Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “network acceleration”

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.

At least 667 records · Page 37
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

FPGAs-as-a-Service Toolkit (FaaST)

Computing needs for high energy physics are already intensive and are expected to increase drastically in the coming years. In this context, heterogeneous computing, specifically as-a-service computing, has the potential for significant gains over traditional computing models. Although previous studies and packages in the field of heterogeneous computing have focused on GPUs as accelerators, FPGAs are an extremely promising option as well. A series of workflows are developed to establish the performance capabilities of FPGAs as a service. Multiple different devices and a range of algorithms for use in high energy physics are studied. For a small, dense network, the throughput can be improved by an order of magnitude with respect to GPUs as a service. For large convolutional networks, the throughput is found to be comparable to GPUs as a service. This work represents the first open-source FPGAs-as-a-service toolkit.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2023

Residual loss due to magnetic flux trapping in niobium superconducting microwave cavities

One of the significant source of radio frequency loss in superconducting microwave cavities is the trapped residual magnetic field during the cooldown due to the incomplete Messier effect. The sensitivity of flux trapping depends on the distribution of impurities, network of dislocations and defects which pin vortices. Recent advancement in cavity processing techniques resulted in very high quality factors with impurity doping, however the rf loss related to the flux trapping increases. High quality factor in SRF cavities is not only limited to particle accelerators but also emerging as an application to quantum computing and quantum information science. In this presentation, we present the results of measurements of flux trapping sensitivity on several 1.3 GHz elliptical accelerating cavities with different grain size, impurity content and surface preparations.

Dhakal, P.↗
DOE OSTI2022

Rolling Optimization of Transmission Network Recovery and Load Restoration Considering Hybrid Wind-Storage System and Cold Load Pickup

A common solution to deal with the stochasticity introduced by fast-ramping wind power integration is to equip wind farms (WFs) with energy storage systems (ESSs) to formulate hybrid WF-ESSs. In addition to leveling off wind power fluctuations during normal operations, a hybrid WF-ESS can be a flexible power source to accumulate the power system restoration. In this paper, we propose a rolling optimization model for transmission network recovery and load restoration considering the contributions of WF-ESSs. The proposed model is formulated as a mixed integer linear programming problem that simultaneously optimizes the amount and location of restorable load blocks as well as the restoration lines. The cold load pickup features of interrupted loads considering the outage duration are modeled in detail. A chance-constrained method is employed to deal with the uncertainty of wind power, and a rolling horizon-based framework is adopted to reduce the influence of forecast error. Case studies are conducted on both New England 39-bus system and part of a provincial power system in China. The results show that the load restoration process can be significantly accelerated by employing the proposed method and contributions of hybrid WF-ESSs to power system restoration are validated.

chance-constrained optimization↗
DOE OSTI2025

Breaking the barrier of human-annotated training data for machine learning-aided plant research using aerial imagery

Machine learning (ML) can accelerate biological research. However, the adoption of such tools to facilitate phenotyping based on sensor data has been limited by (i) the need for a large amount of human-annotated training data for each context in which the tool is used and (ii) phenotypes varying across contexts defined in terms of genetics and environment. This is a major bottleneck because acquiring training data is generally costly and time-consuming. This study demonstrates how a ML approach can address these challenges by minimizing the amount of human supervision needed for tool building. A case study was performed to compare ML approaches that examine images collected by an uncrewed aerial vehicle to determine the presence/absence of panicles (i.e. “heading”) across thousands of field plots containing genetically diverse breeding populations of 2 Miscanthus species. Automated analysis of aerial imagery enabled the identification of heading approximately 9 times faster than in-field visual inspection by humans. Leveraging an Efficiently Supervised Generative Adversarial Network (ESGAN) learning strategy reduced the requirement for human-annotated data by 1 to 2 orders of magnitude compared to traditional, fully supervised learning approaches. The ESGAN model learned the salient features of the data set by using thousands of unlabeled images to inform the discriminative ability of a classifier so that it required minimal human-labeled training data. This method can accelerate the phenotyping of heading date as a measure of flowering time in Miscanthus across diverse contexts (e.g. in multistate trials) and opens avenues to promote the broad adoption of ML tools.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

