Search NASA⌕ Search

SEARCH · Search NASA

Results for “forecast accuracy”

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 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Dead Fuel Moisture Content Reanalysis Dataset for California (2000–2020)

This study presents a novel reanalysis dataset of dead fuel moisture content (DFMC) across California from 2000 to 2020 at a 2 km resolution. Utilizing a data assimilation system that integrates a simplified time-lag fuel moisture model with 10-h fuel moisture observations from remote automated weather stations (RAWS) allowed predictions of 10-h fuel moisture content by our method with a mean absolute error of 0.03 g/g compared to the widely used Nelson model, with a mean absolute error prediction of 0.05 g/g. For context, the values of DFMC in California are commonly between 0.05 g/g and 0.30 g/g. The presented product provides gridded hourly moisture estimates for 1-h, 10-h, 100-h, and 1000-h fuels, essential for analyzing historical fire activity and understanding climatological trends. The methodology presented here demonstrates significant advancements in the accuracy and robustness of fuel moisture estimates, which are critical for fire forecasting and management.

Farguell, Angel (ORCID:000000032395220X)↗
DOE OSTI2024

Visual and genetic stock identification of a test fishery to forecast Columbia River spring Chinook salmon stocks 2 weeks into the future

Abstract Modern fisheries management strives to balance opposing goals of protection for weak stocks and opportunity for harvesting healthy stocks. Test fisheries can aid management of anadromous fishes if they can forecast the strength and timing of an annual run with adequate time to allow fisheries planning. Integration of genetic stock identification (GSI) can further maximize utility of test fisheries by resolving run forecasts into weak‐ and healthy‐stock subcomponents. Using 5 years (2017–2022) of test fishery data, our study evaluated accuracy, resolution, and lead time of predictions for stock‐specific run timing and abundance of Columbia River spring Chinook salmon ( Oncorhynchus tshawytscha ). We determined if this test fishery (1) could use visual stock identification (VSI) to forecast at the coarse stock resolution (i.e., classification of “lower” vs. “upriver” stocks) upon which current management is based and (2) could be enhanced with GSI to forecast at higher stock resolution. VSI accurately identified coarse stocks (83.3% GSI concordance), and estimated a proxy for abundance (catch per unit effort, CPUE) of the upriver stock in the test fishery that was correlated ( R 2 = 0.90) with spring Chinook salmon abundance at Bonneville dam (Rkm 235). Salmon travel rates (~8.6 Rkm/day) provided predictions with 2‐week lead time prior to dam passage. Importantly, GSI resolved this predictive ability as finely as the hatchery broodstock level. Lower river stock CPUE in the test fishery was correlated with abundance at Willamette Falls (Rkm 196, R 2 = 0.62), but could not be as finely resolved as achieved for upriver stocks. We described steps to combine VSI and GSI to provide timely in‐season information and with prediction accuracy of ~12.4 mean absolute percentage error and high stock resolution to help plan Columbia River mainstem fisheries.

Hess, Jon E.↗
DOE OSTI2025

Performance evaluation of automated data-driven feature extraction and selection methods for practical and scalable building energy consumption prediction models

Here, this study quantifies the impact of automated feature engineering methods (feature extraction and selection) on the quality and accuracy of machine learning models that predict building energy consumption. The case study compares model performance for three main scenarios: baseline (no feature extraction and selection), feature extraction only, and feature extraction combined with feature selection (filter and/or wrapper methods) for fully trained machine learning models for 200 metered/sub-metered energy measurements across 118 real buildings. For consistency, the same machine learning model architecture (a black box deep learning neural network with probabilistic forecast output) was used for all scenarios. Based on results, all feature engineering methods provided noticeable prediction accuracy improvements (e.g., 29%-68% median prediction improvement) compared to baseline scenarios. However, in this application, feature selection methods provide little practical value due to their limited performance gains and high computational cost. Smarter algorithm development supported by better computational environments will be needed before feature selection methods can reliably and efficiently improve predictive model performance.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Active learning strategy for high fidelity short-term data-driven building energy forecasting

