Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Operator 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 1,531 records · Page 85
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Adaptive Load Shedding for Grid Emergency Control via Deep Reinforcement Learning

Emergency control, typically such as under-voltage load shedding (UVLS), is broadly used to grapple with low voltage and voltage instability issues in real-world power systems under contingencies. However, existing emergency control schemes are rule-based and cannot be adaptively applied to uncertain and floating operating conditions. Here, we propose an adaptive UVLS algorithm for emergency control via deep reinforcement learning (DRL) and expert systems. We first construct dynamic components for picturing the power system operation as the environment. The transient voltage recovery criteria, which poses time-varying requirements to UVLS, is integrated into the states and reward function to advise the learning of deep neural networks. The proposed method has no tuning issue of coefficients in reward functions, and this issue was regarded as a deficiency in the existing DRL-based algorithms. Case studies illustrate that the proposed method outperforms the traditional UVLS relay in both the timeliness and efficacy for emergency control.

24 POWER TRANSMISSION AND DISTRIBUTION↗
NASA NTRS1989

SSME fault monitoring and diagnosis expert system

An expert system, called LEADER, has been designed and implemented for automatic learning, detection, identification, verification, and correction of anomalous propulsion system operations in real time. LEADER employs a set of sensors to monitor engine component performance and to detect, identify, and validate abnormalities with respect to varying engine dynamics and behavior. Two diagnostic approaches are adopted in the architecture of LEADER. In the first approach fault diagnosis is performed through learning and identifying engine behavior patterns. LEADER, utilizing this approach, generates few hypotheses about the possible abnormalities. These hypotheses are then validated based on the SSME design and functional knowledge. The second approach directs the processing of engine sensory data and performs reasoning based on the SSME design, functional knowledge, and the deep-level knowledge, i.e., the first principles (physics and mechanics) of SSME subsystems and components. This paper describes LEADER's architecture which integrates a design based reasoning approach with neural network-based fault pattern matching techniques. The fault diagnosis results obtained through the analyses of SSME ground test data are presented and discussed.

Ali, Moonis↗
NASA NTRS2026

Deep RL for Fast Long-Horizon Operations Scheduling on NASA's Carruthers Geocorona Observatory Mission

Spacecraft operations scheduling is a highly constrained, long-horizon combinatorial optimization problem that traditionally relies on heuristics, constraint programming, or manual planning. We present a scalable deep reinforcement learning framework developed and deployed for NASA’s Carruthers Geocorona Observatory mission. Our framework introduces a macro-action abstraction known as activity blocks coupled with dynamic action-masking to navigate the intractably large search space and strictly enforce complex power, thermal, and instrument constraints. The resulting architecture generates globally feasible schedules with overwhelming probability, establishes operational trust, and executes a full training cycle in under six hours, circumventing the need for policy robustness by enabling rapid, on-demand retraining. Further, resulting schedules outperform baseline heuristics in scheduled science quality. The deep reinforcement learning framework was deployed as the default operational scheduler for the Carruthers Geocorona Observatory mission from the outset of the mission, demonstrating that deep reinforcement learning can be trusted for real spacecraft operations under complex, evolving constraints.

Geocorona↗
DOE OSTI2019

Ubiquitous Traffic Volume Estimation through Machine Learning Procedure

High-quality traffic volume data is critical for transportation planning, operations, and travel-energy calculations. However, vehicle count data from roadside sensors is very sparse owing to high capital cost of installing, and on-going maintenance of electronic sensor equipment. This research effort aims to bring to market a first of its kind traffic volume data product that provides accurate estimates of traffic volumes across the entire road network for all times (24x7) and all locations (100% coverage). This product greatly improves the coverage and quality of traffic volume information by combining commercial probe traffic data with traditional traffic measurement using state-of-the-art machine learning. Commercial probe traffic data are derived from vehicles that self-report their position and speed as well as crowd sourced data from smartphone applications.

