Artificial intelligent based energy scheduling of steel mill gas utilization system towards carbon neutrality

Steel industry contributes significantly to the world economy, but is highly energy intensive and CO2 intensive since the coal-based blast furnace route is dominant in steelmaking. Besides efficient utilization of the steel mill gases for power and heat supply, deploying technologies of carbon capture, utilization and renewable power is in urgent need for the transition of the steel industry towards carbon neutrality. To attain this goal, this paper develops a low-carbon steel mill gas utilization system with the integration of solvent-based carbon capture, methanol production based carbon utilization and renewable power. An artificial intelligent based optimal scheduling is then proposed to coordinate the interactions among gas, heat, electricity and carbon under variant weather and load conditions. Gradient boosted regression trees with Bayesian optimization is exploited to identify efficient surrogate models for the complex devices within the system. Heuristic search algorithm of particle swarm optimization is applied to find the low-carbon and economical scheduling within the entire scheduling period. Case studies show that the optimal scheduling can unlock complementary advantages among renewable energy, carbon capture and utilization, leading to 97% renewable energy curtailment reduction, 62% CO2 emission reduction and 126 tons of methanol production in 24 h. Sensitivity analyses are carried out to investigate the effects of additional coal consumption, renewable power installed capacity, CO2 emission penalty coefficient and CO2 capture constraint mode, providing broader insight into the operation of the steel mill gas utilization system towards carbon neutrality.