Reconstruçãao 3D de subestações de energia para monitoramento remoto com arquitetura Edge-fog com estudo de integração ao contexto 5G

Abstract

The need for automation and remote monitoring in power substations has been a growing topic in the literature and practical applications, often dealing with a tedious task in an energized environment, at risk to the operator’s health, or even difficult to access. One way to approach this problem is to reproduce the scenario and its components in a virtual environment, through 3D reconstruction of the real entities of the substation. Currently, this scanning is performed largely in a non-optimized way, relying on embedded processing or only data acquisition to be worked offline in the cloud. These applications reduce, or even make it impossible to request, in realtime, the state of the substation components, in addition to not being scalable and efficient. Faced with this challenge, this work proposes two solution fronts: the use of classical techniques optimized by the parallel processing of point cloud data through an iterative octree; and an edge-fog computing architecture, which allows distribution of computational load and storage among all the entities in the solution, which in turn are allocated in processing layers (edge and fog), something unprecedented in the literature for this specific application. The 5G network application is presented along with a brief discussion on how this technology can benefit and allow the overall distributed processing. A robot is developed to collect data from images and point clouds. All algorithmic and mathematical processes are allocated and detailed for both layers. As a result, the work brings a complete analysis of the benefits of using architecture against a local and traditional approach, based on edge. The scalability of the solution is demonstrated at the end, in a real application environment, and we conclude with how much a private 5G network would benefit this solution in the required remote monitoring scenario.