30 DIRECT ENERGY CONVERSION↗
NASA NTRS2010

Space Operations for a New Era of Exploration Launch Vehicles

NASA's Constellation Program is depending on the Ares Projects to deliver the crew and cargo launch capabilities needed to send human explorers to the Moon and beyond. Ares I and V will provide the core space launch capabilities needed to continue providing crew and cargo access to the International Space Station (ISS), and to build upon the U.S. history of human space exploration. Since 2005, Ares has made substantial progress on designing, developing, and testing the Ares I crew launch vehicle and has continued its in-depth studies of the Ares V cargo launch vehicles. The combined Ares I/Ares V architecture has been designed to reduce the complexity and labor intensity of ground operations for America's next journeys beyond low-Earth orbit (LEO). A deliberate effort is being made to ensure a high level of system operability to significantly increase safety and system availability as well as reduce recurring costs for this new launch vehicle. The Ares Projects goal is to instill operability as part of the vehicles requirements development, design, and operations. This simplicity will come from using simpler, proven engine designs, as in the case of the J-2X upper stage engine and RS-68 engine; improving existing hardware, as in the case of the Shuttle-heritage 5-segment solid rocket motor; and using common propulsion and instrument unit elements between Ares I and Ares V. Furthermore, lessons learned while developing Ares I will be applied directly to Ares V operations. In 2009, the Ares Projects plan to conduct the first flight test of Ares I, designated Ares I-X. Ares I-X preparations have already prompted changes to the vehicle stacking and launch infrastructure at Kennedy Space Center (KSC), including removing Shuttle-specific fixtures from the Vehicle Assembly Building (VAB) to accommodate Ares I-style stacking operations, new firing room computers and infrastructure in the VAB Launch Control Center, and new lightning protection system towers at Launch Complex 39B to accommodate the greater height of Ares I-X. In addition to lessons learned from the stacking of Ares I-X, the flight test itself promises to yield important data and operations lessons for assembling, launching, and flying Ares I.

Cook, Stephen A.↗
DOE OSTI2020

Panel Session 117: Panel: Operational Excellence Through a Strong Safety Culture (R9.1)

This panel focused on the what the Department of Energy (DOE) and its contractors are doing to promote safety and excellence. The Integrated Safety Management System (ISMS) is crucial, but additional support provided by the Energy Facility Contractors Group (EFCOG), DOE, and environmental, safety, and health (ES and H) professionals in the field are equally imperative. The panelists discussed the importance of sharing lessons learned as well as are developing performance metrics to track operational and safety successes. This panel focused on current trends in building a strong operational safety culture and will include results from recent national workshops. Safety culture is the foundation of strong operational performance. Panel discussion included best practices from industry, DOE, and US NRC and also review tools being used to measure and evaluate safety culture programs. Panelists with presentations: Strong Safety Culture Promotes Operational Excellence (Kliss McNeel, Patricia Hughes); Safety Culture (Patricia Allen); DOE Safety Culture Improvement (Julie Goeckner); Supporting Safety Culture on the Front Line (Jan Preston); Interdependence of Safety Culture and Organizational Performance (Rizwan Shah)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
NASA NTRS2015

Commercial Orbital Transportation Services (COTS) Program Lessons Learned

This report has been developed by the National Aeronautics and Space Administration (NASA) Human Exploration and Operations Mission Directorate (HEOMD) Risk Management team in close coordination with the COTS Program. This document provides a point-in-time, cumulative, summary of actionable key lessons learned derived from the design project. Lessons learned invariably address challenges and risks and the way in which these areas have been addressed. Accordingly the risk management thread is woven throughout the document.

COTS↗
DOE OSTI2025

VoroClust

SAND2025-11465O VoroClust, also known as Voronoi Clustering, is a fast, density-based unsupervised clustering algorithm applicable to high-resolution and high-dimensional data. It operates as quickly as distance-based clustering methods while effectively capturing complex regional geometries, matching the performance of current density-based methods. VoroClust employs a data-centered sphere cover to reduce computational demands while preserving data topology. It propagates clusters outward from local density peaks. Although supervised machine learning is powerful for applications like image classification and segmentation, it requires comprehensive, consistent datasets, which many applications lack. Unsupervised clustering algorithms analyze the structure of each dataset rather than relying on similarities with other examples, making them well-suited for practical applications with insufficient or inappropriate data for supervised learning. Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.

Ebeida, Mohamed [Sandia National Lab. (SNL-CA), Li↗
NASA NTRS1991

Lessons learned in the development of the STOL intelligent tutoring system

Lessons learned during the development of the NASA Systems Test and Operations Language (STOL) Intelligent Tutoring System (ITS), being developed at NASA Goddard Space Flight Center are presented. The purpose of the intelligent tutor is to train STOL users by adapting tutoring based on inferred student strengths and weaknesses. This system has been under development for over one year and numerous lessons learned have emerged. These observations are presented in three sections, as follows. The first section addresses the methodology employed in the development of the STOL ITS and briefly presents the ITS architecture. The second presents lessons learned, in the areas of: intelligent tutor development; documentation and reporting; cost and schedule control; and tools and shells effectiveness. The third section presents recommendations which may be considered by other ITS developers, addressing: access, use and selection of subject matter experts; steps involved in ITS development; use of ITS interface design prototypes as part of knowledge engineering; and tools and shells effectiveness.

