Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “network performance evaluation”

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 1,369 records · Page 76
All sourcesNASA NTRSDOE OSTI
NASA NTRS2016

Soil Moisture Active Passive Mission L4_SM Data Product Assessment (Version 2 Validated Release)

During the post-launch SMAP calibration and validation (Cal/Val) phase there are two objectives for each science data product team: 1) calibrate, verify, and improve the performance of the science algorithm, and 2) validate the accuracy of the science data product as specified in the science requirements and according to the Cal/Val schedule. This report provides an assessment of the SMAP Level 4 Surface and Root Zone Soil Moisture Passive (L4_SM) product specifically for the product's public Version 2 validated release scheduled for 29 April 2016. The assessment of the Version 2 L4_SM data product includes comparisons of SMAP L4_SM soil moisture estimates with in situ soil moisture observations from core validation sites and sparse networks. The assessment further includes a global evaluation of the internal diagnostics from the ensemble-based data assimilation system that is used to generate the L4_SM product. This evaluation focuses on the statistics of the observation-minus-forecast (O-F) residuals and the analysis increments. Together, the core validation site comparisons and the statistics of the assimilation diagnostics are considered primary validation methodologies for the L4_SM product. Comparisons against in situ measurements from regional-scale sparse networks are considered a secondary validation methodology because such in situ measurements are subject to up-scaling errors from the point-scale to the grid cell scale of the data product. Based on the limited set of core validation sites, the wide geographic range of the sparse network sites, and the global assessment of the assimilation diagnostics, the assessment presented here meets the criteria established by the Committee on Earth Observing Satellites for Stage 2 validation and supports the validated release of the data. An analysis of the time average surface and root zone soil moisture shows that the global pattern of arid and humid regions are captured by the L4_SM estimates. Results from the core validation site comparisons indicate that "Version 2" of the L4_SM data product meets the self-imposed L4_SM accuracy requirement, which is formulated in terms of the ubRMSE: the RMSE (Root Mean Square Error) after removal of the long-term mean difference. The overall ubRMSE of the 3-hourly L4_SM surface soil moisture at the 9 km scale is 0.035 cubic meters per cubic meter requirement. The corresponding ubRMSE for L4_SM root zone soil moisture is 0.024 cubic meters per cubic meter requirement. Both of these metrics are comfortably below the 0.04 cubic meters per cubic meter requirement. The L4_SM estimates are an improvement over estimates from a model-only SMAP Nature Run version 4 (NRv4), which demonstrates the beneficial impact of the SMAP brightness temperature data. L4_SM surface soil moisture estimates are consistently more skillful than NRv4 estimates, although not by a statistically significant margin. The lack of statistical significance is not surprising given the limited data record available to date. Root zone soil moisture estimates from L4_SM and NRv4 have similar skill. Results from comparisons of the L4_SM product to in situ measurements from nearly 400 sparse network sites corroborate the core validation site results. The instantaneous soil moisture and soil temperature analysis increments are within a reasonable range and result in spatially smooth soil moisture analyses. The O-F residuals exhibit only small biases on the order of 1-3 degrees Kelvin between the (re-scaled) SMAP brightness temperature observations and the L4_SM model forecast, which indicates that the assimilation system is largely unbiased. The spatially averaged time series standard deviation of the O-F residuals is 5.9 degrees Kelvin, which reduces to 4.0 degrees Kelvin for the observation-minus-analysis (O-A) residuals, reflecting the impact of the SMAP observations on the L4_SM system. Averaged globally, the time series standard deviation of the normalized O-F residuals is close to unity, which would suggest that the magnitude of the modeled errors approximately reflects that of the actual errors. The assessment report also notes several limitations of the "Version 2" L4_SM data product and science algorithm calibration that will be addressed in future releases. Regionally, the time series standard deviation of the normalized O-F residuals deviates considerably from unity, which indicates that the L4_SM assimilation algorithm either over- or under-estimates the actual errors that are present in the system. Planned improvements include revised land model parameters, revised error parameters for the land model and the assimilated SMAP observations, and revised surface meteorological forcing data for the operational period and underlying climatological data. Moreover, a refined analysis of the impact of SMAP observations will be facilitated by the construction of additional variants of the model-only reference data. Nevertheless, the “Version 2” validated release of the L4_SM product is sufficiently mature and of adequate quality for distribution to and use by the larger science and application communities.

