Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Online Learning”

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 505 records · Page 28

Wikipedia for SmallSats: The SSRI Knowledge Base

The Small Satellite Reliability Initiative (SSRI), in conjunction with NASA’s Small Spacecraft Systems Virtual Institute (S3VI), has developed the SSRI Knowledge Base: a comprehensive, publicly available online tool that aims to improve mission confidence for future small spacecraft. This tool provides vetted, high-quality sources of information on key elements of a successful mission. This content includes best practices and lessons learned from previous missions and Knowledge Base “resources”. These resources include publicly available SSRI-generated content in addition to existing guides, publications, standards, software tools, websites, and books. Acknowledging the constant increase and varied sources of small spacecraft knowledge, along with the value and depth of experience that exists beyond any single development team, the SSRI Knowledge Base is designed to be an open, collaborative platform. User participation is essential to the refinement and growth of this tool. Therefore, users are encouraged to rate resources and to contribute content. The SSRI Knowledge Base is hosted at https://www.nasa.gov/smallsat-institute/ssri-knowledge-base. This webinar will discuss the motivation for and development of the SSRI Knowledge Base, demonstrate the existing tool, and outline plans for further development.

SmallSats↗

Data-Driven Robust M-LS-SVR-Based NARX Modeling for Estimation and Control of Molten Iron Quality Indices in Blast Furnace Ironmaking

Optimal operation of an industrial blast furnace (BF) ironmaking process largely depends on a reliable measurement of molten iron quality (MIQ), which is not feasible using the conventional sensors. This paper proposes a novel data-driven robust modeling method for online estimation and control of MIQ indices. First, to completely capture the nonlinear dynamics of the BF process, the nonlinear autoregressive exogenous (NARX) model is constructed for the MIQ indices. Then, considering that the standard least-squares support vector regression (LS-SVR) cannot directly cope with the multi-output problem, the multi- task transfer learning is proposed to design a novel multi-output LS-SVR (M-LS-SVR) for learning of the NARX model. Furthermore, a novel M-estimator is proposed to reduce the interference of outliers and improve the robustness of the M-LS- SVR model. Since the weights of different outlier data are properly given by the weight function, their corresponding contributions on modeling can be properly distinguished, thus a robust modeling result can be achieved. Finally, a novel multi- objective evaluation index on modeling performance is developed by comprehensively considering the root mean square error of modeling and the correlation coefficient on trend fitting, based on which the NSGA-II algorithm is used to globally optimize the model parameters. Both experiments using industrial data and industrial applications illustrate that the proposed method can eliminate the adverse effect caused by the fluctuation of data in BF process efficiently, which indicates its stronger robustness and higher accuracy. Moreover, control testing shows that the developed model can be well applied to realize data-driven control of the BF process.

blast cavities, Control↗

Harnessing large language models’ zero-shot and few-shot learning capabilities for regulatory research

Abstract Large language models (LLMs) are sophisticated AI-driven models trained on vast sources of natural language data. They are adept at generating responses that closely mimic human conversational patterns. One of the most notable examples is OpenAI's ChatGPT, which has been extensively used across diverse sectors. Despite their flexibility, a significant challenge arises as most users must transmit their data to the servers of companies operating these models. Utilizing ChatGPT or similar models online may inadvertently expose sensitive information to the risk of data breaches. Therefore, implementing LLMs that are open source and smaller in scale within a secure local network becomes a crucial step for organizations where ensuring data privacy and protection has the highest priority, such as regulatory agencies. As a feasibility evaluation, we implemented a series of open-source LLMs within a regulatory agency’s local network and assessed their performance on specific tasks involving extracting relevant clinical pharmacology information from regulatory drug labels. Our research shows that some models work well in the context of few- or zero-shot learning, achieving performance comparable, or even better than, neural network models that needed thousands of training samples. One of the models was selected to address a real-world issue of finding intrinsic factors that affect drugs' clinical exposure without any training or fine-tuning. In a dataset of over 700 000 sentences, the model showed a 78.5% accuracy rate. Our work pointed to the possibility of implementing open-source LLMs within a secure local network and using these models to perform various natural language processing tasks when large numbers of training examples are unavailable.

