Mechanism-data hybrid-driven modeling of energy consumption and carbon emission for converter steelmaking process optimization

Iron and steel metallurgy is a high-energy consumption and high-carbon emission industry with global focus. The energy consumption and carbon emission of converter steelmaking are widely concerned as it is the core process of steel production. Existing work has mainly focused on modeling a single energy medium for converters and optimizing energy consumption and carbon emission from the perspective of the entire iron and steel metallurgical process. However, this research has failed to delve deeply into the energy-saving and carbon reduction potential of key steel production processes. Therefore, this research proposes an energy-saving and carbon reduction optimization method for converter based on mechanism-data hybrid modeling. The energy consumption and carbon emission characteristics of converter are first analyzed. An extreme gradient boosting combined with whale optimization algorithm method (WOA-XGBoost) is then used to establish mechanism-data hybrid-driven energy consumption and carbon emission models for converter. Furthermore, a process parameter optimization model that comprehensively considers converter energy consumption and carbon emission is constructed. An improved non-dominant sorting hiking optimization algorithm (INSHOA) is used to solve the optimization problem. The case study shows that the R² values of the WOA-XGBoost model are 0.931 and 0.967, respectively, which have superior modeling accuracy compared with other methods. The optimization results demonstrate that compared with the empirical scheme, the proposed optimization model reduces the energy consumption, material carbon emission, and total carbon emission of the converter by 7.78 %, 9.60 %, and 35.16 %, respectively. Moreover, the proposed INSHOA shows superior algorithmic performance.