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.
With the introduction of smart devices and rise in distributed, networked energy systems, diligent focus on cybersecurity for an evolving grid is crucial. Watch this video to learn more about what NREL is doing to mitigate threats to today’s energy infrastructure and provide a pathway to a more secure and resilient energy future.
Abstract The utility of discriminative supervised learning models built using multiple training-data sources is investigated for hidden crack localization in concrete. Feed-forward neural network (FFNN) is chosen as the model architecture, and transfer learning is used to assimilate the information obtained from different sources (computational physics simulations and laboratory experiments). The labeled training data consists of values of a damage index and the known locations of hidden cracks. The classification models need to learn how the presence of damage (hidden cracks) affects the damage index at different sensors for different test conditions. To this end, diagnostic FFNN models are built by sequentially adding and training new hidden layers to assimilate labeled information from computer models (different model geometries, test conditions, crack lengths, crack locations) and laboratory experiments on a plain cement slab. These transfer learning-based models are then used to localize damage in concrete specimens that reflect real-world conditions (i.e., specimens with steel reinforcement and randomly distributed aggregate). The actual damage state in these specimens is determined by extracting cores and performing petrographic studies on the extracted cores. The damage probability estimated by transfer learning-based models is compared with the petrographic damage rating index (DRI) to identify the most suitable approach to train the diagnostic models. The transfer learning-based diagnostic methodology shows promise and could be used in various structural health monitoring applications, where sufficient labeled data are typically not available from a single data source.
Machine learning has revolutionized drug discovery by enabling the exploration of vast, uncharted chemical spaces essential for discovering novel patentable drugs. Despite the critical role of human G protein-coupled receptors in FDA-approved drugs, exhaustive in-distribution drug-target interaction testing across all pairs of human G protein-coupled receptors and known drugs is rare due to significant economic and technical challenges. This often leaves off-target effects unexplored, which poses a considerable risk to drug safety. In contrast to the traditional focus on out-of-distribution exploration (drug discovery), we introduce a neighborhood-to-prediction model termed Chemical Space Neural Networks that leverages network homophily and training-free graph neural networks with labels as features. We show that Chemical Space Neural Networks’ ability to make accurate predictions strongly correlates with network homophily. Thus, labels as features strongly increase a machine learning model’s capacity to enhance in-distribution prediction accuracy, which we show by integrating labeled data during inference. We validate these advancements in a high-throughput yeast biosensing system (3773 drug-target interactions, 539 compounds, 7 human G protein-coupled receptors) to discover novel drug-target interactions for FDA-approved drugs and to expand the general understanding of how to build reliable predictors to guide experimental verification.
We investigate the potential of machine learning (ML) methods to model small-scale galaxy clustering for constraining Halo Occupation Distribution (HOD) parameters. Our analysis reveals that while many ML algorithms report good statistical fits, they often yield likelihood contours that are significantly biased in both mean values and variances relative to the true model parameters. This highlights the importance of careful data processing and algorithm selection in ML applications for galaxy clustering, as even seemingly robust methods can lead to biased results if not applied correctly. ML tools offer a promising approach to exploring the HOD parameter space with significantly reduced computational costs compared to traditional brute-force methods if their robustness is established. Using our ANN-based pipeline, we successfully recreate some standard results from recent literature. Properly restricting the HOD parameter space, transforming the training data, and carefully selecting ML algorithms are essential for achieving unbiased and robust predictions. Among the methods tested, artificial neural networks (ANNs) outperform random forests (RF) and ridge regression in predicting clustering statistics, when the HOD prior space is appropriately restricted. We demonstrate these findings using the projected two-point correlation function (w p (r p )), angular multipoles of the correlation function (ξ ℓ (r)), and the void probability function (VPF) of Luminous Red Galaxies from Dark Energy Spectroscopic Instrument mocks. Our results show that while combining w p (r p ) and VPF improves parameter constraints, adding the multipoles ξ 0 , ξ 2 , and ξ 4 to w p (r p ) does not significantly improve the constraints.
The integration of power electronics-based interfaces presents challenges due to the absence of detailed models and the high computational complexity. Generic models used in system studies lack accuracy in capturing converter dynamics. This paper proposes a data-driven approach developed from experimental setup data. This approach enhances accuracy in photovoltaic inverter modeling. We used two types of PV inverters in the experiment. The recorded experimental data undergo processing through a machine learning model. Results from the model trained through machine learning is also presented.