Biochemistry & Molecular Biology↗

Navigating the Noise: Bringing Clarity to ML Parameterization Design With O $\boldsymbol{\mathcal{O}}$(100) Ensembles

Abstract Machine‐learning (ML) parameterizations of subgrid processes (here of turbulence, convection, and radiation) may one day replace conventional parameterizations by emulating high‐resolution physics without the cost of explicit simulation. However, uncertainty about the relationship between offline and online performance (i.e., when integrated with a large‐scale general circulation model) hinders their development. Much of this uncertainty stems from limited sampling of the noisy, emergent effects of upstream ML design decisions on downstream online hybrid simulation. Our work rectifies the sampling issue via the construction of a semi‐automated, end‐to‐end pipeline for size ensembles of hybrid simulations, revealing important nuances in how systematic reductions in offline error manifest in changes to online error and online stability. For example, removing dropout and switching from a Mean Squared Error to a Mean Absolute Error loss both reduce offline error, but they have opposite effects on online error and online stability. Other design decisions, like incorporating memory, converting moisture input from specific humidity to relative humidity, using batch normalization, and training on multiple climates do not come with any such compromises. Finally, we show that ensemble sizes of may be necessary to reliably detect causally relevant differences online. By enabling rapid online experimentation at scale, we can empirically settle debates regarding subgrid ML parameterization design that would have otherwise remained unresolved in the noise.

Lin, Jerry [Department of Earth System Sciences Un↗

PMU-Based Monitoring and Mitigation of Delayed Voltage Recovery Using Admittances

This project analyzes the delayed voltage recovery phenomenon by simplifying the Western Electricity Coordinating Council (WECC) composite load model using the load admittance and the thermal relay dynamics. From this analysis, a closed form expression approximating the recovery time is derived and the key load parameters impacting the behavior of voltage recovery are identified. A monitoring scheme based on the measured load admittance is then proposed to detect the onset of stalling, even in the presence of voltage oscillations, and estimate the duration of the delayed voltage recovery. A mitigation scheme utilizing smart thermostats and offline learning is also derived to ensure that the voltage recovers to the pre-contingency voltage within a specified time. Both the monitoring and mitigation schemes only need local measurements at a substation making them promising for online applications. Results for the monitoring and mitigation schemes are described in detail for the IEEE 162 bus system validating the various assumptions used for the analysis and establishing the connection between the delayed voltage recovery phenomenon and load admittance.

24 POWER TRANSMISSION AND DISTRIBUTION↗

NEML2: A High Performance Library for Constitutive Modeling

NEML2, the New Engineering Material model Library, version 2, is an offshoot of NEML, an earlier material modeling code developed at Argonne National Laboratory. NEML2 extends the key philosophy of its predecessor, i.e., material models are flexible, modular, and can be built from smaller blocks. It also provides modern features that do not exist in the framework of its predecessor such as material model vectorization, automatic differentiation, device-portable just-in-time compilation, operator fusion, lazy tensor evaluation, etc. Moreover, NEML2 can seamlessly integrate with the popular machine learning package PyTorch to take advantage of modern and fast-growing machine learning techniques. In this fiscal year, the development of core library features and capabilities are complete. The purpose of this report is not to serve as a verbatim copy of the software API reference (which is available online at https://reverendbedford.github.io/neml2/). Instead, this report documents the motivation, implementation, design choices, and usage of each core capability as well as their applications in solving practical engineering problems. This report is compiled based on the NEML2 major release 2.0.0.

36 MATERIALS SCIENCE↗

Optimization and stabilization of Fermilab Booster using hybrid Bayesian/RL framework

PIPII project will raise Fermilab Booster intensity and ramp rate. Beam losses will limit average power and are hard to simulate. Presently, Booster uses operator-guided empirical tuning. This task is challenging due to high dimensionality, multiple objectives, critical safety constraints, and drifts. We developed a synergistic suite of Bayesian optimization (BO) and reinforcement learning (RL) tools to optimize and stabilize beam losses. First, active learning was used to build a rough model. Data was collected parasitically using two novel safety constraint types – nonlinear input space restrictions (based on optics model), and uncertainty constraints (to stop bad steps/beam aborts). We then applied online multi-objective BO with scalarized objectives and fitting to improve/rebalance losses, increasing safety margins by 25%. Using BO model as a safety veto, we tried several on/off-policy RL agents for long term stabilization; SAC had best performance. We found that adding contextual (state) information further improved performance, eventually integrating key knobs like linac phase and temperature into the parameter space. Long term testing is ongoing to enable operational use.

