基于卷积神经网络的数字化智能电站数据安全评估模型  

Data Security Evaluation Model of Digital Intelligent Power Station Based on Convolutional Neural Network

作　　者：孙勇 王卓瑜 郑凤柱 张兴彬 李悦 SUN Yong;WANG Zhuo-yu;ZHENG Feng-zhu;ZHANG Xing-bin;LI Yue(State Grid Xinyuan Group Co.,Ltd.,Beijing 100052 China;Beijing Guodiantong Network Technology Co.,Ltd.,Beijing 100070;Henan Luoning Pumped Storage Co.,Ltd.,Luoyang 471700 China;Beijing Haizhi Xingtu Technology Co.,Ltd.,Beijing 100083 China)

机构地区：[1]国网新源集团有限公司,北京100052 [2]北京国电通网络技术有限公司,北京100070 [3]河南洛宁抽水蓄能有限公司,河南洛阳471700 [4]北京海致星图科技有限公司,北京100083

出　　处：《自动化技术与应用》2024年第12期16-19,24,共5页Techniques of Automation and Applications

基　　金：国网新源集团有限公司科技项目资助(SGXYKJ-2021-133)。

摘　　要：为了确保数字化智能电站的安全运行,提出基于卷积神经网络的数字化智能电站数据安全评估模型。通过传感器获取数字化智能电站的异常数据,并对数据进行预处理。运用希尔伯特黄变换提取异常数据中综合电流的瞬时频率,将其作为数字化智能电站数据异常特征。结合卷积神经网络的异常数据特征分类训练诊断异常数据,依据诊断结果进行数字化智能电站数据安全评估。实验表明:该模型评估准确率与容错率高,减少了评估的能量消耗,节约评估成本,可以实现异常数据的及时告警并输出故障原因,便于后续维修工作顺利进行。In order to ensure the safe operation of digital intelligent power station,a data security evaluation model of digital intelligent power station based on convolutional neural network is proposed.The abnormal data of the digital intelligent power station is obtained by sensors,and the data is preprocessed.The instantaneous frequency of the comprehensive current in the abnormal data is extracted by using Hilbert Huang transform,which is taken as the abnormal characteristics of the digital intelligent power station data.The abnormal data is trained and diagnosed by combining the abnormal data characteristics of convolutional neural network,according to the diagnosis results,the data security assessment of digital intelligent power station is carried out.The experiment shows that the model has high evaluation accuracy and fault tolerance,reduces the energy consumption of evaluation,saves the evaluation cost,and can realize the timely alarm of abnormal data and output the fault cause,which is convenient for the subsequent maintenance work.

关 键 词：卷积神经网络 智能电站 安全评估模型 数据采集 异常数据 告警模块 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]