The quality of a data-driven model is heavily dependent on the quality of data. Data from building operation often have data bias problems, which means that the data sample is collected in a way that some members of the intended data population are less likely to be included than others. Data-driven energy forecasting models built on such data hence are biased and could lead to large forecasting errors. Active learning—an effective method to defying data bias—is rarely studied or applied in the area of data-driven building energy forecasting modeling. This paper attempts to fill this gap and explores the application of active learning in data-driven building energy forecasting. The developed strategy in this paper efficiently generate informative training data within a time budget and uses block design to passively consider weather disturbances. The developed active learning strategy is applied and evaluated in both virtual and real-building testbeds against traditional data-driven methods. Via these virtual and real-building evaluation cases, we have demonstrated that the data bias problem typically exists in building operation data is resolved by applying the developed active learning strategy. Furthermore, building energy forecasting models trained from data generated from the active learning strategy have shown improved performances in both model accuracy and model extendibility perspectives. The effectiveness of the block design module is also validated to effectively consider the impact of weather conditions on active learning design.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2021

A Short-Term Solar Forecasting Platform Using a Physics-Based Smart Persistence Model and Data Imputation Method

Electrical energy plays vital role in our socio-economic activity and therefore ensuring the reliability of the electric grid, from the generation, transmission and distribution level is critical. In order to maintain the power system parameter viz., frequency, voltage, etc., optimally, balancing of generation and consumption is very much essential. However, solar energy is infirm power by nature this is due to cloud cover / other local phenomena. Hence, Photovoltaic (PV) power generation brings a significant challenge to the grid operator due to the variability of the solar energy. The complexity of this challenge in terms of planning and dispatch ability of PV resources, aggravates with the high penetration of solar energy into the electric grid. In this setting, reliable solar radiation forecasting models based on accurate and quality input data become essential. In order to develop a suitable model for predicting solar radiation, quality historical / real time measurement is also needed. Under this study NIWE and NREL jointly developed / tested short-term solar forecasting frameworks using a smart persistence and physics-based smart persistence models for intra-hour forecasting of solar radiation (PSPI) and benchmarked 9 different data imputation techniques in 15 Solar Radiation Resource Assessment (SRRA) stations, located at different parts of India. During any measurement campaign, due to various technical reasons, we may miss few observations. However, the missing observation often reduce the performance of any forecasting model. Therefore, suitable data imputation method would assist us to obtain continuous observation of solar radiation. A station-by-station and method-by-method analysis was carried out to understand the performance of each model. Based on our analysis, among all the data imputation methods, the Kalman data imputation method is better for Indian Weather condition. In addition, Kalman StructTS, Linear, Stine and Arima methods yield slightly inferior accuracy compared to Kalman, but outperform the other methods. The extended solar radiation data are used by solar forecasting models to provide the prediction of solar radiation at 15 SRRA stations. As far as short term forecasting model is concerned, the PSPI model outperforms the Smart Persistence model. However, the forecast error is increases with the forecasting horizon.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2024

Offshore application of landslide susceptibility mapping using gradient-boosted decision trees: a Gulf of Mexico case study

Abstract Among natural hazards occurring offshore, submarine landslides pose a significant risk to offshore infrastructure installations attached to the seafloor. With the offshore being important for current and future energy production, there is a need to anticipate where future landslide events are likely to occur to support planning and development projects. Using the northern Gulf of Mexico (GoM) as a case study, this paper performs Landslide Susceptibility Mapping (LSM) using a gradient-boosted decision tree (GBDT) model to characterize the spatial patterns of submarine landslide probability over the United States Exclusive Economic Zone (EEZ) where water depths are greater than 120 m. With known spatial extents of historic submarine landslides and a Geographic Information System (GIS) database of known topographical, geomorphological, geological, and geochemical factors, the resulting model was capable of accurately forecasting potential locations of sediment instability. Results of a permutation modelling approach indicated that LSM accuracy is sensitive to the number of unique training locations with model accuracy becoming more stable as the number of training regions was increased. The influence that each input feature had on predicting landslide susceptibility was evaluated using the SHapely Additive exPlanations (SHAP) feature attribution method. Areas of high and very high susceptibility were associated with steep terrain including salt basins and escarpments. This case study serves as an initial assessment of the machine learning (ML) capabilities for producing accurate submarine landslide susceptibility maps given the current state of available natural hazard-related datasets and conveys both successes and limitations.