Accurate prediction of the optimal catalytic temperature ( T opt ) of enzymes is vital in biotechnology, as enzymes with high T opt values are desired for enhanced reaction rates. Recently, a machine learning method (temperature optima for microorganisms and enzymes, TOME) for predicting T opt was developed. TOME was trained on a normally distributed data set with a median T opt of 37 °C and less than 5% of T opt values above 85 °C, limiting the method’s predictive capabilities for thermostable enzymes. Due to the distribution of the training data, the mean squared error on T opt values greater than 85 °C is nearly an order of magnitude higher than the error on values between 30 and 50 °C. Here, we apply ensemble learning and resampling strategies that tackle the data imbalance to significantly decrease the error on high T opt values (>85 °C) by 60% and increase the overall R 2 value from 0.527 to 0.632. The revised method, temperature optima for enzymes with resampling (TOMER), and the resampling strategies applied in this work are freely available to other researchers as Python packages on GitHub.
The primary purpose of this report is to provide an overview of the advancement in artificial intelligence and machine learning (AI/ML) technologies and their applications in power systems. It offers a foundation for understanding the transformative role of AI/ML in power systems and aims to stimulate further research and development in this area. This report begins with a historical perspective of AI/ML technologies, then explores their advancement to today’s prominence. The document highlights key contributors to the success of AI/ML technologies, including increased computational power, greater data availability, innovative algorithms, and advanced tools. It further introduces various AI/ML techniques, including supervised, unsupervised and reinforcement learning, graph neural networks, and generative AI. It also emphasizes the critical importance of ensuring the safety, security, and trustworthiness of these AI/ML techniques within this sector. The report reviews the recent representative advancements in various power system applications enhanced by AI/ML techniques, underscoring key developments and their transformative impact as evidenced by numerous studies. It also explores both the opportunities and challenges associated with the application of AI/ML technologies to improve power system applications. While the report extensively covers AI/ML applications in power systems, focusing primarily on the technical and operational aspects, it may not thoroughly explore the sociopolitical, economic, and broader regulatory implications of AI/ML integration in power systems. AI/ML techniques hold significant potential for enhancing power system applications; however, they are not omnipotent. It is crucial to acknowledge their limitations and understand that they may not be able to address all challenges in the power system domain. Various factors must be considered that influence the implementation, adoption, and effectiveness of AI/ML solutions, including but not limited to safety, security, transparency, and trustworthiness. Additionally, the incorporation of advanced human–machine interfaces is essential, as it enables humans to validate the effectiveness of AI/ML solutions while remaining actively engaged, fostering trust in AI/ML deployment. Finally, the report summarizes AI/ML research activities supported by the Department of Energy (DOE) Office of Electricity (OE) through the Advanced Grid Modeling (AGM) program. The work aligns with the interests and mission of DOE-OE AGM, with the report serving as a resource for identifying existing progress and for pinpointing future applications within AI/ML that need further exploration and support.
In this study, we present a novel algorithm for determining directionality in 2D distributions of discrete data. We compare a reference dataset with a known direction to a measured dataset with an unknown direction by the Frobenius norm of the difference (FND) to find the unknown direction. To generalize this concept, we develop a continuous Frobenius norm of the difference (CFND) as a continuous analog of the FND and derive its analytical expression. By relating fitted and normalized 2D Gaussian distributions, we show that the CFND approximates the FND, and we validate this relationship with computer simulations. We find that a first-order approximation of the CFND between two similar Gaussian distributions takes the form of an absolute sine function, offering a simple analytical form with potential for specialized applications in segmented inverse beta decay (IBD) neutrino detectors, astronomy, machine learning, and more. Although this method may easily extend to 3D scalar fields, our focus here is on 2D real-valued fields as it directly applies to directionality. Our methodology consists of modeling a 2D Gaussian distribution, binning the data into a histogram, and encoding it as a square matrix. Rotating this matrix around its geometric center and comparing it to a measured dataset using the FND gives us rotational data that we fit with an absolute sine function. The location of the minimum of this fit is the angle closest to the true angle of the direction in the measured dataset. We present the derivation and discuss initial applications of the CFND in our novel algorithm, demonstrating its success in approximating directionality in 2D distributions.
