Search NASA⌕ Search

SEARCH · Search NASA

Results for “property predictions”

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 415 records · Page 23
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Steric and Electronic Influence of Excited-State Decay in Cu(I) MLCT Chromophores

For the past eleven years, a dedicated effort in our research group focused on fundamentally advancing the photophysical properties of cuprous bis-phenanthroline-based metal-to-ligand charge transfer (MLCT) excited states. We rationalized that by gaining control over the numerous factors limiting the more widespread use of Cu I MLCT photosensitizers, they would be readily adopted in numerous light-activated applications given the earth-abundance of copper and the extensive library of 1,10-phenanthrolines developed over the last century. Significant progress has been achieved by recognizing valuable structure-property concepts developed by other researchers in tandem with detailed ultrafast and conventional time-scale investigations, in-silico-inspired molecular designs to predict spectroscopic properties, and applying novel synthetic methodologies. Ultimately, we achieved a plateau in exerting cooperative steric influence to control Cu I MLCT excited state decay. This led to combining sterics with π-conjugation and/or inductive electronic effects to further exert control over molecular photophysical properties. The lessons gleaned from our studies of homoleptic complexes were recently extended to heteroleptic bis(phenanthrolines) featuring enhanced visible light absorption properties and long-lived room-temperature photoluminescence. Furthermore, this Account navigates the reader through our intellectual journey of decision-making, molecular and experimental design, and data interpretation in parallel with appropriate background information related to the quantitative characterization of molecular photophysics using Cu I MLCT chromophores as prototypical examples.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

Comparative Studies of the Structural and Transport Properties of Molten Salt FLiNaK Using the Machine-Learned Neural Network and Reparametrized Classical Forcefields

Despite surging interest in molten salt reactors and thermal storage systems, knowledge of the physicochemical properties of molten salts are still inadequate due to demanding experiments that require high temperature, impurity control, and corrosion mitigation. Therefore, the ability to predict these properties for molten salts from first-principles computations is urgently needed. Herein, we developed and compared a machine-learned neural network force field (NNFF) and a reparametrized rigid ion model (RIM) for a prototypical molten salt LiF–NaF–KF (FLiNaK). We found that NNFF was able to reproduce both the structural and transport properties of the molten salt with first-principles accuracy and classical-MD computational efficiency. Furthermore, the correlation between the local atomic structures and the dynamics was identified by comparing with RIMs, suggesting the significance of polarization of anions implicitly embedded in the NNFF. Furthermore, this work demonstrated a computational framework that can facilitate the screening of molten salts with different chemical compositions, impurities, and additives, and at different thermodynamic conditions suitable for the next-generation nuclear reactors and thermal energy storage facilities.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Multitude Characterization and Prediction of DOE Advanced Biofuels Properties

Advanced multitude of experiments ranging from the liquid fuel to combustion are conducted on Co-OPTIMA fuels to aid the characterization of the fuels. The series of targeted experiments characterized Co-OPTIMA fuel spray atomization, flame topology, flame speed, autoignition, volatility, viscosity, soot/coking, and compatibility. The fuels are selected and prioritized based on input from national lab members. The research characterized and predicted biomass-based, low greenhouse gas fuels and blends combustion, autoignition, and physical properties of mixtures of identified compounds at engine-relevant conditions, in particular those properties that blend non-linearly. The main challenge of Co-OPTIMA is the evaluation of a variety of biofuels and blends in all the reaction conditions that might be encountered in new high-efficiency engines. Despite the improved high-throughput experimental techniques, it seems unlikely that all of the performance metrics could be measured for all relevant petroleum derived, bio-derived molecules and mixtures, and reaction conditions. A series of targeted experiments ranging from the liquid fuel to the combustion process is required, and to extract the maximum information from each experiment. These targeted experiments evaluated how a specific fuel will perform in an engine. The series of targeted experiments are as follows: (a) Spray Atomization, Vaporization and Droplet Formation (b) Combustion Flame and Local Fuel/Air Image-Based Measurements (c) Laminar Flame Speed Measurements (d) Autoignition and Soot Measurements (e) Synchrotron Coupled Fundamental Autoignition Experiments (f) Fuel Coking and Hot Surface Deposit g) Fuel Volatility Measurements (h) Viscosity Measurements (i) Seal Flexible Fuel Compatibility These experimental processes provide an essential pathway for the prediction of fuel behaviors in engines and systematic process for fuel down select.

