Multi-level Hierarchical Clustering Algorithm For Energy-theft Detection in Smart Grid Networks

Without a good and a robust infrastructure, it is difficult for electricity providers to satisfy the increased demand needs, and to have an effective management of the electricity grid, from generation, transmission, and distribution. Several studies have shown that the Smart Grid represents a necessary evolution of current electrical infrastructures. By using communication technology like the Wireless Sensor Networks (WSNs), it is possible to effectively improve the reliability and robustness of electrical networks for better control, monitoring, and management visibility. The only concern is that existing grids often present infrastructure constraints in the implementation of these solutions. Moreover, the very exorbitant cost should not be overlooked, depending on the choice of technology. In agreement with the Chadian electricity company (SNE), this work presents a self-stabilizing hierarchical algorithm taking into account the electrical network. The proposed architecture considers two levels: the Home Area Networks (HAN) layer and Neighborhood Area Networks (NAN) layer. In essence, we define three types of nodes: Cluster Head (CH), Simple Node (SN) and Home Node (HN). Extensive simulations using the OMNeT++ simulator reveal that the proposed method can efficiently identify the fraudulent users in the system.