Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “parameter sampling”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,153 records · Page 64
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

In situ synchrotron X-ray diffraction analysis of ultrasonically assisted microstructure refinement during laser melting process

This work investigates the effect of ultrasonic treatment (UST) on the microstructure of AA4043 alloy during laser melting process. Here, the study utilizes in situ synchrotron X-ray diffraction (SXRD) and postmortem electron backscatter diffraction (EBSD) techniques to examine the ultrasonic grain refinement mechanism. The SXRD diffraction patterns exhibit higher continuity of major diffraction rings in the UST-test pattern, indicating a more refined grain structure. The presence of new diffraction spots in the post-test diffraction pattern with UST suggests the breakdown of epitaxial growth of columnar dendrites during solidification. Thermal profiles determined by analyzing the lattice parameter evolution show a faster cooling rate with UST. The inverse pole figures and aspect ratio distribution from EBSD reveal that the UST samples exhibit a refined, more equiaxed grain structure along the weld centerline. UST sample also develops more fine-sized low angle grain boundaries (LAGBs). These features offer better hot-cracking resistance in aluminum alloys.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Machine Learning-Guided Optimization of SABRE Hyperpolarization for α-Ketoglutarate in Acetone–Water

Signal amplification by reversible exchange (SABRE) is a hyperpolarization method that polarizes target nuclei of metabolites quickly and efficiently. Recent SABRE advances, including Ace-SABRE, yield biocompatible, aqueous solutions of hyperpolarized markers for metabolic monitoring. Building on recent advancements, expanding the substrate scope of Ace-SABRE is desirable. However, SABRE polarization is sensitive to many different parameters; therefore, traditional optimization approaches are experimentally time-consuming. In this proof-of-concept application of machine learning (ML), Bayesian optimization (BO) is used for four important input parameters to model the complex SABRE dynamics while saving experimental time. The presented ML model also provides chemical insights that enable predictions of sample compositions for increased polarization levels. In this paper, we transition from an original average free polarization of p = ∼0.90% to a maximum observed free polarization of p = ∼6.6% for 1- 13 C alpha-ketoglutarate (AKG) with 13 C at natural abundance, utilizing both direct outputs as well as chemical insights revealed by the ML model.

Catalysts↗
DOE OSTI2025

Efficient Measurement of Length Distribution of 1D Nanoparticles in Solution via Optical Polarimetry

The efficient measurement of the length distribution of nanotubes, nanowires, and other one-dimensional (1D) nanoparticles in solution is important to enable their incorporation into materials and devices and to optimize their processing for properties of interest, such as thermal/electrical conductivity or mechanical strength, in suspensions and composites. We report an electric-field (E-field)-assisted optical-polarimetry technique to measure the length distribution of ensembles of high-aspect-ratio particles in dilute suspension. The degree of alignment of polarizable 1D particles suspended in a fluid under Brownian motion explicitly depends on the E-field strength and the particle length. We show that it is possible to extract the length distribution of 1D nanoparticles suspended in an insulating fluid by applying a range of E-fields and using optical polarimetry to measure the corresponding alignment order parameter. Notably, the method is relatively insensitive to the diameter of the 1D particles, which can be poorly known or vary within a sample. The technique is validated with silver nanowires and carbon nanotubes of known lengths, as well as polymer-depletion-length-separated single-wall carbon nanotube samples with length distributions independently measured with analytical ultracentrifugation. Finally, we demonstrate the ability of the optical-polarimetry technique to quantify changes in the length distribution of ultranarrow, sub-nanometer-diameter single-wall carbon nanotubes under different types and durations of ultrasonication. Within its range of applicability (polarizable 1D nanoparticles in the 0.5 to 15 μm length range, constrained by the voltage stability of the media and the suspended particles), the E-field-assisted optical-polarimetry method is a particularly efficient and accurate method to measure the length distribution of nanowires and nanotubes in suspension.

