Research on Construction Technology of Generalized Load Model Analysis Library for Space-Time Dimensions

In the field of electric marketing, with the continuous deepening of the interactive use of intelligent power, users' willingness to autonomy, choice and participation is becoming more and more urgent, and users have gradually become an important subject of power system operation and interoperability. Different types of users present different characteristics for electricity demand, industrial users pay attention to energy efficiency services, power saving consultation, demand side response, distributed energy access, etc.; commercial users pay attention to integrated energy management, new energy power generation solutions, energy saving solutions Etc. Therefore, analyzing the user's electricity usage and mastering the user's electricity consumption characteristics based on electric big data is of great significance for improving the power supply service level, building the power enterprise brand, and practicing the Internet + marketing service. It is urgent to apply artificial intelligence and data analysis technology to the field of power load characteristic analysis, and to build a general-purpose load feature model library to provide shared and scalable model support for various business application scenarios. Application to household change relationship identification, serial household string identification, source network friendly interaction, precise load control, user differentiated service, time-of-use price design, user-level load forecasting, demand response potential assessment, power stealing identification, default power use Business scenarios such as energy saving, industrial production anomaly monitoring, and user portrait portrayal.