Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Learning curves”

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 433 records · Page 24
All sourcesNASA NTRSDOE OSTI
NASA NTRS2018

Flight Demonstration of Novel Atmospheric Satellite Concept

The major focus of the Phase II effort described herein is to develop and demonstrate an aircraft capable of autonomously sailing (i.e., to cruise without propulsion or external assistance), and thereby prove that the dual-aircraft platform (DAP) atmospheric satellite concept is potentially viable. This sailing mode of flight was identified as the number-1 enabling technology required for the stratospheric DAP concept (also known as Stratosat) in the NIAC (NASA Innovative Advanced Concept) Phase I effort. No scientific demonstration of this technology has ever been done or documented to our knowledge. This report describes efforts to take a major step towards the sailing mode of flight capability using a single aircraft connected by cable to a moving ground vehicle which uses sufficient crosswind to cruise without propulsion while "pulling" the ground vehicle forward (i.e., without external assistance). The development of a prototype aircraft is described in terms of novel and key hardware and software elements. A specialized prototype aircraft is described, including a novel cable release mechanism, novel "lateron" control surfaces, and a highly-accurate onboard wind measurement system. Additionally, a novel means to safely connect the aircraft to the moving ground vehicle is described involving a fishing rod/reel and integrated load cell. All of these devices were designed and developed in-house and validated in flight testing. Software is developed to provide look-up tables that give the flight condition targets (i.e., 3-D position relative to ground vehicle, forward speed, aircraft orientation, etc.), based on current wind speed and direction. These tables are successfully validated in flight simulation and used onboard the aircraft. High fidelity analysis of the aircraft aerodynamics are described - required to produce accurate target sailing flight conditions. A novel wind tunnel measurement technique is developed to accurately assess the aerodynamics of the ultra-thin cable. A new specialized flight simulator is described which is utilized to develop and verify the flight software required onboard the aircraft, and to support training of pilots for flying the aircraft while tethered to a ground vehicle. The DAP flight simulator was developed within the Matlab-Simulink framework and included detailed treatment of aircraft/cable aerodynamics, cable dynamics, experimentally-derived propeller-motor thrust curves, actuator responsiveness, and realistic air turbulence. The specialized formation flight controller algorithm, developed using this flight simulator, and onboard the aircraft is described. Finally, a novel auto-tuning software is described and verified within the flight simulator that is shown to refine the sailing flight condition targets during flight using an optimization technique involving doublet maneuvers. Virtual flights using the auto-tuning software indicate that the prototype aircraft should be able to reach and hold sailing conditions despite moderate levels of turbulence provided there is sufficient mean wind available. An overview of the flight testing program is provided. Hundreds of short flights were conducted, primarily using a “dead” short runway at Deland Municipal Airport which permitted use of a moving ground vehicle. Additional flight tests at Space Florida’s Shuttle Landing Facility are also described. First year results from these tests in which the aircraft is controlled manually, demonstrated that excessive flight testing would be required for a pilot to learn to sail with visual cues. However, second year results from autonomous flight these tests included successful demonstration of the closed-loop autonomous formation flight capability (i.e., autonomously determine, reach, and hold the required 3-D location relative to the ground vehicle required for sailing). The next step of using the auto-tune software to autonomously refine the aircraft orientation targets to finally achieve sailing remains the primary goal of future work.

Flight↗
NASA NTRS2001

Use of Pressure Sensitive Paint for Diagnostics in Turbomachinery Flows With Shocks

The technology of pressure sensitive paint (PSP) is well established in external aerodynamics. In internal flows in narrow channels and in turbomachinery cascades, however, there are still unresolved problems. In particular, the internal flows with complex shock structures inside highly curved channels present a challenge. It is not always easy and straightforward to distinguish between true signals and "ghost" images due to multiple internal reflections in narrow channels. To address some of the problems, investigations were first carried out in a narrow supersonic channel of Mach number 2.5. A single wedge or a combination of two wedges were used to generate a complex shock wave structure in the flow. The experience gained in a small supersonic channel was used for surface pressure measurements on the stator vane of a supersonic throughflow fan. The experimental results for several fan operating conditions are shown in a concise form, including performance map points, midspan static tap pressure distributions, and vane suction side pressure fields. Finally, the PSP technique was used in the NASA transonic flutter cascade to compliment flow visualization data and to acquire backwall pressure fields to assess the cascade flow periodicity. A summary of shortcomings of the pressure sensitive paint technology for internal flow application and lessons learned are presented in the conclusion of the paper.