09 BIOMASS FUELS↗
DOE OSTI2023

Building materials genome from ground‐state configuration to engineering advance

Individual phases are commonly considered as the building blocks of materials. However, the accurate theoretical prediction of properties of individual phases remains elusive. The top-down approach by decoding genomic building blocks of individual phases from experimental observations is nonunique. The density functional theory (DFT), as a state-of-the-art solution of quantum mechanics, prescribes the existence of a ground-state configuration at 0 K for a given system. It is self-evident that the ground-state configuration alone is insufficient to describe a phase at finite temperatures as symmetry-breaking non-ground-state configurations are excited statistically at temperatures above 0 K. Our multiscale entropy approach (recently terms as Zentropy theory) postulates that the entropy of a phase is composed of the sum of the entropy of each configuration weighted by its probability plus the configurational entropy among all configurations. Consequently, the partition function of each configuration in statistical mechanics needs to be evaluated by its free energy rather than total energy. The combination of the ground-state and symmetry-breaking non-ground-state configurations represents the building blocks of materials and can be used to quantitatively predict free energy of individual phases with the free energy of each configuration predicted from DFT as well as all properties derived from free energy of individual phases.

CALPHAD↗
DOE OSTI2021

Elevating density functional theory to chemical accuracy for water simulations through a density-corrected many-body formalism

Abstract Density functional theory (DFT) has been extensively used to model the properties of water. Albeit maintaining a good balance between accuracy and efficiency, no density functional has so far achieved the degree of accuracy necessary to correctly predict the properties of water across the entire phase diagram. Here, we present density-corrected SCAN (DC-SCAN) calculations for water which, minimizing density-driven errors, elevate the accuracy of the SCAN functional to that of “gold standard” coupled-cluster theory. Building upon the accuracy of DC-SCAN within a many-body formalism, we introduce a data-driven many-body potential energy function, MB-SCAN(DC), that quantitatively reproduces coupled cluster reference values for interaction, binding, and individual many-body energies of water clusters. Importantly, molecular dynamics simulations carried out with MB-SCAN(DC) also reproduce the properties of liquid water, which thus demonstrates that MB-SCAN(DC) is effectively the first DFT-based model that correctly describes water from the gas to the liquid phase.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Towards atom-level understanding of metal oxide catalysts for the oxygen evolution reaction with machine learning

Abstract Green hydrogen production is crucial for a sustainable future, but current catalysts for the oxygen evolution reaction (OER) suffer from slow kinetics, despite many efforts to produce optimal designs, particularly through the calculation of descriptors for activity. In this study, we develop a dataset of density functional theory calculations of bulk and surface perovskite oxides, and adsorption energies of OER intermediates, which includes compositions up to quaternary and facets up to (555). We demonstrate that per-site properties of perovskite oxides such as Bader charge or band center can be tuned through element substitution and faceting, and develop a machine learning model that accurately predicts these properties directly from the local chemical environment. We leverage these per-site properties to identify promising perovskites with high theoretical OER activity. The identified design principles and promising materials provide a roadmap for closing the gap between current artificial catalysts and biological enzymes such as photosystem II.

