SEARCH · Search NASA
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.
AI is increasingly used in critical infrastructure, but it's often a “black box” that is difficult to understand. AIBOMs, or AI Bill of Materials, are one way to increase AI transparency and accountability. They can also be used in automated cyber defense. AIBOMs are a new concept with challenges to widespread adoption, but their potential in cybersecurity makes them a worthwhile investment.
The study examines the impact of variable air volume (VAV) damper stuck faults on the system operation, building indoor conditions, and reheating energy consumption. This study includes both experimental and simulation studies for five test scenarios, including a fault-free scenario. We implemented VAV damper stuck fault through the building automation system (BAS). Results show that a damper stuck in a high opening position (60% damper opening) results in supplying an excessive amount of cold air from the rooftop unit (RTU) to the conditioned zone, increasing reheating energy consumption. The results of this research can serve as a foundational resource for developing fault detection algorithms.
Automated text summarization (ATS) is crucial for collecting specialized, domain-specific information. Zero-shot learning (ZSL) allows large language models (LLMs) to respond to prompts on information not included in their training, playing a vital role in this process. This study evaluates LLMs' effectiveness in generating accurate summaries under ZSL conditions and explores using retrieval augmented generation (RAG) and prompt engineering to enhance factual accuracy and understanding. We combined LLMs with summarization modeling, prompt engineering, and RAG, evaluating the summaries using the METEOR metric and keyword frequencies through word clouds. Results indicate that LLMs are generally well-suited for ATS tasks, demonstrating an ability to handle specialized information under ZSL conditions with RAG. However, web scraping limitations hinder a single generalized retrieval mechanism. While LLMs show promise for ATS under ZSL conditions with RAG, challenges like goal misgeneralization and web scraping limitations need addressing. Future research should focus on solutions to these issues.
The development of novel instrumentation requires an iterative cycle with three stages: design, prototyping, and testing. Recent advancements in simulation and nanofabrication techniques have significantly accelerated the design and prototyping phases. Nonetheless, detector characterization continues to be a major bottleneck in device development. During the testing phase, a significant time investment is required to characterize the device in different operating conditions and find optimal operating parameters. The total effort spent on characterization and parameter optimization can occupy a year or more of an expert's time. In this work, we present a novel technique for automated sensor calibration that aims to accelerate the testing stage of the development cycle. This technique leverages closed-loop Bayesian optimization (BO), using real-time measurements to guide parameter selection and identify optimal operating states. We demonstrate the method with a novel low-noise CCD, showing that the machine learning-driven tool can efficiently characterize and optimize operation of the sensor in a couple of days without supervision of a device expert.
In this study, YOLO (You Only Look Once), a well-known object detection and image segmentation model, is used to detect qubits on a quantum chip. The model is trained by exposing it to various images of objects of interest and by tweaking various training parameters to allow the model to learn from a limited and highly specialized dataset of SEM imagery. The model was provided with numerous images of qubits and various prominent features of note located on a qubit. These detections were then integrated into a simple GUI, allowing for real-time feedback alongside SEM usage. This will allow us to create various functions enabling automated functions like chip-wide scanning, auto-focusing, and image correction.
The Deep Underground Neutrino Experiment (DUNE) is a flagship international collaboration designed to study neutrinos—tiny, nearly massless particles that may hold answers to fundamental questions about the Universe. Fermilab’s Robotic Test Stand (RTS) plays a critical role in ensuring the quality of approximately 50,000 Application-Specific Integrated Circuit (ASIC) chips that will be used in DUNE’s massive liquid argon detectors. These electronics will be inside the cryostat; therefore, they will need to have a high yield of working chips and low noise. To improve the automation and reliability of the RTS, this project focused on designing and implementing a Python-based finite state machine (FSM) to manage chip handling workflows. The FSM was developed as a modular software framework to coordinate robotic arm movements, manage chip tray positions, and monitor system states during testing. Key features include robust error handling routines, a pause/resume system for safe mid-cycle interruptions, and a simulation mode for iterative testing without hardware dependencies. The system was designed to prepare for seamless integration with RTS hardware components such as the robotic arm and vision system. This integration will streamline collaboration and enable efficient deployment of updates across the six institutions performing testing. The outcomes of this internship contribute to Fermilab’s mission to advance high-energy physics and support the DOE’s national goals by directly improving the testing of equipment to be used in DUNE. The project also provided valuable experience in software design and contributing to the success of DUNE.