SMAP L4_SM↗
DOE OSTI2021

MIRACL Co‐Simulation platform for control and operation of distributed wind in microgrid

Abstract This paper presents a co‐simulation platform (CSP) developed as a part of the “Microgrids, Infrastructure Resilience and Advanced Controls Launchpad (MIRACL)” project, hereafter called MIRACL‐CSP, to allow simulation‐based testing, demonstration, and evaluations of distributed wind under various grid operating conditions. MIRACL‐CSP provides modular interactions among the power distribution system, distributed wind, and utility decision‐making framework. A hierarchical engine for large‐scale infrastructure co‐simulations (HELICS) is used as the core engine of MIRACL‐CSP to establish time and information coordination among the MIRACL‐CSP modules. In this study, MIRACL‐CSP performance is demonstrated using the IEEE 123‐node test distribution grid modelled in GridLAB‐D and the utility decision support application modelled in Python. The functionality of MIRACL‐CSP is demonstrated through various grid operational scenarios in a microgrid and networked microgrid environment. A co‐simulation study is described that demonstrated MIRACL‐CSP capabilities for the microgrid operational scenario in the IEEE 123‐node test system. MIRACL‐CSP is a generic platform that facilitates distributed wind research for different test systems, applications, and valuation of distributed wind.

17 WIND ENERGY↗
DOE OSTI2022

Fault Injection for TensorFlow Applications

As machine learning (ML) has seen increasing adoption in safety-critical domains (e.g., autonomous vehicles), the reliability of ML systems has also grown in importance. While prior studies have proposed techniques to enable efficient error-resilience (e.g., selective instruction duplication), a fundamental requirement for realizing these techniques is a detailed understanding of the application’s resilience. In this work, we present TensorFI 1 and TensorFI 2, high-level fault injection (FI) frameworks for TensorFlow-based applications. TensorFI 1 and 2 are able to inject both hardware and software faults in any general TensorFlow 1 and 2 program respectively. Both are configurable FI tools that are flexible, easy to use, and portable. They can be integrated into existing TensorFlow programs to assess their resilience for different fault types (e.g., bit-flips in particular operations or layers). We use the TensorFI 1 and TensorFI 2 to evaluate the resilience of 12 and 10 ML programs written in TensorFlow, including DNNs used in the autonomous vehicle domain. The results give us insights into why some of the models are more resilient. We also measure the performance overheads of the two injectors, and present 4 case studies, two for each tool, to demonstrate their utility.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Full‐Waveform Simulation of Infrasound Propagation in the Atmosphere: A Case Study of the 2023 April 20 SpaceX Starship Explosion

Infrasound, low‐frequency sound below 20 Hz, has been a key technology to monitor explosion events in the atmosphere. The International Monitoring System (IMS) of the Comprehensive Nuclear‐Test‐Ban Treaty Organization provides the means for continuous monitoring of infrasonic events worldwide. Infrasonic techniques for event location and size estimation can also complement other observational techniques for the detection and characterization of the entry of asteroids or large meteoroids. In this study, we describe the detection capability of IMS infrasound stations for an explosive event in the middle of the atmosphere. Full‐waveform simulations are performed with the specification of atmospheric conditions and incorporated into the event location and explosion yield estimation. We applied it to the 2023 April 20 SpaceX Starship explosion at 29 km altitude. Starship is a super heavy‐lift space vehicle constructed by SpaceX and known as the largest and most powerful rocket ever built. The Starship explosion created huge pressure disturbances in the atmosphere, and its infrasound was detected by the IMS arrays in North America between 2000 and 4000 km. Independent observational data and available ground‐truth information provide a rare opportunity to evaluate the monitoring capability of the IMS network for elevated sources in the atmosphere. We also demonstrate the capability of full‐waveform simulation for infrasound wavefield characterization and prediction to improve event location and yield estimation.

