摘要
针对当前电力信息系统安全状态受威胁严重的问题,提出一种基于卷积神经网络模型的电力信息安全状态监测系统。系统结合电力信息系统特征,应用安全感知单元、安全状态监测模型和数据防篡改模型保障电力信息系统的稳定运行和数据安全。系统通过加入可视化模块和报警模块来设计安全感知报警单元,采取基于STM32F103VCT6微控制器实现多通道模数转换器的控制和应用,能够对多种传感器输出的信号进行转换,提高了即时监测能力。该系统还利用卷积分解技术和深度可分离技术对卷积神经网络进行改进,构建各个通道之间的数据联系,通过安全状态监测模型发挥卷积神经网络的空间优势。试验结果显示,所提研究系统的采样时延最低为132 ms,安全状态评估准确率最大为100%,可实现高效、精确的系统安全状态监测。
A power information security status monitoring system based on convolutional neural network model was proposed to address the serious threat to the current security status of power information systems.The system combined the characteristics of the power information system and applied security perception units,security status monitoring models,and data tamper prevention models to ensure the stable operation and data security of the power information system.The system designed a security perception alarm unit by adding visualization modules and alarm modules.It adopted a STM32F103VCT6 microcontroller to achieve control and application of multi-channel analog-to-digital converters,which can convert signals output by multiple sensors and improve real-time monitoring capabilities.The system also utilized convolutional decomposition and deep separable technology to improve convolutional neural networks,constructing data connections between various channels,and utilizing the spatial advantages of convolutional neural networks through security state monitoring models.The experimental results show that the minimum sampling delay of the research system is 132 ms,and the maximum accuracy of security state assessment is 100%,which can achieve efficient and accurate system security state monitoring.
作者
刘立亮
文涛
叶磊
Liu Liliang;Wen Tao;Ye Lei(State Grid Anhui Electric Power Co.,Ltd.,Xuancheng Power Supply Company,Xuancheng Anhui 242099,China)
出处
《电气自动化》
2024年第5期11-14,共4页
Electrical Automation
关键词
电力信息
安全状态监测
安全感知单元
可视化模块
卷积神经网络
卷积分解
power information
security status monitoring
security perception unit
visualization module
convolutional neural network
convolutional decomposition
