摘要
针对智能电网大数据分析与决策面临的计算效率低,精度较差的问题,文中提出一种基于互信息关联分析的大数据分析决策方法。通过计算电网大数据与配电项目评估指标之间的互信息量大小,可以有效评估项目质量。另外,为了提高项目评估的准确性,提出利用多隐层神经网络算法对判决门限进行训练,通过多层神经元节点预测误差回传反馈,迭代调节神经元权值,得到最终的判决门限。通过系统实现验证了文中所提方法可以有效降低智能电网大数据的数据处理时延,并提高配电项目评估精度,为实现智能电网大数据分析与决策提供有力的工具。
Aiming at the problem of low efficiency and low precision in big data analysis and decision-mak-ing of smart grid,this paper proposes a method of big data analysis and decision-making based on mutual information association analysis.By calculating the mutual information between the big data of power grid and the evaluation index of distribution project,the project quality can be effectively evaluated.Inaddition,inorder to improve the accuracy of project evaluation,a multi-hidden layer neural network algorithm is proposed to train the decision threshold.The final decision threshold is obtained by feedback of prediction errors from multi-layer neuron nodes,and the weights of neurons are adjusted iteratively.The system imple-mentation verifies that the proposed method can effectively reduce the delay of big data processing in smart grid,and improve the accuracy of distribution project evaluation,which provides a powerful tool for large data analysis and decision-making in smart grid.
作者
陈钦柱
符传福
韩来君
CHEN Qin-zhu;FU Chuan-fu;HAN Lai-jun(Institute of Electric Power Science,Hainan Power Grid Co.Ltd.,Haikou 570311,China)
出处
《电子设计工程》
2020年第6期30-34,共5页
Electronic Design Engineering
基金
国家科技支撑计划课题(2013BAA01B03)。
关键词
智能电网
大数据
互信息
项目评估
神经网络
smart grid
big data
mutual information
project evaluation
neural network