ABSTRACT Dynamical mass estimates of simple systems such as globular clusters (GCs) still suffer from up to a factor of 2 uncertainty. This is primarily due to the oversimplifications of standard dynamical models that often neglect the effects of the long-term evolution of GCs. Here, we introduce a new approach to measure the dynamical properties of GCs, based on the combination of a deep-learning framework and the large amount of data from direct N-body simulations. Our algorithm, π-DOC (Predicting Images for the Dynamics Of stellar Clusters) is composed of two convolutional networks, trained to learn the non-trivial transformation between an observed GC luminosity map and its associated mass distribution, age, and distance. The training set is made of V-band luminosity and mass maps constructed as mock observations from N-body simulations. The tests on π-DOC demonstrate that we can predict the mass distribution with a mean error per pixel of 27 per cent, and the age and distance with an accuracy of 1.5 Gyr and 6 kpc, respectively. In turn, we recover the shape of the mass-to-light profile and its global value with a mean error of 12 per cent, which implies that we efficiently trace mass segregation. A preliminary comparison with observations indicates that our algorithm is able to predict the dynamical properties of GCs within the limits of the training set. These encouraging results demonstrate that our deep-learning framework and its forward modelling approach can offer a rapid and adaptable tool competitive with standard dynamical models.
Partitioned Global Address Space (PGAS) models exhibit tremendous promise in developing efficient and productive distributed-memory parallel applications. They have been used extensively in scientific computations due to conveniently offering a ``shared-memory''-like model and convenient interfaces that separate communication with synchronization. Traditionally, PGAS communication models have been applied to dense/contiguously distributed data, but most modern applications depict varied levels of sparsity. Existing PGAS models require certain adaptations to support distributed sparse computations, since associated computations often require matrix arithmetic, in addition to data movement. The Global Arrays toolkit from Pacific Northwest National Laboratory (PNNL) is one of the earliest PGAS models to combine one-sided data communication and distributed matrix operations and is still used in the popular NWChem quantum chemistry suite. Recently, we have expanded the Global Arrays toolkit to support common sparse operations, like sparse matrix-dense matrix multiplies (SpMM), sparse matrix-sparse matrix multiplication (SpGEMM) and Sampled Dense-Dense Matrix Multiplication (SDDMM). As it turns out, these operations are the bedrock of sparse Deep Learning (DL); sparse deep neural networks and Graph Neural Networks (GNNs) have gained increasing attention recently in achieving speedups on training and inference with reduced memory footprints. Unlike scientific applications in High Performance Computing (HPC), modern (distributed-memory capable) DL toolkits often rely on non-standardized and closed-source vendor software optimizations, creating challenges in software-hardware co-design at scale. Our goal is to support a variety of distributed-memory sparse matrix operations and helper functions in the newly created Sparse Global Arrays (SGA), such that it is possible to build portable and productive Machine Learning scenarios for algorithm/software and hardware codesign purposes. Contemporary data-parallel schemes for training/inference are undergoing a major overhaul since model replication limits scalability and causes resource inefficiencies. As such, we have adopted tensor parallelism in decomposing the model and inputs, to mitigate memory issues. Current implementation is built on top of MPI and uses CPUs to maximize the portability across the platforms.
This is the presentation prepared for the ARMA 2025 (59th US Rock Mechanics/Geomechanics Symposium) Conference held in Santa Fe, New Mexico, June 8-11, 2025. Accurate mapping and quantification of these networks are essential to ensure the integrity of CO2 storage reservoirs, understand and reduce potential leakage, and maintain long-term environmental safety. This study presents a novel machine learning-driven approach, integrated with geomechanical analysis, to quantify fracture networks and assess their spatial distribution during CO2 injection. This paper combines microseismic monitoring data with principles of hydraulic diffusivity and geomechanical analysis to characterize reservoir scale fracture network. The novelty of our approach lies in its capacity to assimilate time-dependent pressure data and microseismicity into a cohesive framework, which not only identifies microseismic triggering fronts but also tracks fracture distribution during active injection. Besides, leveraging image log data and analysis our approach also provides another angle of the insights to solidate the fracture networks understanding and geomechanical impacts. Key results from our study include the detection of over 100 distinct fracture clusters across the injection site, with fracture orientations strongly correlated with the prevailing in-situ stress field.
