SEARCH · Engineering Papers
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.
A key obstacle hampering fielding of AI planning applications is the considerable expense of developing, verifying, updating, and maintaining the planning knowledge base (KB). Planning systems must be able to compare favorably in terms of software lifecycle costs to other means of automation such as scripts or rule-based expert systems.
The Advanced Technology Microwave Sounder (ATMS) is a new satellite microwave sounding sensor designed to provide operational weather agencies with atmospheric temperature and moisture profile information for global weather forecasting and climate applications. A TMS will continue the microwave sounding capabilities first provided by its predecessors, the Microwave Sounding Unit (MSU) and Advanced Microwave Sounding Unit (AMSU). The first flight unit was launched a year ago in October, 2011 aboard the Suomi-National Polar-Orbiting Partnership (S-NPP) satellite, part of the new Joint Polar-Orbiting Satellite System (JPSS). Microwave soundings by themselves are the highest-impact input data used by Numerical Weather Prediction models; and A TMS, when combined with the Cross-track Infrared Sounder (CrIS), forms the Cross-track Infrared and Microwave Sounding Suite (CrIMSS). The microwave soundings help meet sounding requirements under cloudy sky conditions and provide key profile information near the surface. ATMS was designed & built by Aerojet Corporation in Azusa, California, (now Northrop Grumman Electronic Systems). It has 22 channels spanning 23-183 GHz, closely following the channel set of the MSU, AMSU-AI/2, AMSU-B, Microwave Humidity Sounder (MHS), and Humidity Sounder for Brazil (HSB). It continues their cross-track scanning geometry, but for the first time, provides Nyquist sample spacing. All this is accomplished with approximately V. the volume, Y, the mass, and Y, the power of the three AMSUs. A description will be given of its performance from its first year of operation as determined by post-launch calibration activities. These activities include radiometric calibration using the on-board warm targets and cold space views, and geolocation determination. Example imagery and zooms of specific weather events will be shown. The second ATMS flight model is currently under construction and planned for launch on the "Jl" satellite of the JPSS program in approximately 2016. Additional units are expected on the J2 and 13 satellites, as well as potentially on future European METOP satellites.
Very High Level design methods emphasize automatic transfer of requirements to formal design specifications, and/or may concentrate on automatic transformation of formal design specifications that include some semantic information of the system into machine executable form. Very high level design methods range from general domain independent methods to approaches implementable for specific applications or domains. Applying AI techniques, abstract programming methods, domain heuristics, software engineering tools, library-based programming and other methods different approaches for higher level software design are being developed. Though one finds that a given approach does not always fall exactly in any specific class, this paper provides a classification for very high level design methods including examples for each class. These methods are analyzed and compared based on their basic approaches, strengths and feasibility for future expansion toward automatic development of software systems.
Large Language Models (LLMs) are emerging as powerful AI tools capable of transforming how building information is collected, processed, analyzed, and applied across diverse research areas. Their capabilities can help building operators, facility managers and other stakeholders such as designers, architects and engineers by providing actionable insights for decision-making across planning, construction, operations, and maintenance of buildings and facilities. This paper explores ten key questions concerning the role of LLMs in shaping sustainable, intelligent, and human-centric buildings. From fundamental definitions to advanced applications, we examine how LLMs facilitate decision-making across the life cycle of buildings and energy systems. LLMs can enhance life cycle assessments (LCA), building energy simulations, and real-time data integration, empowering more efficient and adaptive human-AI environments. They can also contribute to streamlining regulatory compliance, improving post-occupancy evaluations, and fostering more inclusive and participatory design processes. Additionally, this paper addresses the ethical challenges posed by LLMs, such as bias, data privacy, and environmental impacts, and explores their potentials in advancing intelligent digital twins (DT) for ongoing building operations and maintenance. Built upon our applied research using LLMs and the review of tools, datasets, and research gaps, we provide a forward-looking perspective on how LLMs can drive innovation, collaboration, and productivity in the built environment while supporting ethical and effective implementation.
This study will provide comprehensive artificial intelligence (AI)-based solution tools for network security, malware prevention, and sensor data anomaly detection for distributed energy resource (DER) research, development, and demonstration. DER technologies are energy systems (e.g., solar panels, wind turbines, and energy storage systems) that are often connected to the internet and thus vulnerable to cyberattacks. Cybersecurity should be of primary concern for DERs, which is why we propose an integrated multi-layer cyber-defense system for DERs. This system encompasses risk assessments, network security, malware prevention, and detection of anomalies in the sensor data. Implementation of a comprehensive risk assessment with an overview of the model architecture should be the primary step, and should include the potential impact of experiencing, at a given time, one or more cyberattacks on the system. The second step is to ensure that the network security includes firewalls, intrusion detection, and malware prevention. The third step is to provide solution tools that enable sensor data anomaly detection for DERs. By incorporating these considerations into DER research, development, and demonstration, organizations can help ensure the safety and security of their systems and protect against potential cyberattacks.
