Inverse design of operational parameters for iron-rich phase separation from solid waste melts in induction furnace through transport phenomena analysis and multi-objective optimization

Abstract

Improving the removal efficiency of iron-rich impurities and the uniformity of component distribution is an indispensable step in recycling aluminosilicate solid waste to prepare the raw materials of glass ceramics, which can be achieved by smelting reduction separation in the induction furnace. In this work, a multi-physics real-time coupling model that integrates carbothermal reduction and interfacial mass transfer between two-phase melts in the induction furnace was proposed to improve the conventional methods that simulate the interface profile and multi-physics separately. By solving the numerical model, the reduction degree, energy consumption and component distribution can be predicted and the characteristics of mass-heat-component transport and interphase mass transfer can be obtained. For the collaborative optimization of multiple parameters, including current, frequency, and melting duration, an inverse design method that integrates CFD-ANN-NSGA-II was proposed to minimize the iron oxide (FeO) residual and energy consumption, and maximize the uniformity of components in the glass phase simultaneously. The optimal operating parameters set were determined by analysis of the Pareto front and the selection criteria. Further analysis of the Pareto solution set provides a fundamental basis for regulating the operational parameters of different reduction degrees. Overall, the inverse design method that synthesized the proposed numerical model is an effective way of predicting and optimizing the operating parameters of induction heating furnaces of iron removal from solid wastes.