Dyer, Alec S. (ORCID:0000000219813904)↗
DOE OSTI2025

Geothermal well testing pressure prediction by using a hybrid transformer model system: FORGE well use case

Geothermal has huge potential to become an indispensable component in achieving the goal of sustainable energy economy, given its capability to provide consistent baseload power to the electric grid. Injection tests are crucial in geothermal energy system as they naturally help to evaluate reservoir properties, understand fluid flow and even enhance reservoir performance. In this research, we developed a hybrid model system that integrates machine learning (ML) regression, a physics-based mathematical model, and transformer deep learning. Trained and validated using FORGE injection test dataset, this system can forecast the pressure variations both upward and downward over time. The pressure prediction achieved prediction accuracy within 3-6% variance of true pressure values. The system can significantly save time and reduce costs by testing only a few cycles and then using model predictions for further analysis, instead of conducting additional real injection cycle tests. The developed model system also holds promise for designing injection test processes and maintaining well production in geothermal energy. Presented at the IMAGE ‘25 Conference led by Shell.

FORGE↗
DOE OSTI2021

Predicting concrete compressive strength using hybrid ensembling of surrogate machine learning models

This study aims to implement a hybrid ensemble surrogate machine learning technique in predicting the compressive strength (CS) of concrete, an important parameter used for durability design and service life prediction of concrete structures in civil engineering projects. For this purpose, an experimental database consisting of 1030 records has been compiled from the machine learning repository of the University of California, Irvine. The database was used to train and validate four conventional machine learning (CML) models, namely Artificial Neural Network (ANN), Linear and Non-Linear Multivariate Adaptive Regression Splines (MARS-L and MARS-C), Gaussian Process Regression (GPR), and Minimax Probability Machine Regression (MPMR). Subsequently, the predicted outputs of CML models were combined and trained using ANN to construct the Hybrid Ensemble Model (HENSM). It is observed that the proposed HENSM produces higher predictive accuracy compared to the CML models used in the present study. The predictive performance of all models for CS prediction was compared using the testing dataset and it is found that the HENSM model attained the highest predictive accuracy in both phases. Based on the experimental results, the newly constructed HENSM model is very potential to be a new alternative in handling the overfitting issues of CML models and hence, can be used to predict the concrete CS, including the design of less polluting and more sustainable concrete constructions.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Tracing and Forecasting Metabolic Indices of Cancer Patients Using Patient-Specific Deep Learning Models

We develop a patient-specific dynamical system model from the time series data of the cancer patient’s metabolic panel taken during the period of cancer treatment and recovery. The model consists of a pair of stacked long short-term memory (LSTM) recurrent neural networks and a fully connected neural network in each unit. It is intended to be used by physicians to trace back and look forward at the patient’s metabolic indices, to identify potential adverse events, and to make short-term predictions. When the model is used in making short-term predictions, the relative error in every index is less than 10% in the L ∞ norm and less than 6.3% in the L 1 norm in the validation process. Once a master model is built, the patient-specific model can be calibrated through transfer learning. As an example, we obtain patient-specific models for four more cancer patients through transfer learning, which all exhibit reduced training time and a comparable level of accuracy. This study demonstrates that this modeling approach is reliable and can deliver clinically acceptable physiological models for tracking and forecasting patients’ metabolic indices.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2021

Gaussian Process Regression for Aggregate Baseline Load Forecasting

Demand response (DR) is one of the most effective ways to maintain the reliability and improve the flexibility of power systems. Accurate forecasts of baseline loads are essential for DR programs. In the era of big data, machine learning-based approaches present a unique opportunity for baseline load forecasting. Thus, this paper presents a machine learning-based approach using a relatively less explored algorithm, Gaussian process regression (GPR), to forecast aggregate baseline loads. As such, a dataset was generated using a set of EnergyPlus simulations. Using the generated dataset, a GPR-based forecasting model was developed. In addition, support vector regression (SVR)-, artificial neural network (ANN)-, and averaging-based models were developed as baseline models for comparison. These models were compared in terms of accuracy, simplicity, and integrity. The prediction performance of the models showed that the GPR-based model is more accurate and reliable than the others. Such high performance shows the potential of the GPR in baseline load forecasting. GPR, therefore, can be used for DR applications.