Kuklev, Nikita [Fermilab]↗

Tethys Knowledge Management System: Working to Advance the Marine Renewable Energy Industry

Purpose: This paper discusses an ongoing process to assist with the advancement of worldwide understanding of the environmental effects of marine renewable energy (MRE) through a public, online knowledge management system known as Tethys (https://tethys.pnnl.gov). Understanding potential environmental impacts can reduce expensive monitoring requirements and promote growth of this emerging industry. Approach: Tethys acts as the central repository for information in a landscape where literature and knowledge are scattered and difficult to find, providing access for an international audience of stakeholders and the public. An intuitive and flexible design has been developed over 8 years of operation, pairing content curation with synthesis and engagement to enhance the impact of knowledge management. Findings: Tethys has become internationally recognized and trusted as an honest broker of information. A community of participants that use Tethys resources and contributes content has grown to almost 50,000 people around the world, annually. Engineering and siting challenges continue to stall the deployment and establishment of the MRE industry, but progress has been made towards retiring environmental risks and facilitating progression in deploying and operating devices. Practical Implications: The need for knowledge management is particularly important when information is scattered, as is often the case with emerging industries. Tethys serves as an example of intentional design and intuitive organization for knowledge management. Value of Paper: This paper discusses focused knowledge management methods with international cooperation and public engagement to advance an emerging technical industry.

Whiting, Jonathan M.↗

Machine learning in nuclear materials research

Nuclear materials are often demanded to function for extended time in extreme environments, including high radiation fluxes with associated transmutations, high temperature and temperature gradients, mechanical stresses, and corrosive coolants. They also have a wide range of microstructural and chemical makeups, resulting in multifaceted and often out-of-equilibrium interactions. Machine learning (ML) is increasingly being used to tackle these complex time-dependent interactions and aid researchers in developing models and making predictions, sometimes with better accuracy than traditional modeling that focuses on one or two parameters at a time. Conventional practices of acquiring new experimental data in nuclear materials research are often slow and expensive, limiting the opportunity for data-centric ML, but new methods are changing that paradigm. Here we review high-throughput computational and experimental data approaches, especially robotic experimentation and active learning that is based on Gaussian process and Bayesian optimization. We show ML examples in structural materials (e.g., reactor pressure vessel (RPV) alloys and radiation detecting scintillating materials) and highlight new techniques of high-throughput sample preparation and characterizations, and automated radiation/environmental exposures and real-time online diagnostics. Herein, this review suggests that ML models of material constitutive relations in plasticity, damage, and even electronic and optical responses to radiation are likely to become powerful tools as they develop. Finally, we speculate on how the recent trends of using natural language processing (NLP) to aid the collection and analysis of literature data, interpretable artificial intelligence (AI), and the use of streamlined scripting, database, workflow management, and cloud computing platforms that will soon make the utilization of ML techniques as commonplace as the spreadsheet curve-fitting practices of today.

36 MATERIALS SCIENCE↗

Accurate prediction of mega-electron-volt electron beam properties from UED using machine learning

To harness the full potential of the ultrafast electron diffraction (UED) and microscopy (UEM), we must know accurately the electron beam properties, such as emittance, energy spread, spatial-pointing jitter, and shot-to-shot energy fluctuation. Owing to the inherent fluctuations in UED/UEM instruments, obtaining such detailed knowledge requires real-time characterization of the beam properties for each electron bunch. While diagnostics of these properties exist, they are often invasive, and many of them cannot operate at a high repetition rate. Here, we present a technique to overcome such limitations. Employing a machine learning (ML) strategy, we can accurately predict electron beam properties for every shot using only parameters that are easily recorded at high repetition rate by the detector while the experiments are ongoing, by training a model on a small set of fully diagnosed bunches. Applying ML as real-time noninvasive diagnostics could enable some new capabilities, e.g., online optimization of the long-term stability and fine single-shot quality of the electron beam, filtering the events and making online corrections of the data for time-resolved UED, otherwise impossible. This opens the possibility of fully realizing the potential of high repetition rate UED and UEM for life science and condensed matter physics applications.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