Geosciences↗
DOE OSTI2023

Design and implementation of I/O performance prediction scheme on HPC systems through large-scale log analysis

Abstract Large-scale high performance computing (HPC) systems typically consist of many thousands of CPUs and storage units used by hundreds to thousands of users simultaneously. Applications from large numbers of users have diverse characteristics, such as varying computation, communication, memory, and I/O intensity. A good understanding of the performance characteristics of each user application is important for job scheduling and resource provisioning. Among these performance characteristics, I/O performance is becoming increasingly important as data sizes rapidly increase and large-scale applications, such as simulation and model training, are widely adopted. However, predicting I/O performance is difficult because I/O systems are shared among all users and involve many layers of software and hardware stack, including the application, network interconnect, operating system, file system, and storage devices. Furthermore, updates to these layers and changes in system management policy can significantly alter the I/O behavior of applications and the entire system. To improve the prediction of the I/O performance on HPC systems, we propose integrating information from several different system logs and developing a regression-based approach to predict the I/O performance. Our proposed scheme can dynamically select the most relevant features from the log entries using various feature selection algorithms and scoring functions, and can automatically select the regression algorithm with the best accuracy for the prediction task. The evaluation results show that our proposed scheme can predict the write performance with up to 90% prediction accuracy and the read performance with up to 99% prediction accuracy using the real logs from the Cori supercomputer system at NERSC.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1992

Description of the SSF PMAD DC testbed control system data acquisition function

The NASA LeRC in Cleveland, Ohio has completed the development and integration of a Power Management and Distribution (PMAD) DC Testbed. This testbed is a reduced scale representation of the end to end, sources to loads, Space Station Freedom Electrical Power System (SSF EPS). This unique facility is being used to demonstrate DC power generation and distribution, power management and control, and system operation techniques considered to be prime candidates for the Space Station Freedom. A key capability of the testbed is its ability to be configured to address system level issues in support of critical SSF program design milestones. Electrical power system control and operation issues like source control, source regulation, system fault protection, end-to-end system stability, health monitoring, resource allocation, and resource management are being evaluated in the testbed. The SSF EPS control functional allocation between on-board computers and ground based systems is evolving. Initially, ground based systems will perform the bulk of power system control and operation. The EPS control system is required to continuously monitor and determine the current state of the power system. The DC Testbed Control System consists of standard controllers arranged in a hierarchical and distributed architecture. These controllers provide all the monitoring and control functions for the DC Testbed Electrical Power System. Higher level controllers include the Power Management Controller, Load Management Controller, Operator Interface System, and a network of computer systems that perform some of the SSF Ground based Control Center Operation. The lower level controllers include Main Bus Switch Controllers and Photovoltaic Controllers. Power system status information is periodically provided to the higher level controllers to perform system control and operation. The data acquisition function of the control system is distributed among the various levels of the hierarchy. Data requirements are dictated by the control system algorithms being implemented at each level. A functional description of the various levels of the testbed control system architecture, the data acquisition function, and the status of its implementationis presented.

Baez, Anastacio N.↗
NASA NTRS1998

A Self-Tuning Kalman Filter for Autonomous Spacecraft Navigation

Most navigation systems currently operated by NASA are ground-based, and require extensive support to produce accurate results. Recently developed systems that use Kalman Filter and Global Positioning System (GPS) data for orbit determination greatly reduce dependency on ground support, and have potential to provide significant economies for NASA spacecraft navigation. Current techniques of Kalman filtering, however, still rely on manual tuning from analysts, and cannot help in optimizing autonomy without compromising accuracy and performance. This paper presents an approach to produce a high accuracy autonomous navigation system fully integrated with the flight system. The resulting system performs real-time state estimation by using an Extended Kalman Filter (EKF) implemented with high-fidelity state dynamics model, as does the GPS Enhanced Orbit Determination Experiment (GEODE) system developed by the NASA Goddard Space Flight Center. Augmented to the EKF is a sophisticated neural-fuzzy system, which combines the explicit knowledge representation of fuzzy logic with the learning power of neural networks. The fuzzy-neural system performs most of the self-tuning capability and helps the navigation system recover from estimation errors. The core requirement is a method of state estimation that handles uncertainties robustly, capable of identifying estimation problems, flexible enough to make decisions and adjustments to recover from these problems, and compact enough to run on flight hardware. The resulting system can be extended to support geosynchronous spacecraft and high-eccentricity orbits. Mathematical methodology, systems and operations concepts, and implementation of a system prototype are presented in this paper. Results from the use of the prototype to evaluate optimal control algorithms implemented are discussed. Test data and major control issues (e.g., how to define specific roles for fuzzy logic to support the self-learning capability) are also discussed. In addition, architecture of a complete end-to-end candidate flight system that provides navigation with highly autonomous control using data from GPS is presented.