Chemistry↗
DOE OSTI2020

Elastic Properties of Plasma-Exposed Tungsten Predicted by Molecular-Dynamics Simulations

We report results of systematic molecular-dynamics computations of the elastic properties of single-crystalline tungsten containing structural defects, voids and overpressurized He nanobubbles, related to plasma exposure of tungsten serving as a plasma- facing component (PFC) in nuclear fusion devices. Our computations reveal that the empty voids are centers of dilatation resulting in development of tensile stress in the tungsten matrix, whereas He-filled voids (nanobubbles) introduce compressive stress in the plasma-exposed tungsten. We find that the dependence of the elastic moduli of plasma-exposed tungsten, namely, the bulk, Young, and shear modulus, on its void fraction follows a universal exponential scaling relation. We also find that the elastic moduli of plasma-exposed tungsten soften substantially as a function of He content in the tungsten matrix, following an exponential scaling relation; this He-induced exponential softening is in addition to the softening caused in the matrix with increasing temperature. A systematic characterization of the dependence of the elastic moduli on the He bubble size reveals that He bubble growth affects significantly both the bulk modulus and the Poisson ratio of plasma-exposed tungsten, while its effect on the Young and shear moduli of the plasma-exposed material is weak. Our findings contribute directly to the development of a structure-properties database that is required for the predictive modeling of the dynamical response of PFCs in nuclear fusion devices.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Utilization of Synthetic Near-Infrared Spectra via Generative Adversarial Network to Improve Wood Stiffness Prediction

Near-infrared (NIR) spectroscopy is widely used as a nondestructive evaluation (NDE) tool for predicting wood properties. When deploying NIR models, one faces challenges in ensuring representative training data, which large datasets can mitigate but often at a significant cost. Machine learning and deep learning NIR models are at an even greater disadvantage because they typically require higher sample sizes for training. In this study, NIR spectra were collected to predict the modulus of elasticity (MOE) of southern pine lumber (training set = 573 samples, testing set = 145 samples). To account for the limited size of the training data, this study employed a generative adversarial network (GAN) to generate synthetic NIR spectra. The training dataset was fed into a GAN to generate 313, 573, and 1000 synthetic spectra. The original and enhanced datasets were used to train artificial neural networks (ANNs), convolutional neural networks (CNNs), and light gradient boosting machines (LGBMs) for MOE prediction. Overall, results showed that data augmentation using GAN improved the coefficient of determination (R 2 ) by up to 7.02% and reduced the error of predictions by up to 4.29%. ANNs and CNNs benefited more from synthetic spectra than LGBMs, which only yielded slight improvement. All models showed optimal performance when 313 synthetic spectra were added to the original training data; further additions did not improve model performance because the quality of the datapoints generated by GAN beyond a certain threshold is poor, and one of the main reasons for this can be the size of the initial training data fed into the GAN. LGBMs showed superior performances than ANNs and CNNs on both the original and enhanced training datasets, which highlights the significance of selecting an appropriate machine learning or deep learning model for NIR spectral-data analysis. The results highlighted the positive impact of GAN on the predictive performance of models utilizing NIR spectroscopy as an NDE technique and monitoring tool for wood mechanical-property evaluation. Further studies should investigate the impact of the initial size of training data, the optimal number of generated synthetic spectra, and machine learning or deep learning models that could benefit more from data augmentation using GANs.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2020

Consistency-Enhanced Evolution for Variable Selection Can Identify Key Chemical Information from Spectroscopic Data

In the last few decades, spectroscopic techniques such as near-infrared (NIR) spectroscopy have gained wide applications in several industries, such as the pharmaceutical, agricultural, oil, and gas industries. As a result, various soft sensors have been developed to predict sample properties from spectroscopic readings. Because the spectroscopic readings at different wavelengths, especially at the adjacent wavelengths, are highly correlated, it has been shown that variable selection could significantly improve a soft sensor’s prediction performance while reducing the model complexity. To improve the prediction performance, most variable selection methods focus on identifying the variables (i.e., wavelengths or wavelength segments) that are strongly correlated with the dependent variable. Although many successful applications have been reported, these variable selection methods do have their limitations. Specifically, the selected wavelengths sometimes show little connection to the chemical bounds or functional groups presenting in the sample. In addition, the selected variables can be quite sensitive to the choice of the training samples. In this work, we address these limitations from a different perspective: if a variable selection algorithm can identify the truly relevant input variables, it should consistently identify the same subset of variables regardless of the choice of the training samples. Therefore, we propose a variable selection method that aims to improve the consistency of variable selection resulting from different training samples. Furthermore, the new algorithm is termed consistency-enhanced evolution for variable selection (CEEVS). To demonstrate the performance and robustness of CEEVS, we compare the proposed method with three representative variable selection methods using five published NIR data sets. These case studies clearly demonstrate that by improving the variable selection consistency, we can not only achieve improved prediction performance, but also identify key chemical information from spectroscopic data.

