Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Machine Learning Models”

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 1,063 records · Page 59
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

A rule-free workflow for the automated generation of databases from scientific literature

Abstract In recent times, transformer networks have achieved state-of-the-art performance in a wide range of natural language processing tasks. Here we present a workflow based on the fine-tuning of BERT models for different downstream tasks, which results in the automated extraction of structured information from unstructured natural language in scientific literature. Contrary to existing methods for the automated extraction of structured compound-property relations from similar sources, our workflow does not rely on the definition of intricate grammar rules. Hence, it can be adapted to a new task without requiring extensive implementation efforts and knowledge. We test our data-extraction workflow by automatically generating a database for Curie temperatures and one for band gaps. These are then compared with manually curated datasets and with those obtained with a state-of-the-art rule-based method. Furthermore, in order to showcase the practical utility of the automatically extracted data in a material-design workflow, we employ them to construct machine-learning models to predict Curie temperatures and band gaps. In general, we find that, although more noisy, automatically extracted datasets can grow fast in volume and that such volume partially compensates for the inaccuracy in downstream tasks.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Towards Fast and Accurate Predictions of Radio Frequency Power Deposition and Current Profile via Data-driven Modeling

Three machine learning techniques (multilayer perceptron, random forest, and Gaussian process) provide fast surrogate models for lower hybrid current drive (LHCD) simulations. A single GENRAY/CQL3D simulation without radial diffusion of fast electrons requires several minutes of wall-clock time to complete, which is acceptable for many purposes, but too slow for integrated modeling and real-time control applications. The machine learning models use a database of 16,000+ GENRAY/CQL3D simulations for training, validation, and testing. Latin hypercube sampling methods ensure that the database covers the range of 9 input parameters ($n_{e0}$, $T_{e0}$, $I_p$, $B_t$, $R_0$, $n_{||}$, $Z_{eff}$, $V_{loop}$, $P_{LHCD}$) with sufficient density in all regions of parameter space. The surrogate models reduce the inference time from minutes to ~ms with high accuracy across the input parameter space.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2025

Classification of Cloud Particle Imagery and Thermodynamics (COCPIT): A New Databasing Tool for the Characterization of Cloud Particle Images Captured During DOE Field Campaigns

The Department of Energy for decades has explored the earth system and atmosphere through research and deployment of in-situ and remote sensing platforms during field campaigns. Among these datasets exists a vast supply of cloud particle images that provide visual insight into the complex microphysics in the clouds that span our globe. The millions of images collected over decades of deployments provides a unique opportunity to further our understanding of our atmosphere down to the crystal size. This work over the past 5 years has sought to organize these images into digestible datasets that can then be used by scientists to further our understanding of microphysics. A machine learning model was developed that categorizes over 1.5 million images across 11 weather events with over 90% accuracy according to particle type. The database was then extended to include dimensional characteristics of the particle as well as co-location of environmental properties, such as temperature and water content. Then, to initialize the connection between these data and our understanding of how crystals form and grow, weather research and forecasting simulations were run to generate the growth histories of the classified crystals. This research culminates with 2 databases per event: (1) a database of all classified crystals and their dimensional and environmental properties and (2) simulated growth histories of each crystal. Finally, a user interface was created to allow researchers to explore data statistics.

54 ENVIRONMENTAL SCIENCES↗
NASA NTRSDate not supplied

Imbalanced Multi-layer Cloud Classification with Advanced Baseline Imager (ABI) and CloudSat/CALIPSO Data

Clouds at different altitudes play different roles in Earth’s climate. Comprehensive understanding of overlapping clouds is important for climate and weather prediction. The East Pacific region is where El Ni˜no and La Ni˜na originate and where multi-layer clouds frequently occur. The overlap of clouds at different altitudes in this region increases the classification complexity for cloud-based climatological studies. Unlike prior work in cloud layer classification that assumes single layer or two-layer of clouds, in this work, we consider multi-layer cloud classification with 8 cloud-level classes (clear-sky, high, middle, low, high+middle, high+low, middle+low, high+middle+low). We develop and analyze machine learning models on features extracted from satellite images from the East Pacific regions collected by GOES Advanced Baseline Imager (ABI). These are used to classify CloudSat/CALIPSO observed multi-layer clouds. Due to the imbalanced nature of the data, we investigate the adoption of conventional resampling methods, as well as deep learning methods with data augmentation. In our experiments, we utilize the random forest classifier and Multilayer perceptron classifier with data augmentation methods to reduce the class imbalance during training. With these approaches, we achieve a classification accuracy of 83.6% without exploiting any ancillary information.

