Search NASA⌕ Search

SEARCH · Search NASA

Results for “dependable computing”

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 505 records · Page 28
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Optical neural engine for solving scientific partial differential equations

Abstract Solving partial differential equations (PDEs) is the cornerstone of scientific research and development. Data-driven machine learning (ML) approaches are emerging to accelerate time-consuming and computation-intensive numerical simulations of PDEs. Although optical systems offer high-throughput and energy-efficient ML hardware, their demonstration for solving PDEs is limited. Here, we present an optical neural engine (ONE) architecture combining diffractive optical neural networks for Fourier space processing and optical crossbar structures for real space processing to solve time-dependent and time-independent PDEs in diverse disciplines, including Darcy flow equation, the magnetostatic Poisson’s equation in demagnetization, the Navier-Stokes equation in incompressible fluid, Maxwell’s equations in nanophotonic metasurfaces, and coupled PDEs in a multiphysics system. We numerically and experimentally demonstrate the capability of the ONE architecture, which not only leverages the advantages of high-performance dual-space processing for outperforming traditional PDE solvers and being comparable with state-of-the-art ML models but also can be implemented using optical computing hardware with unique features of low-energy and highly parallel constant-time processing irrespective of model scales and real-time reconfigurability for tackling multiple tasks with the same architecture. The demonstrated architecture offers a versatile and powerful platform for large-scale scientific and engineering computations.

Tang, Yingheng (ORCID:0009000153622546)↗
DOE OSTI2025

Designing molecular qubits: computational insights into first-row and group 6 transition metal complexes

In the realm of optically addressable qubits, a previously synthesized and characterized Cr( IV ) pseudo-tetrahedral complex, featuring four strongly donating ligands surrounding the chromium center, has demonstrated potential as a qubit candidate. This study proposes analogs of this complex through a metal substitution strategy, extending the investigation to different complexes based on metal centers selected from first-row and Group 6 transition metals. Computational modeling based on multiconfigurational methods CASPT2 and MC-PDFT was utilized to calculate energy gaps between ground and excited electronic spin states, and zero-field splitting parameters. Simulations were applied to each equilibrium geometry and related deformations based on vibrational modes. All results align with previous experimental findings, but also show that qubits based on V and Ti centers could be more electronically stable than the Cr one, suggesting a lower electronic features dependency from their related geometry. In some cases geometrical deformations provide changes in relative energy gaps between triplet and singlet excited state, that could potentially swap, offering a different initialization process, and some inspiration for ligand design based on such deformations. Additionally, this study identifies an unsynthesized Ti( II ) compound as a promising candidate for molecular qubits. This finding highlights the role of computational multireference methods in the rational design of qubit systems.

Sauza-de la Vega, Arturo [Univ. of Chicago, IL (Un↗
DOE OSTI2024

Exact enforcement of temporal continuity in sequential physics-informed neural networks

The use of deep learning methods in scientific computing represents a potential paradigm shift in engineering problem solving. One of the most prominent developments is Physics-Informed Neural Networks (PINNs), in which neural networks are trained to satisfy partial differential equations (PDEs). While this method shows promise, the standard version has been shown to struggle in accurately predicting the dynamic behavior of time-dependent problems. To address this challenge, methods have been proposed that decompose the time domain into multiple segments, employing a distinct neural network in each segment and directly incorporating continuity between them in the loss function of the minimization problem. In this work we introduce a method to exactly enforce continuity between successive time segments via a solution ansatz. This hard constrained sequential PINN (HCS-PINN) method is simple to implement and eliminates the need for any loss terms associated with temporal continuity. The method is tested for a number of benchmark problems involving both linear and non-linear PDEs. Examples include various first order time dependent problems in which traditional PINNs struggle, namely advection, Allen–Cahn, and Korteweg–de Vries equations. Furthermore, second and third order time-dependent problems are demonstrated via wave and Jerky dynamics examples, respectively. Notably, the Jerky dynamics problem is chaotic, making the problem especially sensitive to temporal accuracy. Finally, the numerical experiments conducted with the proposed method demonstrated superior convergence and accuracy over both traditional PINNs and the soft-constrained counterparts.

