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.
An oral presentation as ESRI user conference 2024 on a dashboard to support just community transitions for Carbon Capture and Storage. This work is part of BIL supported EDX4CCS task on CCS-EJ-SJ Database and application.
Traditional methods for identifying “hit” molecules from a large collection of potential drug-like candidates rely on biophysical theory to compute approximations to the Gibbs free energy of the binding interaction between the drug and its protein target. These approaches have a significant limitation in that they require exceptional computing capabilities for even relatively small collections of molecules. Increasingly large and complex state-of-the-art deep learning approaches have gained popularity with the promise to improve the productivity of drug design, notorious for its numerous failures. However, as deep learning models increase in their size and complexity, their acceleration at the hardware level becomes more challenging. Hyperdimensional Computing (HDC) has recently gained attention in the computer hardware community due to its algorithmic simplicity relative to deep learning approaches. The HDC learning paradigm, which represents data with high-dimension binary vectors, allows the use of low-precision binary vector arithmetic to create models of the data that can be learned without the need for the gradient-based optimization required in many conventional machine learning and deep learning methods. This algorithmic simplicity allows for acceleration in hardware that has been previously demonstrated in a range of application areas (computer vision, bioinformatics, mass spectrometery, remote sensing, edge devices, etc.). To the best of our knowledge, our work is the first to consider HDC for the task of fast and efficient screening of modern drug-like compound libraries. We also propose the first HDC graph-based encoding methods for molecular data, demonstrating consistent and substantial improvement over previous work. We compare our approaches to alternative approaches on the well-studied MoleculeNet dataset and the recently proposed LIT-PCBA dataset derived from high quality PubChem assays. We demonstrate our methods on multiple target hardware platforms, including Graphics Processing Units (GPUs) and Field Programmable Gate Arrays (FPGAs), showing at least an order of magnitude improvement in energy efficiency versus even our smallest neural network baseline model with a single hidden layer. Our work thus motivates further investigation into molecular representation learning to develop ultra-efficient pre-screening tools. We make our code publicly available at https://github.com/LLNL/hdbind.
The commercial-scale deployment of floating offshore wind (FOW) projects is expected to take place in a diverse range of sites that may differ significantly from existing fixed-bottom projects. FOW farms are particularly sensitive to the water depth and the meteorological and oceanographic (metocean) and geotechnical conditions at the project site due to the wave-induced system motions and loads as well as the anchoring system constraints imposed by the seafloor conditions. Uncertainty around the site conditions will permeate through all aspects of project design, leading to suboptimal and overly conservative designs, increased costs, and adversely affected performance. As FOW expands into a global industry, metocean and geotechnical conditions will increasingly vary for projects located in different geographic regions or in far-from-shore, deep-water sites. This study represents the outputs of work package 1 of International Energy Agency Wind Task 49, which focuses on the integrated design of floating wind arrays. The primary goal of this study is to establish the type of parameters and constraints required to characterize FOW array reference sites; provide a realistic and publicly available set of reference site conditions to the FOW community as a baseline set of data for individual research projects; identify and categorize any critical gaps in the existing data or methodologies required to define reference site characteristics; and inform and support the design of reference FOW arrays. A building block concept was developed for synthesizing reference sites for the design of FOW arrays. The building blocks include three classes of site conditions focusing on the techno-economic design of FOW projects: metocean conditions, seabed conditions, and coastal infrastructure. All reference site data produced and collected in this study are publicly available.
