Research On Fast Query Mechanism Of Mass Data Cluster Monitoring For Railway Power Supply Based On Inverted Index

With the increasing scale of high-speed railway in China,the intelligent level of railway power supply system is improving.The integrated dispatching and monitoring system has been gradually applied to all important high and low voltage circuits in the power supply system of high-speed passenger dedicated line.The number of monitoring and control terminals of power supply dispatching monitoring system has greatly increased.Besides,the running speed and frequency of high-speed train are much higher than those of ordinary trains,and the running parameters of railway power supply system change frequently.The sampling frequency of the monitoring and control terminal of the power supply dispatching monitoring system also increases,which makes the collection information of the railway power supply dispatching monitoring system increase exponentially.The amount of monitoring data acquired by the zhejiang-jiangxi railway 10 kV power dispatching monitoring system in only 2 months is more than 300 G.According to this calculation,the amount of monitoring data in 1 year is 1.8T,and the total storage can reach 27 T after 15 years of service.Traditional data processing technology mainly uses relational database for storage management,which has the problem of limited capacity and poor scalability.The traditional relational database is slow in responding to the query of massive monitoring data of over ten million levels,which is easy to cause the scheduling interface card screen and affect the real-time processing of scheduling information.In serious cases,it may even lead to delayed,missed or even lost critical fault information,threatening driving safety.Therefore,it is urgent to study the efficient query and response technology of railway power dispatching monitoring and propose a new fast processing method of massive monitoring data.Aiming at the quick query and response problem of railway power dispatching monitoring system,a distributed cluster database of railway power dispatching monitoring is established in this paper.This distributed cluster database is based on HBase database,facing massive monitoring data.And it has much more storage capacity and scalability than traditional relational database.Considering the low efficiency of HBase database query by non-primary key,which can’t meet the engineering application requirements of railway power supply dispatching monitoring system query monitoring data by non-primary key keywords such as station and equipment number,this paper designs a secondary index structure based on theinverted index structure of railway power supply monitoring data.It changes data access structure,generates inverted index of monitoring data,uses non-primary key positioning,and completes secondary index of non-primary key data.Experimental results show that this method meets the requirement of rapid data retrieval in railway power dispatching monitoring system.The hotspot cache technology is integrated into the distributed cluster database of railway power dispatching monitoring to further improve the query efficiency.In addition,the hotspot cache technology was improved,and the access sequence coding was designed in the cache update and replacement link of the hotspot cache of inverted index,and the access sequence of the records of the hotspot cache of inverted index was managed and sorted.It avoids the operation on the timestamp of the cache system attributes and improves the performance of multi-thread read and write of the hotspot cache method.Through the test of engineering examples,it is verified that the improved hotspot cache replacement method can further improve the query efficiency of railway power supply dispatching monitoring system.