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.
This is a presentation to the INL AI/ML Activities Workshop in May 2023 discussing the application of AI/ML in nuclear operating experience program and low power shutdown initiating event study.
Recent developments in artificial intelligence (AI) and machine learning (ML), implemented through self-driving laboratories (SDLs), are rapidly creating unprecedented opportunities for the accelerated discovery and optimization of materials. This paper provides a joint analysis of SDLs from both academic and industry perspectives, highlighting the importance of integrating human intelligence in these systems. It discusses the necessity of careful planning in SDL design across physical, data, and workflow dimensions, including instrumental setup, experimental workflow, data management, and human–SDL interaction. The significance of integrating human input within SDLs, especially as the focus shifts from individual tools and tasks to the creation and management of complex workflows, is emphasized. The paper stresses on the crucial role of reward function design in developing forward-looking workflows and examines the interplay between hardware evolution, ML application across chemical processes, and the influence of reward systems in research. Ultimately, the article advocates for a future where SDLs blend human intuition in hypothesis formulation with AI's precision, speed, and data-handling capabilities.
Big data is a computing term used to refer to large and complex data sets, typically consisting of terabytes or more of diverse data that is produced rapidly. The analysis of such complex data sets requires advanced analysis techniques with the capacity to identify patterns and abstract meanings from the vast data. The field of data science combines computer science with mathematics/statistics and leverages artificial intelligence, in particular machine learning, to analyze big data. This field holds great promise for behavior analysis, where both clinical and research studies produce large volumes of diverse data at a rapid pace (i.e., big data). This article presents basic lessons for the behavior analytic researchers and clinicians regarding integration of data science into the field of behavior analysis. We provide guidance on how to collect, protect, and process the data, while highlighting the importance of collaborating with data scientists to select a proper machine learning model that aligns with the project goals and develop models with input from human experts. Here, we hope this serves as a guide to support the behavior analysts interested in the field of data science to advance their practice or research, and helps them avoid some common pitfalls.
Two improvements to current requirements analysis practices are suggested: domain modeling, and the systematic application of analysis heuristics. Domain modeling is the representation of relevant application knowledge prior to requirements specification. Artificial intelligence techniques may eventually be applicable for domain modeling. In the short term, however, restricted domain modeling techniques, such as that in JSD, will still be of practical benefit. Analysis heuristics are standard patterns of reasoning about the requirements. They usually generate questions of clarification or issues relating to completeness. Analysis heuristics can be represented and therefore systematically applied in an issue-based framework. This is illustrated by an issue-based analysis of JSD's domain modeling and functional specification heuristics. They are discussed in the context of the preliminary design of simple embedded systems.
Power system fundamental frequency-dependent protection decisions and control decisions are increasingly common in the distribution system space. Understanding the algorithms used in frequency measurements is critical to understanding the decisions made by the many intelligent electronic devices in the power system. These decisions are critically important for reliability analysis for interconnected power systems as they bear directly on resource planning to mitigate under-frequency conditions that would result in under-frequency load shedding (UFLS). This work aims to improve the understanding of frequency measurement algorithms, evaluate the intelligent electronic devices that use frequency measurements for decision making, and understand the algorithms' direct impact on frequency related decision making. First, this work presents background information on the use of frequency measurements in protection logic, specifically UFLS. Second, the paper presents a relay hardware evaluation test bed used to detect and protect systems from under-frequency events. Finally, the paper presents the dynamic events used to evaluate commercially available, off-the-shelf relay equipment and the results from the relay evaluation.
Charged track reconstruction is a critical task in nuclear physics experiments, enabling the identification and analysis of particles produced in high-energy collisions. Machine learning (ML) has emerged as a powerful tool for this purpose, addressing the challenges posed by complex detector geometries, high event multiplicities, and noisy data. Traditional methods rely on pattern recognition algorithms like the Kalman filter, but ML techniques, such as neural networks, graph neural networks (GNNs), and recurrent neural networks (RNNs), offer improved accuracy and scalability. By learning from simulated and real detector data, ML models can identify and classify tracks, predict trajectories, and handle ambiguities caused by overlapping or missing hits. Moreover, ML-based approaches can process data in near-real-time, enhancing the efficiency of experiments at large-scale facilities like the Large Hadron Collider (LHC) and Jefferson Lab (JLAB). As detector technologies and computational resources evolve, ML-driven charged track reconstruction continues to push the boundaries of precision and discovery in nuclear physics. In these proceedings, we highlight advancements in charged track identification leveraging Artificial Intelligence within the CLAS12 detector, achieving a notable enhancement in experimental statistics compared to traditional methods. Additionally, we showcase real-time event reconstruction capabilities, including the inference of charged particle properties, such as momentum, direction, and species identification, at speeds matching data acquisition rates. These innovations enable the extraction of physics observables directly from the experiment in real-time.
