Research On Data Processing And Analysis For Electrical Equipment Condition Monitoring Using Hadoop

With the gradual development of smart grid and high demand on electric device safequantity, electrical equipment condition monitoring data under this environment willincrease enormously within a short interval with more and more devices, makingtraditional state monitoring methods confronted with many difficulties. How to eveulatequickly and efficiently with big data has become a new topic. After the papers on cloudcomputing were published by Google, cloud computing platform Hadoop,which hasgreat advantage of mass storage and computing, high scalability and reliabilitycharacteristics, sprung up outstandingly. Now it has become a new approach to solve thegrid problem mentioned above.Firstly, this paper introduces the current status of the equipment conditionmonitoring with details, in which the efficiency problems are analyzed specifically.Secondly, this paper summarizes the researches on the cloud computing, especially on theapplication of Hadoop. By taking the sharing and use of the state monitoring data as thespecific application requirement and considering the difficulties in the storage and inquiryproduced by the demand of high real-time capability and reliability of online monitoringdata, the Hadoop platform is applied to the state monitoring of electrical equipments,which makes the reliable storage and efficient inquiry of mass electrical data possible.Considering the application requirements of the domestic grid companies and theadvantage of the Hadoop platform, a data storage and state monitoring model based onHadoop is proposed, which meets the high requirements of the equipments statemonitoring for the smart grid. The implementing method of this model is introduced bythe illustration of its overall structure and design of each algorithm sub-module. Thismodel solves relative problems and realizes the reliable storage and efficient inquiry ofthe state monitoring.This paper introduces the building of Hadoop cluster,optimization of the Hadoopcluster,MapReduce program design pattern,the monitoring algorithm achieve.We assessthe MapReduce program efficiency and accuracy with simulation of relatityenvironment.Proved that our design’s effectiveness and correctness.The topic provides a new method of thinking for the data storage and inquiry of electrical equipment condition monitoring for the next generation, lays the application ofHadoop used in smart grid.