论文标题
一种基于深度学习的检测方法，用于联合完整性可用性网络攻击电源系统
A Deep Learning based Detection Method for Combined Integrity-Availability Cyber Attacks in Power System
论文作者
论文摘要
作为地球上最大，最复杂的系统之一，Power Grid（PG）的操作和控制已成为对物理和网络层的复合分析，这使其很容易受到经济和安全考虑的攻击。最近已经提出了一种新型的攻击，即作为数据完整性可利用性攻击的组合攻击，攻击者可以同时操纵和盲目对SCADA系统进行一些测量，从而误导控制操作并保持隐身性。与传统的FDIA相比，这种综合攻击可以进一步复杂化并消除基于模型的检测机制。为了检测这种攻击，本文提出了一种新型的随机降解LSTM-AE（LSTMRDAE）框架，其中可以明确捕获测量的空间 - 周期性相关性，并且不可接受的数据被随机掉落层对抗。评估了所提出的算法，并在各种看不见的攻击尝试下在标准IEEE 118-BUS系统上验证了性能。
As one of the largest and most complex systems on earth, power grid (PG) operation and control have stepped forward as a compound analysis on both physical and cyber layers which makes it vulnerable to assaults from economic and security considerations. A new type of attack, namely as combined data Integrity-Availability attack, has been recently proposed, where the attackers can simultaneously manipulate and blind some measurements on SCADA system to mislead the control operation and keep stealthy. Compared with traditional FDIAs, this combined attack can further complicate and vitiate the model-based detection mechanism. To detect such attack, this paper proposes a novel random denoising LSTM-AE (LSTMRDAE) framework, where the spatial-temporal correlations of measurements can be explicitly captured and the unavailable data is countered by the random dropout layer. The proposed algorithm is evaluated and the performance is verified on a standard IEEE 118-bus system under various unseen attack attempts.