1D nanoparticles↗
DOE OSTI2024

Evaluating the risk of data loss due to particle radiation damage in a DNA data storage system

DNA data storage is a potential alternative to magnetic tape for archival storage purposes, promising substantial gains in information density. Critical to the success of DNA as a storage media is an understanding of the role of environmental factors on the longevity of the stored information. In this paper, we evaluate the effect of exposure to ionizing particle radiation, a cause of data loss in traditional magnetic media, on the longevity of data in DNA data storage pools. We develop a mass action kinetics model to estimate the rate of damage accumulation in DNA strands due to neutron interactions with both nucleotides and residual water molecules, then utilize the model to evaluate the effect several design parameters of a typical DNA data storage scheme have on expected data longevity. Finally, we experimentally validate our model by exposing dried DNA samples to different levels of neutron irradiation and analyzing the resulting error profile. Our results show that particle radiation is not a significant contributor to data loss in DNA data storage pools under typical storage conditions.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Dark Energy Survey Year 3 results: Imprints of cosmic voids and superclusters in the Planck CMB lensing map

ABSTRACT The CMB lensing signal from cosmic voids and superclusters probes the growth of structure in the low-redshift cosmic web. In this analysis, we cross-correlated the Planck CMB lensing map with voids detected in the Dark Energy Survey Year 3 (Y3) data set (∼5000 deg2), expanding on previous measurements that used Y1 catalogues (∼1300 deg2). Given the increased statistical power compared to Y1 data, we report a 6.6σ detection of negative CMB convergence (κ) imprints using approximately 3600 voids detected from a redMaGiC luminous red galaxy sample. However, the measured signal is lower than expected from the MICE N-body simulation that is based on the ΛCDM model (parameters Ωm = 0.25, σ8 = 0.8), and the discrepancy is associated mostly with the void centre region. Considering the full void lensing profile, we fit an amplitude $A_{\kappa }=\kappa _{{\rm DES}}/\kappa _{{\rm MICE}}$ to a simulation-based template with fixed shape and found a moderate 2σ deviation in the signal with Aκ ≈ 0.79 ± 0.12. We also examined the WebSky simulation that is based on a Planck 2018 ΛCDM cosmology, but the results were even less consistent given the slightly higher matter density fluctuations than in MICE. We then identified superclusters in the DES and the MICE catalogues, and detected their imprints at the 8.4σ level; again with a lower-than-expected Aκ = 0.84 ± 0.10 amplitude. The combination of voids and superclusters yields a 10.3σ detection with an Aκ = 0.82 ± 0.08 constraint on the CMB lensing amplitude, thus the overall signal is 2.3σ weaker than expected from MICE.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

Is Terzan 5 the remnant of a building block of the Galactic bulge? Evidence from APOGEE

ABSTRACT It has been proposed that the globular cluster-like system Terzan 5 is the surviving remnant of a primordial building block of the Milky Way bulge, mainly due to the age/metallicity spread and the distribution of its stars in the α–Fe plane. We employ Sloan Digital Sky Survey data from the Apache Point Observatory Galactic Evolution Experiment to test this hypothesis. Adopting a random sampling technique, we contrast the abundances of 10 elements in Terzan 5 stars with those of their bulge field counterparts with comparable atmospheric parameters, finding that they differ at statistically significant levels. Abundances between the two groups differ by more than 1σ in Ca, Mn, C, O, and Al, and more than 2σ in Si and Mg. Terzan 5 stars have lower [α/Fe] and higher [Mn/Fe] than their bulge counterparts. Given those differences, we conclude that Terzan 5 is not the remnant of a major building block of the bulge. We also estimate the stellar mass of the Terzan 5 progenitor based on predictions by the Evolution and Assembly of GaLaxies and their Environments suite of cosmological numerical simulations, concluding that it may have been as low as ∼3 × 108 M⊙ so that it was likely unable to significantly influence the mean chemistry of the bulge/inner disc, which is significantly more massive (∼1010 M⊙). We briefly discuss existing scenarios for the nature of Terzan 5 and propose an observational test that may help elucidate its origin.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2020

Accurate and efficient description of interacting carriers in quantum nanostructures by selected configuration interaction and perturbation theory

