熔盐-液体金属相互溶解度的规律性

REGULARITIES OF MUTUAL SOLUBILITY OF MOLTEN SALT-LIQUID METAL SYSTEMS

应用原子参数（阴阳离子半径，价电子数，表征金属元素价电子云密度的参数）-人工神经网络方法研究了熔盐-液体金属系的互溶度，会溶温度及偏晶点成分的规律．研究结果表明，金属元素的愈大，熔盐在液体金属中溶解度愈小训练后的神经网络可以预报熔盐-液体金属系的会溶温度和偏晶点成分．

Atomic parameter-artificial neural network method has been applied to study the regularities of the mutual solubility, the consolute temperatures, and the composition of monotectic points of molten salt-liquid metal systems. It has been found that larger melt's is correspond to lower mutual solubility. Trained artifical neural can be used to predict the consolute tempertures and the composition of monotectic points of molten salt-liquid metal systems. The error of the results of computerized prediction is rather small.