This work seeks to advance the state of the art in HPC I/O performance analysis and interpretation. In particular, we demonstrate effective techniques to: (1) model output performance in the presence of I/O interference from production loads; (2) build features from write patterns and key parameters of the system architecture and configurations; (3) employ suitable machine learning algorithms to improve model accuracy. We train models with five popular regression algorithms and conduct experiments on two distinct production HPC platforms. We find that the lasso and random forest models predict output performance with high accuracy on both of the target systems. We also explore use of the models to guide adaptation in I/O middleware systems, and show potential for improvements of at least 15% from model-guided adaptation on 70% of samples, and improvements up to 10× on some samples for both of the target systems.
A multiple high-order derivative analysis algorithm has been developed which can automatically extract absorption band positions from low-quality reflectance spectra with little degredation of accuracy. Overlapping bands with comparable widths and intensities can be resolved whose centers are as close as 0.3-0.5 W, with safer resolution limits of 0.6-1.0 W band center separations suggested for overlapping bands that are dissimilar. The segment length for smoothing is continually adjusted to about 0.5 W to minimize signal distortion, and a spectral pattern recognition algorithm predicts the signal spectrum and calculates approximate W across the spectrum using its second derivative. A single-pass cubic spline is applied to the smoothed data, and a sliding segment sixth-order polynomial is fit to the spectrum, with the length of the segment being continuously locally adjusted to 1.0 W across the spectrum. Good reliability and consistency of the algorithm is demonstrated with application to laboratory and earth-based telescope spectra.
Until recently microscope users in space relied on traditional microscopy techniques that required manual operation of the microscope and recording of observations in the form of written notes, drawings, or photographs. This method was time consuming and required the return of film and drawings from space for analysis. No real-time data analysis was possible. Advances in digital and video technologies along with recent developments in article intelligence will allow future space microscopists to have a choice of three additional modes of microscopy: remote coaching, remote control, and automation. Remote coaching requires manual operations of the microscope with instructions given by two-way audio/video transmission during critical phases of the experiment. When using the remote mode of microscopy, the Principal Investigator controls the microscope from the ground. The automated mode employs artificial intelligence to control microscope functions and is the only mode that can be operated in the other three modes as well. The purpose of this presentation is to discuss the advantages and disadvantages of the four modes of of microscopy and how the IMIS, a proposed intelligent microscope imaging system, can be used as a model for developing and testing concepts, operating procedures, and equipment design of specifications required to provide a comprehensive microscopy/imaging capability onboard Space Station Freedom.
The Spacecraft Health Automated Reasoning Prototype (SHARP) is a system designed to demonstrate automated health and status analysis for multi-mission spacecraft and ground data systems operations. Telecommunications link analysis of the Voyager 2 spacecraft is the initial focus for the SHARP system demonstration which will occur during Voyager's encounter with the planet Neptune in August, 1989, in parallel with real time Voyager operations. The SHARP system combines conventional computer science methodologies with artificial intelligence techniques to produce an effective method for detecting and analyzing potential spacecraft and ground systems problems. The system performs real time analysis of spacecraft and other related telemetry, and is also capable of examining data in historical context. A brief introduction is given to the spacecraft and ground systems monitoring process at the Jet Propulsion Laboratory. The current method of operation for monitoring the Voyager Telecommunications subsystem is described, and the difficulties associated with the existing technology are highlighted. The approach taken in the SHARP system to overcome the current limitations is also described, as well as both the conventional and artificial intelligence solutions developed in SHARP.