This project, funded by the US Department of Energy’s Solar Energy Technologies Office (SETO), focused on the operation of microgrids as a coordinated network. The primary objective, which was successfully achieved, was to develop both control strategies and hardware solutions to support the resilient and efficient operation of networked microgrids with high solar penetration. The work was structured around the following four main tasks: • Development of distributed and scalable optimization algorithms for AC-coupled networked microgrids. • Design and implementation of a novel DC interconnection hardware to enable precise power exchange between microgrids. • Laboratory operational validation of the developed technologies using 480 V testbeds and commercially available hardware. • Field operational validation of the complete solution in Adjuntas, Puerto Rico, interconnecting two kW-scale, split-phase microgrids of Casa Pueblo’s microgrids. This project addressed multiple technical challenges across the domains of optimization, control, hardware interconnection, and protection. One of its key contributions was delivering tangible, real-world solutions for networking microgrids. In contrast to purely theoretical or simulation-based work, this project included full-scale hardware operational validation both in the lab and in the field. The work conducted as part of this project—in collaboration with the University of Puerto Rico; the University of Tennessee, Knoxville; the University of Central Florida; and Casa Pueblo—has advanced the state of the art in networked microgrids. Key contributions include the development of distributed control strategies, practical solutions for real-world implementation challenges, and the introduction of a novel DC interlink approach for microgrid interconnection. The project featured both laboratory and field validation using commercial off-the-shelf components. The field deployment successfully validated that a group of microgrids can operate in a coordinated manner, enabling precise power flow between systems and mutual support during extreme events. This project resulted in 15 journal publications and 15 conference papers; 5 graduate students and 15 undergraduate students were supported. The codes of distributed optimization and forecasting were made open-source through OSTI.gov for distributed optimization and forecasting. All the publications are available in the ORNL-hosted project landing page. The DC interlink with state-of-charge balancing control was operationally validated in Adjuntas by interconnecting two real-world, 240 V split-phase microgrids. To the best knowledge of the team, this represents the first operational validation of AC microgrids interconnected via DC-interlinks. As a culmination of this project, a follow-on grant was awarded to support the technology transfer of the distributed optimization framework to a commercial microgrid controller, Stellar Edge, developed by the California-based company New Sun Road.
Prior work in FY24 developed an adversarial AI agent aid in path analysis of physical protection systems. This agent, trained using a model-based reinforcement learning algorithm, was able to successfully learn the most vulnerable path in facilities. It was able to extend the current state of practice for physical protection design by exhibiting dynamic behavior based on current environmental conditions. Whereas PathTrace largely performs a static, graph-based analysis, the AI agent was able to make decisions based on relative position in the facility, current conditions (was the adversarial agnet discovered?), and proximity to secondary targets. The agent demonstrated some novel capabilities, but had limitations that need to be resolved before it can be used for production purposes. For example, the adversarial agent generalizes poorly and takes a relatively long time to train. Nonetheless, there is still considerable promise for developing the adversarial agent further in order to explore even richer, more dynamic behaviors (e.g., adversary motivations, environmental debris, and more). This work considers a complementary idea; development of a planning agent. The planning agent is envisioned as an auto-complete-like tool that can help accelerate security system design by human experts. The agent would respect existing barriers and sensors placed by a human expert while offering cost-effective suggestions (i.e., implicitly balancing effectiveness with cost) to improve the design. The goal is for this agent to be part of an expert’s toolbox, not to totally upend the current state-of-practice, or to displace human experts. The ultimate goal would be concurrent training of both the adversarial and planning agent together, to learn entirely through self-play. This would represent an entirely new way of performing system deign. We selected a hierarchical, model-based reinforcement learning algorithm to serve as the planning agent. This is an extension of concepts used in the prior FY24 adversarial agent work. There, we had a single agent acting an environment. Here, we have two different sub-agents (policies), working together, to form a complete agent. There is a manager policy, which can select abstract goals on slower time scales, and a worker, which performs primitive actions to reach goals selected by the manager. It is worth noting that this class of algorithm is challenging to work with. From our understanding, our work is one of the first successful uses of model-based reinforcement learning (MBRL) in nuclear energy1 , and likely the first hierarchical model-based reinforcement learning application in nuclear energy. Further, this work is one of the first known attempts to apply AI to perform a design tasks in nuclear energy. Consequently, there were significant implementation challenges and the bulk of the work was focused on successful implementation and algorithm design. The results presented here are very low technology readiness level as a consequence of the lack of related literature, but still represent a significant step forward in the pursuit of applied AI for design.