This paper summarizes an automated framework for generating high fidelity fluid data and provides example cases of four generalized and simplified (cosine-based) roughness geometries.
In the advanced reactor domain, a flexible and scalable software/hardware infrastructure is crucial for integrating and validating various control technologies. This study used the Microreactor Automated Control System (MACS) hardware platform as a testbed. MACS was originally designed to mirror Idaho National Laboratory (INL)'s Microreactor Applications Research Validation and Evaluation (MARVEL), a 85-kW thermal fission microreactor. It features control drums for simulated reactivity control; lights that function as a surrogate reactor core, with the brightness being proportional to the reactor power; and light sensors that emulate neutron detectors. To transform MACS into a physical twin of MARVEL for evaluating control methods, the Control and Optimization Modular Modeling Application for Nuclear Deployment (COMMAND) software was employed. This software integrated the hardware with two models of the MARVEL core, based on Reactor Excursion and Leak Analysis Program (RELAP5-3D) and Monte Carlo N-Particle (MCNP) models. The study aimed to demonstrate the gap between control theory and actual practice—a gap that often necessitates empirical adjustments such as control gain retuning, filters, time discretization, and integrator anti-windup measures. Controllers were developed based on increasingly complex simulations without hardware, starting from the base MARVEL model and then introducing actuator saturation constraints and sensor noise. The final control strategy was then tested using MACS, and a comparative performance analysis was conducted.
Reliable digital instrumentation and control systems (DI&C) are integral for sustaining the continued operation of nuclear power plants. These systems ensure that nuclear reactors operate safely, efficiently, and within regulatory requirements. Yet, the cost of designing and licensing new nuclear DI&C can be prohibitively expensive. Under the U.S. Department of Energy Light Water Reactor Sustainability Program, Idaho National Laboratory has developed a framework for supporting the risk-informed design of DI&C systems by offering methods to support the identification, quantification, and evaluation of risks for various DI&C design architectures. The framework indicates potential software failure modes and provides pathways for quantifying the potential for these software failures, including common cause failures. Using the framework’s systematic approach, challenges for assessing risks within new and existing nuclear DI&C systems can be reduced. Nevertheless, the current framework can be further improved using the convenience of automation. This paper introduces the development of Software for the Hazard Identification and Evaluation of Digital Systems (SHIELDS). SHIELDS is an engineering software package that enables the identification, elimination, and mitigation of potential risks and reduces the burden of deploying reliable DI&C systems. This work introduces plans and techniques to digitize and improve the manual risk assessment modules of the framework. These improvements will save time and increase the repeatability and usability of the framework, making it more accessible to a wider range of users. Ultimately, this introduces SHIELDS and how its modules support efficient development of safe and reliable DI&C systems.
Operational safety in mission-critical environments requires AI systems that are accurate, interpretable, and resistant to hallucination. We present an agentic Retrieval-Augmented Generation (RAG) framework, REFSafe, for grounded hazard analysis and automated safety report generation. The system integrates Large Language Models (LLMs) with structured operational data, historical incident repositories, policy documents, and external authoritative sources. Through iterative agentic reasoning, the framework retrieves, verifies, and synthesizes evidence prior to generation, enforcing citation-backed outputs with explicit source attribution (documents, links, and prior events) to ensure traceability and trust. To mitigate hallucinations and unsupported claims, all risk assessments and forecasts are constrained to retrieved evidence, with confidence signals derived from retrieval relevance and source consistency. A transparent pipeline enables subject matter experts (SMEs) to validate predictions, and provide structured feedback, forming a continuous performance calibration loop. Preliminary deployment demonstrates improved reliability in hazard detection and safety/vulnerability report generation. This work advances trustworthy, evidence-grounded AI for predictive safety intelligence in mission-critical operations.
Orthogonal defect classification (ODC) is a method used to categorize software defects, providing valuable insights into the development process. This study focuses on automating the classification of software bug reports into different ODC defect types using few shot learning, a machine learning approach that requires minimal labeled data. Previous research has manually classified bug reports or used traditional machine learning algorithms like linear support vector machine, achieving limited success. Our approach uses few shot learning to improve classification accuracy and efficiency. The results show a harmonic mean of recall and precision (i.e., the F1 score) of around 0.6 which is a performance improvement over previous methods. The results highlight the potential benefit of few shot learning techniques and their application in enhancing the safety and reliability of nuclear digital instrumentation and control (DI&C) systems. Future work will explore incorporating advanced techniques to supplement the model's training data and achieve better results.
