基于改进ELM算法的水电站数字孪生系统仿真实验  

Simulation experiment of digital twin system for hydropower station based on improved ELM algorithm

作　　者：胡恢武 HU Huiwu(Xiangbiling Hydropower Station of Guizhou Jinyuan Weining Energy Co.,Ltd.,State Power Investment Group,Bijie 553107,Guizhou China)

机构地区：[1]国家电投集团贵州金元威宁能源股份有限公司象鼻岭水电站,贵州毕节553107

出　　处：《粘接》2024年第7期185-188,192,共5页Adhesion

摘　　要：为提高水电站数字孪生系统信息物理融合网络安全检测的准确率,提出一种基于改进极限学习机(ELM)的检测方法。以ELM算法为基础检测算法,通过改进其加权方式,同时引入头脑风暴优化算法(BSO)确定ELM的最佳输入权重和隐含层偏置参数;最后,将改进的ELM算法应用于水电站CPS网络安全检测中。仿真结果表明,改进ELM算法在数字孪生的CPS网络安全检测中具有明显的优势,检测准确率、检测率、召回率更高,误报率更低,且检测时间仅为0.39 s,表现出良好的实时性。改进的BSO-ELM方法可实现水电站数字孪生系统CPS网络的安全检测,提高数字孪生系统的安全性。In order to improve the accuracy of cyber⁃physical fusion network security detection of digital twin system of hydropower station,a detection method based on improved Extreme Learning Machine(ELM)was proposed.Based on the ELM algorithm,the weighting method was improved,and the brainstorming optimization algorithm(BSO)was introduced to determine the optimal input weight and hidden layer bias parameters of ELM.The improved ELM algorithm was applied to the CPS network security detection of hydropower stations.The simulation results showed that the improved ELM algorithm had obvious advantages in the CPS network security detection of digital twins,with higher detection accuracy,detection rate and recall rate,lower false alarm rate,and the detection time of 0.39s,showing good real⁃time performance.The improved BSO-ELM method can realize the security detection of the CPS network of the digital twin system of hydropower station and improve the security of the digital twin system.

关 键 词：数字孪生系统 CPS网络 网络安全检测 ELM算法 BSO算法 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置] TM622[自动化与计算机技术—控制科学与工程]