Mathematical modelling of sensitization and desensitization in austenitic alloys using Gaussian-based peak models

Abstract This study presents four Gaussian-based mathematical models for accurately simulating the sensitization and desensitization processes in austenitic alloys. By incorporating temperature and time as variables, these models aim to predict the Degree of Sensitization (DOS) and generate Time-Temperature-Sensitization (TTS) diagrams. Validation using experimental datasets shows that the models perform well across various datasets, minimizing over-fitting when peak sensitization data is adequately provided. The models also introduce a statistical index to assess accuracy, combining fitting parameters systematically. This work offers a comprehensive tool for investigating localized corrosion phenomena and supports the development of strategies to control chromium depletion during heat treatments, relevant for stainless steel and nickel-based alloys.