42 ENGINEERING↗
DOE OSTI2020

Experimental Study of the Fundamental Properties of Warm Dense Mixtures

The aim of the proposed research was to provide a multi-scale study of the properties of warm dense hydrocarbons, by studying the thermodynamic properties through the equation of state and the microscopic properties by x-ray scattering. Understanding the fundamental properties of warm dense mixtures is an intellectual challenge due to the complexity of the system. Unlike liquids or gases, where constituent particle interactions occur through collisions or bonding between valence electrons, in these strongly coupled systems the atoms will be partially ionized and compressed so tightly together that interactions between the inner core electrons can play a role in the systems chemistry. Advances in computational capabilities and development of new theoretical models have been used to predict the properties of mixtures but there is currently no experimental data of the fundamental interaction between the particles in mixtures to test these predictions against. The proposed research was to experimentally investigate the interaction of the different species in hydrocarbons by measuring the compressibility of substances with different carbon and hydrogen ratios and the complexity of the microscopic interactions through elastic and inelastic x-ray scattering. To study the bulk properties of warm dense hydrocarbons we established the Warm Dense Matter Research Laboratory (WDMRL) in the Institute for Shock Physics at Washington State University. The goal was for experiments in the WDMRL to determine shock loading conditions of interest in the hydrocarbon mixtures. We planned on using Hugoniot EOS measurements and a range of carbon and hydrogen concentrations to determine conditions when the EOS of the mixture varied significantly from that of the classical mixing model. Even though shock transit measurements through aluminum foils suggested pressures upto 400GPa, the experiments in the WDMRL were unsuccessful in getting usable shockwave compression data above 100GPa in polystyrene which was below the pressure of interest for hydrocarbon mixtures (>200GPa). To complete the project, we tested a technique using x-ray phase contrast imaging to map the location of tracer layers in a test sample of polycarbonate to record the material motion in dynamically compressed samples. A technique that will be useful for future warm dense matter experiments. These experiments used <200nm gold layers in polycarbonate samples to measure the material velocity and shock speed using x-ray phase contrast imaging. These results were compared to continuum surface measurements performed at ISP and show that the tracer layer technique can measure hydrodynamic properties accurately in dynamically compressed materials.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2026

The DREAMS Project: Disentangling the Impact of Halo-to-halo Variance and Baryonic Feedback on Milky Way Satellite Galaxies

We analyze the properties of satellite galaxies around 1024 Milky Way–mass hosts from the DREAMS Project, simulated within a ΛCDM cosmology. Utilizing the TNG galaxy-formation model, the DREAMS simulations incorporate both baryonic physics and cosmological uncertainties for a large sample of galaxies with diverse environments and formation histories. We investigate the relative impact of the physical uncertainty from the galaxy-formation model on predicted satellite properties using four metrics: the satellite stellar mass function, radial distribution, inner slope of the dark matter density profile, and stellar half-light radius. We compare these predictions to observations from the SAGA survey and the DREAMS N-body simulations and find that uncertainties from baryonic physics modeling are subdominant to the scatter arising from halo-to-halo variance. Where baryonic modeling does affect satellites, the supernova wind energy has the largest effect on the satellite properties that we investigate. Specifically, increased supernova wind energy suppresses the stellar mass of satellites and results in more extended stellar half-light radii. The adopted wind speed has only a minor impact, and other astrophysical and cosmological parameters show no measurable effect. Our findings highlight the robustness of satellite properties against uncertainties in baryonic physics modeling.

