Prediction of Austenite Decomposition During Cooling of Low and Medium-Carbon Low-Alloy Steels

摘要

Based on experimental continuous cooling transformation (CC1) diagrams, an artificial neural network has been trained to predict the amount of microstructural constituents and hardness resulting from austenitizing and controlled cooling at rates from 150 to 0.05 K/s. It can be applied to quenching and normalizing processes of low-alloy steels with 0.05-0.30 C% and 0.4-2.0% Mn with possible microalloying additions. Examples of the good predictive capabilities (about 15 Vickers standard deviation for hardness and a few percent for volume fraction of phases) are reported.