42 ENGINEERING↗
DOE OSTI2025

Pressure–Modulated Luminescence Enhancement and Quenching in a Hydrogen–Bonded Organic Framework

Light emission in the solid state is central for illumination, sensing, and imaging applications. Unlike luminescence in dilute solutions, where the excited states are unimolecular in nature, intermolecular interaction plays a significant role in the quantum yield of solid-state luminophores, manifested as competing aggregation-caused quenching (ACQ) and aggregation-induced enhancement (AIE). Both effects are extensively studied in various systems; however, it remains unclear how their competition depends on molecular conformation and intermolecular stacking. Here the direct observation of pressure-modulated AIE-ACQ competition in a crystalline hydrogen-bonded organic framework (HOF) is reported. Using in situ spectroscopies and computational modeling, the intramolecular vibration and intermolecular π–π stacking directly responsible for the non-radiative decay of the excited state are identified. The extent of these two contributions is modulated by hydrostatic pressure and guest molecules in the HOF pores. Furthermore, this work demonstrates a physically neat model system to understand and control solid-state luminescence, and a potential material platform for piezoluminescent sensing.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2026

Elastic strain engineering of lattice thermal conductivity of silicon: An ab-initio study

Silicon (Si) is the most essential material in the semiconductor industry. It is important to manage the thermal properties of crystalline Si. Elastic strain engineering (ESE) has proven to be an effective tool in controlling the electrical conductivity of Si in strained-silicon technology; its effects on the thermal conductivity of silicon, therefore, warrants careful investigation. The ESE effect is much more pronounced for nanostructured materials due to the ultralarge elastic strains (on the order of 10%) achievable at the nanoscale. In this work, the lattice thermal conductivity (κ L ) of Si under hydrostatic, biaxial, and uniaxial strain states is studied with ab-initio simulations, and the values of strain-dependent κ L compare well with experimental results and existing molecular dynamics simulations. To understand the mechanisms of strain-modulated κ L , the phonon bands, scattering rate, and Grüneisen parameters of phonon modes are computed. It is shown that strain can significantly change the anharmonicity of the crystal system, thus changing phonon scattering rates and κ L . Our results demonstrate that ESE can reduce silicon κ L by up to approximately 90%. Furthermore, uniaxial and biaxial strains can induce highly anisotropic thermal conductivity in Si, with relative variations up to 58.5% and 14.5%, respectively.

Anisotropy in Thermal Conductivity↗
DOE OSTI2026

Collins asymmetries for pion-in-jet production in polarized ℓp collisions at the EIC

We study Collins azimuthal asymmetries for pion-in-jet production in polarized lepton-proton collisions, extending previous analyses of polarized pp scattering to a complementary and theoretically simpler process. We keep adopting a simplified transverse momentum dependent (TMD) approach, with a collinear configuration for the initial state, and employ the transversity and Collins fragmentation functions as extracted from semi-inclusive deep inelastic scattering and $e^{+}e^{-}$ annihilation processes. We then compute azimuthal asymmetries for the Electron-Ion Collider (EIC) kinematics, both within a leading order (LO) approach and by including quasireal photon exchange in the Weizsäcker-Williams approximation. Although this contribution is relevant in the whole kinematical range explored, it does not spoil the dominance of quark-initiated channels, leaving only a marginal role to their gluon counterparts. In this respect, Collins asymmetries in lepton-proton processes allow for a much clearer access to the transversity distribution, including its sea-quark component. As we will argue, a comparison with future EIC data could represent a further step in testing the hypothesis of the universality of the Collins function as well as of the TMD factorization for this class of processes.

Azimuthal asymmetries↗
DOE OSTI2025

Learning to Simulate Aerosol Dynamics with Graph Neural Networks

Aerosol effects on climate, weather, and air quality depend on characteristics of individual particles, which are tremendously diverse and change in time. Particle-resolved models are the only models able to capture this diversity in particle physiochemical properties, and these models are computationally expensive. As a strategy for accelerating particle-resolved microphysics models, we introduce Graph-based Learning of Aerosol Dynamics (GLAD) and use this model to train a surrogate of the particle-resolved model PartMC-MOSAIC. GLAD implements a Graph Network-based Simulator (GNS), a machine learning framework that has been used to simulate particle-based fluid dynamics models. In GLAD, each particle is represented as a node in a graph, and the evolution of the particle population over time is simulated through learned message passing. Here, we demonstrate our GNS approach on a simple aerosol system that includes condensation of sulfuric acid onto particles composed of sulfate, black carbon, organic carbon, and water. A graph with particles as nodes is constructed, and a graph neural network (GNN) is then trained using the model output from PartMC-MOSAIC. The trained GNN can then be used for simulating and predicting aerosol dynamics over time. Results demonstrate the framework's ability to accurately learn chemical dynamics and generalize across different scenarios, achieving efficient training and prediction times. We evaluate the performance across four scenarios, highlighting the framework's robustness and adaptability in modeling aerosol microphysics and chemistry.