machine learning↗
NASA NTRSDate not supplied

Satellite Based Precipitation Estimation in Orographic Regions within the Southwestern United States

Predicting precipitation-induced landslides requires accurate estimation of orographic precipitation. Many research studies have been done to estimate orographic precipitation using measurement/estimation methods that include rain gauges, ground-based radar, satellite-based estimates, and modeling. Each method has strengths and weaknesses, but none have been able to fully resolve orographic precipitation. The Integrated Multi-satellitE Retrievals for Global Precipitation Measurement Mission (IMERG) early run product provides precipitation estimates at a 0.1° spatial resolution, at half-hour time scales, with only a 4-hour latency. This product is currently used for the global landslide hazard assessment, but there are known issues in mountainous terrain that the IMERG algorithm has not been able to fully resolve; and the sparse gauge density in these regions makes it even more difficult. In this study, precipitation events were identified using high temporal resolution (5-15 minute) precipitation observations from rain gauges in mountainous terrain in the southwestern United States. The brightness temperature from several infrared (IR) bands from the Geostationary Operational Environmental Satellite (GOES) 16 satellite were used to estimate precipitation with a k-nearest neighbor machine learning model. Compared to IMERG, the IR estimates from GOES-16 performed better at predicting gauge-identified precipitation events. Additionally, the IR-only based estimates were able to estimate precipitation when IMERG failed to detect precipitation, false negative events. While additional analysis is needed, results indicate the need for better integration of IR observations for more accurate precipitation estimation in mountainous regions.

Jessica Sutton↗
DOE OSTI2023

Extension of Large Fire Emissions From Summer to Autumn and Its Drivers in the Western US

Abstract Burned areas in the western US have increased ten‐fold since 1980s, which are attributable to multiple factors, including increasing heat, changing precipitation patterns, and extended drought. To better understand how these factors contribute to large fire emissions (gridded monthly fire emissions >95th percentile of all the fire emissions in the western US; 0.009 Gg/month), we build a machine learning model to predict fire emissions (PM 2.5 ) over the western US at 0.25° resolution, interpreted using explainable artificial intelligence (XAI). From the predictor contributions derived from XAI, we conduct k‐means clustering analysis to identify four clusters of predictor variables representing different drivers of large fire emissions. The four clusters feature the contributions of fuel load (Cluster 1) and different levels of dryness (Cluster 2–4), controlled by fuel moisture, drought condition, and fire‐favorable large‐scale meteorological patterns featuring high temperature, high pressure, and low relative humidity. In the past two decades, large fire emissions peak in summer. However, large fire emissions increased significantly in September and October in 2010–2020 relative to 2000–2009, extending the peak large fire emissions from summer to autumn. The larger enhancements of large fire emissions during autumn compared to summer are contributed by decreased fuel moisture, along with more frequent concurrent fire‐favorable large‐scale meteorological patterns and drought. These results highlight fuel drying as a common driver supported by multiple drivers, such as warmer temperature and more frequent synoptic patterns favorable for fires, in increasing the autumn risk of large fire emissions across the western US.

54 ENVIRONMENTAL SCIENCES↗
NASA NTRS2024

Low-Cost Sensor Performance Intercomparison, Correction Factor Development, and 2+ Years of Ambient PM2.5 Monitoring in Accra, Ghana

Particulate matter air pollution is a leading cause of global mortality, particularly in Asia and Africa. Addressing the high and wide-ranging air pollution levels requires ambient monitoring, but many low- and middle-income countries (LMICs) remain scarcely monitored. To address these data gaps, recent studies have utilized low-cost sensors. These sensors have varied performance, and little literature exists about sensor intercomparison in Africa. By colocating 2 QuantAQ Modulair-PM, 2 PurpleAir PA-II SD, and 16 Clarity Node-S Generation II monitors with a reference-grade Teledyne monitor in Accra, Ghana, we present the first intercomparisons of different brands of low-cost sensors in Africa, demonstrating that each type of low-cost sensor PM2.5 is strongly correlated with reference PM2.5, but biased high for ambient mixture of sources found in Accra. When compared to a reference monitor, the QuantAQ Modulair-PM has the lowest mean absolute error at 3.04 μg/m3, followed by PurpleAir PA-II (4.54 μg/m3) and Clarity Node-S (13.68 μg/m3). We also compare the usage of 4 statistical or machine learning models (Multiple Linear Regression, Random Forest, Gaussian Mixture Regression, and XGBoost) to correct low-cost sensors data, and find that XGBoost performs the best in testing (R2: 0.97, 0.94, 0.96; mean absolute error: 0.56, 0.80, and 0.68 μg/m3 for PurpleAir PA-II, Clarity Node-S, and Modulair-PM, respectively), but tree-based models do not perform well when correcting data outside the range of the colocation training. Therefore, we used Gaussian Mixture Regression to correct data from the network of 17 Clarity Node-S monitors deployed around Accra, Ghana, from 2018 to 2021. We find that the network daily average PM2.5 concentration in Accra is 23.4 μg/m3, which is 1.6 times the World Health Organization Daily PM2.5 guideline of 15 μg/m3. While this level is lower than those seen in some larger African cities (such as Kinshasa, Democratic Republic of the Congo), mitigation strategies should be developed soon to prevent further impairment to air quality as Accra, and Ghana as a whole, rapidly grow.

