Research On Power Distribution Monitoring Information Cluster Processing Technology Based On Impala RIA SVG

Compared with traditional distribution networks,distributed smart distribution networks are based on data transmission and big data processing,including efficient transmission of data and control commands between multiple power supply units,such as condition monitoring of transformer and distribution equipment.Electrical information collection,distribution automation and other engineering application data.The data collected according to the specified time sequence and can reach the processing link in sub-second order can be called quasi real time data if the measurement accuracy is satisfied.The quasi real time data of distribution network can display the data of current,voltage,active power,reactive power,power factor,power extremum on a given line,and then list the overload equipment data in the single line diagram,so,It is an important guarantee to obtain efficient distribution network management and control decision.In order to obtain the running state information of the equipment accurately,there are more and more collection points of the equipment in the station under test.The normal dispatching automation system of the station contains hundreds of thousands of collection points,distribution of electricity,and the data center will reach millions or even tens of millions of levels.State Grid has launched a research project on quasi-real-time data platforms and provides quasi-real-time data services through access to more than 28 m data points.Considering that intelligent distribution network needs panoramic data acquisition,quasi real-time information,such as voltage,current and equipment status,which are collected in time series,has become the data base to support the reliable operation of smart distribution network.How to quickly query and process quasi-real-time data of distribution network has become the technical bottleneck of intelligent distribution network dispatching decision.Aiming at the fast query processing of quasi real time big data in distribution network,the academic circles deal with the fast query processing of massive quasi real time data of intelligent dispatching and monitoring by speeding up the dispatch monitoring interface and data server.The railway monitoring data is stored in parallel to the special HBase big data cluster server.Compared with the conventional relational database,it has better scalability and storage capability,but the efficiency of HBase query by non-primary row key is low.High latency on the server side.In order to solve the problem of big data processing,the Apache Foundation has released a new data engine,Impala component,which avoids the MapReduce startup overhead through daemons.If the AJAX data caching engine is integrated into the Impala daemon,It is possible to avoid reboot time consuming and Impala can accessthe data stored in HDFS sequentially,which is more efficient than random access of HBase.The research of integrating the AJAX interface data engine into the Impala daemon and the HDFS sequential read access interface provides a new technical means to solve the fast query of massive monitoring data in intelligent scheduling.Considering that railway distribution network not only has the characteristics of more measuring points and fast change in general power distribution network,but also has higher requirements of delay processing.Distribution big data is embedded into MPP query engine by using rich network component container and big data secondary index mechanism.A new fast RWI query method for distribution network data is proposed.A large number of quasi-real-time data can be quickly queried in scheduling and monitoring applications by using the Impala data daemon.Taking tens of millions of real time data derived from the project of railway 10 kV distribution network monitoring system as an example,the loading test and cluster query performance test are carried out.