Materials selection of furnace linings with multi-component refractory ceramics based on an evolutionary screening procedure

Abstract

Materials selection of multi-component systems is a challenging task, which is usually not properly tackled in furnace linings (FL) design. In an attempt to generate a systematic approach to select FL ceramic materials, an evolutionary screening procedure (ESP) is proposed in this paper, where a multi-objective genetic algorithm coupled with a finite element model aims to transform a list of potentially feasible ceramic candidate materials into an optimized set of multi-component systems simultaneously computing external lining temperatures and costs. An ESP-based selection methodology is proposed and studied in light of a case-study concerning an electric resistance furnace. A customized initial screening was carried out to select potential single candidate materials to be used. Next, the ESP was used to select optimized lining configurations while discarding infeasible combinations, such as those in which one of the materials maximum allowed temperature had been breached. Afterwards, TOPSIS (a multi-criteria decision-making approach) was applied to rank the optimized results according to some furnace preference scenarios. The ESP-based selection diminished a universe of 1.9×105 lining configurations into a few units to be investigated in more details by simulating only 3.8% of them, thus requiring reduced amount of time for completing this task. The adopted quantitative strategy uses commonly used computational techniques and could be applied to metallurgical ladle linings, be adapted to select raw materials or resolve other complex decision-making problems in the ceramic community.