The Southwest Regional Partnership on Carbon Sequestration (SWP) is one of 7 regional partnerships formed in 2003 under the U.S. Department of Energy’s (DOE) Regional Carbon Sequestration Partnerships (RCSPs) initiative. The overall purpose of the initiative was to help determine and implement the technology, infrastructure, and regulations most appropriate to promote carbon storage in different regions of the country. Covering Arizona, Colorado, New Mexico, Oklahoma, Utah, and parts of Texas, Wyoming, and Kansas, the SWP evaluated regional carbon storage and utilization potential and focused on technologies and sites that could complement the region’s strong position in energy production. The project progressed through three phases: • Phase I (2003–2005): Characterized regional geologic formations and CO 2 sources, assessed sequestration potential, and identified pilot test sites. • Phase II (2005–2013): Conducted small-scale field tests to validate sequestration methods, including geologic and terrestrial projects. • Phase III (2008–2022): Demonstrated large-scale CO 2 injection at a commercial oil field to test monitoring, verification, and long-term storage strategies. This report covers Phase III. The final project site, the Farnsworth Unit (FWU) in Texas, provided real-world testing of reservoir characterization, monitoring, and risk evaluation tools and processes that could be used in any commercial scale carbon capture, utilization, and storage (CCUS) project. Extensive data collection and analysis helped refine best practices for reservoir characterization, injection monitoring, and storage verification. The SWP contributed to national databases, DOE best practice manuals, and regional geological assessments to support future sequestration efforts. Key lessons learned include the importance of robust data management, strategic site selection, regulatory navigation, and effective industry collaboration. The project’s findings will inform ongoing and future carbon storage initiatives. Task 1 (Regional Characterization) • The SWP continued to participate in national outreach efforts and NATCARB. • The SWP evaluated multiple potential sites before selecting the FWU as the primary field test location. Task 2 (Public Outreach and Education) • The SWP contributed to national databases, DOE best practice manuals, and regional geological assessments to support future sequestration efforts. Task 3 (Permitting and Regulatory Compliance) • The SWP ensured compliance with federal and state regulations, including National Environmental Policy Act (NEPA) requirements. • The SWP obtained all necessary permits for drilling, injection, and monitoring activities. Task 4 (Site Characterization and Planning) • The SWP developed work plans for four key activities: characterization, simulation, monitoring and verification, and risk evaluation. • The SWP collected and synthesized legacy data from multiple sources to build initial static geological models and dynamic reservoir models demonstrating project feasibility. • The SWP conducted an initial risk evaluation and developed mitigation plans. Task 5 (Field Operations and Data Collection) • The SWP drilled, logged, and cored three characterization wells to gather critical subsurface data. • The SWP conducted multiple geophysical surveys, including 3D seismic, crosswell seismic, and vertical seismic profiling, to improve reservoir characterization. Task 6 (Monitoring and Verification) • The SWP performed extensive geological characterization using data from characterization wells and seismic surveys. • The SWP established a surface monitoring network to track CO 2 flux in soil gas, groundwater chemistry, and near-surface atmospheric CO 2 levels. • The SWP built and refined reservoir models to study the effects of relative permeability on simulation behavior and improve calibration with experimental data. Task 7 (Risk Assessment and Model Refinement) • The SWP conducted multiple studies to evaluate reservoir integrity, predict CO 2 plume behavior and improve predictive modeling capabilities. • The SWP refined geological models and used them to enhance the accuracy of simulation models. • The SWP continued quantitative risk assessment of top-ranked risks and strengthened the link between qualitative and quantitative risk methodologies.