We propose a distributed agent architecture (DAA) that can support a variety of paradigms based on both traditional real-time computing and artificial intelligence. DAA consists of distributed agents that are classified into two categories: reactive and cognitive. Reactive agents can be implemented directly in Ada to meet hard real-time requirements and be deployed on on-board embedded processors. A traditional real-time computing methodology under consideration is the rate monotonic theory that can guarantee schedulability based on analytical methods. AI techniques under consideration for reactive agents are approximate or anytime reasoning that can be implemented using Bayesian belief networks as in Guardian. Cognitive agents are traditional expert systems that can be implemented in ART-Ada to meet soft real-time requirements. During the initial design of cognitive agents, it is critical to consider the migration path that would allow initial deployment on ground-based workstations with eventual deployment on on-board processors. ART-Ada technology enables this migration while Lisp-based technologies make it difficult if not impossible. In addition to reactive and cognitive agents, a meta-level agent would be needed to coordinate multiple agents and to provide meta-level control.
Here, we develop and test an artificial intelligence (AI)-based approach to monitor major Brazilian aquifers. The approach combines Gravity Recovery and Climate Experiment (GRACE) data and ground-based hydrogeological measurements from Brazil’s Integrated Groundwater Monitoring Network at hundreds of wells distributed in twelve aquifers across the country. We tested model ensembles based on three AI approaches: Extreme Gradient Boost, Light Gradient Boosting Model and CatBoost, followed by a Linear Regression (LR) step. The approach is further boosted with wavelet and seasonal decomposition processes applied to GRACE data. To determine the AI-based model’s sensitivity to data availability, we propose four experiments combining hydrogeological measurements from different aquifers. Groundwater storage estimates from the Global Land Data Assimilation System (GLDAS) are used as benchmark. A sensitivity analysis shows that the LR-based model ensemble is the best suited and to reproduce groundwater storage change in all studied Brazilian aquifers. Results show that the proposed approach outperforms GLDAS in all experiments, with an RMSE value of 2.68cm for the experiment that covers all monitored wells in Brazil. GLDAS resulted in RMSE=6.76cm. Using our AI model outputs, we quantified the groundwater storage change of two major aquifers, Urucuia and Bauru-Caiuá, over the past two decades: -31km 3 and -6km 3 , respectively. Water loss is driven by a prolonged drought across most of the country and intensification of groundwater pumping for irrigation. This study demonstrates that combining satellite data and AI can be a cost-effective alternative to monitor poorly equipped aquifers at the continental scale, with possible global replicability.
Deep learning (DL) workloads and their performance at scale are becoming important factors to consider as we design, develop and deploy next-generation high-performance computing systems. Since DL applications rely heavily on DL frameworks and underlying compute (CPU/GPU) stacks, it is essential to gain a holistic understanding from compute kernels, models, and frameworks of popular DL stacks, and to assess their impact on science-driven, mission-critical applications. At Oak Ridge Leadership Computing Facility (OLCF), we employ a set of micro and macro DL benchmarks established through the Collaboration of Oak Ridge, Argonne, and Livermore (CORAL) to evaluate the AI readiness of our next-generation supercomputers. In this paper, we present our early observations and performance benchmark comparisons between the Nvidia V100 based Summit system with its CUDA stack and an AMD MI100 based testbed system with its ROCm stack. We take a layered perspective on DL benchmarking and point to opportunities for future optimizations in the technologies that we consider.
Abstract Deep learning is a powerful tool to represent subgrid processes in climate models, but many application cases have so far used idealized settings and deterministic approaches. Here, we develop stochastic parameterizations with calibrated uncertainty quantification to learn subgrid convective and turbulent processes and surface radiative fluxes of a superparameterization embedded in an Earth System Model (ESM). We explore three methods to construct stochastic parameterizations: (a) a single Deep Neural Network (DNN) with Monte Carlo Dropout; (b) a multi‐member parameterization; and (c) a Variational Encoder Decoder with latent space perturbation. We show that the multi‐member parameterization improves the representation of convective processes, especially in the planetary boundary layer, compared to individual DNNs. The respective uncertainty quantification illustrates that methods (b) and (c) are advantageous compared to a dropout‐based DNN parameterization regarding the spread of convective processes. Hybrid simulations with our best‐performing multi‐member parameterizations remained challenging and crash within the first days. Therefore, we develop a pragmatic partial coupling strategy relying on the superparameterization for condensate emulation. Partial coupling reduces the computational efficiency of hybrid Earth‐like simulations but enables model stability over 5 months with our multi‐member parameterizations. However, our hybrid simulations exhibit biases in thermodynamic fields and differences in precipitation patterns. Despite this, the multi‐member parameterizations enable improvements in reproducing tropical extreme precipitation compared to a traditional convection parameterization. Despite these challenges, our results indicate the potential of a new generation of multi‐member machine learning parameterizations leveraging uncertainty quantification to improve the representation of stochasticity of subgrid effects.