Rose, Jonah C. [Princeton University, NJ (United S↗
DOE OSTI2022

A Multi-Scale Computational Platform for Predictive Modeling of Corrosion in Al-Steel Joints (Final Report)

The research team proposed to develop innovative multi-scale models to predict corrosion and the resulting mechanical performances in aluminum-steel joints. The methods of joining considered are resistance spot welding, self-piercing riveting, and rivet-welding, all suitable for mass production applications. The multi-scale models integrate high throughput first-principle calculations based on density functional theory (DFT), high throughput calculation of phase diagrams (CALPHAD) modeling, and finite element method (FEM) simulations. These models are to be validated through laboratory experiments. Furthermore, the models are available as open source so as to enable scientists and engineers in the community to adapt and contribute to the development and application. The approaches rely on the research team’s extensive experience on the prediction of properties of individual phases at finite temperatures and variable compositions through DFT calculations, and our broad expertise on dissimilar material joining and their corrosion. The proposed computational framework enables high throughput computations for improved predictions of corrosion and the associated mechanical performance in dissimilar material joints, resulting in significant reduction in computational time needed by the current state-of-the-art methods. With the participation of researchers from three universities, an auto manufacturer, two manufacturing technology/equipment suppliers, and a software developer/vendor, the interdisciplinary research team applies the technical development on both phase-based modeling and laboratory experiments into the automobile body joining processes for validation and technology demonstration. The global cost of corrosion was estimated at about 3.4% of the global GDP in 2013. By using available corrosion control practices, it is estimated a saving between 15-35% of the cost of corrosion. In the U.S., more than $276 billion is spent repairing corrosion damage. Prediction of the corrosion and its impact on performance of the dissimilar material joints is critical for reducing the massive number of the current corrosion-based recalls for automobiles. Thus, the project goal is to develop models to enable predictive maintenance and end-of-life planning of multi-metal joints with risk of corrosion under different conditions such as exposure to high temperatures in summer and salt solutions in winter, quantified through its pH. An academia-industry consortium led by the University of Michigan and including Pennsylvania State University, University of Illinois Urbana-Champaign, University of Georgia, General Motors Company, Livermore Software Technology Corporation, and Optimal Process Technologies, LLC. created multi-scale models for prediction of corrosion in aluminum-steel joint structures such of them used in vehicle subassemblies – chassis and transmission systems. Starting from the first principle calculations, the team developed mathematical and data-driven models to predict the metallic components, which are formed during joining of two metals, for example aluminum and steel - a lightweight multilateral system which is currently used in more than 60% car bodies. These models were used for simulating chemical reactions that are happening when the joining metallic components are exposed to high temperatures and different pH values. The team was able to predict how the corrosion installs on the metallic components and how they lead to a sudden failure of components in cars. Newly developed machine learning algorithms combining Science, Technology, Engineering and Math disciplines, advanced finite element simulation and experimental validations have been integrated in a platform for prediction of the corrosion evolution and prediction the failure of joints under mechanical loadings and fatigue. Moreover, based on machine learning and inverse analysis, the team proposed solutions for designing new metallic alloys less susceptible to corrosion when joining multi-material assembles. An average of 4% error compared with experiments was achieved for the most common joints that are used in vehicle subassemblies.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Lessons Learned in Employing Data Analytics to Predict Oxidation Kinetics and Spallation Behavior of High-Temperature NiCr-Based Alloys

Machine learning (ML) can offer many advantages in predicting material properties over traditional materials development methods based solely on limited experimental investigations or physical-based simulations with the capability to reduce development cost, risk, and time. However, so far, limited efforts have been made to predict alloy oxidation kinetics and spallation behavior via ML due to the lack of consistently measured and sufficient experimental data and the inherent complexity in oxidation behavior of multicomponent high-temperature alloys. A previous study reported the ability of ML to predict oxidation kinetics of NiCr-based alloys as a function of alloy composition and operating conditions. Here, the performance of a ML model in predicting rate constants and spallation probability was evaluated in light of the roles of the data distribution of the experimental dataset (data analytics), the alloy composition, the exposure environment and the chosen oxidation approach to extracting kinetic values from the measured mass changes (but using either a simple parabolic law or a statistical cyclic oxidation model). Potential strategies to improve the predictions and enhance the extrapolative capability of the previously trained model will be discussed.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Iron telluride ladder compounds: Predicting the structural and magnetic properties of BaFe 2 Te 3

