Molten salt phase diagrams calculation using artificial neural network or pattern recognition-bond parameters Part 1. The prediction of the phase diagrams of binary molten salt systems

Artificial neural network or pattern recognition together with chemical bond parameters method has been used to classify and predict the characteristics of the phase diagrams of binary molten salt systems. These characteristics are the formability, the chemical stoichiometry, the melting type and the melting point or decomposition temperature of intermediate compound and the formability of solid solution or eutectic mixture. The molten salt systems studied are some halide compounds such as MeX-Me'X, MeX-REX3 and MeX-Me'X-4(Me, Me' denote metallic elements, RE rare earth, X halogen) systems. The mathematical models obtained from the experimental data of the known phase diagrams were used to predict the properties of the unknown phase diagrams.