Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “learning efficiency”

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 1,657 records · Page 92

Fractal Dimension Calculation for Big Data Using Box Locality Index

The box-counting approach for fractal dimension calculation is scaled up for big data using a data structure named box locality index (BLI). The BLI is constructed as key-value pair style with the key (an integer array) indexing the location of a “box” (i.e., a gird cell on the multi-dimensional space embedding a given dataset) and the value counting the number of data points inside the box (called “box occupancy”). Compared with the traditionally used E-dim tree, the BLI avoids complex hierarchical structure and encodes only necessary information required by box-counting approach for fractal dimension calculation. Moreover, the key-value pair nature of BLI, together with the fact that the box occupancy is aggregatable, grants box-counting approach the needed scalability for fractal dimension calculation of big data using distributed computing techniques (e.g., MapReduce and Spark). Taking the advantage of the BLI, MapReduce and Spark methods for fractal dimension calculation of big data are developed, which conduct box-counting for each grid level as a cascade of MapReduce/Spark jobs in a bottom-up fashion. In an empirical validation, the MapReduce and Spark methods demonstrated good effectiveness and efficiency in fractal calculation of a big synthetic dataset. In summary, this work provides an efficient solution for the estimation of intrinsic dimension of big data that is essential for many machine learning methods and data analytics.

Fractal dimension, Intrinsic dimension, Box-counti↗

Partners for life: building microbial consortia for the future

We report new technologies have allowed researchers to better design, build, and analyze complex consortia. These developments are fueling a wider implementation of consortium-based bioprocessing by leveraging synthetic biology, delivering on the field’s multitudinous promises of higher efficiencies, superior resiliency, augmented capabilities, and modular bioprocessing. Here we chronicle current progress by presenting a range of screening, computational, and biomolecular tools enabling robust population control, efficient division of labor, and programmatic spatial organization; furthermore, we detail corresponding advancements in areas including machine learning, biocontainment, and standardization. Additionally, we show applications in myriad sectors, including medicine, energy and waste sustainability, chemical production, agriculture, and biosensors. Concluding remarks outline areas of growth that will promote the utilization of complex community structures across the biotechnology spectrum.

59 BASIC BIOLOGICAL SCIENCES↗

Neural operators for stochastic modeling of nonlinear structural system response to natural hazards

Traditionally, neural networks have been employed to learn the mapping between finite-dimensional Euclidean spaces. However, recent research has opened up new horizons, focusing on the utilization of deep neural networks to learn operators capable of mapping infinite-dimensional function spaces. Here, in this work, we employ two state-of-the-art neural operators, the deep operator network (DeepONet) and the Fourier neural operator (FNO) for the prediction of the nonlinear time history response of structural systems exposed to natural hazards, such as earthquakes and windstorms. Specifically, we propose two architectures, a self-adaptive FNO and a fast Fourier transform-based DeepONet (DeepFNOnet), where we employ a FNO beyond the DeepONet to learn the discrepancy between the ground truth and the solution predicted by the DeepONet. To demonstrate the efficiency and applicability of the architectures, two problems are considered. In the first, we use the proposed model to predict the seismic nonlinear dynamic response of a six-story shear building subject to stochastic ground motions. In the second problem, we employ the operators to predict the wind-induced nonlinear dynamic response of a high-rise building while explicitly accounting for the stochastic nature of the wind excitation. In both cases, the trained metamodels achieve high accuracy while being orders of magnitude faster than their corresponding high-fidelity models.

DeepONet↗

Learning-Based Demand Response in Grid Interactive Buildings via Gaussian Processes

This paper presents a predictive controller for a grid-interactive multi-zone building where the temperature dynamics are learned via Gaussian Process (GP) regression. We investigate the development of a learning-based predictive control with two main objectives: (i) continuously learn the temperature dynamics of the building based on data; and, (ii) use the learned dynamics to solve a multi-objective predictive control problem to guarantee occupants' comfort and energy efficiency during normal conditions and demand response events. We leverage the probabilistic non-parametric properties of GPs to estimate the (unknown) non-linear temperature dynamics of the building and to incorporate the uncertainty of those predictions in a multi-objective optimization problem. The GP-based predictive control is solved via a zero-order primal-dual projected-gradient algorithm. We evaluate numerically the performance of the proposed controller using a five-zone commercial building.