Manual labeling for machine learning tasks such as image classification is tedious and labor-intensive; as a result, scientific datasets suitable for deep learning applications are scarce and limited. While data augmentation techniques have shown promise for extending image datasets, very little work has been done to understand the impact of combining multiple augmentation methods sequentially or the limits of their effectiveness when combined. Our work addresses this gap by examining how standard and combinatorial data augmentation affects the performance of machine learning models when trained on small datasets for label classification tasks. For our analysis, we generate single, double and quadruple-augmented datasets for a microscopy image classification task using six standard augmentation methods, and compare the resultant improvements observed in binary classification accuracy with three standard image classification models (DenseNet169, MobileNetV2, ResNet101V2). Our experiments show a non-monotonic relationship between the number of simultaneous augmentation methods and classification accuracy, indicating that there is a trade-off between the degree of augmentation and the model performance. These findings suggest that the optimal number of augmentation methods will vary by domain and use case. We also find that the order in which augmentation methods are applied to a limited dataset matters when combining augmentation schemes, with our use case showing performance differences up to 2.6% when the augmentation order is reversed for double-augmented datasets. Our work offers insights to the limits of data augmentation when working on image classification tasks with limited datasets.
Isotopic measurements of the minor isotopes of uranium ( 233 U, 234 U, 236 U) are indispensable measures for material accountability and processing history but can be particularly challenging to accurately measure due to the very low abundance of the minor isotopes relative to 238 U, or 235 U in the case of highly enriched uranium (HEU). The goal of this project is to quantify detection limits for minor uranium isotope species at environmental levels (e.g 236 U/ 238 U < 10 -7 ) utilizing an ATONA-outfitted instrument and the Total Evaporation Thermal Ionization Mass Spectrometry (TE-TIMS) method. In the first half of Fiscal Year 2025, we set out to identify a suitable experimental material (Task 1) and to optimize our TE-TIMS measurement routine (Task 2). We have successfully identified a material and are in the process of getting an aliquot shipped to LANL. Simultaneously, we began working on our measurement routine utilizing CRM 112A. The results of each task are discussed in detail below.
Surrogate model development is a critical step for uncertainty quantification or other sample-intensive tasks for complex computational models. Here, in this work, we develop a multi-output surrogate form using a class of neural networks (NNs) that employ shortcut connections, namely Residual NNs (ResNets). ResNets are known to regularize the surrogate learning problem and improve the efficiency and accuracy of the resulting surrogate. Inspired by the continuous, Neural ODE analogy, we augment ResNets with weight parameterization strategy with respect to ResNet depth. Weight-parameterized ResNets regularize the NN surrogate learning problem and allow better generalization with a drastically reduced number of learnable parameters. We demonstrate that weight-parameterized ResNets are more accurate and efficient than conventional feed-forward multi-layer perceptron networks. We also compare various options for parameterization of the weights as functions of ResNet depth. We demonstrate the results on both synthetic examples and a large scale earth system model of interest.
The classical way of studying the rainfall‐runoff processes in the water cycle relies on conceptual or physically‐based hydrologic models. Deep learning (DL) has recently emerged as an alternative and blossomed in the hydrology community for rainfall‐runoff simulations. However, the decades‐old Long Short‐Term Memory (LSTM) network remains the benchmark for this task, outperforming newer architectures like Transformers. In this work, we propose a State Space Model (SSM), specifically the Frequency Tuned Diagonal State Space Sequence (S4D‐FT) model, for rainfall‐runoff simulations. The proposed S4D‐FT is benchmarked against the established LSTM and a physically‐based Sacramento Soil Moisture Accounting model under in‐sample and out‐of‐sample simulation setups across 531 watersheds in the contiguous United States (CONUS). Results show that S4D‐FT is able to outperform the LSTM model across diverse regions under both simulation setups, especially for regions that feature snowmelt‐driven or intermittent flow regimes. In contrast, S4D‐FT tends to underperform in flashier, high‐magnitude flow regimes, likely due to its global state‐space convolution computation that emphasizes slow, storage‐driven dynamics, which makes it less effective at picking up short bursts and noisy spikes in the data. In summary, our pioneering introduction of the S4D‐FT for rainfall‐runoff simulations challenges the dominance of LSTM in the hydrology community and expands the arsenal of DL tools available for hydrological modeling.
