Design And Implementation Of Scene Analysis Subsystem For Big Data In Power Grid

The distribution network is at the end of the grid system.With the promotion of distribution automation construction and automatic collection of electricity information,the informatization and interaction of the grid are becoming more and more automated.During the operation of the smart grid,a large amount of user basic data is stored,and the data generated by different users directly has a subtle connection.Data mining of the user's electricity load data through the clustering algorithm can find the characteristics of electricity consumption of typical users,but it is difficult to automate the analysis of the results obtained.In order to solve these problems,this paper design and implement a scene analysis subsystem for data mining clustering results.The system has support for custom models,can automatically generate parsing text and other characteristics of the data can also carry a large amount of computing requests,according to the characteristics of the tag automatically generates the scene analysis results.The system uses distributed storage and a SPARK-based computing platform based on the characteristics of big data in design and implementation.In this paper,firstly,the background and significance of the big data scenario analysis subsystem proposed by power grid are expounded.Relevant technologies related to its realization are expounded,including the development background of smart grid and related algorithms in machine learning.Then,the author analyzes the requirements of subsystems users and analyzes their functional requirements and non-functional requirements and studies the key issues involved in functional requirements.Constructing a scenario of user behavior analysis and negative evaluation of investment in power grid.And explain the data model involved in the system and the relationship between them.Then based on the previous research conclusions,the author analyzes and designs the overall architecture of the grid big data scene analysis subsystem and the interaction between different levels.The implementation of the system key modules involved in the typical use process is described in detail.Subsequently,the system was deployed and tested,and the system was tested to verify the rationality and effectiveness of the system.Finally,the research prospects are put forward in the imperfect areas of the system.