Parallel Diagnosing Method On Big Monitoring Data Of Electric Power Equipment Based On Storm Framework

With the rapid development of smart grid,the power industry has entered the "era of big data." In the gradual construction of smart grid,the real-time capability and reliability of data processing needs to be further improved.In recent years,cloud computing technology has developed rapidly and various distributed processing platforms have emerged,providing a practical computing framework for the processing of massive data.Storm is a distributed real-time processing framework.Having good real-time response capability,high reliability and high scalability,the framework,is widely used,providing a new research idea for the real-time fault diagnosis of massive grid equipment eigenvector data.In this paper,DGA data of dissolved gas in transformer oil is used to test respectively the accuracy and performance with the following two diagnostic models.A flow data processing model based on extreme learning machine is established,and the parallel classification of data flow is implemented under Storm platform.At the same time,in order to deal with the phenomenon of drift in the classification of flow data,the online learning module is added on the basis of fault diagnosis module.The two modules are connected through the memory database Redis,and the fault diagnosis module operates in parallel.By testing its classification accuracy,the result meets the actual application requirements.By testing the throughput and processing latency of cluster data,the experiment shows that the throughput of the flow calculation can be increased and the data processing latency can be decreased by setting the parallelism of the components and other related configurations reasonably.A hybrid clustering algorithm based on fuzzy c-means algorithm is proposed.This algorithm is deployed on Storm platform.First of all,data flow is received and grouped through the data access module,and the grouped data are normalized.The initial clustering center is selected by subtrative clustering method.K-means algorithm iterates continuously after obtaining the initial cluster center so as to obtain a better average cluster center.FCM receives this cluster center for clustering.This model realizes the parallel fault diagnosis of eigenvector data.The results show that the diagnostic accuracy of this method satisfies the needs of fault diagnosis.By comparing the execution time of several methods with different data sizes,it is shown that the proposed hybrid clustering algorithm has better performance.