Lepicovsky, Jan↗
NASA NTRS2006

Estimation of Stability and Control Derivatives of an F-15

A technique for real-time estimation of stability and control derivatives (derivatives of moment coefficients with respect to control-surface deflection angles) was used to support a flight demonstration of a concept of an indirect-adaptive intelligent flight control system (IFCS). Traditionally, parameter identification, including estimation of stability and control derivatives, is done post-flight. However, for the indirect-adaptive IFCS concept, parameter identification is required during flight so that the system can modify control laws for a damaged aircraft. The flight demonstration was carried out on a highly modified F-15 airplane (see Figure 1). The main objective was to estimate the stability and control derivatives of the airplane in nearly real time. A secondary goal was to develop a system to automatically assess the quality of the results, so as to be able to tell a learning neural network which data to use. Parameter estimation was performed by use of Fourier-transform regression (FTR) a technique developed at NASA Langley Research Center. FTR is an equation- error technique that operates in the frequency domain. Data are put into the frequency domain by use of a recursive Fourier transform for a discrete frequency set. This calculation simplifies many subsequent calculations, removes biases, and automatically filters out data beyond the chosen frequency range. FTR as applied here was tailored to work with pilot inputs, which produce correlated surface positions that prevent accurate parameter estimates, by replacing half the derivatives with predicted values. FTR was also set up to work only on a recent window of data, to accommodate changes in flight condition. A system of confidence measures was developed to identify quality-parameter estimates that a learning neural network could use. This system judged the estimates primarily on the basis of their estimated variances and of the level of aircraft response. The resulting FTR system was implemented in the Simulink software system and auto-coded in the C programming language for use on the Airborne Research Test System (ARTS II) computer installed in the F-15 airplane. The Simulink model was also used in a control room that utilizes the Ring Buffered Network Bus hardware and software, making it possible to evaluate test points during flights. In-flight parameter estimation was done for piloted and automated maneuvers, primarily at three test conditions. Figure 2 shows results for pitching moment due to symmetric stabilator actuations for a series of three pitch doublet maneuvers (in a doublet maneuver, a command to change attitude in a given direction by a given amount is followed immediately by a command to change attitude in the opposite direction by the same amount). A time window of 5 seconds was used. The portions of the curves shown in red are those that passed the confidence tests. The technique showed good convergence for most derivatives for both kinds of maneuvers - typically within a few seconds. The confidence tests were marginally successful, and it would be necessary to refine them for use in an IFCS.

Smith, Mark↗
NASA NTRS2023

Prediction of Stiffness and Fatigue Lives of Polymer Matrix Composite Laminates Using Artificial Neural Networks

Machine learning (ML) models are increasingly being used in many engineering fields due to the advancements in ML algorithms and availability of high-speed computing power. One of the most popular ML class of models is artificial neural networks (ANN). ML is increasingly being used in the design and analysis of composite materials and structures, specifically in the constitutive modeling of composite materials with the focus on greatly accelerating multiscale analyses of composite materials and structures through development of surrogate models. Towards that end, both Python and MATLAB-based neural nets have been developed to predict initial stiffness and fatigue life of an eight-ply symmetric polymer matrix composite laminate. Two types of neural networks, a Multilayer Perceptron (MLP) and a Recurrent Neural Network (RNN), have been developed for both platforms. Results show that the both neural net types can provide an excellent estimate of initial stiffness as well as fatigue life of eight-ply symmetric polymer matrix composite laminate. RNNs are better able to capture the shape of the fatigue curve of a laminate. This tool can be very useful for system level studies to obtain an estimate of desired properties and life of PMC composite laminates. The associated surrogate models could also be used in composite multiscale analyses to replace the actual physics-based calculations at lower scales and thereby significantly increase the computational efficiency of such analyses and thus make multiscale analyses a viable industrial tool for large scale structural problems.