In this paper, we present a method to accurately and efficiently calculate many-body states of interacting carriers in quantum nanostructures based on a combination of iterative selection of configurations and perturbation theory. This method enables investigations of large excitonic complexes and multielectron systems with near full configuration interaction accuracy, even though only a small subspace of the full many-body Hilbert space is sampled, thus saving orders of magnitudes in computational resources. Important advantages of this method are that the convergence is controlled by a single parameter, the threshold, and that ground and excited states can be treated on an equal footing. On the example of InAsP nanowire quantum dots described using a million-atom tight-binding approach, we demonstrate the extreme efficiency of the method by numerical studies of large excitonic complexes filling up to the fourth electronic shell. We find that the method generally converges fast as a function of the threshold, profiting from a significant acceleration due to the perturbative corrections. The role of the choice of single-particle basis states is discussed. It is found that the algorithm converges faster in the Hartree-Fock basis only for highly charged systems, where Coulomb repulsion dominates. Finally, based on the observation that second-order perturbative energy corrections only depend on off-diagonal elements of the many-body Hamiltonian, we present a way to accurately calculate many-body states that requires only a relatively small number of Coulomb matrix elements.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Monte Carlo control loops for cosmic shear cosmology with DES Year 1 data

Weak lensing by large-scale structure is a powerful probe of cosmology and of the dark universe. This cosmic shear technique relies on the accurate measurement of the shapes and redshifts of background galaxies and requires precise control of systematic errors. Monte Carlo control loops (MCCL) is a forward modeling method designed to tackle this problem. It relies on the ultra fast image generator (UFig) to produce simulated images tuned to match the target data statistically, followed by calibrations and tolerance loops. Here, we present the first end-to-end application of this method, on the Dark Energy Survey (DES) Year 1 wide field imaging data. We simultaneously measure the shear power spectrum $C_ℓ$ and the redshift distribution $n(z)$ of the background galaxy sample. The method includes maps of the systematic sources, point spread function (PSF), an approximate Bayesian computation (ABC) inference of the simulation model parameters, a shear calibration scheme, and a fast method to estimate the covariance matrix. We find a close statistical agreement between the simulations and the DES Y1 data using an array of diagnostics. In a nontomographic setting, we derive a set of $C_ℓ$ and $n(z)$ curves that encode the cosmic shear measurement, as well as the systematic uncertainty. Following a blinding scheme, we measure the combination of $Ω_m$, $σ_8$, and intrinsic alignment amplitude $A_{IA}$, defined as $S_8D_{IA}=σ_8(Ω_m/0.3)^{0.5}D_{IA}$, where $D_{IA}=1-0.11(A_{IA}-1)$. We find $S_8D_{IA}=0.8954_{-0.039}^{+0.054}$, where systematics are at the level of roughly 60% of the statistical errors. We discuss these results in the context of earlier cosmic shear analyses of the DES Y1 data. Our findings indicate that this method and its fast runtime offer good prospects for cosmic shear measurements with future wide-field surveys.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2020

A Step-Down Test Procedure for Wavelet Shrinkage Using Bootstrapping

Wavelet thresholding (or shrinkage) attempts to remove the noises existing in the signals while preserving inherent pattern characteristics in the reconstruction of true signals. For data-denoising purpose, we present a new wavelet thresholding procedure which employs the step-down testing idea of identifying active contrasts in unreplicated fractional factorial experiments. The proposed method employs bootstrapping methods to a step-down test for thresholding wavelet coefficients. By introducing the concept of a false discovery error rate in testing wavelet coefficients, we shrink the wavelet coefficients with p -values higher than the error rate. The error rate controls the expected proportion of wrongly accepted coefficients among chosen wavelet coefficients. Bootstrap samples are used to approximate the p -value for computational efficiency. We also present some guidelines for selecting the values of hyper-parameters which affect the performance in the step-down thresholding procedure. Based on some common testing signals and an air-conditioner sounds example, the comparison of our proposed procedure with other thresholding methods in the literature is performed. The analytical results show that the proposed procedure has a potential in data-denoising and data-reduction in a variety of signal reconstruction applications.

