Events
Place: CALPHAD XLVII Conference, Juriquilla, Querétaro, México
Irina Roslyakova
Setareh Zomorodpoosh
Abdulmonem Obaied
Brandon Bocklund, Department of Materials Science and Engineering, Pennsylvania State University, University Park, USA
Richard Otis, Engineering and Science Directorate, California Institute of Technology, Pasadena, CA, USA
Zi-Kui Liu, Department of Materials Science and Engineering, The Pennsylvania State University, University Park, USA
The aim of the third generation of CALPHAD databases is to develop an accurate physically-based description of Gibbs energies of pure elements from 0K up to high temperatures that can guarantee a robust prediction of thermodynamic properties of binary, ternary and high-order systems within an entire temperature range. Recently, one of the successful attempts in this direction has been performed by Roslyakova et al. [1]. A novel thermodynamic modeling strategy of stable, metastable and liquid phases is proposed based on automated segmented regression (SR) approach. The SR model [1] contains only 5 parameters to be estimated from experimental data, which can be weighted in fully automated way based on application of unequal k-fold cross-validation method [2]. Moreover, some of the unknown models parameters are associated with physically-based constants and therefore can be compared with available experimental data. The segmented regression model has been successfully applied for 11 pure elements: Al, Cr, Fe [1], Ge, Ir, Mo, Nb, Re, W, Ni, Ta, binary Cr-Nb [2] and ternary Mo-Nb-Ta systems. In this work the application of automated proposed modeling approach has been extended to other 7 elements, such as Ag, Cu, Mg, Bi, Te, Si, Pb and some of their compounds, such as Cr2Nb [3] and Cr2Ta. Comparison with available experimental data showed a good agreement in low and high temperatures for all considered pure elements (see Figure 1).
References:
[1] Roslyakova, I., et al. „Modeling of Gibbs energies of pure elements down to 0K using segmented regression”, CALPHAD Journal, 55 (2016).
[2] Zomorodpoosh S., et al, Statistical approach for automated weighting of experimental data and outlier detection”, CALPHAD 2018 conference (accepted talk).
[3] Jiang, Y., et al., “Thermodynamic reassessment of binary Cr-Nb system down to 0 K”, CALPHAD 62 (2018), 109-118.