Composite↗
DOE OSTI2021

Machine learning-based microstructure prediction during laser sintering of alumina

Abstract Predicting material’s microstructure under new processing conditions is essential in advanced manufacturing and materials science. This is because the material’s microstructure hugely influences the material’s properties. We demonstrate an elegant machine learning algorithm that faithfully predicts the microstructure under new conditions, without the need of knowing the governing laws. We name this algorithm, RCWGAN-GP, which is regression-based conditional generative adversarial networks with Wasserstein loss function and gradient penalty. This algorithm was trained with experimental SEM micrographs from laser-sintered alumina under various laser powers. The RCWGAN-GP realistically regenerates the SEM micrographs under the trained laser powers. Impressively, it also faithfully predicts the alumina’s microstructure under unexplored laser powers. The predicted microstructure features, including the morphology of the sintered particles and the pores, match the experimental SEM micrographs very well. We further quantitatively examined the prediction accuracy of the RCWGAN-GP. We trained the algorithm with computer-created micrograph datasets of secondary-phase growth governed by the well-known Johnson–Mehl–Avrami (JMA) equation. The RCWGAN-GP accurately regenerates the micrographs at the trained time series, in terms of the grains’ shapes, sizes, and spatial distributions. More importantly, the predicted secondary phase fraction accurately follows the JMA curve.

08 HYDROGEN↗
DOE OSTI2020

Exoplanets in the Antarctic Sky. IV. Dual-band Photometry of Variables Found by the CSTAR-II Commissioning Survey at the North Sky

From the experiences learned in three decades of exoplanet search, wide-field transit surveys have proven to be one of the most effective ways to detect exoplanets. Wide field of view, however, suffers from high false-positive rates caused by blended eclipsing binaries. The chromaticity in eclipse depth is an effective feature to distinguish low-depth eclipsing binaries from transiting exoplanets, making multiple-band photometry follow-up advantageous before a target is passed onto more expensive spectroscopic follow-up. Moreover, a multiple-band photometric survey is itself a powerful method to find and vet planetary candidates and narrow down the candidate list of high-priority targets. In this work, we report the first results of a dual-band (Sloan-g and -i) wide-field photometry survey—the Chinese Small Telescope ARray II (CSTAR-II), an updated version of the original CSTAR. As a key component of the Chinese Exoplanet Searching Program from Antarctica, CSTAR-II has been tested thoroughly at a remote arctic site near Mohe during the winter of 2014. In total, 13,531 light curves with the best overall photometric precision of ∼3 mmag were extracted from 7721 stars in the Sloan-g and -i bands. Using a robust method, we have detected 63 variables, of which 48 are newly discovered. The dual-band photometric results as well as the stellar properties of the detected sources are provided in this work.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2026

Neural Posterior Estimation for Scalable and Accurate Inverse Parameter Inference in Li-Ion Batteries

Diagnosing the internal state of Li-ion batteries is critical for battery research, operation of real-world systems, and prognostic evaluation of remaining lifetime. By using physics-based models to perform probabilistic parameter estimation via Bayesian calibration, diagnostics can account for the uncertainty due to model fitness, data noise, and the observability of any given parameter. However, Bayesian calibration in Li-ion batteries using electrochemical data is computationally intensive even when using a fast surrogate in place of physics-based models, requiring many thousands of model evaluations. A fully amortized alternative is neural posterior estimation (NPE). NPE shifts the computational burden from the parameter estimation step to data generation and model training, reducing the parameter estimation time from minutes to milliseconds, enabling real-time applications. The present work shows that NPE can infer parameters equally or more accurately than Bayesian calibration, even if it leads to higher voltage reconstruction errors. We also demonstrate that the higher computational costs for data generation are tractable even in high-dimensional cases (ranging from 6 to 27 estimated parameters). The NPE method also offers several interpretability advantages over Bayesian calibration, such as local parameter sensitivity to specific regions of the voltage curve. The NPE method is demonstrated using an experimental fast charge dataset, with parameter estimates validated against measurements of loss of lithium inventory and loss of active material. The implementation is made available in a companion repository (https://github.com/NatLabRockies/BatFIT).

