Prediction of Steel Transformation Temperatures Using Thermodynamic Modeling and Design of Experiments (DOE)

Abstract

This paper presents a methodology in order to predict A1, solidus and liquidus temperatures using a relatively simple approach. The proposed approach is based on the combined use of the thermodynamic software Thermo-Calc and the composite centered design of experiments (DOE) method. Four important alloying elements (C, Ni, Mn and Cr) were considered in the DOE. The impact of each alloying element on the transformation temperatures was determined and discussed. It was found that carbon has the most important impact on solidus and liquidus whereas Ni, Mn, and Cr have a significant impact on A1. The proposed models were generated using Analysis of Variance (ANOVA) method. A good agreement between experimental and predicted results was found with a maximum error of 1.1 % for transformation temperatures. Furthermore, the proposed models were validated using a large amount of experimental data published in the literature with a maximum error equal to 7.8 %.