Humidity↗
DOE OSTI2025

Establishing nationwide power system vulnerability index across US counties using interpretable machine learning

Power outages have become increasingly frequent, intense, and prolonged in the US due to climate change, aging electrical grids, and rising energy demand. However, largely due to the absence of granular spatiotemporal outage data, we lack data-driven evidence and analytics-based metrics to quantify power system vulnerability. This limitation has hindered the ability to effectively evaluate and address vulnerability to power outages in US communities. Here, in this work, we collected ∼179 million power outage records at 15-min intervals across 3022 US contiguous counties (96.15 % of the area) from 2014 to 2023. We developed a power system vulnerability assessment framework based on three dimensions (intensity, frequency, and duration) and applied interpretable machine learning models (XGBoost and SHAP) to compute Power System Vulnerability Index (PSVI) at the county level. Our analysis reveals a consistent increase in power system vulnerability across the US counties over the past decade. We identified 318 counties across 45 states as hotspots for high power system vulnerability, particularly in the West Coast (California and Washington), the East Coast (Florida and the Northeast area), the Great Lakes megalopolis (Chicago-Detroit metropolitan areas), and the Gulf of Mexico (Texas). Our heterogeneity analysis indicates that urban counties and those located along regional transmission boundaries tend to exhibit significantly higher vulnerability. Our results highlight the significance of the proposed PSVI for evaluating the vulnerability of communities to power outages. The findings underscore the widespread and pervasive impact of power outages across the country and offer crucial insights to support infrastructure operators, policymakers, and emergency managers in formulating policies and programs aimed at enhancing the resilience of the US power infrastructure.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2024

Autonomous Output‐Oriented Aerosol Jet Printing Enabled by Hybrid Machine Learning

Additive manufacturing (AM) is rapidly revolutionizing modern manufacturing with recent progress in advanced printing methods and improved properties of printed materials. However, traditional AM methods are limited by their input‐oriented nature, which demands tedious trial‐and‐error tuning of printing parameters to achieve desired output properties. Here, in this work, an output‐oriented artificial intelligence‐integrated AM (AIAM) method is reported that enables an user to specify desired output properties while the printer autonomously discovers the optimal input printing parameters by integrating hybrid machine learning models and in situ measurements. Based on a predictive mapping between the input printing parameters and the output properties of interests established with <20 experiments designed by active learning, inverse design tasks are performed to intelligently generate the printing parameter settings that lead to desired outcomes using reinforcement learning. This method is demonstrated by autonomous aerosol jet printing (AJP) of conductive polymer films and achieving user‐defined electrical resistances with an ultralow error of 3.7%. The AIAM method, with its output‐oriented nature, holds the potential to significantly improve the autonomy, predictability, efficiency, and accessibility of the AM processes, which will unlock new possibilities in the autonomous and intelligent printing of a broad range of functional materials and devices.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Critical statistical assessment of data in metal additive manufacturing

Obtaining high quality data reflecting the relationships between the additive manufacturing (AM) process parameters, material microstructure and mechanical properties is crucial for the use of machine learning in AM. A database of over 4,000 data entries of metal AM was created thanks to a large number of literature studies on key process parameters and indicators of build quality. Meta-analysis reveals critical biases in the literature. Firstly, majority of studies report only high quality builds, these imbalances in reporting result in weak correlation between process parameters, properties and consolidation, limiting the ability of machine learning models to generalize beyond optimized conditions. Nevertheless, the trained models accurately predict yield strength ($R^2 = 0.85$), suggesting that certain process–property relationships are effectively captured within these models. Secondly, quantitative microstructural data are largely absent, limiting the learning of the microstructure-mechanical properties relationships. Finally, current process window identification is based largely on the consolidation, despite significant uncertainty in its measurement. It is important to identify the process map on the basis of not only the consolidation, but also mechanical behaviour under loading. Such a identification shows that 316 L and Inconel have much larger process map (i.e. highly printable) in comparison to the AlSi10Mg and Ti6Al4V.

