Currently the power quality monitoring data have exhibited a massiveness tendency.If the data are stored in the relational database,problems such as large storage space,low query access speed and poor scalability will be caused.By analyzing the data access characteristics and hardware environment of the power quality monitoring system,this paper proposes a data management scheme for massive power quality monitoring data based on the distributed file system.The monitoring data of different power quality indices are stored in the compressed files;the heterogeneous distributed servers of the existing monitoring sub-stations and related systems are used as file servers to store the original monitoring data files,the server of the monitoring master station is used as master server for storing data characteristic value and file indexes,it is also used for unified management of file resources.The scheme takes full advantage of each server”s storage space and network bandwidth, and saves the storage space and improves the access speed.It also has high reliability and scalability.For example,when the 180 days”data from 100 monitoring sites are stored,the storage space of this scheme is only 2.28% of the traditional scheme;and when the 180 days”three-phase voltage root mean square(RMS)value data of a certain order harmonic are queried,this scheme”s acceleration ratio is about 16.49.The reliability and practicality of this scheme are proved through application in the Sichuan power quality integrated data platform.%目前，电能质量监测数据已经呈现海量化的趋势，如果仅用关系数据库存储，将带来存储占用空间大、存取速度慢、可扩展性差等问题。文中通过分析现有电能质量监测系统中的数据存取特征和硬件环境，提出了一种基于分布式文件系统的海量电能质量监测数据管理方案。此方案将不同电能质量指标的历史监测数据分别压缩后存储在文件中；利用现有监测子站以及相关系统的分布式异构服务器作为文件服务器以存储数据文件；利用监测主站服务器作为主服务器，保存数据特征值和文件索引，并对文件资源进行统一管理。此方案充分利用了各服务器的存储空间和网络带宽，节约了存储空间，提高了存取效率，具有较高的可靠性和可扩展性。以存储100个监测点180 d数据为例，此方案存储空间占用仅为传统关系数据库方案的2.28%；以检索某个监测点180 d的5次谐波三相电压幅值数据为例，此方案加速比约为16.49倍。在四川电能质量一体化数据平台中的成功应用证明了此方案的可靠性和实用性。
展开▼