Seamster, Thomas↗
DOE OSTI2021

Anatomy of Continuous Mars SEIS and Pressure Data from Unsupervised Learning

The seismic noise recorded by the Interior Exploration using Seismic Investigations, Geodesy, and Heat Transport (InSight) seismometer (Seismic Experiment for Interior Structure [SEIS]) has a strong daily quasi-periodicity and numerous transient microevents, associated mostly with an active Martian environment with wind bursts, pressure drops, in addition to thermally induced lander and instrument cracks. That noise is far from the Earth’s microseismic noise. Quantifying the importance of nonstochasticity and identifying these microevents is mandatory for improving continuous data quality and noise analysis techniques, including autocorrelation. Cataloging these events has so far been made with specific algorithms and operator’s visual inspection. We investigate here the continuous data with an unsupervised deep-learning approach built on a deep scattering network. This leads to the successful detection and clustering of these microevents as well as better determination of daily cycles associated with changes in the intensity and color of the background noise. We first provide a description of our approach, and then present the learned clusters followed by a study of their origin and associated physical phenomena. We show that the clustering is robust over several Martian days, showing distinct types of glitches that repeat at a rate of several tens per sol with stable time differences. We show that the clustering and detection efficiency for pressure drops and glitches is comparable to or better than manual or targeted detection techniques proposed to date, noticeably with an unsupervised approach. Finally, here we discuss the origin of other clusters found, especially glitch sequences with stable time offsets that might generate artifacts in autocorrelation analyses. We conclude with presenting the potential of unsupervised learning for long-term space mission operations, in particular, for geophysical and environmental observatories.

58 GEOSCIENCES↗
DOE OSTI2022

Challenges and Lessons Learned from an Analysis of Three Zero Energy Buildings

Zero energy buildings, or zero energy ready buildings, which are designed and operated by public or private commercial property owners can play an important role in reducing carbon emissions. This paper discusses lessons learned and key takeaways from an in-depth analysis of three zero energy buildings that took an integrated design and construction approach to significantly reduce energy use. Two of these projects are new construction and one is a retrofit to zero energy. Findings are based on project literature review, data analysis, and in-depth interviews with the building design teams and staff who were involved during the design, construction, and operation of the buildings. The paper addresses value proposition and cost data in such a way that other building owners can replicate the strategies and technology solutions in response to regulatory mandates or organizational goals. It also provides details on operational improvements taken at each zero energy building to enhance energy performance and increase the potential for load flexibility, and discusses challenges and lessons gleaned from design teams and building staff. The findings serve as a reference for building owners, designers, engineers, contractors, or others interested in, or involved with, the design, construction, or use of new or existing buildings. The paper also includes recommended pathways for widespread adoption of zero energy strategies that can be applied in various locations.

ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATION↗
DOE OSTI2020

Predicting Solar Plant Generation with Machine Learning Techniques

Accurately predicting power generation for PV sites is critical for prioritizing relevant operations & maintenance activities, thereby extending the lifetime of a system and improving profit margins. A number of factors influence power generation at PV sites, including local weather, shading and soiling losses, design of modules, DC mismatches, and degradation over time. Other external factors such as curtailment and grid outages can also have a notable impact on power generation. Machine learning techniques can be used to provide more accurate predictions of PV power production by accounting for important weather and climate information neglected by current industry methods. This article will cover the deficiencies of those methods and will show how machine learning can dramatically improve power generation predictions.

14 SOLAR ENERGY↗
DOE OSTI2017

Scaling Up Data-Parallel Analytics Platforms: Linear Algebraic Operation Cases

Linear algebraic operations such as matrix manipulations form the kernel of many machine learning and other crucial algorithms. Scaling up as well as scaling out such algorithms are key to supporting large scale data analysis that require efficient processing over millions of data samples. To this end, we present, ARION, a hardware acceleration based approach for scaling-up individual tasks of Spark, a popular data-parallel analytics platform. We support both linear algebraic operations of between two dense matrices, and between sparse and dense matrices in distributed environments. ARION provides a flexible control of acceleration according to matrix density, along with efficient scheduling based on runtime resource utilization. We demonstrate the benefit of our approach for general matrix multiplication operations over large matrices with up to four billion elements by using Gramian matrix computation that is commonly used in machine learning. Experiments show that our approach achieves more than 2× and 1.5× end-to-end performance speedups for dense and sparse matrices, respectively, and up to 57.04× faster computation compared to MLlib, a state of the art Spark-based implementation. This work is sponsored in part by the NSF under the grants: CNS-1565314, CNS-1405697, and CNS-1615411. The manuscript has been authored by UT-Battelle, LLC under Contract No. DE-AC05-00OR22725 with the U.S. Department of Energy. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a non-exclusive, paid-up, irrevocable, world-wide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doe-public-access-plan). This research used resources of the Oak Ridge Leadership Computing Facility at the Oak Ridge National Laboratory, which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC05-00OR22725.

