Raw materials consumption reduction for practical electric arc furnace steelmaking: a data association rules mining approach with improved evaluation indicator

Reducing raw materials consumption (RMC) in electric arc furnace (EAF) steelmaking process is beneficial to the reduction in resource and energy consumption. The conventional indicator of evaluating RMC only focuses on EAF inputs and outputs, neglecting the associations between smelting operations and RMC. Traditional methods of reducing RMC rely on manual experience and lack a standard operation guidance. A method based on association rules mining and metallurgical mechanism (ARM-MM) was proposed. ARM-MM proposed an improved evaluation indicator of RMC and the indicator independently showed the associations between smelting operations and RMC. On the basis, 1265 heats of real EAF data were used to obtain the operation guidance for RMC reduction. According to the ratio of hot metal (HM) in charge metals, data were divided into all dataset, low HM ratio dataset, medium HM ratio dataset, and high HM ratio dataset. ARM algorithm was used in each dataset to obtain specific operation guidance. The real average RMC under all dataset, medium HM ratio dataset, and high HM ratio dataset was reduced by 279, 486, and 252 kg/heat, respectively, when obtained operation guidance was applied.