Search for: All records

This paper presents a deep learning based multi-label attack detection approach for the distributed control in AC microgrids. The secondary control of AC microgrids is formulated as a constrained optimization problem with voltage and frequency as control variables which is then solved using a distributed primal-dual gradient algorithm. The normally distributed false data injection (FDI) attacks against the proposed distributed control are then designed for the distributed gener-ator's output voltage and active/reactive power measurements. In order to detect the presence of false measurements, a deep learning based attack detection strategy is further developed. The proposed attack detection is formulated as a multi-label classification problem to capture the inconsistency and co-occurrence dependencies in the power flow measurements due to the presence of FDI attacks. With this multi-label classification scheme, a single model is able to identify the presence of different attacks and load change simultaneously. Two different deep learning techniques are compared to design the attack detector, and the performance of the proposed distributed control and the attack detector is demonstrated through simulations on the modified IEEE 34-bus distribution test system. 

Real-world networked systems often show dynamic properties with continuously evolving network nodes and topology over time. When learning from dynamic networks, it is beneficial to correlate all temporal networks to fully capture the similarity/relevance between nodes. Recent work for dynamic network representation learning typically trains each single network independently and imposes relevance regularization on the network learning at different time steps. Such a snapshot scheme fails to leverage topology similarity between temporal networks for progressive training. In addition to the static node relationships within each network, nodes could show similar variation patterns (e.g., change of local structures) within the temporal network sequence. Both static node structures and temporal variation patterns can be combined to better characterize node affinities for unified embedding learning. In this paper, we propose Graph Attention Evolving Networks (GAEN) for dynamic network embedding with preserved similarities between nodes derived from their temporal variation patterns. Instead of training graph attention weights for each network independently, we allow model weights to share and evolve across all temporal networks based on their respective topology discrepancies. Experiments and validations, on four real-world dynamic graphs, demonstrate that GAEN outperforms the state-of-the-art in both link prediction and node classification tasks.