TRMM Precipitation Application Examples Using Data Services at NASA GES DISC

Data services to support precipitation applications are important for maximizing the NASA TRMM (Tropical Rainfall Measuring Mission) and the future GPM (Global Precipitation Mission) mission's societal benefits. TRMM Application examples using data services at the NASA GES DISC, including samples from users around the world will be presented in this poster. Precipitation applications often require near-real-time support. The GES DISC provides such support through: 1) Providing near-real-time precipitation products through TOVAS; 2) Maps of current conditions for monitoring precipitation and its anomaly around the world; 3) A user friendly tool (TOVAS) to analyze and visualize near-real-time and historical precipitation products; and 4) The GES DISC Hurricane Portal that provides near-real-time monitoring services for the Atlantic basin. Since the launch of TRMM, the GES DISC has developed data services to support precipitation applications around the world. In addition to the near-real-time services, other services include: 1) User friendly TRMM Online Visualization and Analysis System (TOVAS; URL: http://disc2.nascom.nasa.gov/Giovanni/tovas/); 2) Mirador (http://mirador.gsfc.nasa.gov/), a simplified interface for searching, browsing, and ordering Earth science data at GES DISC. Mirador is designed to be fast and easy to learn; 3) Data via OPeNDAP (http://disc.sci.gsfc.nasa.gov/services/opendap/). The OPeNDAP provides remote access to individual variables within datasets in a form usable by many tools, such as IDV, McIDAS-V, Panoply, Ferret and GrADS; and 4) The Open Geospatial Consortium (OGC) Web Map Service (WMS) (http://disc.sci.gsfc.nasa.gov/services/wxs_ogc.shtml). The WMS is an interface that allows the use of data and enables clients to build customized maps with data coming from a different network.

Liu, Zhong↗

The Structure and Tra c Flow Anatomy of the Planet-Scale Urban Vehicular Mobility

A data-driven realistic design and evalua- tion of vehicular mobility has been particularly chal- lenging due to a lack of large-scale real-world mea- surements in the research community. Current research methodologies rely on articial scenarios, random con- nectivity, and use small and biased samples. In this pa- per, we perform a combined study to learn the struc- ture and connectivity of urban streets and modeling and characterization of vehicular trac densities on them. Our dataset is a collection of more than 222 thousand routes and 25 million vehicular mobility images from 1091 online web cameras located in six dierent re- gions of the world. Our results centered around four major observations: i. study shows that driving routes and visiting locations of regions demonstrate power-law distribution, indicating a planned or recently designed road infrastructure; ii. we represent regions by network graphs in which nodes are camera locations and edges are urban streets that connect the nodes. Such represen- tation exhibits small world properties with short path lengths and large clustering coecient; iii. trac densi- ties show 80% temporal correlation during several hours of a day; iv. modeling trac densities against known theoretical distributions show less than 5% deviation for heavy-trailed models such as log-logistic and log- gamma distributions. We believe this work will provide a much-needed contribution to the research community for design and evaluation of future vehicular networks and smart cities.

Thakur, Gautam↗

Image-based novel fault detection with deep learning classifiers using hierarchical labels

One important characteristic of modern fault classification systems is the ability to flag the system when faced with previously unseen fault types. This work considers the unknown fault detection capabilities of deep neural network-based fault classifiers. Specifically, we propose a methodology on how, when available, labels regarding the fault taxonomy can be used to increase unknown fault detection performance without sacrificing model performance. To achieve this, we propose to utilize soft label techniques to improve the state-of-the-art deep novel fault detection techniques during the training process and novel hierarchically consistent detection statistics for online novel fault detection. Lastly, we demonstrated increased detection performance on novel fault detection in inspection images from the hot steel rolling process, with results well replicated across multiple scenarios and baseline detection methods.

42 ENGINEERING↗

Crowdsourcing quality control for Dark Energy Survey images

We have developed a crowdsourcing web application for image quality control employed by the Dark Energy Survey. Dubbed the "DES exposure checker", it renders science-grade images directly to a web browser and allows users to mark problematic features from a set of predefined classes. Users can also generate custom labels and thus help identify previously unknown problem classes. User reports are fed back to hardware and software experts to help mitigate and eliminate recognized issues. We report on the implementation of the application and our experience with its over 100 users, the majority of which are professional or prospective astronomers but not data management experts. We discuss aspects of user training and engagement, and demonstrate how problem reports have been pivotal to rapidly correct artifacts which would likely have been too subtle or infrequent to be recognized otherwise. We conclude with a number of important lessons learned, suggest possible improvements, and recommend this collective exploratory approach for future astronomical surveys or other extensive data sets with a sufficiently large user base. We also release open-source code of the web application and host an online demo versionat http://des-exp-checker.pmelchior.net

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Regional surrogates for predictive control of digital twins

