Optimisation of the operating parameters of a curved continuous casting mould under electromagnetic stirring (EMS) using the response surface method

In order to investigate and mitigate carbon segregation behaviour in the curved mould region during the continuous casting of 46MnVS5 steel, this study developed a multiphysics coupled model for an R10m curved mould. Four operating variables, namely current intensity, frequency, casting speed, and cooling intensity, were systematically manipulated to investigate their individual and interactive effects on carbon distribution and segregation behaviour in the curved mould under electromagnetic stirring. Based on a Box-Behnken design, response surface methodology was employed to establish predictive models for the positive segregation index (PSI) and negative segregation index (NSI), which were taken as the optimisation responses for carbon macrosegregation characteristics at the mould outlet. It shows that increasing the intensity from 200 A to 400 A reduced the PSI by 0.86%, with the optimal NSI of 0.7924 achieved at 300 A. Increasing the frequency from 2 Hz to 4 Hz decreased the PSI by 0.54%, with the minimum NSI of 0.7924 observed at 3 Hz. Casting speed had the most significant effect on segregation: increasing the speed from 2.4 m·min −1 to 3.6 m·min −1 reduced the positive and negative segregation indices by 9.94% and 6.94%, respectively. Although rising the cooling intensity effectively suppressed positive segregation, it concurrently aggravated negative segregation. Through multi-objective optimisation, the optimal process parameters were obtained, including a current intensity of 351.062 A, a frequency of 3.867 Hz, a casting speed of 3.6 m·min −1 , and a cooling intensity of 1200 kW·m −2 . The relative influence of each parameter on segregation decreased in the following order: casting speed > cooling intensity > current intensity > frequency. These findings provide a theoretical basis and practical guidance for controlling carbon segregation at the mould outlet.