This is the conference paper accompanying an oral presentation at the ARMA 2025 (59th US Rock Mechanics/Geomechanics Symposium) Conference held in Santa Fe, New Mexico, June 8-11, 2025. Accurate mapping and quantification of these networks are essential to ensure the integrity of CO2 storage reservoirs, understand and reduce potential leakage, and maintain long-term environmental safety. This study presents a novel machine learning-driven approach, integrated with geomechanical analysis, to quantify fracture networks and assess their spatial distribution during CO2 injection. This paper combines microseismic monitoring data with principles of hydraulic diffusivity and geomechanical analysis to characterize reservoir scale fracture network. The novelty of our approach lies in its capacity to assimilate time-dependent pressure data and microseismicity into a cohesive framework, which not only identifies microseismic triggering fronts but also tracks fracture distribution during active injection. Besides, leveraging image log data and analysis our approach also provides another angle of the insights to solidate the fracture networks understanding and geomechanical impacts. Key results from our study include the detection of over 100 distinct fracture clusters across the injection site, with fracture orientations strongly correlated with the prevailing in-situ stress field.
Nowadays, data generated by large-scale scientific experiments are on the scale of petabytes per month. These data are transferred through dedicated high-bandwidth networks across distributed collaboration sites for processing, storage, and analysis. Like general purpose networks, research networks experience attacks. However, monitoring traffic in such high-bandwidth network environments is challenging with conventional deep packet inspection (DPI)- based network intrusion detection systems (NIDS). Recently, machine learning (ML) algorithms have shown potential in detecting anomalous patterns with better generalizability and faster speed. Compared to the general purpose networks, traffic movement in research network is less volatile and more likely to benefit from the ML-based anomaly analysis. We present ThunderSecure, an efficient AI-powered system for detecting anomalies in 100G research networks. ThunderSecure implements a high-throughput packet processing infrastructure using multi-cores and GPUs. It takes time-series statistics from a variety of network measurements and uses 2D Convolutional Neural Network (CNN) to explore both spatial and temporal dependencies in network data stream. Patterns of science data traffic are learned with a one-class neural network, which blends an Adversarial Autoencoder (AAE) with Gaussian mixture density estimation. Testing traffic flows exhibiting significant deviations from the learned baseline of normality are marked as anomalies. We trained ThunderSecure on hundreds of billions of science data packets recorded from a 100G research network at Fermi National Accelerator Laboratory. The detection performance was evaluated on traffic captured from the same research network days and weeks after the training with different types of attack flows injected. To the best of our knowledge, this is the first ML work in the area of network anomaly detection that has been validated on such extreme scale datasets. Results show that ThunderSecure can recognize science data traffic that are captured long after the training and make nearly certain detection on those with anomalous flows injected, even in the case when the anomaly-to-normal mixing ratio is 0.1%.
Imaging is a dominant strategy for data collection in neuroscience, yielding stacks of images that often scale to gigabytes of data for a single experiment. Machine learning algorithms from computer vision can serve as a pair of virtual eyes that tirelessly processes these images, automatically detecting and identifying microstructures. Unlike learning methods, our Flexible Learning-free Reconstruction of Imaged Neural volumes (FLoRIN) pipeline exploits structure-specifc contextual clues and requires no training. This approach generalizes across diferent modalities, including serially-sectioned scanning electron microscopy (sSEM) of genetically labeled and contrast enhanced processes, spectral confocal refectance (SCoRe) microscopy, and high-energy synchrotron X-ray microtomography (μCT) of large tissue volumes. We deploy the FLoRIN pipeline on newly published and novel mouse datasets, demonstrating the high biological fdelity of the pipeline’s reconstructions. FLoRIN reconstructions are of sufcient quality for preliminary biological study, for example examining the distribution and morphology of cells or extracting single axons from functional data. Compared to existing supervised learning methods, FLoRIN is one to two orders of magnitude faster and produces high-quality reconstructions that are tolerant to noise and artifacts, as is shown qualitatively and quantitatively.
Imaging is a dominant strategy for data collection in neuroscience, yielding stacks of images that often scale to gigabytes of data for a single experiment. Machine learning algorithms from computer vision can serve as a pair of virtual eyes that tirelessly processes these images, automatically detecting and identifying microstructures. Unlike learning methods, our Flexible Learning-free Reconstruction of Imaged Neural volumes (FLoRIN) pipeline exploits structure-specific contextual clues and requires no training. This approach generalizes across different modalities, including serially-sectioned scanning electron microscopy (sSEM) of genetically labeled and contrast enhanced processes, spectral confocal reflectance (SCoRe) microscopy, and high-energy synchrotron X-ray microtomography (μCT) of large tissue volumes. We deploy the FLoRIN pipeline on newly published and novel mouse datasets, demonstrating the high biological fidelity of the pipeline’s reconstructions. FLoRIN reconstructions are of sufficient quality for preliminary biological study, for example examining the distribution and morphology of cells or extracting single axons from functional data. Compared to existing supervised learning methods, FLoRIN is one to two orders of magnitude faster and produces high-quality reconstructions that are tolerant to noise and artifacts, as is shown qualitatively and quantitatively.
