Modelagem de power grids utilizando multi-aspects graphs

Abstract

The electrical grids consist of interconnected meshes responsible for the transportation and delivery of electricity to consumers. These networks have generation, transmission, and distribution points. These generation points are connected through transmission lines. The transmission grid created can be divided into subsystems, which prevent general failures, improve maintenance, and also provide safer service. In Brazil, there is the SIN (National Interconnected System), an electrical system that interconnects almost all regions of the country, formed by transmission and distribution networks, thus connecting generation plants to consumers. Due to the need for energy in people’s daily lives, this system must be modeled so that there are no unnecessary expenses for the installation of equipment and connection of networks, and so that it can serve customers without losses. These networks must be constructed considering the entire system, not just the connection from one point to another. Thus, this grid becomes complex, and its constant review becomes necessary, whether due to the creation of new lines, the removal of old lines/problems, or other reasons. Therefore, it is important to model this system to facilitate these processes and interactions between them. Current models often do not demonstrate the difference between a real network and a computational model of a network. A current approach that can be used to model this power grid is to use the concept of Multi-Aspect Graph (MAG), a model capable of representing multi-layered, time-variable, or high-level networks by dividing them into a list of aspects in a set with each edge of the network. This model has similarities with the traditional graph model but, compared to traditional graphs, can abstract higher-order networks, among other peculiarities. The network model proposed in this work allows for the separate analysis of generation points from substations, thus providing a better understanding of the operation of a real electrical network. Unlike other approaches, the current work emphasizes the importance of analyzing the different aspects of the network, as well as demonstrating the difference between a real network and network data, showing that for data, nodes with few connections are of little relevance, but for a real network, they can be essential.