Amasyali, Kadir↗
DOE OSTI2022

Convolutional neural networks for intra-hour solar forecasting based on sky image sequences

Accurate and timely solar forecasts play an increasingly critical role in power systems. Compared to longer forecasting timescales, very short-term solar forecasting has lagged behind in both research and practice. In this paper, we propose deep convolutional neural networks (CNNs) to provide operational intra-hour (10-minute-ahead to 60-minute-ahead) solar forecasts. We develop two CNN structures inspired by a widely-used CNN architecture. The CNNs are tailored to our solar forecasting regression tasks and rely solely on sky image sequences. Case studies based on six years of data (over 150,000 data points) demonstrate that the best CNN model has forecast skill scores of 20%-39% over the naive persistence of cloudiness benchmark, even at these very short timescales. The CNNs also have consistently superior performance when compared to shallow machine learning models with meteorological predictors, where the improvement averages around 7%. The sensitivity analyses show that the sky image length, resolution, and weather conditions have impacts on the deep learning model accuracy. In our intra-hour problem with specific setups, two sky images with a 10-minute 128 x 128 resolution yield the most accurate forecasts. Current limitations, future work, and deployment challenges and solutions are also discussed.

14 SOLAR ENERGY↗
DOE OSTI2026

NLML: A Deep Neural Network Emulator for the Exact Nonlinear Interactions in a Wind Wave Model

Nonlinear wave interactions describe the resonant energy transfer between wave components, playing a fundamental role in the evolution of ocean wave spectra. Nonlinear wave interactions significantly influence wave growth and development, making them essential for accurate wave modeling. However, resolving the full six-dimensional Boltzmann integral of the exact nonlinear wave interactions (Webb-Resio-Tracy method, WRT) is computationally expensive, limiting its application in real-time operational wave forecasting and for research purposes. Current approximations, such as the Discrete Interaction Approximation (DIA), prioritize computational speed over accuracy, resulting in significant errors in wave mean parameters. Here, we introduce NLML, a machine learning (ML) emulator designed to approximate the exact nonlinear wave interactions within WAVEWATCH III (WW3), with the goal of achieving the accuracy of WRT while maintaining the stability and computational speed of DIA. By leveraging GPU capabilities such as half precision inference, we achieved substantial speedups, up to 136x mathematical equation faster than the WRT and only a modest 1.04x mathematical equation slowdown relative to DIA, while achieving 2x mathematical equation the accuracy of DIA in global wave spectral energy and mean wave parameters, with up to 7x mathematical equation higher accuracy in some regions. Unlike previous ML approaches, NLML maintained inherent stability throughout model integration in a standalone, year-long WW3 simulation, without requiring additional constraints. Our new ML parameterization bridges the gap between accuracy and efficiency, offering a promising alternative for improving wave modeling in operational settings and research purposes.

16 TIDAL AND WAVE POWER↗
DOE OSTI2022

Multi-Fidelity Stochastic Economic Dispatch for Operating Low Carbon Power Grids

Grid operators can address the inherently stochastic nature of renewables by solving a two-stage stochastic programming model that minimizes the cost of dispatch decisions while accounting for the complex grid dynamics. It is common to use a sample average approximation to estimate the expectation of the second stage costs in this model. However, the large sample count needed for numerical accuracy makes effective modeling large-scale electric grids computationally intractable. We introduce a control variate multi-fidelity estimator for the second-stage recourse that enables high quality dispatch decisions in real-time with a reduced computational burden. We obtain a hierarchy of model fidelities by linearizing the AC power flow system representation to DC power flow, and by relaxing transmission and voltage network constraints. We evaluate the performance of our proposed method on a synthetic grid with 73 buses against a deterministic baseline with persistence forecast and a high-fidelity reference. Our analysis shows a computational speed-up of 7.62x with a minimal loss in accuracy. The multi-fidelity method is well suited to fidelity combinations that use a simplified network topology in their lower fidelity model and is an attractive option for applications where accurate grid modeling needed on a limited computational budget.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2022