aerosol chemistry dynamics↗
DOE OSTI2026

Computational Design of Interlayers for Thermally Stable Compositionally Graded Coatings on Nickel Alloys

To extend the service life of Ni-based superalloys, refractory metal coatings are often used. However, direct bonding between metals with dissimilar crystal structure promotes brittle intermetallic phase formation. This work presents a computational thermodynamic framework for high throughput design of functionally graded interlayers to suppress deleterious phases that may form at the interlayer. The Thermo-Calc software package was used to screen candidate metallic interlayer elements based on stability of solid-solution phases. Vanadium was identified as a promising interlayer due to its consistent suppression of intermetallic phases. Temperature-dependent phase diagram mapping between 600 and 1000 °C guided selection of a compositional pathway that significantly reduced intermetallic formation compared to directly joining the Ni-based and Nb refractory alloys. Time–temperature–transformation analysis was performed to assess whether equilibrium-predicted phases are kinetically accessible along regions of the graded path where non-solid-solution phases are not fully suppressed. The methodology was further applied to additional Ni-based alloy and coating systems, illustrating its transferability as an approach for rapid computational design of graded interlayers in dissimilar high-temperature materials.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Bayesian calibration and uncertainty quantification of a rate-dependent cohesive zone model for polymer interfaces

In this work we present a rate-dependent cohesive zone model for the fracture of polymeric interfaces and performs a Bayesian calibration, an uncertainty quantification, and a sensitivity analysis for the model. The proposed cohesive zone model accounts for both reversible elastic and irreversible rate-dependent separation sliding deformation at the interface. The viscous dissipation due to the irreversible opening at the interface is modeled using elastic-viscoplastic kinematics that incorporates the effects of strain rate. Inverse calibration of parameters for such complex models through trial and error is challenging due to the large number of parameters of the model. Moreover, the calibrated parameter values are often non-unique and uncertain when the available experimental data is limited. To tackle this challenge, we employ a Bayesian calibration approach to identify parameters from experimental data, the resulting parameters significantly enhance the accuracy of the model. To quantify the uncertainty associated with the inverse parameter estimation, a modular Bayesian approach is employed to calibrate the unknown model parameters, accounting for the parameter uncertainty of the cohesive zone model. The advantages of the Bayesian calibration over a deterministic parameter fit are demonstrated. Further, to quantify the model uncertainties, such as incorrect assumptions or missing physics, a discrepancy function is introduced, which significantly improves the model’s prediction. Finally, the total uncertainty of the model is quantified in a predictive setting. A sensitivity analysis is performed to assess how changes in the input variables of the model affect the peak load, facilitating the identification of a concise set of highly influential parameters. The present approach can be used for calibration and uncertainty quantification for other complex computational mechanics models. It should also facilitate the designing of interface materials under uncertainty.

42 ENGINEERING↗
DOE OSTI2026

Optimization of Processing, Microstructure, and Hardness of an Al–Ce–Ni–Mn–Zr Alloy With Laser Additive Manufacturing

Here, this study examines the processing behavior, microstructure, surface roughness, and hardness properties of an aluminum alloy containing 8.2 Ce, 4.5 Ni, 0.5 Mn, and 0.7 Zr (wt%) fabricated using laser powder bed fusion. Sixty samples were produced across a range of laser powers, scan speeds, and hatch spacings to evaluate their effect on porosity, hardness, and microstructural features. Porosity was measured using X-ray computed tomography, while microstructure and surface roughness were characterized by scanning electron (SEM) and laser confocal microscopy. High dense and cracking-free Al–Ni–Ce alloy was successfully manufactured. Porosity showed a U-shaped dependence on energy input, increasing under both insufficient and excessive melting conditions. Hardness increased with cooling rate due to finer cellular structures and solute redistribution. A general statistical model was developed to capture the relationships between processing parameters and material response. Results identify a narrow processing window defined by laser powers between 350 and 370 W, scan speeds from 1400 to 1800 mm/s, and hatch distances between 0.14 and 0.18 mm. Within this window, porosity is minimized (below 0.01%) and hardness is maximized (up to 160 HV), demonstrating that careful control of these parameters enables dense, high strength aluminum components suitable for demanding structural applications.