In Fiscal Year (FY) 2025, the National Laboratory of the Rockies (NLR) continued to advance computing as a cornerstone of energy innovation, expanding the Kestrel high-performance computing (HPC) system to 56 peak petaflops. This growth strengthened Kestrel's role as a national asset for applied energy research, enabling larger, more complex simulations and accelerating the integration of artificial intelligence (AI) methods across the laboratory's computing portfolio. In FY 2025, AI was a component of most projects running on Kestrel, underscoring its central role in modern energy science and engineering. Kestrel supported a broad and diverse set of 507 modeling and simulation projects, engaging 855 researchers across the U.S. Department of Energy's (DOE's) Office of Critical Minerals and Energy Innovation (CMEI) portfolio and other offices, as well as partners from industry, academia, and utilities. These efforts span critical materials discovery, energy systems modeling, grid modernization, advanced manufacturing, and other areas essential to strengthening U.S. energy security and competitiveness. Together, these collaborations produced 708 technical outputs, including 293 peer-reviewed publications, reflecting both the depth and impact of the science enabled by NLR's computing capabilities. This year's report highlights the growing importance and benefit of AI throughout NLR's research programs and features work by early career researchers who are helping shape the future of computing-enabled energy innovation. Explore these sections and the many project successes captured in the pages that follow.
Weather forecasting is critical to both the Space Transportation System (STS) ground operations and the launch/landing activities at NASA Kennedy Space Center (KSC). The current launch frequency places significant demands on the USAF weather forecasters at the Cape Canaveral Forecasting Facility (CCFF), who currently provide the weather forecasting for all STS operations. As launch frequency increases, KSC's weather forecasting problems will be great magnified. The single most important problem is the shortage of highly skilled forecasting personnel. The development of forecasting expertise is difficult and requires several years of experience. Frequent personnel changes within the forecasting staff jeopardize the accumulation and retention of experience-based weather forecasting expertise. The primary purpose of this project was to assess the feasibility of using Artificial Intelligence (AI) techniques to ameliorate this shortage of experts by capturing aria incorporating the forecasting knowledge of current expert forecasters into a Weather Forecasting Expert System (WFES) which would then be made available to less experienced duty forecasters.
The 'Machine Learning Approaches to Predicting Induced Seismicity and Imaging Geothermal Reservoir Properties' project looks to apply machine learning (ML) methods to Microearthquake (MEQ) data for imaging geothermal reservoir properties and forecasting seismic events, in order to advance geothermal exploration and safe geothermal energy production. As part of the project, this submission provides data arrays for 149 microearthquakes between the year 2012 and 2013 at the Newberry EGS Site for use with the Deep Learning Algorithm that has been developed. The data provided includes raw waveform data, location data, normalized waveform data, and processed waveform data. Penn State Geothermal Team has shared the following files from the project: - 149 microearthquakes (MEQs) between 2012 and 2013 at Newberry EGS sites, 'Normalized Waveform Inputs.npz' are normalized waveforms. - labels of 149 MEQs: Processed Waveform Inputs.npz - location labels of 149 MEQs: Location Data.npz Note: .npz is the python file format by NumPy that provides storage of array data.
There has been a surge of interest in applying deep learning in particle and nuclear physics to replace labor-intensive offline data analysis with automated online machine learning tasks. This paper details a novel AI-enabled triggering solution for physics experiments in Relativistic Heavy Ion Collider and future Electron-Ion Collider. The triggering system consists of a comprehensive end-to-end pipeline based on Graph Neural Networks that classifies trigger events versus background events, makes online decisions to retain signal data, and enables efficient data acquisition. Here, the triggering system first starts with the coordinates of pixel hits lit up by passing particles in the detector, applies three stages of event processing (hits clustering, track reconstruction, and trigger detection), and labels all processed events with the binary tag of trigger versus background events. By switching among different objective functions, we train the Graph Neural Networks in the pipeline to solve multiple tasks: the edge-level track reconstruction problem, the edge-level track adjacency matrix prediction, and the graph-level trigger detection problem. We propose a novel method to treat the events as track-graphs instead of hit-graphs. This method focuses on intertrack relations and is driven by underlying physics processing. As a result, it attains a solid performance (around 72% accuracy) for trigger detection and outperforms the baseline method using hit-graphs by 2% higher accuracy.