42 ENGINEERING↗
DOE OSTI2023

AutoTG: Reinforcement Learning-Based Symbolic Optimization for AI-Assisted Power Converter Design

Power converters are pervasive in modern electronic component design. They can be found in all electronic devices from household appliances and cellphone chargers to vehicles. Currently, designing new circuit topologies is hard because it requires human expertise based on experience and is difficult to automate. However, artificial-intelligence-assisted design can significantly facilitate the development of new power converters and/or improve the final result. Intelligently designed highly efficient power converters can have a significant effect on many important attributes, such as power efficiency, layout size, cost, heat dissemination, energy requirements, etc. We propose Autonomous Topology Generator (AutoTG), a reinforcement-learning-based framework that generates power converter topology candidates based on user specifications, optimized for user preferences. By modeling power converter design as a symbolic optimization problem, we sequentially sample components in an autoregressive manner until new topologies are formed, providing both the topology specification and the sizing (magnitude of each component parameter) of the proposed power converter. Here, we provide an empirical evaluation and show that AutoTG is able to generate varied high-efficiency topologies within component restrictions based on user input and show that previously unknown topologies can be found for further evaluation.

(AI)-based design↗
DOE OSTI2020

Modeling thin layers of analytes on substrates for spectral analysis: use of solid/liquid n and k values to model reflectance spectra

Since solids are only sometimes seen en masse in a pure bulk form, and for liquids other than water almost never, a capability to model reflectance spectra from analytes deposited on various substrates would be highly advantageous. If available, the real, n ( ν ) , and imaginary, k ( ν ) , components of the complex refractive index, n n + i k , can be used to simulate infrared spectra, accounting for reflection, refraction, and absorption phenomena as a function of wavelength. We focus on using the Pacific Northwest National Laboratory (PNNL) derived n / k vectors for solid and liquid analytes deposited as thin layers on different types of substrates including conductors, such as aluminum, and inorganic dielectrics, such as glass. The model is an adaptation of the Monte Carlo ray trace modeling program, TracePro, extended through the use of its macrolanguage. The model is tested using thin films of organic liquids including silicone oil and no. 2 diesel fuel, as well as organic solids such as caffeine and acetaminophen on aluminum and glass. The predicted spectra for the solid films were compared to experimental hemispherical reflectance data measured using a Fourier transform spectrometer with an integrating sphere. The thickness of the calculated layer is a parameter for predicting the (transflectance) spectra and is obtained using the areal density measured from gravimetric methods to generate the thin-layer samples. Comparison of the calculated spectra with experimental hemispherical reflectance data shows excellent agreement, indicating promise for the use of measured n / k data to synthesize reference spectral data. In particular, accounting for the inhomogeneity of the deposits greatly improved the match with experimental data. Finally, the theoretical modeling shows that for thicker layers (ca. 20 to 100 μm) of typical organics possessing moderately strong k values, the longwave infrared features are often saturated and better spectral contrast is obtained from the overtone/combination bands in the shortwave infrared.

42 ENGINEERING↗
DOE OSTI2026

Removal of trace gases can both increase and decrease cloud droplet formation

Aerosols consist of liquid or solid particles dispersed in a gas. Aerosol measurements generally rely on drying the particles before quantifying their physicochemical properties. This drying can potentially remove semivolatile compounds from the particles. Here, we show size-resolved cloud condensation nuclei (CCN) measurements quantifying the hygroscopicity parameter in the presence and absence of a denuder. The denuder efficiently removed alkanes and weakly functionalized acids, aldehydes, and alcohols with fewer than 10 carbon atoms from the gas phase. Denuding organic compounds perturbed the CCN-derived hygroscopicity parameter by up to 50%. Denuding either rendered the particles more or less CCN active, and the direction of the effect depended on sample relative humidity and trace gas concentration. The effect was weakest in early spring and strongest in late spring and summer. The measurements demonstrate an unexpectedly strong coupling between the particle and gas phase, influencing CCN activity through either volatilization or surface adsorption, or both.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