Membership inference attacks (MIAs), which enable adversaries to determine whether specific data points were part of a model's training dataset, have emerged as an important framework to understand, assess, and quantify the potential information leakage associated with machine learning systems. Designing effective MIAs is a challenging task that usually requires extensive manual exploration of model behaviors to identify potential vulnerabilities. In this paper, we introduce AutoMIA -- a novel framework that leverages large language model (LLM) agents to automate the design and implementation of new MIA signal computations. By utilizing LLM agents, we can systematically explore a vast space of potential attack strategies, enabling the discovery of novel strategies. Our experiments demonstrate AutoMIA can successfully discover new MIAs that are specifically tailored to user-configured target model and dataset, resulting in improvements of up to 0.18 in absolute AUC over existing MIAs. This work provides the first demonstration that LLM agents can serve as an effective and scalable paradigm for designing and implementing MIAs with SOTA performance, opening up new avenues for future exploration.
With the advent of large language models (LLMs), in both the open source and proprietary domains, attention is turning to how to exploit such artificial intelligence (AI) systems in assisting complex scientific tasks, such as material synthesis, characterization, analysis and discovery. Here, we explore the utility of LLMs, particularly ChatGPT4, in combination with application program interfaces (APIs) in tasks of experimental design, programming workflows, and data analysis in scanning probe microscopy, using both in-house developed APIs and APIs given by a commercial vendor for instrument control. We find that the LLM can be especially useful in converting ideations of experimental workflows to executable code on microscope APIs. Beyond code generation, we find that the GPT4 is capable of analyzing microscopy images in a generic sense. At the same time, we find that GPT4 suffers from an inability to extend beyond basic analyses for more in-depth technical experimental design. We argue that an LLM specifically fine-tuned for individual scientific domains can potentially be a better language interface for converting scientific ideations from human experts to executable workflows. Such a synergy between human expertise and LLM efficiency in experimentation can open new doors for accelerating scientific research, enabling effective experimental protocols sharing in the scientific community.
The CIRCLES project aims to reduce instabilities in traffic flow, which are naturally occurring phenomena due to human driving behavior. Also called “phantom jams” or “stop-and-go waves,” these instabilities are a significant source of wasted energy. Toward this goal, the CIRCLES project designed a control system, referred to as the MegaController by the CIRCLES team, that could be deployed in real traffic. Our field experiment, the MegaVanderTest (MVT), leveraged a heterogeneous fleet of 100 longitudinally controlled vehicles as Lagrangian traffic actuators, each of which ran a controller with the architecture described in this article. The MegaController is a hierarchical control architecture that consists of two main layers. The upper layer is called the Speed Planner and is a centralized optimal control algorithm. It assigns speed targets to the vehicles, conveyed through the LTE cellular network. The lower layer is a control layer, running on each vehicle. It performs local actuation by overriding the stock adaptive cruise controller, using the stock onboard sensors. The Speed Planner ingests live data feeds provided by third parties as well as data from our own control vehicles and uses both to perform the speed assignment. The architecture of the Speed Planner allows for the modular use of standard control techniques, such as optimal control, model predictive control (MPC), kernel methods, and others. The architecture of the local controller allows for the flexible implementation of local controllers. Corresponding techniques include deep reinforcement learning (RL), MPC, and explicit controllers. Depending on the vehicle architecture, all onboard sensing data can be accessed by the local controllers or only some. Likewise, control inputs vary across different automakers, with inputs ranging from torque or acceleration requests for some cars to electronic selection of adaptive cruise control (ACC) setpoints in others. The proposed architecture technically allows for the combination of all possible settings proposed previously, that is {Speed Planner algorithms} × {local Vehicle Controller algorithms} × {full or partial sensing} × {torque or speed control}. As a result, most configurations were tested throughout the ramp up to the MegaVandertest (MVT).