25 ENERGY STORAGE↗
DOE OSTI2021

From Femtoseconds to Gigaseconds: The SolDeg Platform for the Performance Degradation Analysis of Silicon Heterojunction Solar Cells

Heterojunction Si solar cells exhibit notable performance degradation. Here, we modeled this degradation by electronic defects getting generated by thermal activation across energy barriers over time. To analyze the physics of this degradation, we developed the SolDeg platform to simulate the dynamics of electronic defect generation. First, femtosecond molecular dynamics simulations were performed to create a-Si/c-Si stacks, using the machine learning-based Gaussian approximation potential. Second, we created shocked clusters by a cluster blaster method. Third, the shocked clusters were analyzed to identify which of them supported electronic defects. Fourth, the distribution of energy barriers that control the generation of these electronic defects was determined. Fifth, an accelerated Monte Carlo method was developed to simulate the thermally activated time-dependent defect generation across the barriers. Our main conclusions are as follows. (1) The degradation of a-Si/c-Si heterojunction solar cells via defect generation is controlled by a broad distribution of energy barriers. (2) We developed the SolDeg platform to track the microscopic dynamics of defect generation across this wide barrier distribution and determined the time-dependent defect density N(t) from femtoseconds to gigaseconds, over 24 orders of magnitude in time. (3) We have shown that a stretched exponential analytical form can successfully describe the defect generation N(t) over at least 10 orders of magnitude in time. (4) We found that in relative terms, Voc degrades at a rate of 0.2%/year over the first year, slowing with advancing time. (5) We developed the time correspondence curve to calibrate and validate the accelerated testing of solar cells. We found a compellingly simple scaling relationship between accelerated and normal times t normal ∝ t accel T(accel)/T(normal) . (6) We also carried out experimental studies of defect generation in a-Si:H/c-Si stacks. We found a relatively high degradation rate at early times that slowed considerably at longer time scales.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

A Stochastic Approach to Reconstruct Gamma-Ray-burst Light Curves

Gamma-ray bursts (GRBs), as they are observed at high redshift ( z = 9.4), are vital to cosmological studies and investigating Population III stars. To tackle these studies, we need correlations among relevant GRB variables with the requirement of small uncertainties on their variables. Thus, we must have good coverage of GRB light curves (LCs). However, gaps in the LC hinder the precise determination of GRB properties and are often unavoidable. Therefore, extensive categorization of GRB LCs remains a hurdle. We address LC gaps using a stochastic reconstruction, wherein we fit two preexisting models (the Willingale model; W07; and a broken power law; BPL) to the observed LC, then use the distribution of flux residuals from the original data to generate data to fill in the temporal gaps. We also demonstrate a model-independent LC reconstruction via Gaussian processes. At 10% noise, the uncertainty of the end time of the plateau, its correspondent flux, and the temporal decay index after the plateau decreases by 33.3%, 35.03%, and 43.32% on average for the W07, and by 33.3%, 30.78%, 43.9% for the BPL, respectively. The uncertainty of the slope of the plateau decreases by 14.76% in the BPL. After using the Gaussian process technique, we see similar trends of a decrease in uncertainty for all model parameters for both the W07 and BPL models. These improvements are essential for the application of GRBs as standard candles in cosmology, for the investigation of theoretical models, and for inferring the redshift of GRBs with future machine-learning analyses.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS2005

