Latest revision as of 11:04, 23 October 2024

Abstract

This study explores the use of machine learning (ML) models in predicting the macroscopic properties of heterogeneous composites. Traditional micromechanics parameters have limitations, thus ML models with and without feature engineering are utilized. For artificial neural network (ANN) models with feature engineering, microstructural descriptors from SEM images of nickel-based superalloys are used to predict hardness. 10 descriptors are selected to reduce the computational cost of the deep neural network (DNN) with the support of the shallow neural network (SNN), and accuracy is enhanced by incorporating two additional descriptors. The result surpasses existing physics-based models. Models without feature engineering employ a convolutional neural network (CNN) to predict the effective thermal conductivity of thermal insulation composite materials. The CNN model demonstrates accurate predictions for novel microstructures. ML models can achieve more efficient predictions than traditional methods, indicating their potential in advancing materials science. In summary, harnessing artificial intelligence to capture the scattering characteristics of heterogeneous materials enables both DNN and CNN models to achieve more efficient predictions compared to traditional methods. This highlights the potential of machine learning in advancing materials science and expediting the development of materials with desired properties.

Full Paper