Computational Metallurgy applications are intelligent software solutions in the field of metallurgical process modeling. These process models map the different sub-processes of steel making in a holistic approach. Computational Metallurgy applications are useful tools for optimizing the entire production process and for supporting decision-making.
The detection of inefficiencies to minimize production costs is one of the key sucess factors of a smart factory. Next to predictive maintenance, process simulations and the adaptation of process parameters play an essential role in reducing production costs. These two subjects, the process simulation and the adaptation of process parameters, are the focus of our Computational Metallurgy applications.
Generally, on-line models calculate the current state of the metallurgical process in real time on the basis of all necessary process data and in combination with a variety of model parameters. However, a verification of this calculation can only be done when comparing calculation results and measurements (such as measured temperatures and analyses of the melt and slag). If deviations are detected, model parameters have to be adapted (in the simplest case) or model improvements have to be carried out. Subsequently, these changes or adjustments must be verified again. This time-consuming process can often only be done during industrial production. The effects of the modifications are therefore only visible very late, after a sufficient amount of production data has been collected and validated thoroughly. The concept of the digital twin helps to overcome this disadvantage:
By providing all relevant (cyclic) historic process data, these data can be used as input parameters for offline calculations. Both online and offline calculations are based on the same mathematical model core. On the one hand, this approach enables the evaluation and verification of new model parameters based on historic data. On the other hand, model parameters can be adapted quickly and efficiently. Above all, this approach represents the so-called digital twin and the possibility to optimize processes easily and quickly without affecting the current production.