Thermodynamic re-assessment of pure chromium using modified segmented regression model

Abstract

Chromium is considered an essential addition element in multiple industrial applications. Therefore, obtaining an accurate and robust description of its properties is important. Recently, several attempts were made to improve the thermodynamic description of pure elements and binary systems in the frame of the third generation CALPHAD databases. Moreover, it became increasingly important to develop mathematical tools that can allow researchers more access to each phase of the material over the entire temperature range. Thus, the recently proposed segmented regression model was modified using the so-called sigmoid function for the case of pure Cr. A comprehensive analysis of phase equilibria and thermodynamic properties of pure Cr was used to develop an updated unary description from 0 K to 6000 K using the proposed modified segmented regression (MSR) model. In addition, two newly developed machine learning tools were utilized to analyze the datasets used in this assessment to determine any possible outliers.