demand response↗

Simplifying activations with linear approximations in neural networks

A key step in Neural Networks is activation. Among the different types of activation functions, sigmoid, tanh, and others involve the usage of exponents for calculation. From a hardware perspective, exponential implementation implies the usage of Taylor series or repeated methods involving many addition, multiplication, and division steps, and as a result are power-hungry and consume many clock cycles. We implement a piecewise linear approximation of the sigmoid function as a replacement for standard sigmoid activation libraries. This approach provides a practical alternative by leveraging piecewise segmentation, which simplifies hardware implementation and improves computational efficiency. In this paper, we detail piecewise functions that can be implemented using linear approximations and their implications for overall model accuracy and performance gain. Our results show that for the DenseNet, ResNet, and GoogLeNet architectures, the piecewise linear approximation of the sigmoid function provides faster execution times compared to the standard TensorFlow sigmoid implementation while maintaining comparable accuracy. Specifically, for MNIST with DenseNet, accuracy reaches 99.91% (Piecewise) vs. 99.97% (Base) with up to 1.31x speedup in execution time. For CIFAR-10 with DenseNet, accuracy improves to 98.97% (Piecewise) vs. 99.40% (Base) while achieving 1.24x faster execution. Similarly, for CIFAR-100 with DenseNet, the accuracy is 97.93% (Piecewise) vs. 98.39% (Base), with a 1.18x execution time reduction. These results confirm the proposed method’s capability to efficiently process large-scale datasets and computationally demanding tasks, offering a practical means to accelerate deep learning models, including LSTMs, without compromising accuracy.

Activation function↗

Autonomous Synthesis and Inverse Design of Electrochromic Polymers with High Efficiency and Accuracy

Here, the design and synthesis of functional polymers, aimed at targeted properties through specific structures, have long been challenged by their complex and often nonlinear structure–property relationships. Key processes, including knowledge accumulation for predictive design and experimental refinement and validation, are traditionally labor-insensitive and time-consuming, making it difficult to balance accuracy and efficiency. Here, we introduce an accelerated, autonomous system for the on-demand synthesis of electronic polymers that achieves the desired electrochromic functionality with high accuracy and efficiency. Our approach leverages large language model-assisted data mining, a physics-informed copolymer machine learning model, and an AI-driven autonomous robotic workflow in the Polybot lab. Within 72 h, Polybot autonomously synthesized electrochromic polymers (ECPs) with targeted, previously-unreported color values, including green polymers with specific absorption profiles, precisely fine-tuning copolymer structures with a 5% step size in comonomer composition within a three-monomer system. A publicly accessible ECP informatics database has also been created to foster knowledge exchange.

AI-driven Robotic Lab↗

Advancing electrocatalytic nitrogen fixation: insights from molecular systems

Nitrogen fixation has a rich history within the inorganic chemistry community. In recent years attention has (re)focused on developing electrocatalytic systems capable of mediating the nitrogen reduction reaction (N 2 RR). Well-defined molecular catalyst systems have much to offer in this context. This personal perspective summarizes recent progress from our laboratory at Caltech, pulling together lessons learned from a number of studies we have conducted, placing them within the broader context of thermodynamic efficiency and selectivity for the N 2 RR. Here, in particular, proton-coupled electron transfer (PCET) provides an attractive strategy to achieve enhanced efficiency for the multi-electron/proton reduction of N 2 to produce NH 3 (or NH 4 + ), and electrocatalytic PCET (ePCET) via an ePCET mediator affords a promising means of mitigating HER such that the N 2 RR can be achieved in a catalytic fashion.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Adaptive Client Selection in Federated Learning: A Network Anomaly Detection Use Case

Federated Learning (FL) has become a ubiquitous approach for training machine learning models on decentralized data, addressing the myriad privacy concerns inherent in traditional centralized methods. However, the efficiency of FL depends on effective client selection and robust privacy preservation mechanisms. Inadequate client selection may lead to suboptimal model performance, while insufficient privacy measures risk exposing sensitive data. This paper proposes a client selection framework for FL that integrates differential privacy and fault tolerance. Our adaptive approach dynamically adjusts the number of selected clients based on model performance and system constraints, ensuring privacy through calibrated noise addition. We evaluate our method on a network anomaly detection use case using the UNSW-NB15 and ROAD datasets. Results show up to a 7% increase in accuracy and a 25% reduction in training time compared to FedL2P. Moreover, we highlight the trade-offs between privacy budgets and model performance, with higher privacy budgets reducing noise and improving accuracy. Our fault tolerance mechanism, while causing a slight performance drop, enhances robustness to client failures. Statistical validation using Mann-Whitney U tests confirms the significance of these improvements (p < 0.05).