Additive manufacturing↗
DOE OSTI2024

Quantifying the basic reproduction number and underestimated fraction of Mpox cases worldwide at the onset of the outbreak

In 2022, there was a global resurgence of mpox, with different clinical-epidemiological features compared with previous outbreaks. Sexual contact was hypothesized as the primary transmission route, and the community of men having sex with men (MSM) was disproportionately affected. Because of the stigma associated with sexually transmitted infections, the real burden of mpox could be masked. We quantified the basic reproduction number (R 0 ) and the underestimated fraction of mpox cases in 16 countries, from the onset of the outbreak until early September 2022, using Bayesian inference and a compartmentalized, risk-structured (high-/low-risk populations) and two-route (sexual/non-sexual transmission) mathematical model. Machine learning (ML) was harnessed to identify underestimation determinants. Estimated R 0 ranged between 1.37 (Canada) and 3.68 (Germany). The underestimation rates for the high- and low-risk populations varied between 25–93% and 65–85%, respectively. The estimated total number of mpox cases, relative to the reported cases, is highest in Colombia (3.60) and lowest in Canada (1.08). In the ML analysis, two clusters of countries could be identified, differing in terms of attitudes towards the 2SLGBTQIAP+ community and the importance of religion. Given the substantial mpox underestimation, surveillance should be enhanced, and country-specific campaigns against the stigmatization of MSM should be organized, leveraging community-based interventions.

60 APPLIED LIFE SCIENCES↗
NASA NTRSDate not supplied

Detecting Risk and Anomalies in Airplane Dynamics Through Entropic Analysis of Time Series Data

Despite recent efforts to move away from traditional threshold exceedance detection methods for aircraft state monitoring, modern aircraft still rely on safety thresholds to communicate to pilots the identification of an anomaly in the aircraft when a threshold is surpassed. Current anomaly detection methods mainly depend on uninterpretable machine learning models to learn complex patterns and relationships contained in the time series data of aircraft. Although these methods are capable of identifying known anomalies, their deficiency in interpretability presents a challenge when translating them to different aircraft. To overcome this deficiency, entropic analysis of aircraft dynamics seeks to characterize the complexity, or lack thereof, of the aircraft dynamics prior to the development of a risk scenario. This complexity characterization provides a more straightforward summary of state changes in the dynamics of flight variables. To build a foundation for entropic analysis, we analyzed the complexity of unstable approaches, an anomalous event present in many of today’s aviation accidents. The analysis revealed a statistically significant difference in the complexity distribution of flight variables under a stable approach versus an unstable approach. These differences in complexity were especially notable minutes before an approach was identified as unstable. Moreover, the multiscale entropic analysis revealed the presence of signal complexity at multiple time scales across multiple time windows before landing. By capturing state changes and corrections in the aircraft dynamics using entropy, advanced, yet still interpretable, sensor systems based on entropic frameworks from this study can be constructed in the future using classical machine learning approaches.

Risk detection↗
DOE OSTI2026

Early calendar life and health prediction of silicon batteries via machine learning with uncertainty quantification

Lithium-ion batteries with silicon anodes promise high energy density but are limited by calendar lifetime. Reducing the long iteration time to obtain experimental results requires predicting calendar lifetime early in a cell's life. In this study, we demonstrate that lightweight machine learning models with feature engineering can provide calendar lifetime estimates from early electrochemical signals. After 1 month of electrochemical aging, the best models achieve 10% error in calendar-life prediction and can separate "bad" from "good" lifetime cells with a mean F1 score of 0.857. As battery systems exhibit inherent variability, four methods for uncertainty quantification are compared, and confidence intervals are demonstrated with an uncertainty of +-3.6 months in lifetime prediction. A feature importance analysis indicates that early patterns in voltage decay are the strongest indicators of calendar lifetime. Finally, this modeling approach has high error when generalizing to new electrode chemistries or testing conditions but with appropriately low confidence.

