电网企业数据人工智能中台设计

Design of Artificial Intelligence Data Middle Platform for Power Grid Enterprise Data

本文通过分析电网业务实际,构建电网业务模型,通过分析数据接入、数据治理、数据模型、通用分析模型及数据服务的逻辑关系,连接各层级与各业务,提出利用云环境中的大数据人工智能分析工具对数据进行再加工的方法,并在“设计验证”、“试点建设”、“推广建设”三个主要阶段对中台的设计进行了详细定义,通过引入领域驱动思想,构建一个高内聚低耦合的业务子域,通过优化模型、数据共享等措施,搭建基于电网企业数据人工智能中台的应用。本文提出的电网企业数据人工智能中台结构从业务数据安全的角度较传统中台结构安全系数更高,更具有电网业务可持续发展的潜力。本文在大数据人工智能中台搭建完成的基础上,进一步地提出了模型的迭代优化的三种方法,并分别论述了三种模型更新方法的适用范围与优缺点。

By analyzing the reality of power grid business,this paper constructs the power grid business model,connects all levels and businesses by analyzing the logical relationship between data access,data governance,data model,general analysis model and data services,and puts forward the method of reprocessing data by using the big data artificial intelligence analysis tool in the cloud environment.The three main stages of“promotion and construction”define the design of the center platform in detail.By introducing the field driven idea,build a business sub domain with high cohesion and low coupling,and build the application of artificial intelligence middle platform based on power grid enterprise data through optimization model,data sharing and other measures.From the perspective of business data security,the middle platform structure of data artificial intelligence of power grid enterprises proposed in this paper has higher safety coefficient than the traditional middle platform structure,and has more potential for the sustainable development of power grid business.On the basis of building big data AI platform,this paper puts forward three methods of iterative optimization of model,and discusses the application scope,advantages and disadvantages of three methods of model updating.