Salt Kinematics and InSAR

As part of a long-term attempt to learn how the climatic and tectonic signal interact to shape a steady state mountain monitored displacement of a markers in SE termination and also near the summit of a small viscous salt fountain extruding onto the Central plateau of Iran. The marker displacements relate to the first InSAR interferograms of salt extrusion (980913 to 990620) calculated Earth tides, winds, air pressures and temperatures. In the first documented staking exercise, hammered wooden stakes vertically through the surgical marl (c. 1 Ocm deep) onto the top of crystalline salt. These stakes installed in an irregular array elongate E-W along the c.50 m high cliff marking the effective SE terminus of the glacier at Qum Kuh(Centra1 Iran) ,just to the E of a NE trending river cliff about 40 m high. We merely measured the distances between pairs of stakes with known azimuth about 2 m apart to calculate sub horizontal strain in a small part of Qum Kuh. Stakes moved and micro strains for up to 46 pairs of stakes (p strain= ((lengthl-length2)/1engthl) x 10-1) was calculated for each seven stake epochs and plotted against their azimuth on simplified array maps. The data fit well the sine curves cxpected of the maximum and minimum strain ellipses. The first documented stakes located on the SE where the InSAR image show -1 1 to 0 mm pink to purple, 0 to lOmm purple to blue, and show high activity of salt in low activity area of the InSAR image (980913 to 990620).Short term micro strains of stake tie lines record anisotropic expansions due to heating and contraction due to cooling. All epochs changed between 7 to 1 17 days (990928 to000 1 16), showed 200 to 400 micro strain lengthening and shortening. The contraction and extension existed in each epoch, but the final strain was extension in E-W in Epoch land 6, contraction in E-W direction during epochs 2-3-4-5 and 7. The second pair of stakes hammered about 20 cm deep into the deep soils(more than 1 m) , near summit, where the colors change between 19 to 29mm in InSAR image(9809 13 to 990620). Additional information is included in the original abstract.

Aftabi, Pedarm↗
DOE OSTI2024

Data Science Shows that Entropy Correlates with Accelerated Zeolite Crystallization in Monte Carlo Simulations

We have performed a data science study of Monte Carlo simulation trajectories to understand factors that can accelerate formation of zeolite nanoporous crystals, a process that can take days or even weeks. In previous work, Monte Carlo simulations predicted and experiments confirmed that using a secondary organic structure-directing agent (OSDA) accelerates crystallization of all-silica LTA zeolite, with experiments finding a three-fold speedup [PCCP 24, 142-148 (2022)]. However, it remains unclear what physical factors cause the speed-up. Here, we apply data science to analyze the simulation trajectories to discover what drives accelerated zeolite crystallization in Monte Carlo going from a one-OSDA synthesis (1OSDA) to a two-OSDA version (2OSDA). We encoded simulation snapshots using the Smooth Overlap of Atomic Positions approach, which represents all 2- and 3-body correlations within a given cutoff distance. Principal component analyses failed to discriminate datasets of structures from 1OSDA and 2OSDA simulations, while the Support Vector Machine (SVM) approach succeeded at classifying such structures with an area-under-curve (AUC) score of 0.99 (where AUC = 1 is a perfect classification) with all 3-body correlations, and as high as 0.94 with only 2-body correlations. SVM decision functions reveal relatively broad / narrow histograms for 1OSDA / 2OSDA datasets, suggesting that the two simulations differ strongly in information heterogeneity. Informed by these results, we performed pair (2-body) entropy calculations during crystallization, resulting in entropy differences that semi-quantitatively account for the speedup observed in the previous Monte Carlo simulations. We conclude that altering synthesis conditions in ways that substantially changes the entropy of labile silica networks may accelerate zeolite crystallization, and we discuss possible approaches for achieving such acceleration.