Truong, Son H.↗
NASA NTRS2005

Middleware Evaluation and Benchmarking for Use in Mission Operations Centers

Middleware technologies have been promoted as timesaving, cost-cutting alternatives to the point-to-point communication used in traditional mission operations systems. However, missions have been slow to adopt the new technology. The lack of existing middleware-based missions has given rise to uncertainty about middleware's ability to perform in an operational setting. Most mission architects are also unfamiliar with the technology and do not know the benefits and detriments to architectural choices - or even what choices are available. We will present the findings of a study that evaluated several middleware options specifically for use in a mission operations system. We will address some common misconceptions regarding the applicability of middleware-based architectures, and we will identify the design decisions and tradeoffs that must be made when choosing a middleware solution. The Middleware Comparison and Benchmark Study was conducted at NASA Goddard Space Flight Center to comprehensively evaluate candidate middleware products, compare and contrast the performance of middleware solutions with the traditional point- to-point socket approach, and assess data delivery and reliability strategies. The study focused on requirements of the Global Precipitation Measurement (GPM) mission, validating the potential use of middleware in the GPM mission ground system. The study was jointly funded by GPM and the Goddard Mission Services Evolution Center (GMSEC), a virtual organization for providing mission enabling solutions and promoting the use of appropriate new technologies for mission support. The study was broken into two phases. To perform the generic middleware benchmarking and performance analysis, a network was created with data producers and consumers passing data between themselves. The benchmark monitored the delay, throughput, and reliability of the data as the characteristics were changed. Measurements were taken under a variety of topologies, data demands, and data characteristics, using several middleware products. All results were compared to systems using traditional point-to-point socket communication. By comparing performance results under Merent settings, inferences could be made about each middleware's ability to meet certain requirements of the GPM mission. The second phase simulated a middleware-based mission operations center. Actual mission support tools were either used or simulated to create real world demands on the middleware. Network and computer demands were watched closely to verify that no specific idiosyncrasies of mission operations activities would prove unsupportable by the middleware. In our presentation, we will address some commonly accepted misconceptions concerning middleware in mission support architectures. Specifically, we will focus on the perception that middleware solutions are too slow or impose too much overhead for real-time mission operations, and that middleware solutions are too expensive for small

Antonucci, Rob↗
DOE OSTI2023

Predicting Biomass Yields of Advanced Switchgrass Cultivars for Bioenergy and Ecosystem Services Using Machine Learning

The production of advanced perennial bioenergy crops within marginal areas of the agricultural landscape is gaining interest due to its potential to sustainably produce feedstocks for biofuels and bioproducts while also improving the sustainability and resilience of commodity crop production. However, predicting the biomass yields of this production system is challenging because marginal areas are often relatively small and spread around agricultural fields and are typically associated with various abiotic conditions that limit crop production. Machine learning (ML) offers a viable solution as a biomass yield prediction tool because it is suited to predicting relationships with complex functional associations. The objectives of this study were to (1) evaluate the accuracy of commonly applied ML algorithms in agricultural applications for predicting the biomass yields of advanced switchgrass cultivars for bioenergy and ecosystem services and (2) determine the most important biomass yield predictors. Datasets on biomass yield, weather, land marginality, soil properties, and agronomic management were generated from three field study sites in two U.S. Midwest states (Illinois and Iowa) over three growing seasons. The ML algorithms evaluated in the study included random forests (RFs), gradient boosting machines (GBMs), artificial neural networks (ANNs), K-neighbors regressor (KNR), AdaBoost regressor (ABR), and partial least squares regression (PLSR). Coefficient of determination (R 2 ) and mean absolute error (MAE) were used to evaluate the predictive accuracy of the tested algorithms. Results showed that the ensemble methods, RF (R 2 = 0.86, MAE = 0.62 Mg/ha), GBM (R 2 = 0.88, MAE = 0.57 Mg/ha), and GBM (R 2 = 0.78, MAE = 0.66 Mg/ha), were the most accurate in predicting biomass yields of the Independence, Liberty, and Shawnee switchgrass cultivars, respectively. This is in agreement with similar studies that apply ML to multi-feature problems where traditional statistical methods are less applicable and datasets used were considered to be relatively small for ANNs. Consistent with previous studies on switchgrass, the most important predictors of biomass yield included average annual temperature, average growing season temperature, sum of the growing season precipitation, field slope, and elevation. This study helps pave the way for applying ML as a management tool for alternative bioenergy landscapes where understanding agronomic and environmental performance of a multifunctional cropping system seasonally and interannually at the sub-field scale is critical.

