skip to main content

Title: Deep Learning-Based Hardness Prediction of Novel Refractory High-Entropy Alloys with Experimental Validation
Hardness is an essential property in the design of refractory high entropy alloys (RHEAs). This study shows how a neural network (NN) model can be used to predict the hardness of a RHEA, for the first time. We predicted the hardness of several alloys, including the novel C0.1Cr3Mo11.9Nb20Re15Ta30W20 using the NN model. The hardness predicted from the NN model was consistent with the available experimental results. The NN model prediction of C0.1Cr3Mo11.9Nb20Re15Ta30W20 was verified by experimentally synthesizing and investigating its microstructure properties and hardness. This model provides an alternative route to determine the Vickers hardness of RHEAs.
; ; ; ; ;
Award ID(s):
Publication Date:
Journal Name:
Page Range or eLocation-ID:
Sponsoring Org:
National Science Foundation
More Like this
  1. Abstract

    The compositional dependence and influence of relaxation state on the deformation behavior of a Pt–Pd-based bulk metallic glasses model system was investigated, where platinum is systematically replaced by topologically equivalent palladium atoms. The hardness and modulus increased with rising Pd content as well as by annealing below the glass transition temperature. Decreasing strain-rate sensitivity and increasing serration length are observed in nano indentation with increase in Pd content as well as thermal relaxation. Micro-pillar compression for alloys with different Pt/Pd ratios validated the greater tendency for shear localization and brittle behavior of the Pd-rich alloys. Based on total scattering experiments with synchrotron X-ray radiation, a correlation between the increase in stiffer 3-atom cluster connections and reduction in strain-rate sensitivity, as a measure of ductility, with Pd content and thermal history is suggested.

  2. null (Ed.)
    ABSTRACT The abundances of neutron (n)-capture elements in the carbon-enhanced metal-poor (CEMP)-r/s stars agree with predictions of intermediate n-density nucleosynthesis, at Nn ∼ 1013–1015 cm−3, in rapidly accreting white dwarfs (RAWDs). We have performed Monte Carlo simulations of this intermediate-process (i-process) nucleosynthesis to determine the impact of (n,γ) reaction rate uncertainties of 164 unstable isotopes, from 131I to 189Hf, on the predicted abundances of 18 elements from Ba to W. The impact study is based on two representative one-zone models with constant values of Nn = 3.16 × 1014 and 3.16 × 1013 cm−3 and on a multizone model based on a realistic stellar evolution simulation of He-shell convection entraining H in a RAWD model with [Fe/H] = −2.6. For each of the selected elements, we have identified up to two (n,γ) reactions having the strongest correlations between their rate variations constrained by Hauser–Feshbach computations and the predicted abundances, with the Pearson product–moment correlation coefficients |rP| > 0.15. We find that the discrepancies between the predicted and observed abundances of Ba and Pr in the CEMP-i star CS 31062−050 are significantly diminished if the rate of 137Cs(n,γ)138Cs is reduced and the rates of 141Ba(n,γ)142Ba or 141La(n,γ)142La increased. The uncertainties of temperature-dependent β-decay rates of the same unstable isotopes have amore »negligible effect on the predicted abundances. One-zone Monte Carlo simulations can be used instead of computationally time-consuming multizone Monte Carlo simulations in reaction rate uncertainty studies if they use comparable values of Nn. We discuss the key challenges that RAWD simulations of i process for CEMP-i stars meet by contrasting them with recently published low-Z asymptotic giant branch (AGB) i process.« less
  3. The microstructure, Vickers hardness, and compressive properties of novel low-activation VCrFeTaxWx (x = 0.1, 0.2, 0.3, 0.4, and 1) high-entropy alloys (HEAs) were studied. The alloys were fabricated by vacuum-arc melting and the characteristics of these alloys were explored. The microstructures of all the alloys exhibited a typical morphology of dendritic and eutectic structures. The VCrFeTa0.1W0.1 and VCrFeTa0.2W0.2 alloys are essentially single phase, consisting of a disordered body-centered-cubic (BCC) phase, whereas the VCrFeTa0.2W0.2 alloy contains fine, nanoscale precipitates distributed in the BCC matrix. The lattice parameters and compositions of the identified phases were investigated. The alloys have Vickers hardness values ranging from 546 HV0.2 to 1135 HV0.2 with the x ranging from 0.1 to 1, respectively. The VCrFeTa0.1W0.1 and VCrFeTa0.2W0.2 alloys exhibit compressive yield strengths of 1341 MPa and 1742 MPa, with compressive plastic strains of 42.2% and 35.7%, respectively. VCrFeTa0.1W0.1 and VCrFeTa0.2W0.2 alloys have excellent hardness after annealing for 25 h at 600–1000 °C, and presented compressive yield strength exceeding 1000 MPa with excellent heat-softening resistance at 600–800 °C. By applying the HEA criteria, Ta and W additions into the VCrFeTaW are proposed as a family of candidate materials for fusion reactors and high-temperature structural applications.
  4. Abstract