We report on our experiences replicating 7.3 petabytes (PB) of Earth System Grid Federation (ESGF) computational simulation data from Lawrence Livermore National Laboratory (LLNL) in California to Argonne National Laboratory (ANL) in Illinois and Oak Ridge National Laboratory (ORNL) in Tennessee—a task motivated by a need for increased reliability, capacity, and performance. This task presented significant challenges: the need to move 29 million files twice under time pressure from aging storage hardware; a source file system bottleneck limiting throughput to 1.5 GB/s; frequent site maintenance windows; and the need for complete reliability at scale. We addressed these challenges using a simple replication tool that invoked Globus to transfer large bundles of files while tracking progress in a database, dynamically rerouting transfers to work around maintenance periods and file system limitations. Under the covers, Globus organized transfers to make efficient use of the high-speed Energy Sciences network (ESnet) and the data transfer nodes deployed at participating sites, and also addressed security, integrity checking, and recovery from a variety of transient failures. This success demonstrates the considerable benefits that can accrue from the adoption of performant data replication infrastructure. The replication tool is available at https://github.com/esgf2-us/data-replication-tools.
Accurate vehicle movement classification through signalized intersections is of paramount importance to the analysis of intersection performance and the optimization of traffic control strategies. Conventional techniques for tracking vehicle turning movements depend on infrastructure-based strategies like human counts, loop detectors, and video analytics, all of which are costly, prone to errors, and spatially constrained. High-frequency trajectory data can be utilized to determine vehicle movement patterns in a scalable and infrastructure-independent method due to the adoption of connected vehicles (CVs). In recent years, several studies have utilized connected vehicle data to generate performance measures. Most of the trajectory-based performance measures approaches, however, require map matching-i.e., extracting geospatial references from maps to identify the movements that individual vehicles make at a signalized intersection. These approaches are often time-consuming and hinder scalability since geographic features need to be provided for an analysis to be conducted. Map matching methods are prone to errors as different map versions change these geographic features. This research presents a novel automatic classification pipeline that uses CV trajectory data to classify vehicle movements at signalized crossings, specifically pass-through left-turn and right-turn maneuvers. The process starts by filtering trips that cross a spatial bounding box that has been defined at the target intersection. Approach and departure headings for each trajectory crossing the boundary are computed and are clustered together to identify dominant movements. The proposed algorithm is used to classify the movement of vehicles at 10 intersections in the state of California, and the results indicate that the algorithm can classify movements at these intersections with varying traffic volumes and road network configurations, all in a map-less framework with no need for conflation of vehicle trajectories to a digital base map.
Current approaches for risk analysis of software vulnerabilities using manual assessment and numeric scoring do not complete fast enough to keep pace with the maintenance work rate to patch and mitigate the vulnerabilities. This paper proposes a new approach to modeling software vulnerability risk in the context of the network environment and firewall configuration. In the approach, vulnerability features are automatically matched up with networking, target asset, and adversary features to determine whether adversaries can exploit a vulnerability. The ability of adversaries to reach a vulnerability is modeled by automatically identifying the network services associated with vulnerabilities through a pipeline of machine learning and natural language processing and automatically analyzing network reachability. Our results show that the pipeline can identify network services accurately. We also find that only a small number of vulnerabilities pose real risks to a system. However, if left unmitigated, adversarial reach to vulnerabilities may extend to nullify the effect of firewall countermeasures.
The need to optimize the processing conditions of additively manufactured (AM) metals and alloys has driven advances in throughput capabilities for material property measurements such as tensile strength or hardness. High-throughput (HT) characterization of AM metal microstructure has fallen significantly behind the pace of property measurements due to intrinsic bottlenecks associated with the artisan and labor-intensive preparation methods required to produce highly polished surfaces. This inequality in data throughput has led to a reliance on heuristics to connect process to structure or structure to properties for AM structural materials. In this study, we show a transformative approach to achieve laser powder bed fusion (LPBF) printing, HT preparation using dry electropolishing and HT electron backscatter diffraction (EBSD). This approach was used to construct a library of > 600 experimental EBSD sample sets spanning a diverse range of LPBF process conditions for AM Kovar. This vast library is far more expansive in parameter space than most state-of-the-art studies, yet it required only approximately 10 labor hours to acquire. Build geometries, surface preparation methods, and microscopy details, as well as the entire library of >600 EBSD data sets over the two sample design versions, have been shared with intent for the materials community to leverage the data and further advance the approach. Using this library, we investigated process–structure relationships and uncovered an unexpected, strong dependence of microstructure on location within the build, when varied, using otherwise identical laser parameters.