This presentation summarizes work completed over the past year for PrOMMiS Task 2.1. The first half of the preseation motivates the need to develop sustainable critical mineral separation technology, and how PSE is particularly well-suited to tackle this problem. The second half of the presentation presents preliminary results for the custom cost model for diafiltration (2.1) using the superstructure flowsheet developed by Carnegie Mellon University. The presentation serves to emphasize the multi-disciplinary and collaborate nature of developing a circular economy of critical minerals and materials.
Fock-basis-encoding qudits allow quantum gates to operate on multiple qudit levels simultaneously. This capability is advantageous for applications involving complex local gate operations, such as the digital quantum simulation of non-Abelian gauge theories. However, compiling unitary operations to native qudit gates remains a challenging task, often requiring substantial computational resources. In this work, we theoretically investigate the direct pulse control of a superconducting radio-frequency cavity coupled to an ancilla transmon for gate synthesis. We will discuss the improvements and limitations applicable to high Fock states. This work paves the way toward more efficient and robust qudit-gate synthesis.
Fock-basis-encoding qudits allow quantum gates to operate on multiple qudit levels simultaneously. This capability is advantageous for applications involving complex local gate operations, such as the digital quantum simulation of non-Abelian gauge theories. However, compiling unitary operations to native qudit gates remains a challenging task, often requiring substantial computational resources. In this work, we theoretically investigate the direct pulse control of a superconducting radio-frequency cavity coupled to an ancilla transmon for gate synthesis. We will discuss the improvements and limitations applicable to high Fock states. This work paves the way toward more efficient and robust qudit-gate synthesis.
Optimizing scientific software is a difficult task because codebases are often large and complex, and performance can depend upon several factors including the algorithm, its implementation, and hardware among others. Causes of poor performance can originate from disparate sources and be difficult to diagnose. Recent years have seen a multitude of work that use large language models (LLMs) to assist in software development tasks. However, these tools are trained to model the distribution of code as text, and are not specifically designed to understand performance aspects of code. In this work, we introduce a reinforcement learning based methodology to align the outputs of code LLMs with performance. This allows us to build upon the current code modeling capabilities of LLMs and extend them to generate better performing code. Here, we demonstrate that our fine-tuned model improves the expected speedup of generated code over base models for a set of benchmark tasks from 0.9 to 1.6 for serial code and 1.9 to 4.5 for OpenMP parallel code.
There is nowadays a consensus among many countries that geological disposal is a favourable solution for the long-term management. Although different host formations and different barrier systems are under consideration around the world, clay-based materials form an important component for waste isolation in most national programmes. Hence, a good comprehension of the effect of gas flow on the hydro-mechanical behaviour of clay-based soils is essential, both at laboratory and field scale. Task B under the international cooperative project DECOVALEX-2023 has recently shown that, after some enhancement, models can be employed to reproduce laboratory scale tests, even with different sample geometries 37 . However, further work is required to understand whether they can be applied to simulate a large-scale experiment. Up-scaling of models for the advective transport of gas through clay-based low permeable material presents a number of problems related to the difficulty in obtaining consistent hydrogeological parameters and constitutive relationships at both laboratory and field scale. Based on a unique dataset from a large-scale gas injection test (Lasgit) performed at the Äspö Hard Rock Laboratory (Sweden), Task B within DECOVALEX-2023 has explored the refinement of these numerical strategies applied to the simulation of gas flow. Work performed within the task reveals that codes do not need to be substantially modified from the laboratory models to reproduce full-scale tests: indeed, model parameters calibrated and validated at laboratory scale have been applied to predict field scale gas flow at Lasgit, including peak gas pressure and injected cumulative gas volume. By means of (1) the introduction of interfaces between blocks to reflect the experimental configuration and the (2) adjustment of some parameters (e.g., higher permeability), the updated models are able to represent most of the key features observed in the experimental data, even at a large scale.