77 NANOSCIENCE AND NANOTECHNOLOGY↗
NASA NTRS2019

Demisability of GFRP and CFRP Components of Reentering Orbital Debris: Phase I Test Results

Observations of surviving reentry debris on the ground and research performed by Hyperschall Technologie Göttingen (HTG) [1] indicated that significantly more glass fiber-reinforced polymer (GFRP) and carbon fiber-reinforced polymer (CFRP) components survive reentry than current models predict. NASA’s Orbital Debris Program Office conducted a series of tests to evaluate the accuracy of material demise models for reentering orbital debris used in NASA’s Object Reentry Survival Analysis Tool (ORSAT) and Debris Assessment Software (DAS). Testing is planned in a multi-phase series to allow for quick quantification of results as well as refinement of methods resulting from lessons learned during early phases. The Phase 1 tests discussed here validated ORSAT models for homogeneous metals, provided an efficient quantification of composite material demisability properties like mass loss rate and overall time to demise, and identified potential failure modes, which are currently not well understood. Phase 2 tests will be used to further understand mass loss rates and modes of both thermal and mechanical failure in composite materials. The authors exposed 95 samples of aluminum, CFRP, Kevlar fiber-reinforced polymer, GFRP, and sheets of G10 fiberglass to conditions approximating the reentry environment using an inductively coupled plasma (ICP) torch facility. The cylindrical CFRP samples were exposed to the atmospheric pressure plasma, at both the end and the midpoint, to investigate the difference in demisability between parts with exposed edges, like panels, and parts with no edges, such as carbonoverwrapped pressure vessels (COPVs). In a non-oxidative environment, no composite materials demised within the 5-minute test time. In the oxidative, elevated heat flux environment, CFRP samples demised between 210 s and 270 s. For the first 100 s of insertion time, most of the mass loss was due to pyrolysis of resin, creating an approximately bi-linear mass-loss rate curve with time. In a non-oxidative environment, carbon filaments were observed to unravel from some of the CFRP end-burned samples; however, this effect did not seem to affect the overall time to demise for the samples significantly. These results indicate that both GFRP and CFRP components survive reentry with significantly more remaining mass than current models predict.

Greene, Benton R.↗
DOE OSTI2023

Prediction of chronic kidney disease progression using recurrent neural network and electronic health records

Chronic kidney disease (CKD) is a progressive loss in kidney function. Early detection of patients who will progress to late-stage CKD is of paramount importance for patient care. To address this, we develop a pipeline to process longitudinal electronic heath records (EHRs) and construct recurrent neural network (RNN) models to predict CKD progression from stages II/III to stages IV/V. The RNN model generates predictions based on time-series records of patients, including repeated lab tests and other clinical variables. Our investigation reveals that using a single variable, the recorded estimated glomerular filtration rate (eGFR) over time, the RNN model achieves an average area under the receiver operating characteristic curve (AUROC) of 0.957 for predicting future CKD progression. When additional clinical variables, such as demographics, vital information, lab test results, and health behaviors, are incorporated, the average AUROC increases to 0.967. In both scenarios, the standard deviation of the AUROC across cross-validation trials is less than 0.01, indicating a stable and high prediction accuracy. Our analysis results demonstrate the proposed RNN model outperforms existing standard approaches, including static and dynamic Cox proportional hazards models, random forest, and LightGBM. The utilization of the RNN model and the time-series data of previous eGFR measurements underscores its potential as a straightforward and effective tool for assessing the clinical risk of CKD patients concerning their disease progression.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2023

Predicting wind-driven spatial deposition through simulated color images using deep autoencoders

