Energy consumption simulation and diagnosis in the sintering process: A mechanism and data-driven approach

The steel industry is a pillar of China's national economy, with energy consumption accounting for more than 10 % of the national total. As a key stage in steel production, the sintering process contributes about 8 % of total energy use. Under the dual carbon goals, improving energy efficiency in sintering has become an urgent challenge. However, existing prediction models often exhibit errors exceeding 10 %, limiting energy-saving research. To address this issue, this study develops a sintering energy consumption model using one year of continuous production data from a steel plant, combining mechanism analysis with a genetic algorithm–optimized long short-term memory network (GA-LSTM). The model achieved a prediction error below 5 % over 30 consecutive days. In addition, an energy diagnostic framework based on the analytic hierarchy process was proposed, analyzing influencing factors across three levels: “energy consumption–energy carriers–operational parameters.” Results show that solid fuel is the dominant energy carrier, contributing 73.36 %, while parameters such as sinter basicity strongly affect fuel use. Proper adjustment can reduce solid fuel consumption by 2.6 kg/t. Considering process stability and parameter coupling, two optimization strategies—uncorrected and corrected—were proposed, reducing energy consumption by 1.85 kgce/t and 1.25 kgce/t, respectively. • Mechanism and data-driven methods for energy consumption simulation are proposed. • A hierarchical energy diagnosis method based on the simulation model is proposed. • Solid fuel was identified as the dominant factor driving fluctuations in sintering energy consumption. • Targeted optimization strategies can reduce energy consumption by 2.57 kg/t and 1.25 kgce/t respectively. • The practical effectiveness of the proposed methods is demonstrated by case study.