Aluminum alloys↗
DOE OSTI2025

Supported Single‐Atom Manganese Catalysts for the Trimerization of Ethylene

Selective ethylene oligomerization via oxidative cyclization, forming metallacyclic intermediates, is typically catalyzed by molecular titanium and chromium complexes to produce butenes, hexenes, or octenes, depending on the supporting ligand framework. However, this mechanism requires significant electron density at the metal active site and is not known to be generalizable to other first-row transition metals. In this work, we computationally investigate the electronic modulation of five transition metals (Mn, Fe, Co, Ni, and Cu) supported on titania (TiO₂) through reductive lithium intercalation to promote selective oligomerization via oxidative cyclization, using density functional theory (DFT). Our findings predict that Mn/LiTiO₂ exhibits high catalytic activity due to the exergonic nature of oxidative cyclization with two ethylene molecules. Additionally, lithium titanate (LiTiO₂) supports enhance catalytic performance compared to TiO₂. Experimental validation confirms that Mn/LiTiO₂ achieves higher conversion rates and improved selectivity toward hexene (C₄:C₆ = 1:2.6). The enhanced activity is attributed to lithiation, which alters the electronic environment around Mn active sites. Mechanistic studies reveal that the formation of a seven-membered ring, a key intermediate for hexene formation, is more favorable on LiTiO₂ than TiO₂. This work provides the first evidence of Mn catalyzing selective ethylene oligomerization via oxidative cyclization in either homogeneous or heterogeneous catalysis.

Kim, Yu Lim [Argonne National Laboratory (ANL), Ar↗
DOE OSTI2024

Understanding the stability of a plastic‐degrading Rieske iron oxidoreductase system

Abstract Rieske oxygenases (ROs) are a diverse metalloenzyme class with growing potential in bioconversion and synthetic applications. We postulated that ROs are nonetheless underutilized because they are unstable. Terephthalate dioxygenase (TPA DO PDB ID 7Q05 ) is a structurally characterized heterohexameric α 3 β 3 RO that, with its cognate reductase (TPA RED ), catalyzes the first intracellular step of bacterial polyethylene terephthalate plastic bioconversion. Here, we showed that the heterologously expressed TPA DO /TPA RED system exhibits only ~300 total turnovers at its optimal pH and temperature. We investigated the thermal stability of the system and the unfolding pathway of TPA DO through a combination of biochemical and biophysical approaches. The system's activity is thermally limited by a melting temperature ( T m ) of 39.9°C for the monomeric TPA RED , while the independent T m of TPA DO is 50.8°C. Differential scanning calorimetry revealed a two‐step thermal decomposition pathway for TPA DO with T m values of 47.6 and 58.0°C (Δ H = 210 and 509 kcal mol −1 , respectively) for each step. Temperature‐dependent small‐angle x‐ray scattering and dynamic light scattering both detected heat‐induced dissociation of TPA DO subunits at 53.8°C, followed by higher‐temperature loss of tertiary structure that coincided with protein aggregation. The computed enthalpies of dissociation for the monomer interfaces were most congruent with a decomposition pathway initiated by β‐β interface dissociation, a pattern predicted to be widespread in ROs. As a strategy for enhancing TPA DO stability, we propose prioritizing the re‐engineering of the β subunit interfaces, with subsequent targeted improvements of the subunits.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

Understanding the Competition between Alcohol Formation and Dimerization during Electrochemical Reduction of Aromatic Carbonyl Compounds