pixelvar79/ESGAN-Flowering-Detection-paper

Machine learning (ML) can accelerate biological research. However, the adoption of such tools to facilitate phenotyping based on sensor data has been limited by (i) the need for a large amount of human-annotated training data for each context in which the tool is used and (ii) phenotypes varying across contexts defined in terms of genetics and environment. This is a major bottleneck because acquiring training data is generally costly and time-consuming. This study demonstrates how a ML approach can address these challenges by minimizing the amount of human supervision needed for tool building. A case study was performed to compare ML approaches that examine images collected by an uncrewed aerial vehicle to determine the presence/absence of panicles (i.e. “heading”) across thousands of field plots containing genetically diverse breeding populations of 2 Miscanthus species. Automated analysis of aerial imagery enabled the identification of heading approximately 9 times faster than in-field visual inspection by humans. Leveraging an Efficiently Supervised Generative Adversarial Network (ESGAN) learning strategy reduced the requirement for human-annotated data by 1 to 2 orders of magnitude compared to traditional, fully supervised learning approaches. The ESGAN model learned the salient features of the data set by using thousands of unlabeled images to inform the discriminative ability of a classifier so that it required minimal human-labeled training data. This method can accelerate the phenotyping of heading date as a measure of flowering time in Miscanthus across diverse contexts (e.g. in multistate trials) and opens avenues to promote the broad adoption of ML tools.

Varela, Sebastian↗
DOE OSTI2025

Accurate and uncertainty-aware multi-task prediction of HEA properties using prior-guided deep Gaussian processes

Surrogate modeling techniques have become indispensable in accelerating the discovery and optimization of high-entropy alloys (HEAs), especially when integrating computational predictions with sparse experimental observations. This study systematically evaluates the training and testing performance of four prominent surrogate models—conventional Gaussian processes (cGP), Deep Gaussian processes (DGP), encoder-decoder neural networks for multi-output regression and eXtreme Gradient Boosting (XGBoost)—applied to a hybrid dataset of experimental and computational properties of the 8-component HEA system Al-Co-Cr-Cu-Fe-Mn-Ni-V. We specifically assess their capabilities in predicting correlated material properties, including yield strength, hardness, modulus, ultimate tensile strength, elongation, and average hardness under dynamic/quasi-static conditions, alongside auxiliary computational properties. The comparison highlights the strengths of hierarchical deep modeling approaches in handling heteroscedastic, heterotopic, and incomplete data commonly encountered in materials science. Our findings illustrate that combined surrogate models such as DGPs infused with machine-learned priors outperform other surrogates by effectively capturing inter-property correlations and by assimilating prior knowledge. This enhanced predictive accuracy positions the combined surrogate models as powerful tools for robust and data-efficient materials design.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Uncertainty Aware Deep Learning for Fault Prediction Using Multivariate Time Series Signals

The superconducting radio-frequency cavities are a crucial component of the Continuous Electron Beam Accelerator Facility (CEBAF) at Jefferson Lab. When a cavity faults, beam delivery to experimental end users is disrupted. Prediction of cavity faults prior to onset is essential to reduce operation and maintenance costs. In this work, a parallel long short-term memory (LSTM)-convolution neural network (CNN)-based deep learning (DL) model is proposed to predict impending faults using pre-fault signals. Further, we introduce an uncertainty quantification approach using Monte Carlo dropout with the LSTM-CNN model to ascertain confidence in the prediction. The model was tested using multivariate time series signals from stable cavity operations and before faults. Initial results show that on the test dataset, the model can identify impending faults before their onset with an average 10-fold cross validation accuracy of 97.39% and a standard deviation of 0.12% using a 100-ms time window. It is also observed that the model performs better as the prediction time moves closer to the fault onset. For additional context, we compare the performance of the model with three machine-learning-based (ML) fault prediction models. Our proposed parallel LSTM-CNN-based DL method shows better performance than the ML-based methods.

