application of bayesian ann and rjmcmc to predict the grain size of hot strip low carbon steels

Abstract

Artificial Neural Network (ANN) and Reversible Jump Markov Chain Monte Carlo (RJMCMC) are used to predict the grain size of hot strip low carbon steels, as a function of steel composition. Results show a good agreement with experimental data taken from Mobarakeh Steel Company (MSC). The developed model is capable of recognizing the role and importance of elements in grain refinement. Furthermore, effects of these elements including manganese, silicon and vanadium are investigated in the present study, which are in good agreement with the literature.