The electrochemical reductive dimerization of small aromatic carbonyl compounds derived from lignocellulosic biomass is a crucial C−C coupling reaction for upgrading small molecules to longchain hydrocarbons, particularly in the synthesis of drop-in sustainable aviation fuels. Although other electrochemical reduction reactions of these reactants (i.e., hydrogenation and hydrogenolysis) have undergone extensive mechanistic investigation, the understanding of dimerization remains relatively underdeveloped. Most importantly, there is a lack of understanding of the selectivity-determining step between dimerization and monomer reduction and critical factors that can affect this step. In this study, we provide a comprehensive mechanistic model to explain the competition between dimerization and monomer reduction of benzaldehyde under various conditions. Our model proposes that the selectivity between dimerization and monomer reduction depends on the competition between desorption of a ketyl radical from the electrode, necessary for dimerization, and further reduction of the ketyl radical to an alcohol on the electrode by proton-coupled electron transfer (PCET). Computationally comparing the adsorption/desorption energy and PCET activation barrier energy is challenging because conventional DFT calculations substantially underestimate the PCET kinetic barriers. In this study, we employed constrained DFTbased configuration interaction (CDFT-CI) to obtain a reliable comparison of these energies. Our mechanistic model was tested and supported by experimental results obtained with four electrodes (Cu, Pb, Bi, graphite), three pH conditions (acidic, neutral, basic), and three potentials. Our study offers a coherent mechanistic foundation that can explain how each of these conditions impacts the desorption and PCET processes and the selectivities for dimerization and alcohol production.

09 BIOMASS FUELS↗
DOE OSTI2025

Multireference diffusion Monte Carlo reaches 2D materials

Abstract Quantum confinement in 2D materials strongly enhances electronic correlation effects. Therefore, predicting the properties of these unique materials, with both a high level of accuracy and computational efficiency, without relying on adjustable parameters or functionals, remains an outstanding theoretical challenge. The majority of theoretical studies are based on the approximations of density functional theory (DFT). The reliability of DFT predictions are heavily dependent on the choice of an approximated exchange-correlation functional. Here, we estimate the magnitude of impact of correlation on the total energy for the quintessential 2D material, graphene, by performing and comparing state-of-the-art selected CI and quantum Monte Carlo extrapolated calculations for a single unit cell at the$$\Gamma$$point. We demonstrate that Self-Healing Diffusion Monte Carlo (SHDMC) obtains a very compact, but high-quality wavefunction for this system that lacks the strong basis set dependence displayed by state of the art quantum chemistry methods. The SHDMC wavefunction is of higher quality compared to that obtained from sCI, in the same orbital basis, while being$$\sim$$ 1000 times smaller in terms of determinant count compared to sCI. We also demonstrate that extrapolating SHDMC results to the infinite determinant limit compares extremely well with complete basis set extrapolated sCI. Our work paves the way for future validation of SHDMC applied to challenging 2D materials.

Science & Technology - Other Topics↗
DOE OSTI2024

R-matrix calculations for opacities: I. Methodology and computations

Abstract An extended version of the R -matrix methodology is presented for calculation of radiative parameters for improved plasma opacities. Contrast and comparisons with existing methods primarily relying on the distorted wave approximation are discussed to verify accuracy and resolve outstanding issues, particularly with reference to the opacity project (OP). Among the improvements incorporated are: (i) large-scale Breit–Pauli R -matrix calculations for complex atomic systems including fine structure, (ii) convergent close coupling wave function expansions for the ( e + ion) system to compute oscillator strengths and photoionization cross sections, (iii) open and closed shell iron ions of interest in astrophysics and experiments, (iv) a treatment for plasma broadening of autoionizing resonances as function of energy-temperature-density dependent cross sections, (v) a ‘top-up’ procedure to compare convergence with R -matrix calculations for highly excited levels, and (vi) spectroscopic identification of resonances and bound ( e + ion) levels. The present R -matrix monochromatic opacity spectra are fundamentally different from OP and lead to enhanced Rosseland and Planck mean opacities. An outline of the work reported in other papers in this series and those in progress is presented. Based on the present re-examination of the OP work, opacities of heavy elements might require revisions in high temperature-density plasma sources.

