OUR SOLUTIONS
Scheduled renovation of overhead distribution infrastructures
Enedis carries out detailed inspections of its lines, to identify damaged equipment and prioritize replacement operations.
The current limitations of human expert processing (slow processing time, limited number of expert technicians able to carry out processing, cost of processing) prevent a significant increase in the number of kilometers visited, thus limiting its ability to prevent incidents before they occur.
How can we automate fast, reliable diagnosis based on analysis of aerial images taken by drone?
Our SaaS and edge solutions
The proposed solution integrates a Cloud services platform and Boxes hosting these services locally. Together, they form a distributed decision-support system, based on Artificial Intelligence and Vision.
Human-Computer Interfaces (HCI) ensure interaction between business experts and AI, both in the learning and recognition phases.
Customer benefits
Fast, high-performance, precise, with reduced production costs
Reduced integration costs with business information systems, thanks to a modular, secure architecture.
Centralize data feedback on a single platform, enabling sharing with stakeholders.
Harmonization of reporting formats used by Business Experts
In a nutshell
The application developed enables you to :
- acquire images of aerial infrastructures, from a drone in flight, in real time,
- automatically analyze videos in real time to detect electrical cables and the various parts connecting the cables to the tower,
- segment images of these elements, in interaction with an operator specialized in diagnostics,
- analyze these parts to diagnose the degree of wear.
Enedis operates Europe's largest electricity network, with 1.4 million kilometers of lines in France. Of these, 230,000 km are overhead medium-voltage lines (attached to poles). Because of its exposure to the vagaries of the weather, this part of the network accounts for 70% of incidents impacting on service quality for our customers. As a result, these lines are regularly inspected in depth by an expert, either on foot, by helicopter or, more recently, by drone. The salient point of drone acquisition is that the expert has to analyze the video after the visit, sometimes frame by frame, to detect anomalies (damaged pole, broken insulator, worn stirrup...): a very time-consuming process. Our intuition was to use artificial intelligence as a tool to help detect these anomalies from drone videos. We turned to A2D to turn this idea into reality. Their approach and expertise enabled us to offer the expert a tailor-made solution: on the one hand, it considerably minimizes his analysis time, and on the other, it doesn't change his habits, since the AI is hidden behind the application he currently uses. Finally, by proposing a solution based on edge-computing, processing is carried out locally, without depending on computer network speeds, while benefiting from and participating in the learning process of the AI matrix. Throughout the development process, Pierre and his team were very pedagogical, regularly informing us of their progress and systematically putting themselves in the user's shoes. The solution resulting from this A2D-Enedis cooperation enables us to meet the challenges of power grid maintenance at a controlled cost.
