Predicting CaO activity in multiple slag system using improved whale optimization algorithm and categorical boosting

Abstract

Abstract The activity of slag components is one of the primary factors influencing the thermodynamic properties of slag. In this study, a feasible model was established to predict the a(CaO) using improved whale optimization algorithm (IWOA) and Categorical Boosting (CatBoost). The effects of other variables on a(CaO) were listed in descending order of influence as follows: w(CaO), w(SiO2), temperature, w(MgO), and w(Al2O3). And the IWOA-CatBoost model achieved the highest R2 value of 0.9200, lowest RMSE of 0.0042, and lowest MAE of 0.0030 in predicting the a(CaO). The performance of the optimal IWOA-CatBoost model was evaluated and compared with that of known models. The results demonstrate that the IWOA-CatBoost model outperformed existing models and methods, such as the Factsage, ion and molecule coexistence theory, and genetic algorithm—backpropagation neural network. The accurate calculation of slag component activity is of great significance to the analysis of the thermodynamic properties of slag. Meanwhile, the approach and algorithm used to develop the a(CaO) prediction model can also be applied to predicting the activity of other slag components or other metallurgical applications (e.g., predicting molten steel temperature, steel composition, and alloy yield).