With the development of the country's economy, the social electricity load continues to rise. The importance of electricity in national economic development and daily life has become increasingly prominent, and the impact of power failures on social life is also increasing. In this case, as a routine task of the terminal business of power companies, there is an increasing need to improve the timeliness and accuracy of power fault repair. At present, there are relatively many weak points in some network grids, massive distribution network equipment makes fault location difficult, and operation, distribution and debugging business data difficult to share simultaneously, resulting in insufficient response to fault repair. This project hopes to improve the efficiency of power grid fault repair through technical transformation and integration of existing power grid resources. The project research uses existing resources for secondary development, explores the usable value of smart meter information collection system data, gives full play to the overall role of the fault repair management system, and realizes rapid location of fault locations and fault types of power grids through information interaction between operation, distribution and dispatching systems. Intelligent diagnosis, thereby further generating reliable and optimized fault solutions, providing a reliable basis for fault repair work, and improving the overall power supply reliability and continuity of the power grid. The innovation points of this project are as follows: 1. Build a data classification and screening channel to conduct autonomous fault detection. Modeling is completed by analyzing the abstract characteristics of abnormal data, and the extreme learning machine algorithm is applied to smart meter data analysis. Using its characteristics of fast learning speed, good generalization performance and excellent modulation mechanism, it quickly classifies massive data and captures abnormal data., to achieve autonomous detection and rapid early warning of faults. 2. Establish a fault model library and build a mapping database to achieve full identification of fault types. Analyze typical faults of various types of equipment in the distribution network, and collect corresponding smart meter abnormal data feature sets. Use this model library for matching and comparison to make fault types clear at a glance. Build a one-to-one mapping relationship between relevant data in the smart meter information collection system and the fault repair management system to realize interactive docking of the systems and provide a basis for subsequent fault judgment and positioning from the data communication level. 3. Propose recovery closed-loop management and active power fault reporting technology to select appropriate topology search strategies based on the inherent logic between data and different power grid topologies, which is more targeted when locating power grid faults. On-site emergency repair personnel can interact with the system and command center in real time through handheld phones, which is conducive to the monitoring and management of the fault recovery process. At present, this project has published 1 scientific and technological paper, obtained 2 invention patents, and applied for 2 invention patents. After novelty search testing by the Shanghai Science and Technology Novelty Search Consulting Center of China Academy of Sciences, the project is novel and its technical level has reached the leading domestic level. This project won the gold medal in the 2018 Shanghai City Excellent Invention Selection Competition. This model has been promoted and applied within the jurisdiction of Shanghai Pudong Power Supply Company, and has achieved significant social and economic benefits. It is planned to be promoted in batches throughout the industry. Since its promotion in 2015, Pudong Power Grid has shortened the fault response time by more than 70000 minutes throughout the year and improved the power supply reliability rate of the entire network by 0.001%. In terms of improving quality and efficiency, the cumulative income brought by new profits and cost savings every year is approximately 1.65 million yuan. While reducing the risk of power grid failure escalation and operation, this project speeds up the response speed of failure repair, improves customers 'power quality and satisfaction, ensures residents' normal production and life, fulfills the social responsibilities of power supply companies, and can effectively improve the public's recognition of power supply companies and enhance the company's brand image.