    In the last decade, much work in atmospheric science has focused on spatial verification (SV) methods for gridded prediction, which overcome serious disadvantages of pixelwise verification. However, neural networks (NN) in atmospheric science are almost always trained to optimize pixelwise loss functions, even when ultimately assessed with SV methods. This establishes a disconnect between model verification during versus after training. To address this issue, we develop spatially enhanced loss functions (SELF) and demonstrate their use for a real-world problem: predicting the occurrence of thunderstorms (henceforth, “convection”) with NNs. In each SELF we use either a neighborhood filter, which highlights convection at scales larger than a threshold, or a spectral filter (employing Fourier or wavelet decomposition), which is more flexible and highlights convection at scales between two thresholds. We use these filters to spatially enhance common verification scores, such as the Brier score. We train each NN with a different SELF and compare their performance at many scales of convection, from discrete storm cells to tropical cyclones. Among our many findings are that (i) for a low or high risk threshold, the ideal SELF focuses on small or large scales, respectively; (ii) models trained with a pixelwise loss function performmore »surprisingly well; and (iii) nevertheless, models trained with a spectral filter produce much better-calibrated probabilities than a pixelwise model. We provide a general guide to using SELFs, including technical challenges and the final Python code, as well as demonstrating their use for the convection problem. To our knowledge this is the most in-depth guide to SELFs in the geosciences.

    Significance Statement

    Gridded predictions, in which a quantity is predicted at every pixel in space, should be verified with spatially aware methods rather than pixel by pixel. Neural networks (NN), which are often used for gridded prediction, are trained to minimize an error value called the loss function. NN loss functions in atmospheric science are almost always pixelwise, which causes the predictions to miss rare events and contain unrealistic spatial patterns. We use spatial filters to enhance NN loss functions, and we test our novel spatially enhanced loss functions (SELF) on thunderstorm prediction. We find that different SELFs work better for different scales (i.e., different-sized thunderstorm complexes) and that spectral filters, one of the two filter types, produce unexpectedly well calibrated thunderstorm probabilities.

    « less
  5. The empirical parameters of mixing enthalpy (ΔHmix), mixing entropy (ΔSmix), atomic radius difference (δ), valence electron concentration (VEC), etc., are used in this study to design a depleted uranium high-entropy alloy (HEA). X-ray diffraction (XRD), scanning electron microscopy (SEM), and transmission electron microscopy (TEM) were used to assess the phase composition. Compression and hardness tests were conducted to select alloy constituents with outstanding mechanical properties. Based on the experimental results, the empirical criteria of HEAs are an effective means to develop depleted uranium high-entropy alloys (DUHEAs). Finally, we created UNb0.5Zr0.5Mo0.5 and UNb0.5Zr0.5Ti0.2Mo0.2 HEAs with outstanding all-round characteristics. Both alloys were composed of a single BCC structure. The hardness and strength of UNb0.5Zr0.5Mo0.5 and UNb0.5Zr0.5Ti0.2Mo0.2 were 305 HB and 1452 MPa, and 297 HB and 1157 MPa, respectively.