Artificial Intelligence‐based determination of fracture toughness and bending strength of silicon nitride ceramics

Abstract

AbstractTwo mechanical properties, fracture toughness (KIC) and bending strength (σ), of silicon nitride (Si3N4) ceramics were determined from their microstructural images via convolutional neural network (CNN) models. The Si3N4 samples used for database were fabricated using various kinds of sintering additives under different process conditions. In total, 330 data sets were prepared and used for building the CNN models for artificial intelligence‐bassed determination of the two mechanical properties and testing the determination accuracy of the trained models. The determination coefficients (R2), which were used as accuracy indices, were approximately 0.85 for KIC and 0.92 for σ. Although both the R2 values were relatively high, the lower value for KIC suggests that it is influenced more by what is little obtained from the microstructural information, such as grain‐boundary characteristics. Furthermore, gradient‐weighted class activation mapping, which can visualize which parts of the image the CNN models focus on, showed that the trained models determined the two mechanical properties based on correct recognition of the microstructural difference among the images.