基于深度学习的电力工程多源异构信息智能评估算法  

Deep learning-based algorithm for intelligent evaluation of multi-source heterogeneous information in power engineering

作　　者：夏青山 角丽萍 陈春丽 XIA Qingshan;JUE Liping;CHEN Chunli(Yunnan Power Grid Co.,Ltd.,Kunming 650000,China)

机构地区：[1]云南电网有限责任公司,云南昆明650000

出　　处：《电子设计工程》2024年第24期91-94,99,共5页Electronic Design Engineering

基　　金：云南电网科技项目(YNKJXM20220164)。

摘　　要：针对电力工程数据缺乏统一规范和标准,且难以深入挖掘分析已有电力工程数据,无法对工程建设提供数据支撑等问题,提出一种面向电力工程的多源异构数据评估方法,用于对相关工程数据进行分析评估。在建立电力工程多源异构数据库的基础上,通过RobustScaler技术对数据进行预处理,使用AdaBoost算法对训练数据集进行学习,得到基于弱回归器的组合策略模型。基于实际电力工程数据对模型进行的仿真试验结果表明,文中所提模型平均绝对误差为3.26%,均方根误差为6.47%,相比同类算法在数据分析的准确程度和稳定性上均具有明显优势。In response to the lack of unified specifications and standards for power engineering data,as well as the difficulty of deep mining and analyzing existing power engineering data,and the inability to provide data support for engineering construction,a multi-source heterogeneous data evaluation method for power engineering data analysis and evaluation is proposed.On the basis of establishing a multi-source heterogeneous database for power engineering,the data is preprocessed using RobustScaler technology,and the AdaBoost algorithm is used to learn the training dataset,resulting in a combination strategy model based on weak regressors.The simulation test results of the model based on actual power engineering data show that the average absolute error of the proposed model is 3.26%,and the root mean square error is 6.47%.Compared with similar algorithms,the proposed model has good advantages in the accuracy and stability of data analysis.

关 键 词：深度学习 电力工程 大数据分析 ADABOOST算法 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论] TN918.4[自动化与计算机技术—计算机科学与技术]