Plug loads are responsible for a significant portion of the energy consumed in commercial buildings, yet their distributed and ever-changing nature makes them one of the most challenging building end uses to manage. Plug load management systems exist today that utilize smart plugs to meter and control devices at the outlet level, however, their uptake has been relatively slow in part due to the significant labor required for installation and maintenance. Learning behavior algorithms and automatic and dynamic load detection have been identified as two technology areas that could accelerate the adoption of plug load management systems by reducing these labor demands and providing additional energy efficiency and non-energy benefits. Learning behavior algorithms learn occupant behavior and adjust plug load management systems accordingly, allowing for the automatic creation of optimized control schedules. Automatic and dynamic load detection allows a plug load management system to identify devices as they are plugged in to a building and keeps the system up to date as devices are moved throughout a building. In this paper, we present our findings with respect to the current state of these two technologies based on a review of existing research and patents, as well as a series of interviews with companies working in the plug load space. We have found that, as of now, no commercialized solutions exist for these plug load technologies and that more work is needed to bring them to market. In addition, we summarize our findings related to the technology challenges, market barriers, drivers, and opportunities for these technologies moving forward.
ABSTRACT Galaxy morphology is a fundamental quantity, which is essential not only for the full spectrum of galaxy-evolution studies, but also for a plethora of science in observational cosmology (e.g. as a prior for photometric-redshift measurements and as contextual data for transient light-curve classifications). While a rich literature exists on morphological-classification techniques, the unprecedented data volumes, coupled, in some cases, with the short cadences of forthcoming ‘Big-Data’ surveys (e.g. from the LSST), present novel challenges for this field. Large data volumes make such data sets intractable for visual inspection (even via massively distributed platforms like Galaxy Zoo), while short cadences make it difficult to employ techniques like supervised machine learning, since it may be impractical to repeatedly produce training sets on short time-scales. Unsupervised machine learning, which does not require training sets, is ideally suited to the morphological analysis of new and forthcoming surveys. Here, we employ an algorithm that performs clustering of graph representations, in order to group image patches with similar visual properties and objects constructed from those patches, like galaxies. We implement the algorithm on the Hyper-Suprime-Cam Subaru-Strategic-Program Ultra-Deep survey, to autonomously reduce the galaxy population to a small number (160) of ‘morphological clusters’, populated by galaxies with similar morphologies, which are then benchmarked using visual inspection. The morphological classifications (which we release publicly) exhibit a high level of purity, and reproduce known trends in key galaxy properties as a function of morphological type at z < 1 (e.g. stellar-mass functions, rest-frame colours, and the position of galaxies on the star-formation main sequence). Our study demonstrates the power of unsupervised machine learning in performing accurate morphological analysis, which will become indispensable in this new era of deep-wide surveys.
We evaluate uncertainty quantification (UQ) methods for deep learning applied to liquid argon time projection chamber (LArTPC) physics analysis tasks. As deep learning applications enter widespread usage among physics data analysis, neural networks with reliable estimates of prediction uncertainty and robust performance against overconfidence and out-of-distribution (OOD) samples are critical for their full deployment in analyzing experimental data. While numerous UQ methods have been tested on simple datasets, performance evaluations for more complex tasks and datasets are scarce. Here we assess the application of selected deep learning UQ methods on the task of particle classification using the PiLArNet monte carlo 3D LArTPC point cloud dataset. We observe that UQ methods not only allow for better rejection of prediction mistakes and OOD detection, but also generally achieve higher overall accuracy across different task settings. We assess the precision of uncertainty quantification using different evaluation metrics, such as distributional separation of prediction entropy across correctly and incorrectly identified samples, receiver operating characteristic curves (ROCs), and expected calibration error from observed empirical accuracy. We conclude that ensembling methods can obtain well calibrated classification probabilities and generally perform better than other existing methods in deep learning UQ literature.