Marfo, William [University of Texas at El Paso,Dep↗

Modular Spiking Neural Circuits for Mapping Long Short-Term Memory on a Neurosynaptic Processor

Due to the distributed and asynchronous nature of neural computation through low-energy spikes, brain-inspired hardware systems offer high energy efficiency and massive parallelism. One such platform is the IBM TrueNorth neurosynaptic system. Recently, TrueNorth compatible representation learning algorithms have emerged, achieving close to the state-of-the-art performance in various data sets. However, its application in temporal sequence processing models, such as recurrent neural networks (RNNs), is still only at the proof of concept level. There is an inherent difficulty in capturing temporal dynamics of an RNN using spiking neurons, which is only exasperated by the hardware constraints in connectivity and synaptic weight resolution. This paper presents a design flow that overcomes these difficulties and maps a special case of recurrent networks called long short-term memory (LSTM) onto a spike-based platform. The framework utilizes various approximation techniques, such as activation discretization, weight quantization, and scaling and rounding, spiking neural circuits that implement the complex gating mechanisms, and a store-and-release technique to enable neuron synchronization and faithful storage. While the presented techniques can be applied to map LSTM to any spiking neural network (SNN) simulator/emulator, here we choose the TrueNorth chip as the target platform by adhering to its hardware constraints. Three LSTM applications, parity check, extended Reber grammar, and question classification, are evaluated. The tradeoffs among accuracy, performance, and energy tradeoffs achieved on TrueNorth are demonstrated. Finally, this is compared with the performance on an SNN platform without hardware constraints, which represents the upper bound of the achievable accuracy.

97 MATHEMATICS AND COMPUTING↗

Data-Driven Optimization of Pixelated CdZnTe Spectrometers for Uranium Enrichment Assay

Here, in recent work [Vavrek et al. (2025)], we developed the performance optimization framework spectre-ml for gamma spectrometers with variable performance across many readout channels. The framework uses non-negative matrix factorization (NMF) and clustering to learn groups of similarly-performing channels and sweep through various learned channel combinations to optimize the performance tradeoff of including worse-performing channels for better total efficiency. In this work, we integrate the pyGEM uranium enrichment assay code with our spectre-ml framework, and show that the U-235 enrichment relative uncertainty can be directly used as an optimization target. We find that this optimization reduces relative uncertainties after a 30 -minute measurement by an average of 20%, as tested on six different H3D M400 CdZnTe spectrometers, which can significantly improve uranium non-destructive assay measurement times in nuclear safeguards contexts. Additionally, this work demonstrates that the spect re-ml optimization framework can accommodate arbitrary end-user spectroscopic analysis code and performance metrics, enabling future optimizations for complex Pu spectra.

Gamma-ray detection↗

Rapid Electrochemical Diagnosis of Battery Health and Safety from Cells to Modules

Rapid electrochemical diagnosis of battery health and failure is critical for ensuring reliable battery performance and battery safety. Traditional battery health diagnostics such as capacity measurements and DC pulse tests are reliable and well-understood, however, these measurements of battery capacity and resistance do not capture all aspects of battery degradation. Other aspects of degradation, such as electrolyte decomposition, lithium-plating, and particle cracking are difficult to detect electrochemically but are crucial to measure to get a full picture of battery safety and flag out potential failures. In this work, lab- and field-aged commercial lithium-ion batteries and modules of various chemistries and formats are tested using a variety of traditional electrochemical characterization methods as well as using 2-minute pseudo-random DC pulse sequences at rest and during charge/discharge. The electrochemical measurements are compared to physical cell measurements, cell efficiency, drive cycle performance, physical and thermal heterogeneity, and qualitative safety metrics using statistical and machine-learning methods to discover if a comprehensive "battery health map" can be accurately identified using only rapid DC measurements.