DOE JGI Metagenome Workflow

The DOE Joint Genome Institute (JGI) Metagenome Workflow performs metagenome data processing, including assembly; structural, functional, and taxonomic annotation; and binning of metagenomic data sets that are subsequently included into the Integrated Microbial Genomes and Microbiomes (IMG/M) (I.-M. A. Chen, K. Chu, K. Palaniappan, A. Ratner, et al., Nucleic Acids Res, 49:D751–D763, 2021, https://doi.org/10.1093/nar/gkaa939) comparative analysis system and provided for download via the JGI data portal (https://genome.jgi.doe.gov/portal/). This workflow scales to run on thousands of metagenome samples per year, which can vary by the complexity of microbial communities and sequencing depth. Here, we describe the different tools, databases, and parameters used at different steps of the workflow to help with the interpretation of metagenome data available in IMG and to enable researchers to apply this workflow to their own data. We use 20 publicly available sediment metagenomes to illustrate the computing requirements for the different steps and highlight the typical results of data processing. The workflow modules for read filtering and metagenome assembly are available as a workflow description language (WDL) file (https://code.jgi.doe.gov/BFoster/jgi_meta_wdl). The workflow modules for annotation and binning are provided as a service to the user community at https://img.jgi.doe.gov/submit and require filling out the project and associated metadata descriptions in the Genomes OnLine Database (GOLD) (S. Mukherjee, D. Stamatis, J. Bertsch, G. Ovchinnikova, et al., Nucleic Acids Res, 49:D723–D733, 2021, https://doi.org/10.1093/nar/gkaa983).

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2020

Effect of Using Regenerated Combined FAU and MOR Zeolites as Catalysts during the Pyrolysis of Kraft Lignin

The SiO 2 /Al 2 O 3 mole ratio, pore size, and acid sites are the key parameters of zeolite’s activity in lignin pyrolysis. In this study, the comparison of individual Y and M zeolites, the combined ‘Y + M’ sample after regeneration, and their effect on lignin pyrolysis were studied in five cycles (regeneration and reuse). The results were explained using Brunauer, Emmet, and Teller (BET), micropore surface area (MSA), and total acid sites (TAS) analyses. In comparison with the individual Y or M zeolite sample, the consistent higher catalytic activities of the combined ‘Y + M’ sample in repeated cycles were observed. Pyrolysis heavy oils were characterized by nuclear magnetic resonance (NMR) spectroscopy and gel permeation chromatography (GPC). The NMR analyses revealed that with increased zeolite regeneration cycles, p-hydroxy phenyl and methoxyl groups increased. Decreases in guaiacyl phenolic hydroxyl were less for the combined ‘Y + M’ sample than the individual Y and M zeolites. Lower weight average (M w ) of heavy oil for the combined ‘Y + M’ sample indicated the enhanced cleavage of lignin structures in pyrolysis. These results support the higher catalytic activity of regenerated zeolites for the combined ‘Y + M’ sample compared with individual Y and M zeolites due to the improved MSA and TAS.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2020

Weak-lensing Analysis of X-Ray-selected XXL Galaxy Groups and Clusters with Subaru HSC Data

We present a weak-lensing analysis of X-ray galaxy groups and clusters selected from the XMM-XXL survey using the first-year data from the Hyper Suprime-Cam (HSC) Subaru Strategic Program. Our joint weaklensing and X-ray analysis focuses on 136 spectroscopically confirmed X-ray-selected systems at 0.031 6 z 6 1.033 detected in the 25 deg2 XXL-N region, which largely overlaps with the HSC-XMM field. With high-quality HSC weak-lensing data, we characterize the mass distributions of individual clusters and establish the concentration–mass (c–M) relation for the XXL sample, by accounting for selection bias and statistical effects and marginalizing over the remaining mass calibration uncertainty. We find the mass-trend parameter of the c–M relation to be β = -0.07 ± 0.28 and the normalization to be c200 = 4.8 ± 1.0 (stat) ± 0.8 (syst) at M 200 = 1014h -1M and z = 0.3. We find no statistical evidence for redshift evolution. Our weaklensing results are in excellent agreement with dark-matter-only c–M relations calibrated for recent ΛCDM cosmologies. The level of intrinsic scatter in c200 is constrained as σ(ln c 200 ) < 24% (99.7% CL), which is smaller than predicted for the full population of ΛCDM halos. This is likely caused in part by the X-ray selection bias in terms of the cool-core or relaxation state. We determine the temperature–mass (TX–M 500 ) relation for a subset of 105 XXL clusters that have both measured HSC lensing masses and X-ray temperatures. The resulting TX–M 500 relation is consistent with the self-similar prediction. Our TX–M 500 relation agrees with the XXL DR1 results at group scales but has a slightly steeper mass trend, implying a smaller mass scale in the cluster regime. The overall offset in the TX–M 500 relation is at the ~ 1.5σ level, corresponding to a mean mass offset of (34 ± 20)%. We also provide bias-corrected, weak-lensing-calibrated M 200 and M 500 mass estimates of individual XXL clusters based on their measured X-ray temperatures.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