Pradhan, A. K. (ORCID:0000000187753643)↗
DOE OSTI2025

Determining spin-dependent light dark matter rates from neutron scattering

The scattering and absorption rates of light dark matter with electron spin-dependent interactions depend on the target’s spin response. We show how this response is encoded by the target’s dynamical magnetic susceptibility, which can be measured using neutron scattering. We directly use existing neutron scattering data to compute the dark matter scattering rate in a candidate target material, finding close agreement with the previous first-principles calculation at MeV dark matter masses. Complementary experiments and measurements can extend the reach of this technique to other dark matter models and masses, and identify promising target materials for future experiments.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2025

Graph-based Reversible Evaluation and Tangents Library

GRETL is a C++ library for evaluation, re-evaluation and algorithmic differentiation of functional operations on an arbitrary computational graph with limited memory usage. Similar to popular machine learning frameworks in Python, like PyTorch and JAX, it tracks and stores both operations and output data as functions are evaluated. Once this composition of functions is built up, the entire chain of operations can be back propagated to compute sensitivities of the final result with respect to any number of inputs. In contrast to most machine learning applications, memory usage becomes the bottleneck for back propagation in many physics applications, especially for time-dependent PDEs. Dynamic check pointing becomes essential. An important distinguishing feature of GRETL is its ability to limit the maximum memory usage by automatically dynamic checkpointing the data output for each graph operation (see Wang, Moin, Iaccarino, 2009). During backpropagation, parts of the graph that are no longer in memory are automatically re-evaluated from upstream checkpointed states as needed for derivative sensitivity calculations (or more precisely, for vector-Jacobian products). GRETL is particularly beneficial for applications, such as coupled multi-physics, where deriving adjoint-based sensitivities and managing checkpoint memory across modules becomes onerous. Cases which can be readily handled by the GRETL library include: different time-integration algorithms per physics (e.g., coupled predictor-corrector algorithms, IMEX, etc.), sub-cycling, asynchronous integrators, state dependent timestep sizes, iterative solvers and coupling algorithms, controller algorithms, and more.

Tupek, MichaelR [Lawrence Livermore National Labor↗
DOE OSTI2024

Performance Comparison of Machine Learning Models for Ultrasonic Nondestructive Evaluation of Alkali-Silica Reaction in Concrete

Alkali-silica reaction (ASR) causes concrete degradation, leading to cracking, rebar corrosion, and reduced structural integrity, which raises safety concerns. Ultrasonic nondestructive evaluation (NDE) effectively assesses concrete properties and monitors ASR progression. However, its deployment and analysis require specialized expertise and subjective interpretation. As computational power increases, artificial intelligence (AI) and machine learning (ML) algorithms are increasingly being used to automate NDE data analysis across various industries for AI-assisted automation. Regulatory agencies are adapting to this technological shift, prompting a need to evaluate current ML technologies’ capabilities and limitations in assessing concrete material properties and damage. This report presents a comparative analysis of four ML regression models for predicting concrete material damage induced by ASR expansion using long-term ultrasonic data monitoring. The models investigated include linear regression (LR), support vector regression (SVR), shallow neural networks (NN), and deep neural networks (DNN). LR, SVR, and shallow NN models use features extracted from ultrasonic signals, whereas the DNN model processes time-domain ultrasonic signals and frequency spectra directly. The study systematically compared the models’ performance from various perspectives, including model input, prediction performance, and generalization ability. The findings indicate significant variability in model performance, with some ML algorithms achieving very high or very low prediction accuracy depending on the preprocessing and feature engineering (extraction and selection) applied. Key insights include the observation that shallow ML models (LR, SVR, and shallow NNs) require meticulous preprocessing and feature extraction to achieve high accuracy. In contrast, the DNN model, although it bypasses the need for feature engineering, necessitates extensive preprocessing to mitigate noise and computational demands. The SVR model emerged as the top performer among the shallow models, and the DNN model exhibited superior performance on specific datasets but struggled with generalization across specimens from different batches. Additionally, the SVR model is sensitive to temperature variations, whereas the DNN model is robust in this regard. Using recurrent neural networks is recommended for future ASR expansion prediction studies. Recurrent neural networks’ inherent ability to capture temporal dependencies and long-term patterns makes them well suited for analyzing sequential ultrasonic monitoring data. Overall, the results and conclusions of this study could provide insights into the capabilities and effectiveness of ML when applied to ultrasonic NDE data and help identify best practices for using ML for ultrasonic NDE of concrete material properties.

36 MATERIALS SCIENCE↗