ADVANCED PROPULSION SYSTEMS,ENERGY STORAGE↗

Subsurface Characterization of Hydraulic Fracture Test Site-2 (HFTS-2), Delaware Basin

Hydraulic Fracturing Test Site-2 (HFTS-2) is a field-based research experiment performed in the Wolfcamp Formation of the Permian (Delaware) Basin. This paper focuses on integration, advanced geological characterization, and 3D subsurface modeling of the comprehensive HFTS-2 dataset. The study showcases a multidisciplinary reservoir characterization approach that incorporates geology, petrophysics, geochemistry, geomechanics, microseismic, and subsurface engineering analysis. Subsurface characterization of organic-rich mudstone formations requires understanding complex hydraulic fracture network growth in relation to inherent lithology, geomechanical properties, and interaction with pre-existing natural fractures. This paper presents a characterization workflow incorporating pre- and post-stimulation subsurface data, unique to the HFTS-2 dataset. The study integrated: (1) rock properties from logs, cores, and thin sections; (2) natural and hydraulic fracture descriptions from cores and image logs; (3) local and regional stresses; (4) geomechanics; (5) microseismic; (6) fiber optic (FO) and bottomhole pressure gauge (BHPG) response; and (7) produced fluids analysis. During a stimulation treatment, creation of the stimulated rock volume (SRV) is influenced by several subsurface factors. Key contributing factors include structural context, stress conditions, lithology, facies architecture, pre-existing natural fractures, and geomechanical properties. The HFTS-2 subsurface data integration indicates that the SRV is comprised of a complex juxtaposition of hydraulic fracture swarms, as evidenced by image logs analysis, core description, and microseismic monitoring. The HFTS-2 microseismic event density was used to generate 3D heat maps that serve as a representative SRV footprint, corroborated by secondary datasets. These maps were further integrated with petrophysical and geomechanical characteristics, as well as responses from FO and BHPG, to estimate the lateral and vertical dimensions of the effective fractures. The geological characterization for the HFTS-2 dataset combined with 3D modeling for petrophysical and geomechanical properties provides a strong foundation for subsurface simulation and optimization studies. Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/21URTC/1-21URTC/D011S005R001/2477501/urtec-2021-5243-ms.pdf/1 by Carol Worster on 28 February 2022 URTeC 5243 The workflow improved our understanding of HFTS-2 hydraulic fracture propagation and characteristics in relation to offset pressure depletion and interaction with pre-existing natural fractures. Analysis showed that fracture geometry varies by stage and by well, and a complex fracture network is generated with varying fracture density. The multidisciplinary workflow presented herein for integration and characterization serves as a foundation to evaluate completion efficiency and estimate areal and vertical stimulation and depletion extent for the project. Furthermore, the workflow and learnings can also be transferred to other unconventional plays.

58 GEOSCIENCES↗

Multi-Physics Simulations for Molten Salt Reactor Evaluation: Chemistry Modeling and Database Development

To aid in design and licensing of molten salt reactors, a framework integrating the complex interaction of reactor neutronics, thermal hydraulics, and chemistry is being developed within the Department of Energy Advanced Reactor Technology Program’s Molten Salt Reactor (MSR) campaign. The challenges of integrating thermochemical and thermophysical behavior into a multi-physics reactor simulations include the following: (1) population of data needed for refinement of current models and development of nonexistent models through experimental measurements, first principles calculations, and development of a machine learning approach (2) thermochemical and thermophysical model development, (3) further development of Thermochimica, an open-source efficient equilibrium solver used to link thermochemical models to the multi-physics code, (4) a framework for integrating kinetic phenomena: nucleation, precipitation, mass/heat transport, and corrosion models, and (5) a computational environment to efficiently utilize the data and models within a multi-physics modeling tool. These challenges are being addressed through a collaboration among Oak Ridge National Laboratory, Argonne National Laboratory, the University of South Carolina, and the University of Ontario Institute of Technology.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

AOI.1 Application of Artificial Intelligence techniques enabling coal fired power plants the ability to achieve higher efficiency, improved availability, and increased reliability of their operations (Final Report)

During this effort, SparkCognition with support from the Electric Power Research Institute (EPRI) was tasked with applying artificial intelligence (AI) to improve the reliability, efficiency, and safety of operations at a coal-fired plant. By implementing AI techniques, like machine learning (ML), it is believed that operators can leverage existing data sources to gain more insights such as advanced warning of machine degradation. With enough lead time, a reliability engineer can take action to minimize, or even avoid, impact to production. To complete this work effort, SparkCognition developed and refined an ML-based model using sensor data for a Steam Turbine unit at a host site. The models were deployed in an online, web-based solution that allows users to visualize model outputs and supporting data. The final solution, based on SparkCognition’s proprietary software platform called SparkPredict®, was shared with EPRI who completed an online evaluation of results to determine the solution’s ability to detect actionable events.

