摘要
Abstract
Cyber-physical power systems(CPPSs)are increasingly exposed to coordinated cyber-physical attacks because power system operation depends heavily on information communication and control functions.Existing defense studies usually focus on physical attacks against transmission lines or cyber attacks against information nodes separately.However,the coupling relationships between the power network and the information network may cause cross-layer failure propagation and amplify attack consequences.In addition,the attacker's available resources are difficult to determine accurately before attacks,so defense strategies based on fixed attack resources may underestimate system risk.To address these issues,a defense resource allocation method for CPPSs under uncertain coordinated attacks is proposed.
Firstly,a CPPS model considering multiple coupling characteristics is established.Structural coupling is used to describe the correspondence between transmission lines and communication links.Energy-supply coupling is introduced to characterize the influence of power node load shedding on the operating state of information nodes.Control coupling is further established to describe the degradation of power node regulation capability caused by information load loss.Based on these coupling relationships,the operating response of CPPSs under physical attacks and distributed denial of service(DDoS)attacks is modeled.Secondly,considering the uncertainty of attack resources,a tri-level coordinated defense-attack-defense(CDAD)optimization model is constructed.The upper level determines the defense strategy,including transmission line hardening and information node protection.The middle level searches for the worst coordinated attack strategy under different attack-resource scenarios.The lower level minimizes the post-attack operating cost under power network constraints,information network constraints,and coupling constraints.Finally,an improved nested column-and-constraint generation(NC&CG)algorithm is proposed.A coarse-solution and fine-solution strategy is adopted to reduce unnecessary high-precision solution of inner subproblems in early iterations,while strict convergence accuracy is restored in the final stage.
Case studies were conducted on the IEEE RTS-79 system and the IEEE 118-bus system.The results show that coordinated cyber-physical attacks caused larger system losses than physical attacks alone.In the IEEE RTS-79 system,when the physical attack resource was 8 and the DDoS attack resource was 2,the load shedding reached 2 580 MW,compared with 1 402 MW under physical attacks alone.The coordinated allocation of physical and cyber defense resources reduced post-attack load shedding;for example,when the physical defense resource was 5,increasing the cyber defense resource from 0 to 3 reduced the load shedding from 725 MW to 508 MW.Sensitivity analysis shows that the system was more sensitive to the control coupling coefficient than to the energy-supply coupling coefficient.In the IEEE 118-bus test scenario,the improved NC&CG algorithm obtained the same objective value as the conventional NC&CG algorithm,while reducing the computation time from 15 184 s to 10 326 s.
The following conclusions can be drawn from the study.(1)The proposed CPPS coupling model can describe the effects of structural coupling,energy-supply coupling,and control coupling on coordinated attack propagation.Ignoring these coupling relationships may lead to underestimation of system vulnerability.(2)The CDAD optimization model can coordinate transmission line hardening and information node protection under uncertain coordinated attacks,thereby reducing post-attack load shedding.(3)The improved NC&CG algorithm improves computational efficiency through the coarse-solution and fine-solution strategy while maintaining final solution accuracy.关键词
电力信息物理系统/多重耦合特性/不确定协同攻击/防御资源优化/改进嵌套列与约束生成算法Key words
Cyber-physical power systems/multiple coupling characteristics/uncertain coordinated attack/defense resource optimization/improved nested column-and-constraint generation algorithm分类
信息技术与安全科学