Rahman, Md Monibor↗
DOE OSTI2020

Generation and evaluation of synthetic patient data

Background: Machine learning (ML) has made a significant impact in medicine and cancer research; however, its impact in these areas has been undeniably slower and more limited than in other application domains. A major reason for this has been the lack of availability of patient data to the broader ML research community, in large part due to patient privacy protection concerns. High-quality, realistic, synthetic datasets can be leveraged to accelerate methodological developments in medicine. By and large, medical data is high dimensional and often categorical. These characteristics pose multiple modeling challenges. Methods: In this paper, we evaluate three classes of synthetic data generation approaches; probabilistic models, classification-based imputation models, and generative adversarial neural networks. Metrics for evaluating the quality of the generated synthetic datasets are presented and discussed. Results: While the results and discussions are broadly applicable to medical data, for demonstration purposes we generate synthetic datasets for cancer based on the publicly available cancer registry data from the Surveillance Epidemiology and End Results (SEER) program. Specifically, our cohort consists of breast, respiratory, and non-solid cancer cases diagnosed between 2010 and 2015, which includes over 360,000 individual cases. Conclusions: We discuss the trade-offs of the different methods and metrics, providing guidance on considerations for the generation and usage of medical synthetic data.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2023

Ab initio study of the structure and properties of amorphous silicon hydride from accelerated molecular dynamics simulations

This paper presents a large-scale ab initio simulation study of amorphous silicon hydride (a-Si 1-x H x ) with an emphasis on the structure and properties of the material across a range of hydrogen concentration by combining accelerated molecular dynamics (MD) simulations with first-principles density-functional calculations. The accelerated MD scheme relied on classical metadynamics, which enabled the development of 2500+ high-quality structural models of a-Si 1-x H x , with system sizes ranging from 150 to 6000 atoms and hydrogen concentrations vary from 6 to 20 at. %. The resulting amorphous networks were found to be completely free from any coordination defects and that they all exhibited a pristine band-gap in their electronic spectrum. The microstructural properties of hydrogen distributions were examined with an emphasis on the presence of isolated and clustered environments of hydrogen atoms. The results were compared with experimental data obtained from X-ray diffraction, infrared spectroscopy and nuclear magnetic resonance studies.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Intelliquench: An Adaptive Machine Learning System for Detection of Superconducting Magnet Quenches

In superconducting magnets, the irreversible transition of a portion of the conductor to resistive state is called a “quench.” Having large stored energy, magnets can be damaged by quenches due to localized heating, high voltage, or large force transients. Unfortunately, current quench protection systems can only detect a quench after it happens, and mitigating risks in Low Temperature Superconducting (LTS) accelerator magnets often requires fast response (down to ms). Additionally, protection of High Temperature Superconducting (HTS) magnets is still suffering from prohibitively slow quench detection. In this study, we lay the groundwork for a quench prediction system using an auto-encoder fully-connected deep neural network. After dynamically trained with data features extracted from acoustic sensors around the magnet, the system detects anomalous events seconds before the quench in most of our data. While the exact nature of the events is under investigation, we show that the system can “forecast” a quench before it happens under magnet training conditions through a randomized experiment. This opens up the way of integrated data processing, potentially leading to faster and better diagnostics and detection of magnet quenches

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2023

Exploring for Superhot Geothermal Targets in Magmatic Settings: 2022 Field Campaign at Newberry Volcano

This paper presents preliminary results from a subset of work carried out as part of a multinational research project entitled DErisking Exploration for multiple geothermal Plays in magmatic ENvironments (DEEPEN), supported by the U.S. Department of Energy (DOE) and Geothermica, a joint effort by EU member states and associated countries. The DEEPEN project will develop a customized approach to exploration for supercritical and superhot geothermal plays in magmatic systems, which will be applied to two demonstration sites. This paper summarizes field activities carried out at the U.S. demonstration site, Newberry Volcano in central Oregon. The objective of this work effort is to refine the subsurface model of Newberry Volcano, with special focus on deeper zones including the magmatic plumbing system and other key geologic elements. New data collection included gravity and wideband magnetotelluric (MT) surveys, as well as reinstallation of a seismic network. The National Renewable Energy Laboratory (NREL) and Enthalpion Energy LLC (Enthalpion) worked with the Deschutes National Forest Fort Rock District to use a low ground disturbance method of MT deployment to collect MT data inside the caldera and other restricted areas inside the National Volcanic Monument. This opened these areas to geophysical exploration for the first time in decades. Sites along and adjacent to the south rim of the volcano constituted the primary survey objectives. A team from Lawrence Berkeley National Laboratory (LBNL), the U.S. Geological Survey (USGS), and AltaRock also began the process of reinstalling the seismic network from the AltaRock enhanced geothermal system (EGS) demonstration in anticipation of further development activities at the site. The data ingestion, reduction, and analysis phase of the project is ongoing. We are currently processing the MT and gravity data and are developing a new, highly GPU-accelerated, 3D joint MT and gravity inversion to better localize the south rim/south flank conductive target and better understand its relationship to deep heat, fluid sources, and surface extrusive features. Joint inversions, which have not yet been undertaken at Newberry, will allow us to obtain constraints on the geologic model that cannot be determined from each method in isolation, improving our ability to image key geologic features at depth.

