Abstract:In order to effectively deal with the risk management of power grids, solve the level problems and barrier of different majors and improve the lean management, it is an urgent to build up the management system of penetration control. Based on the decentralization idea, distributed storage the operation data of power grids, and big data mining, this paper builds up a penetration control mechanism for the coordination evaluation of multi-levels and multi-specialities coordination. The independent management hosts are connected tightly to strengthen the management by carrying out the collaborative security analysis, finding the weak parts and management faults. The FPGrowth and KNN algorithms not only improves storage and retrieval efficiency but also increases classification accuracy. This paper uses them for mining and analyzing blackouts to ensure the judgement accuracy of responsibility subjects. At present, the scheme has been used in Chongqing power grid and played an important role in risk management. It is of referential value for relevant power grid enterprise.