Research And Development Of Microgrid Data Analysis System Based On Microservice Architecture

With the development of the economy,the demand for electricity continues to grow,the structure of the power grid becomes more and more complex,and the traditional power grid structure faces many problems.The smart grid based on the micro grid is developing rapidly.In order to ensure the safe,reliable,economical and efficient operation of the microgrid,it is necessary to provide timely and accurate data analysis result information to visually display the data change trend.Therefore,it is very important to establish a microgrid data analysis system.Microservices create applications around the business domain.These applications can be independently developed,deployed,tested,managed,and updated to decouple service functions,enhance system resiliency,and facilitate expansion and maintenance.Based on microservice architecture and other mobile internet related technologies,this paper designs and implements microgrid-based microgrid data analysis system.In this thesis,photovoltaic power generation is taken as an example to elaborate the data analysis system.The GRU(Gated Recurrent Unit)neural network model is used to predict the short-term photovoltaic power generation,and the experimental comparison analysis is carried out.The experimental results show that the proposed method has a good effect on short-term photovoltaic power generation prediction.Design and implement the state monitoring and fault warning module of the microgrid data analysis system to realize the daily data status monitoring and fault warning of the platform.In this thesis,domain-driven design is used to design system architecture.Then by analyzing the characteristics of microservice cluster load balancing,using the sliding window technology,a load balancing method based on genetic simulated annealing algorithm is designed and implemented to solve the load allocation problem of large-scale request waiting,and to achieve faster request processing on the basis of making full use of resources.Finally,Kubernetes is used to build the platform of the micro-grid data analysis system micro-service,which realizes the registration and discovery of micro-service architecture,elastic scaling,rolling upgrade and rollback,and log monitoring.Compared with the traditional architecture design,the micro-service-based microgrid data analysis system designed and implemented in this paper satisfies the access,monitoring and deployment of micro-grid multi-components,and realizes microgrid power prediction,condition monitoring and early warning.The data analysis has certain application guiding significance for the design and implementation of the microgrid data analysis system.