Improvement in Applicability of Commercially Available Slag Models for Thermodynamic Estimations in Steel Refining Slag

摘要

Most engineering materials are multi-component in nature. Processing of such a material is governed by the thermodynamic properties of its constituents. However, measurement of these thermodynamic parameters is not easy, and becomes increasingly complicated as one moves from binary and ternary to multi-component systems. In the absence of direct measurement, commercially available computer-based packages are increasingly being used by scientists and engineers for estimation of thermodynamic properties in multi-component systems. Users may be tempted to accept such results as adequately reliable. It should be noted that any such package is validated over a limited range of chemical composition, usually where experimentally measured values are available. Its use outside this range brings in the risk of significant error in estimation. Activities of individual constituents were measured experimentally in multi-component oxide systems containing CaO, SiO2, FeO, Fe2O3, MgO, P2O5 and Al2O3. The results were compared with values obtained using three different commercial thermodynamic packages, each based on a separate model and set of assumptions. Significant differences were observed while computing the thermodynamic activities of FeO, Fe2O3 and P2O5. Further, the error in estimation of activity values was found to be non-random. The results indicate that different packages should be used for estimating the thermodynamic properties of different species and individual models require further adjustments for improving the reliability of predictions. The observations are important for selection of appropriate package suitable for a particular material system.