Phase Detection of Sintered Ceramic Materials Using Machine Learning

Lai, John Vin (2026) Phase Detection of Sintered Ceramic Materials Using Machine Learning. Final Year Project (Bachelor), Tunku Abdul Rahman University of Management and Technology.

Text LAI JOHN VIN_Full Text.pdf Restricted to Registered users only Download (24MB)

Abstract

Sintered zirconia ceramics are valued for their strength, heat resistance, and electrical properties. Phase detection in sintered zirconia ceramics helps us to know which crystal phases are present and helps us predict how these materials will behave in real-world uses. However, the phase detection process relies on time-intensive, expert-driven methods like X-ray diffraction (XRD) and Scanning Electron Microscopy (SEM) analysis. This project proposes to train a machine learning (ML) model on SEM images of stabilized zirconia sintered at temperatures ranging from 1300 °C to 1500 °C to classify the main zirconia phases and any secondary phases that may be present. Evaluation performance such as precision, recall, F1-score, and mean average precision with confusion matrices will be used to identify phase distributions, and then quantifying phase fractions to perform linear regression against sintering temperature. This research aims to deliver a validated ML workflow that accelerates phase detection and produces insights into how sintering conditions influence phase composition, supporting faster quality control and future materials research