Identifying Typical Load Curves of Industrial Customers Based on DBSCAN-DPC Dual-Layer Algorithm

With the rapid development of electricity market reforms and smart grid technologies, accurately identifying the typical load curves of industrial customers is crucial for electricity retail companies in optimizing demand response strategies and enhancing market segmentation. A density-based spatial clustering of applications with noise and density peak clustering (DBSCAN-DPC) dual-layer clustering algorithm is proposed in this paper. First, abnormal load curves are detected based on DBSCAN. Then, a multi-dimensional index is extracted to create customer profiles, and DPC is applied for secondary clustering to identify typical load curves and refine segmentation. Finally, case studies on industrial and commercial customers in eastern China demonstrate that the proposed method achieves effective market segmentation of electricity customers.