Quantification of micro-damage evolution process in ceramics through extensive analysis of multi-source heterogeneous data

Abstract

Due to the randomness of hard and brittle material damage evolution, assessing damage and interpreting mechanisms is challenging. This study focuses on analyzing material damage mechanisms by employing extensive multi-source heterogeneous data to extract quantitative information. A machine vision-based method for detecting and assessing defects in ceramics was proposed, achieving 98 % accuracy. Leveraging this method in scratching experiments, the progression of damage and micro-failure mechanisms in Si3N4 was investigated through the analysis of defect area distribution, force and acoustical signals. Results revealed that the distribution of defect areas conforms to a three-parameter Weibull distribution. At scratch depths of 0.01, 0.02 mm, approximately 90 % of damage accumulated within 1e2–1e4 μm2; conversely, 0.03–0.06 mm led to a decrease to 30 – 40 %. Extensive crack propagation occurs within 0.02–0.03 mm, resulting in the intersecting formation of spalling blocks. Damage evolution occurs in three stages: shallow and small cracks with scattered spalling blocks in the first stage, deep and long cracks with continuous and irregularly shaped spalling blocks in the second stage, and nearly no cracks with continuous and regularly shaped blocks in the third stage. This study not only proposes an efficient method to detect and evaluate ceramic defects, but also deeply understands the evolution process of material damage mechanism. Considering defect distribution and statistical trends, this research not only provide guidance for the optimization of machining parameters, especially in the scene of specific precision and high removal rate. At the same time, the damage evolution and failure mechanism of Si3N4 provide useful experimental basis and data support for deepening the understanding of material behavior during ultra-precision grinding.