Disruption Prediction and Avoidance in High Beta Long Pulse KSTAR Plasmas

This research is a concentrated and coordinated experimental and physics analysis program by a small U.S. team, focused on the topical area of “Long Pulse Control” as stated in the proposal solicitation. The overall goal of the research is to produce world-class analysis regarding forecasting and avoidance of disruptions in a long-pulse, superconducting tokamak at high stability parameters (e.g. normalized beta) with direct, quantifiable figures of merit (i.e. plasma disruption rate - “disruptivity”) and its reduction based on improved stability physics and control. The research targets the backbone of the problem – the basic, main issues – to produce this quantitative analysis by: (1) Creating and applying a new physics-based paradigm to understand how to eliminate disruptions: disruption event characterization, and event chain analysis – a direct manifestation of the approach suggested in the DOE Transients Report, (2) Ensuring that the components required to produce high-quality measurements and analysis needed for disruption event characterization for the tokamak in question (here, KSTAR) exist (e.g. pressure profile peaking, high-accuracy pitch angle data, compensated low frequency toroidal mode number n = 1 mode measurement) to create a primarily physics-based disruption forecasting capability, and (3) Addressing disruption avoidance for the first time in KSTAR by attacking primary disruption ‘event chains’, utilizing the most important associated plasma measurements (dynamic n = 1 field). The focus of this work will be avoidance of macroscopic instabilities, especially global modes.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2026

PopGNN: Graph Neural Network-Based Flexible Future Population Forecasting Model

Accurate population forecasts is important to plan critical infrastructure and services, from housing and education to healthcare and transport. However, traditional population prediction studies have only employed traditional machine learning models limited to capture complex spatial interdependencies and patterns. Althogh recently computer vision-based framework was introduced with with promising accuracy, it has critical limitations for real-world planning applications: it function only at fixed spatial resolutions, restricting their use in diverse boundaries such as census tracts, neighborhoods, or administrative zones. Therefore, this study suggests a Graph Neural Network (GNN)-based population prediction framework, called PopGNN. This model recorded remarkable performance compared with state-of-the-art models and traditional baseline models in the grid and administrative boundaries. Furthermore, our framework achieved comparable predictive accuracy to a computer vision-based model in both the South Korea and Tennessee case studies. Consequently, this study is valuable in that a single model can provide accurate population forecasts that address diverse planning demands, ranging from granular grid-level estimates for precise service allocation and facility location planning to aggregate administrative-level forecasts for macro-scale regional policy and resource distribution.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

Bayesian Physics Informed Spatio-Temporal Network for Streamflow Data Imputation

Reliable reconstruction of incomplete streamflow records is critical for improving hydrological forecasting, flood preparedness, and water resource management. However, large observational gaps and uncertainties in governing physical parameters limit the accuracy of traditional statistical and machinelearning imputation frameworks. To address these challenges, we develop a Bayesian Physics-Informed Spatio-Temporal Network (BPI-STNet) that jointly captures spatial and temporal dependencies while enforcing hydrologic consistency through embedded physical constraints. The framework integrates a GraphSAGE-LSTM architecture to model spatial connectivity across gauges and temporal flow dynamics, coupled with a Bayesian update mechanism to estimate uncertain parameters in a simplified water-balance framework. Unlike conventional physics-informed networks that rely on sampling-based posterior estimation, BPI-STNet derives an analytic solution to the inverse problem, allowing closed-form Bayesian updates of uncertain parameters Λ={α,β,k} using Gaussian priors and likelihoods. Applied to daily observations from the Susquehanna River Basin (1980-2022), BPI-STNet achieves substantial improvements over a purely data-driven RGNN baseline, which reduced RMSE by 23 % and MAE by 9 %, and achieving an average NSE values up to 0.96. The results demonstrate that coupling Bayesian inference with physics-informed learning yields physically consistent, uncertainty-aware reconstructions that preserve the temporal persistence and statistical distribution of observed flows. The proposed framework establishes a generalizable paradigm for data-sparse hydrologic systems where both data fidelity and physical interpretability are essential.

Krishnan Kutty Ambika, Anukesh [ORNL] (ORCID:00000↗
DOE OSTI2021

Daily grass reference evapotranspiration with Meteosat Second Generation shortwave radiation and reference ET products

