关键词: 电力, 大数据, 云计算, 用电信息, 数据采集, 数据分析

Abstract: In order to deal with the bottlenecks of the performance and scalability of power data analysis in present big data era and meet the requirements of power production and marking systems, the advantages of cloud computing technology is firstly analyzed, then the architecture and key techniques of Hadoop-based power big data analysis system are proposed. The architecture combines the characteristics of power big data and the emerging cloud computing technology, and includes three key technologies, i.e., the distributed gridfile based multi-dimensional index, the automatic SQL to HiveQL translation tool and the hybrid data storage model, which can support data update operation. Therefore, the traditional power data analysis system can be improved and updated. The application of the system in the power information collection system in Zhejiang indicates that the performance of the system is improved by 5 times with only one-eighth of the hardware configuration as compared with the traditional system. This proves that the cloud computing technology can significantly improve the performance and reduce the cost of big data acquisition and analysis.

Key words: electric power, big data, cloud computing, electricity information, data acquisition, data analysis