25 ENERGY STORAGE↗
NASA NTRSDate not supplied

Algorithmic Detection of Elemental Biosignatures

Machine learning models that classify a sample as indicative or non-indicative of life could play an important role in life-detection missions. Their predictions result from agnostic algorithms and thereby add redundancy to judgements resulting from human expertise. Additionally, their important features can reveal the most informative measurements within the operational constraints of a life-detection mission. The Ladder of Life Detection (Neveu 2018) identifies the need for an understanding of how combinations of multiple biosignatures affect overall confidence. The present work provides a starting point to answer this need, and future work will expand the data types to obtain even more predictive combinations of features. Elemental abundance was chosen as a starting set of features due to its availability in diverse sample types, which are needed to train a generalizable model. A standardized dataset was collected, including 35 non-indicative, e.g., lunar rock, basalt; 19 indicative mixed, e.g., seawater, agricultural soil; 46 indicative non-alive, e.g., coal, chalk; and 10 indicative alive, e.g., biofilm, bacteria. This dataset could be valuable for complementary biosignature research. The samples were standardized to the same limit of detection of a simulated mission scenario. Four classification models were used: k-nearest neighbors (KNN), logistic regression (LR), linear support vector machines (SVM), and Gaussian naïve Bayes (GNB). To obtain feature importances, KNN was run on three principal components of the training data and LR and SVM were run with L1 and L2 regularization. The performances and feature importances of the six model variants on 40:60 train to validation ratios were assessed with Monte Carlo simulations. ROC AUC and mean accuracy scores ranged between 82% - 94%, with sensitivity greater than specificity. For indicative of life predictors, all models had C and Ca as strong and Cl as medium; a majority of models had N, K, and P as medium. For non-indicative of life predictors, all models had Si as strong, and a majority of models had Mg, Al, and Ti as medium. Varied elements were Fe (slightly non-indicative), H (slightly indicative), O (widely varied), Na, Mn, and S. These results serve as a proof of concept and suggest important elemental signals beyond merely the CHNOPS of Earth-based life.

Algorithmic↗
DOE OSTI2024

Hot, cold, or just right? An infrared biometric sensor to improve occupant comfort and reduce overcooling in buildings via closed-loop control

To improve occupant comfort and save energy in buildings, we have developed a closed-loop air conditioning (AC) sensor-controller that predicts occupant thermal sensation from the thermographic measurement of skin temperature distribution, then uses this information to reduce overcooling (cooling-energy overuse that discomforts occupants) by regulating AC output. Taking measures to protect privacy, it combines thermal-infrared (TIR) and color (visible spectrum) cameras with machine vision to measure the skin-surface temperature profile. Since the human thermoregulation system uses skin blood flow to maintain thermoneutrality, the distribution of skin temperature can be used to predict warm, neutral, and cool thermal states. We conducted a series of human-subject thermal-sensation trials in cold-to-hot environments, measuring skin temperatures and recording thermal sensation votes. We then trained random-forest classification machine-learning models (classifiers) to estimate thermal sensation from skin temperatures or skin-temperature differences. The estimated thermal sensation was input to a proportional integral (PI) control algorithm for the AC, targeting a sensation level between neutral and warm. Our sensor-controller includes a sensor assembly, server software, and client software. The server software orients the cameras and transmits images to the client software, which in turn assesses occupant skin temperature distribution, estimates occupant thermal sensation, and controls AC operation. A demonstration conducted in a conference room in an office building near Houston, TX showed that our system reduced overcooling, decreasing AC load by 42% when the room was occupied while improving occupant comfort (fraction of “comfortable” votes) by 15 percentage points.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2024

Coincidence anomaly detection for unsupervised locating of edge localized modes in the DIII-D tokamak dataset

Using supervised learning to train a machine learning model to predict an on-coming edge localized mode (ELM) requires a large number of labeled samples. Creating an appropriate data set from the very large database of discharges at a long-running tokamak, such as DIII-D, would be a very time-consuming process for a human. Considering this need and difficulty, we use coincidence anomaly detection, an unsupervised learning technique, to train an ELM-identifier to identify and label ELMs in the DIII-D discharge database. This ELM-identifier shows, simultaneously, a precision of 0.68 and a recall of 0.63 (AUC is 0.73) on identifying ELMs in example time series pulled from thousands of discharges spanning five years. In a test set of 50 discharges, the algorithm finds over 26 thousand ELM candidates, more than 5 times the existing catalog of ELMs labeled by humans.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2024