09 BIOMASS FUELS↗
DOE OSTI2023

Residual-based error correction for neural operator accelerated infinite-dimensional Bayesian inverse problems

We explore using neural operators, or neural network representations of nonlinear maps between function spaces, to accelerate infinite-dimensional Bayesian inverse problems (BIPs) with models governed by nonlinear parametric partial differential equations (PDEs). Neural operators have gained significant attention in recent years for their ability to approximate the parameter-to-solution maps defined by PDEs using as training data solutions of PDEs at a limited number of parameter samples. The computational cost of BIPs can be drastically reduced if the large number of PDE solves required for posterior characterization are replaced with evaluations of trained neural operators. However, reducing error in the resulting BIP solutions via reducing the approximation error of the neural operators in training can be challenging and unreliable. We provide an a priori error bound result that implies certain BIPs can be ill-conditioned to the approximation error of neural operators, thus leading to inaccessible accuracy requirements in training. To reliably deploy neural operators in BIPs, we consider a strategy for enhancing the performance of neural operators: correcting the prediction of a trained neural operator by solving a linear variational problem based on the PDE residual. We show that a trained neural operator with error correction can achieve a quadratic reduction of its approximation error, all while retaining substantial computational speedups of posterior sampling when models are governed by highly nonlinear PDEs. The strategy is applied to two numerical examples of BIPs based on a nonlinear reaction–diffusion problem and deformation of hyperelastic materials. We demonstrate that posterior representations of the two BIPs produced using trained neural operators are greatly and consistently enhanced by error correction.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Application of unsupervised deep learning to image segmentation and in-situ contact angle measurements in a CO 2 -water-rock system

Rock surface wettability is a critical property that regulates multiphase flows in porous media, which can be quantified using the surface contact angle (CA). X-ray micro-computed tomography (μCT) provides an effective approach to in-situ measurements of surface CAs. However, the CA measurement accuracy depends significantly on the quality of CT image segmentation, which is the clustering of CT pixels into separate phases. Inspired by this, we developed a deep learning (DL)-based CA measurement workflow. Motivated by the recent tremendous progress in unsupervised learning techniques and aiming to avoid expensive manual data annotations, an unsupervised DL pipeline for CT image segmentation was proposed and implemented, which includes unsupervised model training and post-processing. The unsupervised model training was driven by a novel loss function constrained with feature similarity and spatial continuity and implemented by iterative forward and backward paths; the former clustered the pixel-wise feature vectors extracted by convolution neural networks, whereas the latter updated the parameters using gradient descent. An over-segmentation strategy was adopted for model training. The post-processing steps based on agglomerative hierarchical clustering (AHC) were implemented to further merge the over-segmented model output to the desired cluster number, which is intended to improve the efficiency of image segmentation. The developed unsupervised DL pipeline was compared with other commonly-used image segmentation methods using pixel-wise and physics-based evaluation metrics on a synthetic raw-image dataset, which had a known ground truth. The unsupervised DL pipeline showed the best performance. Next, the segmented images were input to an automatic CA measurement tool, and the results were validated by comparisons with manual measurements. The CA values from the manual and automatic measurements showed similar distributions and statistical properties. The automatic measurement demonstrated a wider spectrum because of the much larger number of measurement data points. The primary novelty of the unsupervised DL pipeline developed in this study lies in the novel loss function and the over-segmentation strategy associated with AHC post-processing. Finally, the workflow has been proven an efficient tool for pore-scale wettability characterization, which has a wide range of applications in fundamental studies of multiphase flows in natural porous media, which have critical implications to geological carbon sequestration, hydrocarbon energy recovery, and contaminant transport in groundwater.

42 ENGINEERING↗
DOE OSTI2025

Deployment of neural-network-based neutron microscopic cross sections in the Griffin reactor physics application