Lawrence Livermore National Security, LLC (LLNS) is interested in receiving information about technologies that could be available in the 2029-2030 timeframe that may serve to enable the vision for a Future Generation High Performance Computing (HPC) Center (FG-HPCC) described in this document. The future HPC Center vision has been conceived to meet the future mission needs of the Advanced Simulation and Computing (ASC) Program within the National Nuclear Security Administration (NNSA). LLNS envisions a center composed not of many independent clusters, but of heterogeneous elements accessible to users as a single system. The capabilities will be integrated to create a scalable, flexible, yet tightly coupled computing center capable of integrated HPC, AI, and cloud-like workloads.
This describes a knowledge base (KB) partitioning approach to solve the problem of real-time performance using the CLIPS AI shell when containing large numbers of rules and facts. This work is funded under the joint USAF/NASA Advanced Launch System (ALS) Program as applied research in expert systems to perform vehicle checkout for real-time controller and diagnostic monitoring tasks. The Expert System advanced development project (ADP-2302) main objective is to provide robust systems responding to new data frames of 0.1 to 1.0 second intervals. The intelligent system control must be performed within the specified real-time window, in order to meet the demands of the given application. Partitioning the KB reduces the complexity of the inferencing Rete net at any given time. This reduced complexity improves performance but without undo impacts during load and unload cycles. The second objective is to produce highly reliable intelligent systems. This requires simple and automated approaches to the KB verification & validation task. Partitioning the KB reduces rule interaction complexity overall. Reduced interaction simplifies the V&V testing necessary by focusing attention only on individual areas of interest. Many systems require a robustness that involves a large number of rules, most of which are mutually exclusive under different phases or conditions. The ideal solution is to control the knowledge base by loading rules that directly apply for that condition, while stripping out all rules and facts that are not used during that cycle. The practical approach is to cluster rules and facts into associated 'blocks'. A simple approach has been designed to control the addition and deletion of 'blocks' of rules and facts, while allowing real-time operations to run freely. Timing tests for real-time performance for specific machines under R/T operating systems have not been completed but are planned as part of the analysis process to validate the design.
The intrinsic similarity between the firing of a rule and the firing of a neuron has been captured in this research to provide a neural network development system within an existing production system (CLIPS). A very important by-product of this research has been the emergence of an integrated technique of using rule based systems in conjunction with the neural networks to solve complex problems. The systems provides a tool kit for an integrated use of the two techniques and is also extendible to accommodate other AI techniques like the semantic networks, connectionist networks, and even the petri nets. This integrated technique can be very useful in solving complex AI problems.
Artificial intelligence and machine learning systems have the potential to influence the future design and implementation of cybersecurity systems for the power grid. These systems may enhance the overall operation of the power system by leveraging and making sense of massive amounts of data. However, we must also understand how AI/ML will need to be protected from cyber threat actors. We discuss the existing insights the NREL team has developed using AI/ML systems and then present resources including ESIF and the Cyber Energy Emulation Platform that can be used to generate training data and insights. We end by offering suggestions on priority research paths for AI in cybersecurity.
Advanced composites play a pivotal role in modern engineering, offering exceptional strength-to-weight ratios and tailored properties, essential for various industries. High-rate composite manufacturing techniques have rapid production capabilities, which are essential for meeting the demands of industries requiring cost-saving, efficiency, and quick turnaround times. This chapter explores the Additive Manufacturing- Compression Molding (AM-CM) system developed by Oak Ridge National Laboratory (ORNL) for advanced composites manufacturing. The AM-CM system integrates additive manufacturing with compression molding, facilitating the production of polymer composite parts with superior mechanical properties and meticulously controlled microstructures. This innovative system not only ensures precise material deposition but also operates as a fast composite manufacturing process, enhancing productivity and performance, which are needed attributes across industrial applications. Through comprehensive mechanical testing and microstructural analysis, AM-CM promotes remarkable fiber alignment and reduced porosity in composite parts compared to alternative thermoplastic high-rate composite manufacturing methods. Furthermore, AM-CM enables overmolding reinforcement using continuous carbon fiber and supports selective reinforcement through customizable toolpaths. It also facilitates the production of hybrid materials to achieve tailored mechanical properties. Future advancements in AM-CM technology aim to enhance process efficiency, broaden material versatility, and improve part performance. This involves exploring novel materials, advancing process monitoring, implementing automation technologies, and integrating artificial intelligence (AI) and machine learning (ML) for predictive modeling and real-time optimization in composite manufacturing. These developments will establish the AM-CM system as a transformative technology in composite manufacturing, driving innovation across industries.