Xu, Luna↗
DOE OSTI2026

Post-Event Fault Identification with Machine Learning for Protection System Validation

Power system protection devices have transitioned over the past few decades from mechanical to analog devices, then to solid state and finally digital. Relays and their associated critical network of equipment have significantly increased in complexity. Even internally, relays have gained significant intricacy, with relatively simple overcurrent or differential functions now being assisted by a myriad of other functions. This is necessary as the grid becomes more complex, but it brings increased difficulty in monitoring and upkeep. Misoperation caused by accidental improper relay settings or deliberate malicious actions is a constant challenge faced by all utilities. These improper settings can be difficult to identify and may require exhaustive post-mortem analysis, typically after a major outage event has already occurred. A mechanism is needed for monitoring the behavior of protection systems to validate that their performance falls within expectations. Relays that fail to isolate a fault or trip when there is no system disturbance can be flagged for settings review in situations where this behavior may not have been noticed due to manual restoration or backup protection operations. This work presents a concept for a machine learning (ML) system capable of validating the performance of protection systems by identifying fault events and characterizing protection system responses based solely on available current and voltage measurements. As a first step in its development, an experimental dataset is generated, and a random forest model is implemented with high accuracy in distinguishing four power system scenarios.

24 - POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2025

An analysis of physics limited dispatch of nuclear renewable integrated energy systems using deep reinforcement learning and dynamic modeling

Previous approaches to dispatching nuclear integrated energy systems (NIES) have focused on the profitability and flexibility of these systems to operate on energy grids with highly variable pricing. However, due to the complexity involved in modeling and designing these systems, there has been less emphasis on ensuring that these dispatch strategies are physically achievable. It is imperative to develop methods that allow the system to remain within the desired NIES operating conditions and perform this based on realistic limited forecasted information. This research employs next generation artificial intelligence, namely deep reinforcement learning (DRL), and a dynamic system model written in Modelica to find a safe and profitable dispatch strategy for a solar nuclear hybrid design. The DRL agent is shown to find a novel dispatch strategy that manages both power ramping and power levels while respecting operational limits. This DRL-based dispatch is compared to other dispatching strategies including an optimal design solution from mixed integer linear programming (MILP). It is found that incorporating the physics of such a tightly coupled NIES limits the profitability of the MILP-based dispatch strategy. As a result, the MILP solution overestimates the design’s generated revenue. In contrast, DRL significantly reduces the number of breaches of safe operational conditions during energy arbitrage while maintaining profitability. Furthermore, this work paves the way for a more detailed assessment of NIES profitability and could be used to aid operator decisions on future NIES projects.

14 - SOLAR ENERGY↗
DOE OSTI2017

Foresee: A user-centric home energy management system for energy efficiency and demand response

This paper presents foresee, a user-centric home energy management system that can help optimize how a home operates to concurrently meet users' needs, achieve energy efficiency and commensurate utility cost savings, and reliably deliver grid services based on utility signals. Foresee is built on a multiobjective model predictive control framework, wherein the objectives consist of energy cost, thermal comfort, user convenience, and carbon emission. Foresee learns user preferences on different objectives and acts on their behalf to operate building equipment, such as home appliances, photovoltaic systems, and battery storage. In this work, machine-learning algorithms were used to derive data-driven appliance models and usage patterns to predict the home's future energy consumption. This approach enables highly accurate predictions of comfort needs, energy costs, environmental impacts, and grid service availability. Simulation studies were performed on field data from a residential building stock data set collected in the Pacific Northwest. Results indicated that foresee generated up to 7.6% whole-home energy savings without requiring substantial behavioral changes. When responding to demand response events, foresee was able to provide load forecasts upon receipt of event notifications and delivered the committed demand response services with 10% or fewer errors. Foresee fully utilized the potential of the battery storage and controllable building loads and delivered up to 7.0-kW load reduction and 13.5-kW load increase. As a result, these benefits are provided while maintaining the occupants' thermal comfort or convenience in using their appliances.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