Stellar Abundance Maps of the Milky Way Disk

To understand the formation of the Milky Way's prominent bar it is important to know whether stars in the bar differ in the chemical element composition of their birth material as compared to disk stars. This requires stellar abundance measurements for large samples across the Milky Way's body. Such samples, e.g., luminous red giant stars observed by the Sloan Digital Sky Survey's APOGEE survey, will inevitably span a range of stellar parameters; as a consequence, both modeling imperfections and stellar evolution may preclude consistent and precise estimates of their chemical composition at a level of purported bar signatures, which has left current analyses of a chemically distinct bar inconclusive. Here, we develop a new self-calibration approach to eliminate both modeling and astrophysical abundance systematics among red giant branch (RGB) stars of different luminosities (and hence surface gravity $\mathrm{log} g$). We apply our method to 48,853 luminous APOGEE Data Release 16 RGB stars to construct spatial abundance maps of 20 chemical elements near the Milky Way's mid-plane, covering galactocentric radii of 0 kpc < R GC < 20 kpc. Our results indicate that there are no abundance variations whose geometry matches that of the bar, and that the mean abundance gradients vary smoothly and monotonically with galactocentric radius. We confirm that the high-α disk is chemically homogeneous, without spatial gradients. Furthermore, we present the most precise [Fe/H] versus R GC gradient to date with a slope of – 0.057 ±0.001 dex kpc –1 out to approximately 15 kpc.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

Multi-Fidelity Modeling and Control for Building Temperature Control

The ability to control energy loads such as a building's heating, ventilation, and air conditioning (HVAC) system can help facilitate increased penetration of variable renewable energy sources into the electric grid. To be able to control these HVAC systems more effectively, detailed simulations of the corresponding building physics is becoming increasingly important. These detailed simulations can be complex, nonlinear, and can require immense computational power when used in an advanced control method such as model predictive control (MPC), prompting the need to explore less computationally intensive strategies. In this work, a multi-fidelity approach is proposed to combine samples from a complex, high-fidelity model with a simple, low-fidelity model within the MPC control loop. More specifically, the parameters of a reduced-order, linear building model are periodically updated with knowledge from its high-fidelity counterpart - an EnergyPlus model - in an online fashion using a Gaussian Process surrogate model. Hence, highly accurate predictions of current and future conditions in a building are maintained with a substantially reduced computational burden compared to using the high-fidelity models alone. In other words, this linear parameter varying model preserves the low computational requirements of a low-order linear model while accurately modeling a building's dynamics. This allows a building controller to take highly informed actions without requiring a large computational budget.

building modeling↗
DOE OSTI2021

Machine Learning Augmented Predictive and Generative Model for Rupture Life in Ferritic and Austenitic Steels

The Larson-Miller parameter (LMP) offers an efficient and fast scheme to estimate the creep rupture life of alloy materials for high temperature applications. However, owing to poor generalizability and dependence on the constant C, which is typically not known a-priori, estimations using the Larson-Miller parameter often result in suboptimal performance for a wide range of materials. At best it is useful in comparing alloys of similar composition. In this work, three machine learning (ML) schemes were developed for rupture life prediction for 9-12% Cr ferritic-martensitic steels and austenitic stainless steels, i.e., a hierarchical model to parameterize LMP using the LMP constant C to compute rupture life, a hierarchical model to parameterize both C and LMP to compute rupture life, and a direct prediction of rupture life. Specifically, we show that the third scheme, using a gradient boosting algorithm, can be used to train ML models for very accurate prediction of rupture life in a variety of alloys (Pear-son Correlation Coefficient > 0.9 for 9-12% Cr and > 0.8 for austenitic stainless steels). In addition, the Shapley value was used to quantify feature importance, making the model interpretable by identifying the effect of various features on the model performance. Furthermore, a variational autoencoder-based generative model was built by conditioning on the experimental dataset to sample hypothetical synthetic candidate alloys from the learnt joint distribution not existing in both 9-12 % Cr ferritic-martensitic steel and austenitic stainless steel datasets. Finally, based on the predictive and generative model, a reinforcement learning strategy has been proposed to guide experimentalists into designing better heat resistant alloys.

Mamun, Md Osman G.↗