Here we assess the accuracy of estimating daily grass reference evapotranspiration (PM-ET o ) using daily shortwave radiation (R s ) and reference evapotranspiration (ET REF ) products provided by the Meteosat Second Generation (MSG) geostationary satellite delivered by the Satellite Applications Facility on Land Surface Analysis (LSA-SAF) framework. The accuracy of using reanalysis ERA5 shortwave radiation data (R s ERA5 ) provided by the European Center for Medium-Range Weather Forecasts (ECMWF) is also evaluated. The assessments were performed using observed weather variables at 37 weather stations distributed across continental Portugal, where climate conditions range from semi-arid to humid, and 12 weather stations located in Azores islands, characterized by humid, windy and often cloudy conditions. This study&#x27;s use of data from a variety of climate conditions contributed to a unique and innovative assessment of the usability of LSA-SAF and ERA5 products for ET o estimation. The first assessment focused on comparing LSA-SAF estimates of R s (R s LSA-SAF ) against ground stations (R s ground ). The results showed a good matching between the two R s data sets for continental Portugal but a tendency for R s LSA-SAF to under-estimate R s ground in the cloudy islands of Azores. ET o values computed using R s LSA-SAF data and observed temperature, humidity and wind speed (ET o LSA-SAF ) were then compared with PM-ET o estimates with ground-based data, which were used as benchmark; input data of temperature and humidity needed for PM-ET o were quality checked for surface aridity effects. It was observed that ET o LSA-SAF is strongly correlated with PM-ET o (R 2 > 0.97) for most locations in continental Portugal, with regression coefficient of a linear regression forced to the origin ranging between 0.95 and 1.05, mean root mean square error (RMSE) of 0.13 mm d –1 , and Nash and Sutcliff efficiency of modeling (EF) above 0.95. For most Azores locations, ET o LSA-SAF over-estimated PM-ET o . This is likely a consequence of the high spatio-temporal heterogeneity of weather conditions that occur in these oceanic islands together with the different footprints of satellite (averaged over the pixel) and station observations. Reanalysis ERA5 shortwave radiation data presented similar behavior to the LSA-SAF products, however with slightly lower accuracy. The daily LSA-SAF ET REF product (ET REF LSA-SAF ) was assessed and results have shown a good accuracy of this product, with acceptable RMSE and high EF values, for continental Portugal but a low accuracy for the Azores islands. A simplified bias correction approach was shown to improve both ET o derived from the LSA-SAF products, namely for Azores stations, which seem to be representative of smaller areas. The use of the FAO-PM temperature approach (PMT) was also assessed using the R s LSA-SAF and R s ERA5 data, which showed a superiority of the LSA-SAF product for ET o estimations (ET o PMT LSA-SAF ). No significant differences (p < 0.05) were observed in terms of the median value of the RMSE when adopting ET o PMT and ET REF LSA-SAF . Differently, results showed that using the R s LSA-SAF in the PMT approach (ET o PMT LSA-SAF ) produces significantly better RMSE results than ET o PMT and ET REF LSA-SAF . Overall, the performed assessment allows concluding that the use of R s LSA-SAF, and to a lesser extent the use of the R s ERA5 , highly improves the accuracy of computation of ET o when R s observations are not available, including when only temperature data are accessible. The use of the ET REF LSA-SAF product is a good alternative when observed weather data are not available.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Prediction of carbon nanostructure mechanical properties and the role of defects using machine learning

Graphene-based nanostructures hold immense potential as strong and lightweight materials, however, their mechanical properties such as modulus and strength are difficult to fully exploit due to challenges in atomic-scale engineering. This study presents a database of over 2,000 pristine and defective nanoscale CNT bundles and other graphitic assemblies, inspired by microscopy, with associated stress–strain curves from reactive molecular dynamics (MD) simulations using the reactive INTERFACE force field (IFF-R). These 3D structures, containing up to 80,000 atoms, enable detailed analyses of structure-stiffness-failure relationships. By leveraging the database and physics- and chemistry-informed machine learning (ML), accurate predictions of elastic moduli and tensile strength are demonstrated at speeds 1,000 to 10,000 times faster than efficient MD simulations. Hierarchical Graph Neural Networks with Spatial Information (HS-GNNs) are introduced, which integrate chemistry knowledge. HS-GNNs as well as extreme gradient boosted trees (XGBoost) achieve forecasts of mechanical properties of arbitrary carbon nanostructures with only 3 to 6% mean relative error. The reliability equals experimental accuracy and is up to 20 times higher than other ML methods. Predictions maintain 8 to 18% accuracy for large CNT bundles, CNT junctions, and carbon fiber cross-sections outside the training distribution. The physics- and chemistry-informed HS-GNN works remarkably well for data outside the training range while XGBoost works well with limited training data inside the training range. The carbon nanostructure database is designed for integration with multimodal experimental and simulation data, scalable beyond 100 nm size, and extendable to chemically similar compounds and broader property ranges. The ML approaches have potential for applications in structural materials, nanoelectronics, and carbon-based catalysts.

Winetrout, Jordan J.↗