黄河流域能源富集区工业用水驱动机制研究OA

Uncovering the driving mechanisms of industrial water use change scientifically is fundamental to accu-rately predicting industrial water demand.Aiming at the current problems of unclear driving mechanisms of industrial water use variation and insufficient theoretical support for water demand forecasting,this paper takes Baotou,Ordos and Yulin-energy-rich cities in the Yellow River Basin-as study areas.Big data mining is employed to construct the correlation network between industrial indicators and water use.A non-parametric deep learning model is estab-lished by coupling PSO-SVM-based dominant factor analysis and symbolic regression algorithm,through which the constitutive equation for industrial water use evolution is constructed and its driving mechanisms are revealed.A dual-driven water demand forecasting framework integrating big data mining and physical mechanisms is proposed,and industrial water demand from 2025 to 2035 is projected.Results show that due to disparities in industrial structure,resource endowment and water resource conditions,the evolution mechanisms of industrial water use in different regions exhibit both industrial commonalities and regional heterogeneities.Industrial added value serves as a common driving indicator with a contribution rate above 0.06,while water resource development utilization rate acts as a common restraining indicator with a contribution rate below-0.06.Under the joint effect of multiple driving factors,the future trends of industrial water demand differ substantially among Baotou,Ordos and Yulin:industrial water demand will decrease by 12.2％in Baotou,while increase by 10.5％and 11.6％in Ordos and Yulin,respectively.