Abstract For centuries, scientists have observed nature to understand the laws that govern the physical world. The traditional process of turning observations into physical understanding is slow. Imperfect models are constructed and tested to explain relationships in data. Powerful new algorithms can enable computers to learn physics by observing images and videos. Inspired by this idea, instead of training machine learning models using physical quantities, we used images, that is, pixel information. For this work, and as a proof of concept, the physics of interest are wind-driven spatial patterns. These phenomena include features in Aeolian dunes and volcanic ash deposition, wildfire smoke, and air pollution plumes. We use computer model simulations of spatial deposition patterns to approximate images from a hypothetical imaging device whose outputs are red, green, and blue (RGB) color images with channel values ranging from 0 to 255. In this paper, we explore deep convolutional neural network-based autoencoders to exploit relationships in wind-driven spatial patterns, which commonly occur in geosciences, and reduce their dimensionality. Reducing the data dimension size with an encoder enables training deep, fully connected neural network models linking geographic and meteorological scalar input quantities to the encoded space. Once this is achieved, full spatial patterns are reconstructed using the decoder. We demonstrate this approach on images of spatial deposition from a pollution source, where the encoder compresses the dimensionality to 0.02% of the original size, and the full predictive model performance on test data achieves a normalized root mean squared error of 8%, a figure of merit in space of 94% and a precision-recall area under the curve of 0.93.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Single-Point Extrapolation to the Complete Basis Set Limit through Deep Learning

Machine learning (ML) offers an attractive method for making predictions about molecular systems while circumventing the need to run expensive electronic structure calculations. Once trained on ab initio data, the promise of ML is to deliver accurate predictions of molecular properties that were previously computationally infeasible. In this work, we develop and train a graph neural network model to correct the basis set incompleteness error (BSIE) between a small and large basis set at the RHF and B3LYP levels of theory. Our results show that, when compared to fitting to the total potential, an ML model fitted to correct the BSIE is better at generalizing to systems not seen during training. We test this ability by training on single molecules while evaluating on molecular complexes. We also show that ensemble models yield better behaved potentials in situations where the training data is insufficient. However, even when only fitting to the BSIE, acceptable performance is only achieved when the training data sufficiently resemble the systems one wants to make predictions on. As a result, the test error of the final model trained to predict the difference between the cc-pVDZ and cc-pV5Z potential is 0.184 kcal/mol for the B3LYP density functional, and the ensemble model accurately reproduces the large basis set interaction energy curves on the S66x8 dataset.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

Pulsed Thermal Tomography Nondestructive Examination of Additively Manufactured Reactor Materials and Components (Final Technical Report)

Metal Additive Manufacturing (AM) is a promising method for cost-efficient fabrication of complex shape structures for applications in harsh environment, such as in a nuclear reactor. However, internal defects (pores) occur in high-strength AM alloys, which are manufactured with Laser Powder Bed Fusion (LPBF) AM method. Pulsed Infrared Thermography (PIT) is an efficient nondestructive evaluation (NDE) method to examine actual structures, because this method offers one-sided non-contact measurements, and fast processing of large sample areas. However, imaging of material defects, particularly defects with sizes at microscopic level, is challenging. In this report, we benchmark the performance of several Unsupervised Learning (UL) algorithms designed to enhance imaging of microscopic defects in metals with PIT. UL aims to learn the latent principal patterns (dictionaries) in PIT data to detect defects with minimal human supervision. Performance of Independent Component Analysis (ICA), Sparse Coding (SC), Principal Component Analysis (PCA) and Exploratory Factor Analysis (EFA) was compared using F-score, UL model training time and defects reconstruction time. We obtained the average F-score of 0.75, and a highest F-score of 0.89 for the EFA algorithm. Overall, EFA outperforms other UL algorithms considered in this study. In another approach, we investigate Thermal Tomography (TT), which is a computational method for reconstruction of depth profile of internal material defects from PIT nondestructive evaluation (NDE). TT algorithm obtains depth reconstructions of thermal effusivity, which has been shown to provide visualization of subsurface internals defects in metals. In many applications, one needs to determine the defect shape and orientation from reconstructed effusivity images. Interpretation of TT images is non-trivial because of blurring, which increases with depth due to heat diffusion-based nature of image formation. We have developed a deep learning convolutional neural network (CNN) to classify size and orientation of subsurface material defects in TT images. CNN was trained with TT images produced with computer simulations of 2D metallic structures (thin plates) containing elliptical subsurface voids. Performance of CNN was investigated using test TT images developed with computer simulations of plates containing elliptical defects, and defects with shape imported from scanning electron microscopy (SEM) images. CNN demonstrated the ability to classify radii and angular orientation of elliptical defects in previously unseen test TT images. We have also demonstrated that CNN trained on TT images of elliptical defects is capable of classifying shape and orientation of irregular defects. Training the CNN on irregular defect shapes instead of on elliptical shapes would make the resulting classifications more descriptive of actual defect shapes. However, this requires a much higher volume of SEM images of material defects, which are difficult to obtain because of random occurrence of defects in LPBF. To address this challenge, we developed a generative adversarial network (GAN) to augment the existing dataset of SEM defect images. The GAN model is demonstrated to create novel yet realistic defect shapes that can be used as input for simulated PTT images to train CNN. We also investigate several approaches based on Gaussian Random Circle and Bezier Curves for constructing parametric models of irregular-shape defects.