Supporting ARPA-E Power Grid Optimization (Final Report)

Pacific Northwest National Laboratory (PNNL), Arizona State University (ASU), Georgia Institute of Technology (Georgia Tech), Los Alamos National Laboratory (LANL), National Renewable Energy Laboratory (NREL), Texas A&M University (TAMU), The University of Texas at Austin (UT), and the University of Wisconsin-Madison (UW-M) supported the ARPA-E Grid Optimization (GO) Competition by providing a common problem formulation, data format, datasets, evaluation mechanism, scoring, rules, and results that resulted in the awarding of $\$9.24$ million dollars to teams from academia, industry, and national labs for solving three sets of increasingly difficult non-linear, security- constrained AC Optimal Powerflow (AC-OPF) optimization problems in order to increase the efficiency of the US Electric Grid. It is estimated that a 1% increase in efficiency can save $\$1$ billion. Current industry practices typically use a linear DC model (DC-OPF) in order solve the OPF problem within the time constraints of the operation schedule. The GO Competition challenges the best power engineers, mathematicians, and computer scientists to make possible operational decisions based on accurate physical models. To accomplish this, the GO Competition created a series of Challenges and funded teams to produce the best solver. Challenge 1 was to solve the security constrained Alternating Current Optimal Power Flow (ACOPF) problem. Challenge 2 extended that to by adding adjustable transformer tap ratios, phase shifting transformers, switchable shunts, price-responsive demand, ramp rate constrained generators and loads, and fast-start unit commitment (UC). Furthermore, Challenge 2 was a maximization problem while Challenge 1 was a minimization problem. While Challenge 3 was being developed, the entrants were invited to find better solutions to the Challenge 2 synthetic datasets with no restrictions on time, hardware, or algorithms. The Challenge 2 solutions turned out to be very good. Challenge 3 expanded the Challenge 2 problem further by using multiperiod dynamic markets, including advisory models for extreme weather events, day-ahead markets, and the real-time markets with an extended look-ahead. These problems included active bid-in demand and topology optimization. Together the Challenges used nearly 30 million CPU hours. Since each team was working on the same problem, using the same data, and running on the same hardware, fair comparisons could be drawn as to the best solver. The datasets were varied enough, however, that the best solver for one dataset was not necessarily the best at another, so cumulative scores were used. The process was managed by the PNNL maintained website https://GOCompetition.energy.gov, where Entrants could find information about the problem, the data, the rules, submit their solver for evaluation, and see the scores of all the competing teams on a Leaderboard. Interest was world-wide but only American teams were eligible for prizes. The Competition has produced 34 journal articles 115 papers and been cited over 500 times in the literature, including 12 dissertations (4 from foreign countries; Columbia (2), Germany, and Italy) and 3 from the DOE ExaScale project. Software developed by Pearl Street Technologies for Challenges 1 and 2 is now deployed by Southwest Power Pool (SPP) and Midcontinent Independent Service Operator (MISO). Other teams have received inquiries from venture capitalists. Google DeepMind has thanked the Competition for making the datasets developed for the Competition public. They are using it to train machine learning models. The larger datasets have billions of unknowns to be solved for, but only a small percent matter in the final solution. Knowing what unknowns are important can dramatically speedup the solution.

24 POWER TRANSMISSION AND DISTRIBUTION↗
NASA NTRS2023

Predicting Two-Dimensional Airfoil Performance Using Graph Neural Networks

Computer simulations require the use of meshes to simulate geometries. These meshes capture important geometric features of the design and can be used in machine learning modeling. This report explores the use of graph neural networks (GNNs) to learn features from two-dimensional (2D) airfoil designs represented as a set of nodes connected using edges. This type of network is common in aerospace applications: most geometries are represented as a mesh in order to perform analysis. The objective of this work is to use GNNs to predict the performance of 2D airfoils generated using the program XFOIL. The predicted performance parameters include bulk quantities such as coefficients of lift (C L ), drag (C d , C dp ), moment (C m ), and node-specific quantities such as coefficient of pressure (C p ). In this report, a spline convolutional graph-based neural network is compared with deep learning neural networks to predict both bulk and node-specific quantities. The findings indicate the GNNs are able to predict bulk quantities quite well; however, when the number of outputs is increased, the deep neural network (DNN) proves to be better in its prediction capability. Two different normalization strategies were compared in the training of both GNNs and DNNs: minmax and standard deviation. In both types of networks, standard deviation scaling proved to be the best.

machine learning↗