geophysics↗
DOE OSTI2026

White-Rabbit-Disciplined FPGA Readout for Fermilab Timing Events

Fermilab's accelerator timing links broadcast short event codes to thousands of devices at once, but the links themselves carry no absolute notion of time; that comes separately from a White Rabbit reference. This work builds the piece that ties the two together on a single board. On a Xilinx Kria KR260 (Zynq UltraScale+), the programmable logic decodes a real Fermilab TCLK link, stamps every event with an absolute White-Rabbit \{sec, ns\} UTC time, and reads the timestamped stream out over AXI4-Lite; a thin Linux process on the same die publishes each event into a Redis stream on the control network. To exercise the full chain on one board, the decoded events are re-encoded as gigabit ACLK, transmitted out an SFP+ optical port, looped back over a short fiber jumper, and decoded again on the same timeline, and are additionally mirrored as an ACLK-Lite Manchester waveform for benchtop probing. Across sustained, multi-day testing against real Fermilab TCLK, the pipeline has decoded, timestamped, and published hundreds of millions of events with practically zero loss, and folding the timestamped stream on the 60-second accelerator supercycle recovers the machine's periodic structure directly from the published data.

Rossel, Jacob [Fermilab; UC, Berkeley (main)] (ORC↗
DOE OSTI2026

A Single-Board Fermilab Timing Pipeline on the Xilinx KR260: Decode, Timestamp, Publish, Mirror

Full-width abstract \renewcommand{\maketitlehookd}{% \begin{abstract} \noindent Fermilab's accelerator timing links broadcast short event codes to thousands of devices at once, but the links themselves carry no absolute notion of time; that comes separately from a White Rabbit reference. This work builds the piece that ties the two together on a single board. On a Xilinx Kria KR260 (Zynq UltraScale+), the programmable logic decodes a real Fermilab TCLK link, stamps every event with an absolute White-Rabbit \{sec, ns\} UTC time, and reads the timestamped stream out over AXI4-Lite; a thin Linux process on the same die publishes each event into a Redis stream on the control network. To exercise the full chain on one board, the decoded events are re-encoded as gigabit ACLK, transmitted out an SFP+ optical port, looped back over a short fiber jumper, and decoded again on the same timeline, and are additionally mirrored as an ACLK-Lite Manchester waveform for benchtop probing. Across sustained, multi-day testing against real Fermilab TCLK, the pipeline has decoded, timestamped, and published hundreds of millions of events with practically zero loss, and folding the timestamped stream on the 60-second accelerator supercycle recovers the machine's periodic structure directly from the published data.

Rossel, Jacob [UC, Berkeley; Fermilab] (ORCID:0009↗
DOE OSTI2025

Artificial to Spiking Neural Networks Conversion with Calibration in Scientific Machine Learning

Here, we introduce a method to convert physics-informed neural networks (PINNs), commonly used in scientific machine learning, to spiking neural networks (SNNs), which are expected to have higher energy efficiency compared to traditional artificial neural networks (ANNs). We first extend the calibration technique of SNNs to arbitrary activation functions beyond ReLU, making it more versatile, and we prove a theorem that ensures the effectiveness of the calibration. We successfully convert PINNs to SNNs, enabling computational efficiency for diverse regression tasks in solving multiple differential equations, including the unsteady Navier–Stokes equations. We demonstrate great gains in terms of overall efficiency, including separable PINNs (SPINNs), which accelerate the training process. Overall, this is the first work of this kind and the proposed method achieves relatively good accuracy with low spike rates.