The capability to utilize neural networks to predict macroscopic and microscopic cross section parametric spaces has been developed for the Griffin reactor physics application. The LibTorch interface enables Griffin's MOOSE-based materials to interact with LibTorch-trained models, allowing for the evaluation of complex macroscopic or microscopic cross section spaces, which are then used to evaluate the neutronic properties of the Griffin finite element model. This study benchmarks traditional ISOXML-formatted tabulation libraries against neural network-based models for 279 nuclides on 20,160 grid points for zero-dimensional and two-dimensional reactor models. Benchmark metrics include the fundamental mode eigenvalue, fission and absorption rates, and various temperature coefficients of reactivity (isothermal, fuel, and moderator). From the perspective of storage space, the complete set of LibTorch models uses 11 MB on disk, compared to the 10 GB for the ISOXML multigroup library that covers the same grid space. For the two-dimensional performance case considered in Griffin, the Torch model uses 97% less RAM than the reference ISOXML dataset while runtime increases by a factor of 3 when using the LibTorch model compared to the ISOXML dataset with multi-linear interpolation. The LibTorch model consistently yields errors within 0.01% for most analyzed quantities except for the temperature coefficients of reactivity where the maximum discrepancies are up to 0.3 $\frac{pcm}{K}$. Due to the neural network attempting to best predict quantities with no regard for a positive or negative bias for any given quantity, predictions may experience random fluctuations, resulting in both positive and negative errors. Future work will entail both depletion and coupled transient analysis to determine the predictive capabilities of Griffin with neural network-based cross sections.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
NASA NTRS2004

Wind Tunnel Test of an RPV with Shape-Change Control Effector and Sensor Arrays

A variety of novel control effector concepts have recently emerged that may enable new approaches to flight control. In particular, the potential exists to shift the composition of the typical aircraft control effector suite from a small number of high authority, specialized devices (rudder, aileron, elevator, flaps), toward larger numbers of smaller, less specialized, distributed device arrays. The concept envisions effector and sensor networks composed of relatively small high-bandwidth devices able to simultaneously perform a variety of control functions using feedback from disparate data sources. To investigate this concept, a remotely piloted flight vehicle has been equipped with an array of 24 trailing edge shape-change effectors and associated pressure measurements. The vehicle, called the Multifunctional Effector and Sensor Array (MESA) testbed, was recently tested in NASA Langley's 12-ft Low Speed wind tunnel to characterize its stability properties, control authorities, and distributed pressure sensitivities for use in a dynamic simulation prior to flight testing. Another objective was to implement and evaluate a scheme for actively controlling the spanwise pressure distribution using the shape-change array. This report describes the MESA testbed, design of the pressure distribution controller, and results of the wind tunnel test.

Raney, David L.↗
NASA NTRS2023

Toward Physics-informed Neural Networks for 3D Multi-layer Cloud Mask Reconstruction

Three-dimensional (3D) cloud retrievals are critical for understanding their impact on climate and other applications such as aviation safety, weather prediction, and remote sensing. However, obtaining high-resolution and accurate vertical representation of clouds remains unsolved due to the limitations imposed by satellite instrumentation, viewing conditions, and the complexity of cloud dynamics. Cloud masks are essential for comprehending various cloud vertical properties, but deriving accurate 3D cloud masks from 2D satellite imagery data is a challenging task. To tackle these challenges, we introduce a physics-informed loss function for training deep learning models that can extend 2D cloud images into 3D cloud masks. The proposed loss, called CloudMask Loss, is composed of two domain knowledge-informed loss terms: one for evaluating cloud position and thickness, and the other for measuring the number of layers. By combining these loss terms, we improve the trainability of the deep learning models for more accurate and meaningful results. We apply the proposed loss function to different neural networks and demonstrate significant improvements in multi-layer cloud mask reconstruction. Utilizing the same neural network architecture, our proposed loss outperforms standard binary crossentropy loss in terms of multi-layer cloud classification accuracy, number of layers accuracy, and thickness mean absolute error (MAE). The proposed loss function can be readily integrated into various neural network architectures, resulting in substantial performance gains in 3D cloud mask generation.

multi-layer clouds↗
DOE OSTI2024

OASIS: Offsetting Active Reconstruction Attacks in Federated Learning

Federated Learning (FL) has garnered significant attention for its potential to protect user privacy while enhancing model training efficiency. For that reason, FL has found its use in various domains, from health care to industrial engineering, especially where data cannot be easily exchanged due to sensitive information or privacy laws. However, recent research has demonstrated that FL protocols can be easily compromised by active reconstruction attacks executed by dishonest servers. These attacks involve the malicious modification of global model parameters, allowing the server to obtain a verbatim copy of users' private data by inverting their gradient updates. Tackling this class of attack remains a crucial challenge due to the strong threat model. In this paper, we propose a defense mechanism, namely OASIS, based on image augmentation that effectively counteracts active reconstruction attacks while preserving model performance. We first uncover the core principle of gradient inversion that enables these attacks and theoretically identify the main conditions by which the defense can be robust regardless of the attack strategies. We then construct our defense with image augmentation showing that it can undermine the attack principle. Comprehensive evaluations demonstrate the efficacy of the defense mechanism highlighting its feasibility as a solution.

