable to the full set but are 1 or more orders of magnitude smaller, and notably, the number of terms in the optimized multipass greedy expansions scales approximately linearly with the number of atoms.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2022

Learning Global Proliferation Expertise Evolution Using AI-Driven Analytics and Public Information

Detecting and anticipating global proliferation expertise and capability evolution from unstructured, noisy, and incomplete public data streams is a highly desired, but extremely challenging task. Here, in this article, we present our pioneering data-driven approach to support the non-proliferation mission to detect and explain the evolution of proliferation expertise and capability development globally from terabytes of publicly available information (PAI), focusing on our knowledge extraction pipeline and descriptive analytics. We first discuss how we fuse nine open-source data streams, including multilingual data, to convert 4 TB of unstructured data to structured knowledge and encode dynamically evolving proliferation expertise representations—content and context graphs. For this, we rely on natural language processing (NLP) and deep learning (DL) models to perform information extraction, topic modeling, and distributed text representation (aka embedding) learning. We then present interactive, usable, and explainable descriptive analytics to refine domain knowledge and present it in a human-understandable form. Finally, we introduce future work avenues that will leverage our dynamic knowledge representations and descriptive analytics to enable predictive and prescriptive inferences to achieve real-time domain understanding and contextual reasoning about global proliferation expertise and capability evolution.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗

DOE OSTI2018

Identifying structural changes with unsupervised machine learning methods

In this work, unsupervised machine learning methods are used to identify structural changes using the melting point transition in classical molecular dynamics simulations as an example application of the approach. Dimensionality reduction and clustering methods are applied to instantaneous radial distributions of atomic configurations from classical molecular dynamics simulations of metallic systems over a large temperature range. Principal component analysis is used to dramatically reduce the dimensionality of the feature space across the samples using an orthogonal linear transformation that preserves the statistical variance of the data under the condition that the new feature space is linearly independent. From there, k-means clustering is used to partition the samples into solid and liquid phases through a criterion motivated by the geometry of the reduced feature space of the samples, allowing for an estimation of the melting point transition. This pattern criterion is conceptually similar to how humans interpret the data but with far greater throughput, as the shapes of the radial distributions are different for each phase and easily distinguishable by humans. The transition temperature estimates derived from this machine learning approach produce comparable results to other methods on similarly small system sizes. These results show that machine learning approaches can be applied to structural changes in physical systems.

42 ENGINEERING↗