The complexity of assuring cyber resilience for physical process interactions in connected systems such as energy grids increases dramatically as the coupling between processes becomes more direct and responsive. An example of this growing complexity is provided by Integrated Energy Systems (IES), in which various processes such as nuclear heat generation and commodity production are being directly coupled for increased responsiveness to highly variable signals such as market pricing or electricity demand. As such, the potential attack surface of the coupled processes is larger than the two processes independently. Securing these complex systems requires two-fold monitoring: cybersecure monitoring for potential malicious incursion, and physics monitoring for system tampering. Physics monitoring includes analyzing the behavior of the signals within the system for anomalous behavior. This analysis has been shown to be insufficient if approached by only data-driven machine learning and artificial intelligence (MLAI) techniques or only low-level model comparison. Previous efforts at Purdue University suggested combining high-fidelity models with MLAI algorithms as a basis for a software tool for detecting anomalies in physical processes. This work built on that suggestion, developing an advanced library for signal decomposition and analysis using both MLAI and high-fidelity physics algorithms for greatly improved anomaly detection, especially false data injection. This software can be used as part of a secure imbedded intelligence (SEI) system designed under Consequence-driven Cyber-informed Engineering (CCE) for complex coupled systems. This library established a foundation for online and posteriori analysis of digital signals for the purpose of detecting potential malicious tampering in digital signals representing physical processes. Demonstrations carried out throughout the development highlight the effective use of characterization algorithms to detect signal perturbations, particularly triangle attack-style perturbations, in three wide-ranging applications: seismic monitoring, nuclear thermal hydraulics system simulation, and custom manufacturing.
Introduction: This virtual special issue stems from 36 presentations given at the Fourth Chemical Forensics Symposium held 3 - 4 April 2017 as part of the 253rd American Chemical Society (ACS) National Meeting & Exposition in San Francisco, CA, USA. Since 2011, the ACS Division of Analytical Chemistry has hosted symposia that review developments and topics arising in chemical forensics. Chemical forensics is a developing scientific discipline that aims to attribute a chemical (or mixture) of interest to its source by the analysis of the chemical itself or associated material constituents to address investigative, legal or intelligence questions. “Source” refers to how, where and when a chemical was handled or produced. Given past and recent attacks with chemical warfare agents (CWAs) and other toxic chemicals there is a strong need for research and development in chemical forensics for the purpose of finding those responsible for chemical attacks. Hence, many of the papers in this issue deal with the source attribution of toxic chemicals that can be used as weapons. However, given the needs, principles and methods that are shared across different application areas of material provenance or authentication, papers that deal with the source attribution of other illicitly used chemicals and materials are included.
INTRODUCTION: The development of the Evidence Library for use with the IMPACT (Informing Mission Planning via Analysis of Complex Tradespaces) probability risk assessment tool involved a multilayered, time intensive process of data collection and analysis by subject matter experts from the Exploration Medical Capability (ExMC) Element Clinical and Science Team to produce clinical findings forms (CliFFs) for 120 medical conditions. Artificial Intelligence Large Language Models (LLMs) can be leveraged to facilitate this process, thus reducing labor and time. TOPIC: CliFFs contain information about medical conditions as they pertain to spaceflight. This includes condition definitions, incidence data, crew task impairment estimates caused by conditions, treatment protocols and references to literature used for gathering condition evidence. Guided by the Evidence Library Methods document and the CliFF development instructions, a team has leveraged Microsoft Azure AI services and open-source documentation to construct an AI-assisted automated pipeline for CliFF development. This process is designed to search, retrieve, and evaluate the applicable data, and ultimately generate a completed CliFF. The LLM evaluates the relevance of each of the source materials to spaceflight, either as direct evidence or as an analog. The model extracts keywords and generates brief summaries to enhance search and retrieval in later stages of CliFF development. For instance, it can calculate epidemiological statistical data, such as incidence rates and the likelihood of best or worst-case scenarios. APPLICATION: Large Language Models (LLMs) can efficiently summarize large amounts of text. Leveraging this technology will automate data retrieval and evidence gathering for medical databases, like the IMPACT tool, by aiding in the labor-intensive process of analyzing large bodies of literature and organizing it into a formatted document like a CliFF. This added efficiency will enable expeditious expansion of the Evidence Library with additional medical conditions and update previous CLiFFs as new technology becomes available.
The aging water distribution system in the United States, constructed mainly during the 1970s with some pipes dating back 125 years, is experiencing significant deterioration leading to substantial water losses. Along with the potential for water loss savings, improvements in the distribution system by using leak detection technologies can create net energy and cost savings. In this work, a new framework has been presented to calculate the economic level of leakage within water supply and distribution systems for two primary leak detection technologies (acoustic vs. satellite). In this work, a new framework is presented to calculate the economic level of leakage (ELL) within water supply and distribution systems to support smart infrastructure in smart cities. A case study focused using water audit data from Atlanta, Georgia, compared the costs of two leak mitigation technologies: conventional acoustic leak detection and artificial intelligence–assisted satellite leak detection technology, which employs machine learning algorithms to identify potential leak signatures from satellite imagery. The ELL results revealed that conducting one survey would be optimum for an acoustic survey, whereas the method suggested that it would be expensive to utilize satellite-based leak detection technology. However, results for cumulative financial analysis over a 3-year period for both technologies revealed both to be economically favorable with conventional acoustic leak detection technology generating higher net economic benefits of USD 2.4 million, surpassing satellite detection by 50%. A broader national analysis was conducted to explore the potential benefits of US water infrastructure mirroring the exemplary conditions of Germany and The Netherlands. Achieving similar infrastructure leakage index (ILI) values could result in annual cost savings of $\$4$–$\$4.8$ billion and primary energy savings of 1.6–1.9 TWh. These results demonstrate the value of combining economic modeling with advanced leak detection technologies to support sustainable, cost-efficient water infrastructure strategies in urban environments, contributing to more sustainable smart living outcomes.
