电力统计大数据质量可视化控制方法研究

Research on Visual Control Method of Power Statistics Big Data Quality

为使电力统计大数据具备可读性与应用准确性,提升整个电力企业的运转效率,提出可视化控制方法。通过引入、去除数据变量,建立电力统计大数据的质量控制算法。结合地理信息系统技术与Web端口,构建由Web端口连接地理信息系统中各组成部分的平台。采用B/S架构与设计工具模块等,组成可视化实现单元。基于四层卷积神经网络结构,在非线性映射层前后,添加特征缩小网络层与扩展层,构建六层卷积神经网络结构,输出可视化控制结果。实验结果表明,所提方法有效去除了大部分问题数据,数据质量有显著提升。

In order to make the big data of power statistics readable and accurate,and improve the operation efficiency of the whole power en⁃terprise,a visual control method is proposed.By introducing and removing data variables,the quality control algorithm of power statistical big data is established.Combined with GIS technology and Web port,a platform connecting the components of GIS by Web port is constructed.B/S architecture and design tool module are adopted to form a visual implementation unit.Based on the four layer convolution neural network structure,the feature reduction network layer and extension layer are added before and after the nonlinear mapping layer,the six layer convo⁃lution neural network structure is constructed,and the visual control results are output.Experimental results show that the proposed method ef⁃fectively removes most of the problem data and significantly improves the data quality.