Design And Implementation Of Big Data Analysis Platform For Fan Cluster

In recent years,with the shortage of fossil energy and environmental pollution,countries around the world have paid more and more attention to the development and utilization of renewable energy.Wind power has achieved rapid development in recent years due to its advantages of cleanness,safety and abundant energy storage.With the transformation of China's energy structure,wind energy will play an increasingly important role in China's energy structure.Therefore,it is of great significance to conduct in-depth research on wind power production issues.However,wind power companies still use traditional database technology and data processing technology to conduct big data analysis and research.Due to the lack of special technology and tools,the timeliness and availability of data processing are not strong.The analysis of data is still in its infancy and lacks deep data mining.In order to solve the current data analysis in the wind power industry,the problem of slow response,low real-time performance,and difficulty in processing high-dimensional and complex data is designed.This paper designs and develops a big cluster data analysis platform for wind turbine clusters.The main contents of this article are as follows:(1)The paper analyzes the demand of the wind cluster cluster big data analysis platform,designs the system architecture and technical architecture of the platform based on the micro-service idea,and completes the division of the platform function modules and the design of the communication mode between modules.(2)For the data loss and data anomaly of wind power data,the paper uses k-nearest neighbor algorithm to complete the missing data.The isolated forest algorithm is used to detect the anomaly data,and the effectiveness of the algorithm is verified by an example.And accuracy,while achieving the normalization of wind power data.(3)Based on Spark Streaming,the BP neural network model is parallelized,and the wind turbine power prediction is taken as an example to prove that the parallelized BP neural network model can greatly improve the high-precision prediction.Reduce the time for analysis and prediction.(4)Based on the research of Kubernetes+Spring Cloud microservice architecture,using Docker,Spring Cloud,Vue,Consul and other technologies,the development of the wind turbine cluster big data analysis platform was completed,and the Mesos-based Spark big data processing cluster was built.Used for platform data processing.Finally,the main work of this paper is summarized,and the shortcomings of this paper and the future work direction are prospected.