基于深度神经决策森林的电网数字化转型数据融合方法

Data fusion method for power grid digital transformation based on deep neural decision forest

在电网数字化转型过程中,所需处理的数据种类繁多、规模庞大、数据源分布、格式不一致,导致数据质量差、数据分析和处理效率低。为此,提出基于深度神经决策森林的电网数字化转型数据融合方法。通过电网管理平台(计财域)对电网进行数字化转型,根据数据库功能差异进行分库设计,并进行读写分离设计,更好地支撑平台应用的集中部署、降低主实例事务处理压力。结合深度神经决策森林融合数字化转型后的电网数据,采用卷积神经网络,提取数字化转型后电网数据特征,将其输入由多决策树组合而成的决策森林中,通过数据分类完成数据融合。实验结果表明:本方法能够有效完成电网数字化转型数据融合,融合后数据更易查看,便于后续的应用分析;最终的数据融合准确率达到98.7%,数据损失函数值仅为0.7,且数据融合覆盖度较高,可以提升电网数字化转型过程中的电网数据应用效果。

In the process of digital transformation of power grid, the data needed to be processed are various, large in scale, inconsistent in data source distribution and format, resulting in poor data quality and low efficiency of data analysis and processing. Therefore, a data fusion method for power grid digital transformation based on deep neural decision forest is proposed. The digital transformation of the power grid is carried out through the power grid management platform (accounting domain). The sub database design is carried out according to the functional differences of the database, and the read-write separation design is carried out to better support the centralized deployment of platform applications and reduce the transaction processing pressure of the main instance. Combining the deep neural decision forest to fuse the power grid data after the digital transformation, the convolutional neural network is used to extract the characteristics of the power grid data after the digital transformation, input them into the decision forest composed of multiple decision trees, and complete the data fusion through data classification. The experimental results show that the method in this paper can effectively complete the data fusion of power grid digital transformation, and the fused data is easier to view and facilitate subsequent application analysis;The final data fusion accuracy rate reached 98.7%, the data loss function value was only 0.7, and the data fusion coverage was high, which could improve the grid data application effect in the process of digital transformation of the grid.