36 MATERIALS SCIENCE↗
DOE OSTI2024

A Comparison between Invariant and Equivariant Classical and Quantum Graph Neural Networks

Machine learning algorithms are heavily relied on to understand the vast amounts of data from high-energy particle collisions at the CERN Large Hadron Collider (LHC). The data from such collision events can naturally be represented with graph structures. Therefore, deep geometric methods, such as graph neural networks (GNNs), have been leveraged for various data analysis tasks in high-energy physics. One typical task is jet tagging, where jets are viewed as point clouds with distinct features and edge connections between their constituent particles. The increasing size and complexity of the LHC particle datasets, as well as the computational models used for their analysis, have greatly motivated the development of alternative fast and efficient computational paradigms such as quantum computation. In addition, to enhance the validity and robustness of deep networks, we can leverage the fundamental symmetries present in the data through the use of invariant inputs and equivariant layers. In this paper, we provide a fair and comprehensive comparison of classical graph neural networks (GNNs) and equivariant graph neural networks (EGNNs) and their quantum counterparts: quantum graph neural networks (QGNNs) and equivariant quantum graph neural networks (EQGNN). The four architectures were benchmarked on a binary classification task to classify the parton-level particle initiating the jet. Based on their area under the curve (AUC) scores, the quantum networks were found to outperform the classical networks. However, seeing the computational advantage of quantum networks in practice may have to wait for the further development of quantum technology and its associated application programming interfaces (APIs).

Forestano, Roy T. (ORCID:0000000203552076)↗
DOE OSTI2025

A comparative analysis of YOLOv8 and U-Net image segmentation approaches for transmission electron micrographs of polycrystalline thin films

Metallic thin films offer a platform to experimentally study the dynamics of microstructural evolution, but the required transmission electron microscopy (TEM)-based imaging generates complex images that are challenging to segment and quantify. This work provides a comparative analysis of a new YOLOv8 model and an established U-Net model for bright-field TEM images of polycrystals, employing a framework leveraging physical observables to evaluate performance against two hand-traced benchmark datasets. This methodology obviates the comparison of large, diversely structured, and manually labeled datasets that are required to assess performance on a per-image/per-pixel basis. It is found that the YOLOv8 model, adapted for real-time instance segmentation, has up to 43× faster inferencing (NVIDIA GeForce RTX 4090) compared to U-Net and reconstructs hand-traced grain size distributions (GSDs) with excellent fidelity, finding mean diameter within 3% for grains near an optimal magnification; for grains that deviate from the optimal pixel-diameter, the size of small- (large)-diameter grains is systematically over- (under)-estimated. This is partially mitigated by including scale-aware augmentations during training. Moreover, when the bias is corrected post-inference by a rigid shift in distribution, the YOLOv8 model reproduces ground truth GSDs with exceptional fidelity, with statistical tests indicating <5% probability that the distributions are distinct. Based on ground truth data, calibration curves pertaining to this shift can be constructed for a given model. This issue is not present in the U-Net model’s results, indicating that for quantitative measurements where the true size of objects is of interest, special procedures must be implemented for YOLO-based models.

36 MATERIALS SCIENCE↗