Along with electrical collection scope gradually expanded, especially the propulsion of "full coverage, full collection, full cost controlling" construction work, the soar of data collection capacity in electricity brings great difficulties to data management. One is the storage of massive data. The other is the analysis and application of data. As the support of the acquisition system’s full coverage, deep mining the data of acquisition system, deepening apply the function of acquisition system, strengthen the integration of marketing, distribution and other related systems, to achieve the "four modernizations" goals of automation of electricity copying and checking, fine management of line loss, intelligent service of interaction, and practicability of cost control function, to comprehensively promote the change of the way of marketing development, to improve the level of the company’s power supply service. But the real-time storage of daily data and the analysis and mining of data have become the key of system control and application. Especially the line loss management, load management, on-line monitoring, measurement of qualified voltage power, and the analysis of power supply reliability, all need the storage, rapid analysis and intelligent diagnosis of data, to support the system application.According to current existing bottleneck of storage, calculation and extension in the electric information collection system, actively explore new technology. Using the technique which is based on the interaction of real-time database and cloud computing machine, can realize the transition from the traditional "centralized analysis-> centralized processing-> unified display" to "distributed analysis-> distributed processing-> unified display". Electric information collection database uses real-time data stream processing technology, and the data management uses DataCube distributed infrastructure based on cloud storage file system, to realize complex relational query of relational database. The main work, methods and research content as follows.1) Research on real-time data processing technology based on electricity information collection system to solve a series of high timeliness demands such as real-time efficient storage, real-time analysis, real-time calculation of electricity information.2) Aimed at different types of data, develop interfaces for data exchange and sharing between relational database, real-time database and cloud storage platform.3) Design general data interface to support data analysis and data processing for marketing inspection monitoring system, online intelligent evaluation system of power quality, production and distribution command platform, marketing business application system, and marketing GIS system.4) Build cloud computing service platform with high reliability, stability and expansibility, to support techniques for business function expansion of the electric information collection system.5) Research on real-time online data mining and statistical functions, to provide strong basis for intelligent decision analysis of the orderly use of electricity.The result of implementation of the project as follows.Based on the integrated application of real-time database and cloud computing, cloud real-time storage platform, the system consists of main control node and processing nodes, including storage subsystem, management subsystem and dispatch subsystem. It uses efficient parallel computing technology and advanced index mechanism to implement real-time tasks rapid processing, complex tasks efficient computing and high throughput of big data processing batch tasks. It provides integrated business frame for power companies to help them adjust to changing demands. |