Wilfried Schäfer, Götz Hartmann, Erik Hepp
Over the last decades, many research groups have studied the mould behavior and the solidification in the continuous casting of steel billets. The process was analyzed by various measurement techniques, calculations and/or analytical or numerical simulations. For a better understanding of the detailed phenomena influencing the casting process performance and billet quality, it is possible today to combine newest developments in coupled 3D numerical heat and mass transport simulations with computational optimization methods based on genetic algorithms. In this contribution, the authors show the use of these genetic algorithms for autonomous multi-objective numerical optimization of the continuous casting process of steel billets. A primary objective for the optimization was to find the best possible coupling between casting speed, spray nozzle layout and liquid pool depth. This has to be done while still keeping the surface temperatures within narrow constraints. The overall goal is to increase the productivity and to make no compromise concerning the quality of the final product. The use of a multi-objective optimization algorithm made it possible to follow each of these objectives simultaneously. Further evaluation of the results allows an estimation of the sensitivity of the casting quality to process parameters.