The large amount of hazardous waste produced and discharged in the process of copper smelting has brought serious environmental risks and disposal pressure. However, the migration and transformation behavior of hazardous elements in the copper smelting system is affected by multiple factors, and idealized chemical mechanism studies are difficult to reveal the complex patterns in large-scale production. This study has built a Bayesian network model based on the Expectation-Maximum algorithm and cluster tree propagation algorithm to reveal the key influencing factors and mechanisms of the distribution of hazardous elements in actual large-scale production. The results show that the main operating parameters affecting the flow direction of hazardous elements are blast volume, oxygen supply volume, input copper content, quartz input, etc. The variation patterns of elemental fluxes of arsenic and lead with input variables in all production processes of copper smelting system were obtained. The fluxes of these two hazardous elements alter similarly in most processes. The maximum variations of arsenic and lead content in material flows are up to 48% and 32%, respectively. This study can help to supplement and perfect the theoretical basis of the distribution pattern of hazardous elements and to realize automatic prediction of element flows under different operating conditions.
Read full abstract