电力系统大模型应用中的AI服务器安全性与可控性研究  

Research on the Security and Controllability of AI Server in the Application of Power System Large Model

作　　者：王斯诺 王仑 刘伟 郭敬林 范亚娜 WANG Sinou;WANG Lun;LIU Wei;GUO Jingin;FAN Yana(Beijing Guodiantong Network Technology Co.,Ltd.,Beijing 100085,China)

机构地区：[1]北京国电通网络技术有限公司,北京100085

出　　处：《移动信息》2025年第4期175-177,共3页Mobile Information

摘　　要：随着电力系统的数字化转型,AI服务器在大模型中的应用不断深入,促进了电力系统的智能化。然而,安全性和可控性问题成为新的挑战。文中分析了AI服务器在物理、网络、系统和数据安全方面存在的隐患,探讨了模型、数据与决策的可控性问题,提出了从技术、管理和合作等层面加强AI服务器的安全性与可控性,为电力系统的稳定、安全运行提供了理论支持。With the digital transformation of power systems,the application of AI servers in large models continues to deepen,promoting the intelligence of power systems.However,security and controllability issues have become new challenges.This paper analyzes the hidden dangers of AI servers in physical,network,system and data security,discusses the controllability of models,data and decisions,and proposes to strengthen the security and controllability of AI servers from the aspects of technology,management and cooperation,providing theoretical support for the stable and safe operation of power systems.

关 键 词：电力系统大模型 AI服务器 物理安全 网络安全 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]