PINN↗
DOE OSTI2025

Science acceleration and accessibility with self-driving labs

In the evolving landscape of scientific research, the complexity of global challenges demands innovative approaches to experimental planning and execution. Self-Driving Laboratories (SDLs) automate experimental tasks in chemical and materials sciences and the design and selection of experiments to optimize research processes and reduce material usage. This perspective explores improving access to SDLs via centralized facilities and distributed networks. We discuss the technical and collaborative challenges in realizing SDLs’ potential to enhance human–machine and human–human collaboration, ultimately fostering a more inclusive research community and facilitating previously untenable research projects.

Canty, Richard B. [North Carolina State University↗
DOE OSTI2021

Curating a comprehensive set of enzymatic reaction rules for efficient novel biosynthetic pathway design

Enzyme substrate promiscuity has significant implications for metabolic engineering. The ability to predict the space of possible enzymatic side reactions is crucial for elucidating underground metabolic networks in microorganisms, as well as harnessing novel biosynthetic capabilities of enzymes to produce desired chemicals. Reaction rule-based cheminformatics platforms have been implemented to computationally enumerate possible promiscuous reactions, relying on existing knowledge of enzymatic transformations to inform novel reactions. However, past versions of curated reaction rules have been limited by a lack of comprehensiveness in representing all possible transformations, as well as the need to prune rules to enhance computational efficiency in pathway expansion. To this end, we curated a set of 1224 most generalized reaction rules, automatically abstracted from atom-mapped MetaCyc reactions and verified to uniquely cover all common enzymatic transformations. Additionally, we developed a framework to systematically identify and correct redundancies and errors in the curation process, resulting in a minimal, yet comprehensive, rule set. These reaction rules were capable of reproducing more than 85% of all reactions in the KEGG and BRENDA databases, for which a large fraction of reactions is not present in MetaCyc. Our rules exceed all previously published rule sets for which reproduction was possible in this coverage analysis, which allows for the exploration of a larger space of known enzymatic transformations. By leveraging the entire knowledge of possible metabolic reactions through generalized enzymatic reaction rules, we are able to better utilize underground metabolic pathways and accelerate novel biosynthetic pathway design to enable bioproduction towards a wider range of new molecules.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2023

Optimizing Data Movement for GPU-Based In-Situ Workflow Using GPUDirect RDMA

The extreme-scale computing landscape is increasingly dominated by GPU-accelerated systems. At the same time, in-situ workflows that employ memory-to-memory inter-application data exchanges have emerged as an effective approach for leveraging these extreme-scale systems. In the case of GPUs, GPUDirect RDMA enables third-party devices, such as network interface cards, to access GPU memory directly and has been adopted for intra-application communications across GPUs. In this paper, we present an interoperable framework for GPU-based in-situ workflows that optimizes data movement using GPUDirect RDMA. Specifically, we analyze the characteristics of the possible data movement pathways between GPUs from an in-situ workflow perspective, and design a strategy that maximizes throughput. Furthermore, we implement this approach as an extension of the DataSpaces data staging service, and experimentally evaluate its performance and scalability on a current leadership GPU cluster. The performance results show that the proposed design reduces data-movement time by up to 53% and 40% for the sender and receiver, respectively, and maintains excellent scalability for up to 256 GPUs.

Zhang, Bo↗
DOE OSTI2022

A Julia Framework for Graph-Structured Nonlinear Optimization

Graph theory provides a convenient framework for modeling and solving structured optimization problems. Under this framework, the modeler can arrange/assemble the components of an optimization model (variables, constraints, objective functions, and data) within nodes and edges of a graph, and this representation can be used to visualize, manipulate, and solve the problem. In this work, we present a Julia framework for modeling and solving graph-structured nonlinear optimization problems. Our framework integrates the modeling package Plasmo.jl (which facilitates the construction and manipulation of graph models) and the nonlinear optimization solver MadNLP.jl (which provides capabilities for exploiting graph structures to accelerate solution). We illustrate with a simple example how model construction and manipulation can be performed in an intuitive manner using Plasmo.jl and how the model structure can be exploited by MadNLP.jl. We also demonstrate the scalability of the framework by targeting a large-scale, stochastic gas network problem that contains over 1.7 million variables.

Cole, David↗