Data-driven-based potential assessment of industrial and commercial loads

New power system requires extracting valuable information from massive data, providing utility value for the power system and providing high-quality services to users, which promotes the development of demand response. The method of data-driven-based potential assessment of industrial and commercial loads is proposed in this paper. The raw data is processed through empirical mode decomposition and load clustering techniques to assess the potential of industrial and commercial loads. The load clustering technique for low-frequency components could identify the type of load, and participate in other auxiliary services, such as load decomposition, etc. Furthermore, the clustering results could reflect the long-term response capacity and serve as a potential assessment index for the long-term response time. In addition, the high-frequency components are added considered as detailed information based on low-frequency components, which provides a basis for distinguishing typical users of the same type of load, and could be used as a potential assessment index for short-term response rates. This assessment method will help the power system guide industrial and commercial loads to participate in demand response as needed, formulate more accurate dispatch plans, and provide better services.