电网调度系统网络安全态势感知研究

Research on network security situation awareness of electric power dispatching system

电网调度系统网络安全常采用基于边界保护的物理隔离,此方式无法有效抵抗高级病毒软件的入侵。基于此,构建了基于Q学习算法的网络状态转移算法,预测网络攻击可能的最佳路径。改进了网络安全度量标准,在系统损失、攻击成本、防御成本的基础上,引入防御回报、网络状态转移成本,以静态经济收益作为网络状态转移的判断准则。应用演化博弈理论构建复制动态方程,动态再现攻防双方的对抗行为,计算经济收益动态变化,预测攻击行为的变化,确定最佳防御策略。仿真结果表明基于Q学习算法与演化博弈理论的电网调度网络安全态势感知方案能够有效识别网络攻击的可能路径以及带来的最大威胁,有助于调度人员作出有效决策。

Physical isolation based on boundary protection is often used in network security of power grid dispatching system, which cannot effectively resist the invasion of advanced virus software. On this basis, a network state transition algorithm based on Q learning algorithm is constructed to predict the best possible path of network attack. Based on the system loss, attack cost and defense cost, the defense return and network state transition cost are introduced, and the static economic benefit is taken as the criterion of network state transition. The evolutionary game theory is used to construct replication dynamic equation, which dynamically reproduces the antagonistic behavior of both attackers and defenders, calculates the dynamic change of economic returns, predicts the change of attacking behavior, and determines the best defense strategy. The simulation results show that the security situational awareness scheme of power dispatching network based on Q learning algorithm and evolutionary game theory can effectively identify the possible path of network attack and the greatest threat, and help dispatchers make effective decisions.