Burden Surface Decision Using MODE with TOPSIS in Blast Furnace Ironmkaing

Burden surface distribution plays a key role in achieving an energy-efficient status of blast furnace (BF). However, actual adjustment of burden surface usually depends on the operator's experience when the production status changes. Meanwhile, due to the characteristics of high dimension, strong coupling, and distributed parameters, it is difficult to establish the accurate mechanism model for BF ironmaking process. Considering the aforementioned issues, this paper proposes an integrated multi-objective optimization framework for optimizing burden surface distribution based on the analysis of BF operation characteristics. Firstly, data-driven models are constructed for two objectives, i.e., gas utilization ratio (GUR) and coke ratio (CR), and two constraints using adaptive particle swarm optimization (APSO) based extreme learning machine (ELM), named APSO-ELM. Multi-objective optimization is subsequently carried out between GUR and CR using the multi-objective differential evolution algorithm (MODE) to generate the Pareto optimal solutions. Finally, TOPSIS is applied to select a best compromise solution among the Pareto optimal solutions for this optimization problem. Comprehensive experiments are presented to illustrate the performance of the proposed integrated multi-objective optimization framework. The experimental results demonstrate that the proposed framework can give a reasonable burden surface profile according to the production status changes to guarantee the BF operation more efficient and stable.