Since the discovery of pressure-induced superconductivity in the two-leg ladder system BaFe 2 X 3 (X=S, Se), with the 3d iron electronic density n=6, quasi-one-dimensional iron-based ladders have attracted considerable attention. In this work, we use density-functional theory to predict that the novel n=6 iron ladder BaFe 2 Se 3 could be stable with a similar crystal structure as BaFe 2 Se 3 . Furthermore, our results also indicate that BaFe2Te3 will display a complex 2×2 block-type magnetic order. Due to the magnetic striction effects of this block order, BaFe2Te3 should be a magnetic noncollinear ferrielectric system with a net polarization 0.31μC/cm2. In summary, the similar electronic density and magnetic ground state of Te- and Se-based ladders indicates both should display similar properties. In particular, the physical and structural similarity with BaFe 2 Se 3 suggests that BaFe 2 Se 3 could become superconducting under high pressure.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2024

Alternative CNDOL Fockians for fast and accurate description of molecular exciton properties

CNDOL is an a priori, approximate Fockian for molecular wave functions. In this study, we employ several modes of singly excited configuration interaction (CIS) to model molecular excitation properties by using four combinations of the one electron operator terms. Those options are compared to the experimental and theoretical data for a carefully selected set of molecules. The resulting excitons are represented by CIS wave functions that encompass all valence electrons in the system for each excited state energy. The Coulomb–exchange term associated to the calculated excitation energies is rationalized to evaluate theoretical exciton binding energies. This property is shown to be useful for discriminating the charge donation ability of molecular and supermolecular systems. Multielectronic 3D maps of exciton formal charges are showcased, demonstrating the applicability of these approximate wave functions for modeling properties of large molecules and clusters at nanoscales. This modeling proves useful in designing molecular photovoltaic devices. Our methodology holds potential applications in systematic evaluations of such systems and the development of fundamental artificial intelligence databases for predicting related properties.

Chemistry↗
DOE OSTI2020

Probabilistic modeling and prediction of out-of-plane unidirectional composite lamina properties

Computational simulation provides an efficient means to predict the behavior of customized hybrid material configurations using validated, physics-based models. One limitation to this approach is the quality and quantity of available data to characterize the many constituent input properties. Therefore, a systematic approach to identify the most influential parameters on the hybrid behavior and quantify the corresponding uncertainty in predictive capabilities is required. Here, an approach using Bayesian multimodel inference and imprecise global sensitivity analysis is presented to investigate the effects of sparse constituent data on the prediction of composite material properties. The methodology allows the identification, using quantified uncertainties, of the most influential constituent material parameters for specified homogenized properties. This sensitivity analysis further enables a dimension reduction when assessing the influence of uncertainties on material properties and can be used to inform testing programs of the constituent properties that require additional testing/data collection in order to minimize uncertainty in macro-scale composite properties. The methodology is specifically demonstrated on the prediction and sensitivity analysis of out-of-plane mechanical properties of a unidirectional lamina.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Parton labeling without matching: unveiling emergent labelling capabilities in regression models

Parton labeling methods are widely used when reconstructing collider events with top quarks or other massive particles. State-of-the-art techniques are based on machine learning and require training data with events that have been matched using simulations with truth information. In nature, there is no unique matching between partons and final state objects due to the properties of the strong force and due to acceptance effects. We propose a new approach to parton labeling that circumvents these challenges by recycling regression models. The final state objects that are most relevant for a regression model to predict the properties of a particular top quark are assigned to said parent particle without having any parton-matched training data. This approach is demonstrated using simulated events with top quarks and outperforms the widely used $χ$ 2 method.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
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.↗