Surface tension prediction of high-temperature melts

In this paper the artificial neural network (ANN) was applied to predict th e surface tension of pure liquid metals and Butler's original treatment in conjunction with the CALPHAD (CALculation of PHAse Diagram) technique was e xtended to predict surface tensions of binary, ternary and multi-component high-temperature melts: liquid alloys, molten salt mixtures and molten matt es. The model parameter beta in Butler's equation was discussed for differe nt high-temperature melts. In addition, the geometric models and general so lution model were used for predicting surface tensions of multi-component s ystems from those of sub-binary systems. The importance of the ternary inte raction term in geometric models and general solution model for ternary sys tems far away from regular solutions is emphasized. (C) 2001 Elsevier Scien ce B.V. All rights reserved.