Digital twins of complex systems must involve a model that is fast, generalizable, and usable for real-time control. For example, high-fidelity nonlinear multiphysics simulations can capture laser-material interactions, but are too slow for optimization or model predictive control (MPC). Reduced-order models, used to accelerate such computation, frequently fail to generalize to unseen inputs or control states. We show theoretically that this failure is intrinsic, i.e., that a learned model is non-unique outside the sampled subspace when its low-rank structure arises from limited excitation and clustered eigenvalues, rather than from a user-imposed truncation alone. Motivated by this result, we propose a control-ready regional surrogate-construction framework for both autonomous and nonautonomous dynamics; it employs Koopman lifting to represent nonlinearities, while preserving spatial locality. We illustrate our approach by constructing a control-ready surrogate for the digital twin of a thermal component of additive-manufacturing process. Our surrogate, localized in space through a von Neumann stencil, is learned from noisy high-fidelity simulations that emulate thermal-camera images collected during the manufacturing. It is linear in thermo-physically augmented states so that MPC reduces to a convex quadratic program. The surrogate requires no online correction, generalizes to unseen scan paths and power profiles of the laser, and is more than three orders of magnitude faster than a finite-difference solver. Furthermore, when the MPC sequence computed on the digital twin is applied to this solver, closed-loop temperature regulation is recovered, showing that the surrogate preserves control-relevant input-output behavior.

Data-driven model↗

RADAI: A Large-Scale Realistic Dataset for Radiation Detection Algorithm Development

Open, realistic datasets are essential for developing and benchmarking radiation detection algorithms, yet they remain scarce. The Radiological Anomaly Detection and Identification (RADAI) project was develop to create datasets that meet the training and testing needs for sophisticated radiation detection algorithms. The RADAI dataset is a large-scale synthetic resource that integrates high-fidelity Monte Carlo simulations with realistic urban scenarios to capture both background variability and source signatures. RADAI models construction-material NORM, people and vehicles, urban clutter, and dynamic environmental effects such as cosmic-ray and rain-induced transients, and they provide list-mode detector data with motion and response modeling suitable for algorithm training and evaluation. The RADAI project resulted in three publicly-released complementary datasets together with an online scoring portal for standardized performance assessment and an open software toolkit that supports data access, augmentation, model development, and evaluation. These resources enable reproducible comparisons across methods and promote rigorous studies at the scale required by contemporary machine learning. By grounding algorithm development in realistic, well-documented conditions, RADAI supports progress toward more robust detection, identification, and localization in complex urban environments.

Ghawaly, James M. [Division of Computer Science an↗

Cooperative Research and Development Agreement between University of Genoa and National Energy Technology Laboratory [Abstract]

The production, transportation, storage, and consumption of oil and gas resources rely on a global network of infrastructure, which is maintained and regulated by numerous commercial, non-profit, academia, and local to national government entities. Thus, information about this infrastructure is often highly variable in the quality, quantity and accessibility of the data. However, integrating and standardizing these disparate data sets into a global, open-access assessment and inventory of oil and gas infrastructure can provide a framework to support advanced analyses, including predictions and assessments of global methane infrastructure and emissions risk, identify information gaps, evaluate economic costs, and support a range of decision making needs. At present, there is no single point of access to authoritative, publicly available oil and gas infrastructure data. In addition, conventional online search approaches are increasingly inefficient as the volume and variety of open-data online continues to evolve and increase. Consequently, there exists a great deal of uncertainty surrounding the extent, gaps, and quality of infrastructure data across the world. This six-month effort used a combination of expert driven and machine learning techniques to build a virtual taxonomy of global oil and gas infrastructure to provide an open data foundation for addressing a range of stakeholder questions and needs.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

Operations automation

This is truly the era of 'faster-better-cheaper' at the National Aeronautics and Space Administration/Jet Propulsion Laboratory (NASA/JPL). To continue JPL's primary mission of building and operating interplanetary spacecraft, all possible avenues are being explored in the search for better value for each dollar spent. A significant cost factor in any mission is the amount of manpower required to receive, decode, decommutate, and distribute spacecraft engineering and experiment data. The replacement of the many mission-unique data systems with the single Advanced Multimission Operations System (AMMOS) has already allowed for some manpower reduction. Now, we find that further economies are made possible by drastically reducing the number of human interventions required to perform the setup, data saving, station handover, processed data loading, and tear down activities that are associated with each spacecraft tracking pass. We have recently adapted three public domain tools to the AMMOS system which allow common elements to be scheduled and initialized without the normal human intervention. This is accomplished with a stored weekly event schedule. The manual entries and specialized scripts which had to be provided just prior to and during a pass are now triggered by the schedule to perform the functions unique to the upcoming pass. This combination of public domain software and the AMMOS system has been run in parallel with the flight operation in an online testing phase for six months. With this methodology, a savings of 11 man-years per year is projected with no increase in data loss or project risk. There are even greater savings to be gained as we learn other uses for this configuration.

Boreham, Charles Thomas↗