A Multi-Label Prediction Model for Iron Ore Sintering Process Incorporating Time Features

The steelmaking sintering process is critical for quality control and efficiency in iron production. Addressing the challenge of accurately predicting quality indicators amidst diverse parameters and multiple time scales is essential. In this paper, a multilabel regression model incorporating temporal features is proposed for predicting quality metrics during the sintering process. The model integrates 34 input parameters to predict 7 quality indicators, reflecting the complex nature of the process. By applying the model to real data from a steel plant's No. 2 sintering machine and validating its performance, the study demonstrates the model's effectiveness in multi-label regression tasks. The experimental results show that the model incorporating temporal features performs better in predicting sinter ore quality indicators, effectively capturing complex parameter relationships and providing reliable predictions.