The finite element modeling of an airframe structure requires knowledge of general purpose programs such as NASTRAN as well as a detailed understanding of the airframe structure. Due to the sophistication of general purpose programs such as NASTRAN, a substantial investment in time and effort is required to gain expertise in using them effectively. This paper describes the development of an expert system used in the validation of NASTRAN based finite element models. Experts in the NASTRAN based finite element modeling of airframe structures were interviewed to document, understand, and represent their knowledge and reasoning in the expert system. Finite element stress analysis and internal loads reports generated by the experts were reviewed to determine expert resolution of problem areas. As a result, areas requiring expert assistance in the modeling of the airframe structures were identified. The finite element input data is represented as a set of 'facts'. A rule based representation is used to code the expert knowledge. A hierarchical set of rules are applied. The expert system first acts as an intelligent front end to insure that all the prerequisites needed to perform the analysis are present. This includes material properties, boundary conditions, and connectivity information. The next step examines if incompatible sets of elements are connected. The succeeding step examines if the specific airframe component is modeled by the appropriate set of elements. The next and final step examines if the airframe members used are adequate to represent the anticipated state of stress. The expert system is designed to inform the user the severity of the error, the likely consequence and possible remedial action. The shell used in the development of the expert system is CLIPS. CLIPS contains a forward chaining inference engine based on the Rete algorithm. CLIPS may be implemented on most personal computers as well as mini computers and mainframes. The approach taken is general and can be implemented using commercially available expert system shells.
Uncompetitive antagonists of the N-methyl D-aspartate receptor (NMDAR) have demonstrated therapeutic benefit in the treatment of neurological diseases such as Parkinson's and Alzheimer's, but some also cause dissociative effects that have led to the synthesis of illicit drugs. The ability to generate NMDAR antagonists in silico is therefore desirable for both new medication development and preempting and identifying new designer drugs. Recently, generative deep learning models have been applied to de novo drug design as a means to expand the amount of chemical space that can be explored for potential drug-like compounds. In this study, we assess the application of a generative model to the NMDAR to achieve two primary objectives: (i) the creation and release of a comprehensive library of experimentally validated NMDAR phencyclidine (PCP) site antagonists to assist the drug discovery community and (ii) an analysis of both the advantages conferred by applying such generative artificial intelligence models to drug design and the current limitations of the approach. We apply, and provide source code for, a variety of ligand- and structure-based assessment techniques used in standard drug discovery analyses to the deep learning-generated compounds. Overall, we present twelve candidate antagonists that are not available in existing chemical databases to provide an example of what this type of workflow can achieve, though synthesis and experimental validation of these compounds are still required.
The conclusions of the design research of the song adaptive delta modulator are presented for source encoding voice signals. The variation of output SNR vs input signal power/when 8, 9, and 10 bit internal arithmetic is employed. Voice intelligibility tapes to test the 10-bit system are used. An analysis of a delta modulator is also presented designed to minimize the in-band rms error. This is accomplished by frequency shaping the error signal in the modulator prior to hard limiting. The result is a significant increase in the output SNR measured after low pass filtering.
The development of dynamic dielectric analysis techniques (as a 'smart' sensor for quantitative NDE) and of intelligent closed-loop cure systems is reported. The cure process of both BF3:R-catalyzed and uncatalyzed tetraglycidyl-4,4'-diaminodiphenyl methane (TGDDM)/diamine epoxy resins was studied. Measurements were made over a frequency range of six decades. The resin was monitored continuously throughout the cure process as it changed from a viscous liquid to a highly crosslinked solid. From the frequency dependence of the dielectric loss, the specific conductivity has been determined and shown to directly monitor the viscosity before the gel point is reached. Dielectric master plots of the cure process, analogous to time-temperature superposition plots for rheological data, have been developed.