deep neural networks↗
DOE OSTI2026

Residential Vehicle-to-Home Backup Power Capabilities: Key Findings from a ComEd Beneficial Electrification R&D Pilot

This report summarizes key findings from a collaborative technical study of residential, non-grid-tied vehicle-to-home (V2H) backup power systems in Commonwealth Edison’s (ComEd’s) service territory. The work integrates (1) a feeder-level technoeconomic analysis (TEA) using historical outage-event data and simulated electric-vehicle (EV) driving/charging profiles to estimate potential reliability and customer interruption-cost impacts under V2H and vehicle-to-grid (V2G) adoption scenarios; (2) controlled laboratory performance testing of a representative V2H backup ecosystem to characterize transfer-to-backup behavior, sustained power delivery, efficiency trends, and repeatable reliability limitations; and (3) a cybersecurity assessment aligned with NIST Cybersecurity Framework (CSF) 2.0 and ISO/SAE 21434 to evaluate interface-level risk drivers and identify program-relevant mitigations. Results indicate that V2H can provide measurable resilience value, but outcomes are strongly context dependent on outage patterns and the share of events that are “V2H-applicable.” Typical transfer-to-backup behavior clustered on the order of minutes, but rare long-delay edge cases were observed (including an event approaching 30 minutes) and should be treated as a reliability risk. High-power testing showed that peak-rated output is not necessarily continuously deliverable; stable operation may require operation below nameplate ratings and attention to thermal and installation constraints. The cybersecurity assessment highlights a broad attack surface spanning commissioning, home networks, embedded services, and cloud/OTA pathways, motivating minimum controls for secure onboarding, signed updates, patch cadence, and coordinated vulnerability response for any scaled deployment.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2022

Classifying handedness in chiral nanomaterials using label error robust deep learning

Abstract High-throughput scanning electron microscopy (SEM) coupled with classification using neural networks is an ideal method to determine the morphological handedness of large populations of chiral nanoparticles. Automated labeling removes the time-consuming manual labeling of training data, but introduces label error, and subsequently classification error in the trained neural network. Here, we evaluate methods to minimize classification error when training from automated labels of SEM datasets of chiral Tellurium nanoparticles. Using the mirror relationship between images of opposite handed particles, we artificially create populations of varying label error. We analyze the impact of label error rate and training method on the classification error of neural networks on an ideal dataset and on a practical dataset. Of the three training methods considered, we find that a pretraining approach yields the most accurate results across label error rates on ideal datasets, where size and other morphological variables are held constant, but that a co-teaching approach performs the best in practical application.

36 MATERIALS SCIENCE↗
DOE OSTI2025

A Variational Autoencoder Model Toward Molecular Structure Representation Learning of Fuels

Here, in this work, a Variational Autoencoder (VAE)-based data-driven modeling framework is developed with the overarching goal of enabling fuel design. The VAE model is trained on a large dataset with several chemical species to learn a compressed latent space molecular representation. Chemical structure in the form of Simplified Molecular Input Line Entry System (SMILES) string is fed as input, encoded into the VAE latent space, and decoded back to the SMILES string using Long Short-Term Memory (LSTM) networks. Complexities of the VAE training loss function are thoroughly examined by varying the weightage (beta (𝜷) parameter) of the latent space regularization term, thereby assessing the balance between reconstruction accuracy and validity, and focusing on both accurate molecular structure reconstruction and latent space consistency. Two different strategies for 𝜷 variation are evaluated: linear annealing and cyclic annealing. In addition, the impact of total correlation adjustment and hierarchical priors is also studied with regard to the balance between reconstruction fidelity and latent space regularization, and potential issues such as posterior collapse, over-regularization, and poor disentanglement of latent variables. Overall, the best performance of the model is achieved with hierarchical priors and incrementally increasing 𝜷 from 0 to a threshold value of 0.25 over 75 epochs. The generative VAE model can be readily coupled with Quantitative Structure–Property Relationship (QSPR) analysis to develop an integrated end-to-end framework for fuel-property prediction and molecular design of novel promising fuels.

fuel design↗