News
Have you ever heard of non-technical losses? According to the National Electric Energy Agency (ANEEL), this refers to what we popularly call "gatos" (illegal connections). This type of loss is a major challenge for the sector and, for this reason, Pix Force and Neoenergia have teamed up for an important study aimed at mitigating this problem using artificial intelligence. We present the case study in this text. Enjoy the reading!
Speak with one of our specialists and discover how Pix Force can transform your business
What are non-technical losses?
According to ANEEL, non-technical losses originate from various problems caused in the electrical grid. These losses, also popularly referred to as "gatos" (illegal connections), are associated with the management of energy companies and the socioeconomic conditions of the affected areas. Among the main problems that cause this loss, we can highlight: • Theft
• Clandestine connection
• Grid diversion
• Meter tampering
• Reading, measurement, and billing errors
Also according to the agency, the percentage of non-technical losses has grown in recent years, equaling technical losses in 2020. The same study shows that the impacts are significant for both concessionaires and consumers. In 2020 alone, there was a cost of R$5.6 billion related to these losses. It is important to emphasize that this value ends up affecting the consumer tariff and, therefore, it is necessary to think: how to reduce these losses and optimize the electrical energy industry? Based on this question, Pix Force joined forces with Neoenergia to develop a study to solve this problem. Artificial intelligence in solving problems Far beyond the costs, non-technical losses bring several problems to concessionaires. Among them, we can mention: • Revenue reduction
• Drop in distribution system efficiency
• Equipment overload
• Loss of credibility, quality, and reliability in the market
In partnership with Neoenergia, one of the largest electricity distributors in Brazil, Pix Force participated in the SOFTEX IA² program to develop solutions for the electricity sector. Using the Pix Grid product, we thought of an alternative based on artificial intelligence to inspect, control, and optimize electrical grids. Quote: Read more: Artificial Intelligence in the Electrical Sector
But how does the project work?
More specifically, Pix Force sought a solution based on data collected by Neoenergia and other external sources. Computer vision, remote sensing, and other specialists assisted in the development of a methodology capable of creating a heat map. On this map, it was possible to indicate the probability of energy fraud and theft in rural area networks, the main regions covered by the study. To do this, we started with the automatic search of available satellite images for the segmentation and classification of rural properties, center pivot irrigation systems, and other points of interest. The analyses took into account the distances between the electricity meters in relation to the distribution network and the location of the points of interest. Thus, technical visits could be prioritized through an analysis of the probability of irregularities. The project developed a web platform, which allowed analysts to create field inspection missions based on the heat map, and a mobile application. These make it easier for field agents to receive the defined mission information with the proper geographical guidance. The application also allowed the capture and transmission of photos for targeted decision-making. The tool developed for creating the heat maps involved several technologies. Everything was made possible by capturing Sentinel-2 and CBERS satellite images, data from the Rural Environmental Registry (CAR), and data from the National Water Agency (ANA), in addition to historical data common to Neoenergia's operation. After structuring this data, it was processed by edge detection algorithms and fed into neural networks and classification algorithms trained throughout the months of the projects: • CircleNet performed the detection of the pivots and extracted information about them.
• RandomForest and SVM did the selection and classification of the other features
The selection of the regions of interest was made with the preparation of vector data of water resources and with the pointing of meters using Neoenergia data. These were used for the training sessions that resulted in the generation of heat maps with the probabilities of potential fraud and theft. With this, the methodology used proved to be appropriate for the project's objective, obtaining significant and innovative results for the electricity sector.
Get to know more about Pix Grid
Pix Grid is a solution that uses computer vision to inspect the electricity sector. It starts from some points of interest that help optimize processes and have a better analysis of the entire field. Using artificial intelligence for these activities brings a series of benefits such as: • Increased productivity
• Faster access to data
• Identification of points of interest
• Production of automatic alerts
In addition to monitoring non-technical losses, Pix Grid allows other activities such as the inspection of transmission lines and a pruning prioritization system. Want to know in more detail? Access the website and speak with Pix Force.
Fabio Caraça
Fábio Caraça is the Chief Growth Officer at Pix Force. He leads Pix Force's transformation into a scalable SaaS operation, combining strategic vision, culture, and high-impact execution.
