Minimização do custo global de infraestrutura de comunicação e alocação de PMUs em WAMS utilizando betweenness centrality e dominância em grafos

Abstract

This work presents a method for optimizing the global cost of Communication Infrastructure (CI) and allocation of Phasor Measurement Units (PMUs) in Wide Area Measurement Systems (WAMS). State Estimation (SE) is a tool that helps in the treatment of data obtained from measurements, eliminating gross errors and topology errors. This tool can only be run when the available measurement data is sufficient, therefore there must be a sufficient amount of PMUs in an Electric Power System (EPS) to guarantee SE. Voltage, current and phase measurements on all system buses are obtained in real time from measuring devices installed in the EPS and with the help of the SE, a secure database is provided for decision-making in electrical networks, being very important for the implementation of a Smart Grid. The allocation of PMUs and CI demands a high cost, and CI has a greater impact on the overall cost of the WAMS. The high impact of CI on total cost motivated this new optimization method. The objective of this work was to optimize the CI and the number of PMUs to obtain a lower overall cost, in simulations with the IEEE-14, IEEE-30, IEEE-118 and IEEE-300 networks. Through the knowledge of graph theory and the elaboration of the allocation of PMUs as an Adapted Connected Dominant Set problem, it was possible to optimize the total cost. Thus, a new weighted metric was proposed and used with Betweenness Centrality. In addition to Betweenness Centrality, two algorithms quite common in graph theory were used in the optimization process. Dijkstra’s algorithm that finds the shortest paths between vertices and Kruskal’s algorithm that searches for a minimal spanning tree for a connected weighted graph, were used to compose the simulation. The developed algorithm was able to reduce the overall cost of the SEPs tested.