For the electricity sector, AI goes beyond just a new source of consumption in times of oversupply in Brazil. It is seen as a strategic ally to increase the operational efficiency and resilience of the network in the face of intensifying extreme weather events, which have caused increasing losses. About two years ago, a storm hit São Paulo and left thousands without power. Last year, a flood in Rio Grande do Sul knocked a transmission tower off the ground. Winds of more than 200 km/h in the northern region and fires near transmission line safety lines are also on the list of recent incidents.
With more than 80 factories and 74 thousand kilometers of transmission lines installed in the country, and a presence in the five regions of Brazil, Axia Energia (formerly Eletrobras) invests in initiatives in artificial intelligence, data, industrial, automation, drones and robotics, the company’s main lines with the aim of innovation and adding value. An asset control center was established at the company’s headquarters in Rio de Janeiro. This coincides with the massive installation of sensors and meters in projects to begin analyzing data and begin developing innovative models, such as those studying extreme winds in Brazil. In this wind case, Nvidia’s Earth-2 platform is used.
Through the Climate Center, the company now has the structure to provide daily bulletins, specific alerts and special forecasts. The platform has the ability to not only predict meteorological events, but also indicate which assets may be affected. In this way, the company will be able to prepare in advance to reduce risks and avoid interruptions to the power supply, as well as protect employees in the field. At the beginning of the year, a project to modernize and expand the network capacity and connection to the Tokoroi Hydropower Station was completed.
Thanks to the work of 50 specialists, the factory is 100% connected to Wi-Fi technology. The project implemented approximately 450 access points in all areas of the plant, including: the power plant, water intake, underground passages, transformer yard, spillway, dam, and supervision and command buildings. Therefore, communication with employees does not depend solely on radio.
Data centres, the construction of which must be driven by the advancement of artificial intelligence, pose challenges to the operation and management of assets. These data centers can have elastic demand, can incorporate new resilience mechanisms such as demand response and storage, and can have an impact on neighboring transport networks. To evaluate the impacts, an electric power microgrid was built providing wind and solar energy in a 1 MW data center, installed in Casa Nova (BA), near the Sopradinho hydropower station.
In the distribution sector, it has been applied to a topic that keeps many distributors up at night throughout Brazil: non-technical losses, energy theft, leading to losses for franchisees and reduced revenues for public entities. Energisa has pioneered the use of algorithms to assess anomalous patterns, as credit card issuers do when cards are used infrequently. He improved the system with more artificial intelligence embedded in data analysis. In seven years, losses of 950 gigawatts were avoided, enough to supply a city with 400,000 customers for one year. Today’s non-technical losses are 3.74%, one of the lowest in the country, with customers in the North, Southeast, Midwest and Northeast regions.