As part of their duties, Human Reliability Analysis must often evaluate if crews in nuclear power plants (NPPs) can complete tasks associated with a human-failure event within time limits. For example, the time required in NPP scenarios is determined by systematic and structured walkthroughs, feasibility studies, recorded times from training exercises, and interviews with experienced operators and experts. Typically, a point estimate is derived for the estimate (mean, maximum, or 95th percentile of time required). Using point-estimate values can mask the risk associated with variability among crews, plant conditions and set-up, environmental conditions, and other impact factors under which these actions are executed. While point estimates for time required and time available have served the industry well, without considering the uncertainty they could lead to biased understanding about the risk. The Integrated Human Event Analysis System - General Methodology (IDHEAS-G) model (developed by the US Nuclear Regulatory Commission, NRC) for human error probability calculates human error probability by summing two probabilities: insufficient time and cognitive error. As such, the model takes a more holistic approach by considering the full distributions for time required and time available to calculate the human error probability because the time available to complete the task is insufficient. In this study, we expand on the work of the NRC and discuss methods for estimating these time considerations. For example, for the time required, the impact of Performance Influencing Factors (PIFs) on the distribution was divided into impacts that are aleatory in nature, such as crew-to-crew variability, and those that are epistemic (i.e., the PIFs). Starting with the factors that introduce aleatory uncertainty, a first-order distribution was developed from a large set of time required (i.e., NPP task completion times) data for the range of operator actions that occur in the NPP control room under simulated accident conditions. The first-order distribution can then be adjusted to account for epistemic uncertainty using research associated with the impact of applicable PIFs on the time required. We also develop guidance for analysts to address the probability distributions for the time available. The guidance we developed on how to estimate time required and time available distributions is based on the identification of pertinent research and data, data analyses, and expert knowledge elicitation.
Accurate representation of the multiscale features in spatiotemporal physical systems using vision transformer architectures requires extremely long, computationally prohibitive token sequences. To address this issue, we propose two novel adaptive tokenization schemes that dynamically adjust patch sizes based on local features: one ensures convergent behavior to uniform patch refinement, while the other offers better computational efficiency. Moreover, we present a set of spatiotemporal attention schemes, where the temporal or axial spatial dimensions are decoupled, to evaluate their baseline computational and data efficiencies and to determine whether adaptive tokenization can improve this performance. We assess the performance of the proposed multiscale adaptive model, MATEY, in a sequence of experiments. Compared to a full spatiotemporal attention scheme or a scheme that decouples only the temporal dimension, we find that fully decoupled axial attention is less efficient and expressive, requiring more training time and model parameters to achieve the same accuracy. The experiments on the adaptive tokenization schemes show that, compared to a uniformly refined model, the proposed schemes achieve comparable or improved accuracy at a much lower cost in the tested two-dimensional settings. While the asymptotic analysis suggests the potential for favorable scaling, empirical validation at substantially longer sequence lengths remains to be performed in future work. Finally, we demonstrate in two fine-tuning tasks featuring different physics that models pretrained on PDEBench data outperform the ones trained from scratch, especially in the low data regime with frozen attention.
Symmetric Nonnegative Matrix Factorization (SymNMF) is a technique in data analysis and machine learning that approximates a matrix with a product of a nonnegative, low-rank matrix and it transpose. To design faster and more scalable algorithms for SymNMF we develop two randomized algorithms for its computation. The first method uses randomized matrix sketching to compute an initial low-rank approximation to the input matrix and proceeds to uses this as a low-rank input to rapidly compute a SymNMF. The second methods uses randomized leverage score sampling to approximately solve constrained least squares problems. Many successful methods for SymNMF rely on (approximately) solving sequences of constrained least squares problems. Here, we prove theoretically that leverage score sampling can approximately solve constrained least squares problems to e-accuracy. Finally we demonstrate both methods work in practice by applying them to graph clustering tasks on large real world data sets. These experiments show that our methods approximately maintain solution quality and achieve significant speed ups for both large dense and large sparse problems.