Design And Application Of Cloud Computing Platform Based On Large Data Of Wind Farm SCADA

With the continuous progress of China's green energy strategy,the wind power industry has developed rapidly.Wind farm SCADA system provides a new idea for wind farm operation and maintenance,establishing a remote control center for wind farms and realizing the centralized supervision of wind farm operation,maintenance,management and logistics.It overcomes a series of problems,including the harsh environment of wind farms,scattered sites,large numbers of units,difficulties in operation and maintenance,etc.It has created a operation and maintenance model of “unattended,less people,and regional management”,which greatly promotes the further development of the wind power industry.However,with the explosive growth of wind farm data,the problem of insufficient data storage and processing capacity in traditional wind farm SCADA systems has become increasingly prominent.This article addresses the shortcomings of existing centralized control platforms,taking the advantage of the flexibility of cloud computing platform architecture and paralleled data processing,designing a cloud computing platform based on wind farm SCADA big data,realizing distributed cloud storage and efficient cloud computing of SCADA big data in wind farms,compensating the disadvantages of high cost of data storage,poor data processing capability,and lack of data mining capabilities in traditional wind farm SCADA systems,and effectively acquiring the information implied by wind power big data.Finally,the advantage of the platform is verified through the test.Firstly,based on the overall design framework of the traditional wind farm SCADA system,the master station system of the centralized control center,the wind farm substation system,and the specific functions of the SCADA system of the wind farm.In order to improve the ability of data storage and analysis and provide a basic platform for the future expansion of intelligent application,the design idea of cloud computing platform for wind farm SCADA system is proposed.Secondly,it adopts four related key technologies to build Hadoop 2.0 cloud computing platform,which are distributed file system HDFS,resource and data management YARN,distributed computing model MapReduce,in-memory computing framework Spark.By studying the general technical framework of big data cloud computing platform,a cloud computing platform based on SCADA data of wind farms is designed.The overallarchitecture and internal logic of the platform are described,and the two advantages of data cloud storage and cloud computing are highlighted.Thirdly,designing clustering and regression data mining algorithms in cloud computing platform to finish cleansing and regression prediction model for wind farm SCADA big data and the wind power short-term prediction function is realized.In order to adapt to the cloud computing platform and get clustering and regression results better,two algorithms are parallelization improved by MapReduce.Finally,it explains the building environment of the cloud computing platform and the detailed construction steps.Based on the historical data of SCADA system of a practical wind farm,clustering and regression algorithms are tested on the cloud computing platform.Clustering algorithm is used to compare the efficiency of single machine and cloud computing platform cluster,and the efficiency of MapReduce and Spark framework.The performance and accuracy of the two algorithms are tested and analyzed,and the superiority of the cloud computing platform is verified..