20 FOSSIL-FUELED POWER PLANTS↗

SULI Research Report Paper

The Intelligence and Space Research Group 4 (ISR-4) at the Los Alamos National Laboratory focuses on signal processing and developing space electronics for various national security missions. The Space and Atmospheric Burst Reporting System (SABRS) is a sensing payload designed for nuclear detonation detection. As the sensors from previous payloads become obsolete, the requirement for a novel payload design increases. SABRS Prime is the successor to the previous SABRS payloads and is scheduled for delivery in 2028. As part of the SABRS Prompt Gamma Instrument Interface Board, high frequency signals from silicon photomultiplier tubes need to be split and subsequently transmitted to the Analog to Digital Converter. Prior to developing the Space Grade Instrument Interface Board, evaluation modules with Commercial off-the-shelf (COTS) components are evaluated and tested under rigorous conditions. Verification of power levels and gain specifications are some of the important steps taken during the evaluation stage. Working on the Signal Splitting Board has enabled me to work with various schematic software applications and learn more about the processes behind designing space grade electronic components. In addition, I conducted research on efficiency improvements for space-based power systems. Changes in design approaches were found to be the most significant sources of improvement. Being able to interact with researchers and peers across various scientific disciplines has given me a better insight into my future pursuits. Working in a R&D environment has cemented my decision to work on space applications for Quantum Sensing and pursue a PhD in Electrical Engineering. The SULI internship program at the Los Alamos National Laboratory has helped me develop the skillset of a researcher, understand the rigorous processes behind developing space grade hardware, and improve my critical thinking abilities to support the development of reliable instrumentation for deep space missions.

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEF↗

Modernizing Accelerator Responsiveness and Controls in Operations

Accelerators increasingly use artificial intelligence (AI) and machine learning (ML) software and workflows for a variety of tasks, from optimization to fault detection and recovery. Efficient and sustainable application of these technologies necessitates specialized and facility-specific infrastructure commitments. Accelerator facilities also introduce unique radiation and security hazards, placing additional demands on operational infrastructure. These needs further escalate the prioritization of effective collaboration models and associated funding mechanisms and legal frameworks.

43 PARTICLE ACCELERATORS↗

Latent Stochastic Differential Equations for Modeling Quasar Variability and Inferring Black Hole Properties

Quasars are bright and unobscured active galactic nuclei (AGN) thought to be powered by the accretion of matter around supermassive black holes at the centers of galaxies. The temporal variability of a quasar’s brightness contains valuable information about its physical properties. The UV/optical variability is thought to be a stochastic process, often represented as a damped random walk described by a stochastic differential equation (SDE). Upcoming wide-field telescopes such as the Rubin Observatory Legacy Survey of Space and Time (LSST) are expected to observe tens of millions of AGN in multiple filters over a ten year period, so there is a need for efficient and automated modeling techniques that can handle the large volume of data. Latent SDEs are machine learning models well suited for modeling quasar variability, as they can explicitly capture the underlying stochastic dynamics. In this work, we adapt latent SDEs to jointly reconstruct multivariate quasar light curves and infer their physical properties such as the black hole mass, inclination angle, and temperature slope. Our model is trained on realistic simulations of LSST ten year quasar light curves, and we demonstrate its ability to reconstruct quasar light curves even in the presence of long seasonal gaps and irregular sampling across different bands, outperforming a multioutput Gaussian process regression baseline. Our method has the potential to provide a deeper understanding of the physical properties of quasars and is applicable to a wide range of other multivariate time series with missing data and irregular sampling.

79 ASTRONOMY AND ASTROPHYSICS↗

Automated Proton Track Identification in MicroBooNE Using Gradient Boosted Decision Trees

MicroBooNE is a liquid argon time projection chamber (LArTPC) neutrino experiment that is currently running in the Booster Neutrino Beam at Fermilab. LArTPC technology allows for high-resolution, three-dimensional representations of neutrino interactions. A wide variety of software tools for automated reconstruction and selection of particle tracks in LArTPCs are actively being developed. Short, isolated proton tracks, the signal for low- momentum-transfer neutral current (NC) elastic events, are easily hidden in a large cosmic background. Detecting these low-energy tracks will allow us to probe interesting regions of the proton's spin structure. An effective method for selecting NC elastic events is to combine a highly efficient track reconstruction algorithm to find all candidate tracks with highly accurate particle identification using a machine learning algorithm. We present our work on particle track classification using gradient tree boosting software (XGBoost) and the performance on simulated neutrino data.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