At the end of February, the first semifinal for the 5th edition of the IndesIA Datathon took place. The Datathon is an initiative in which teams, made up of collaborating companies and technological partners, tackle specific challenges set by the companies themselves. Each team sets a challenge and, with the data provided by the associate company, it has the task of solving it within a certain amount of time. These results are then presented and aspects such as technical quality, applicability, and the presentation itself are assessed.

This event is organised by IndesIA, the Industry Association for the Advancement of the Data Economy and Artificial Intelligence, and its mission is to promote the strategic use of data using AI provided by associate companies.

At this edition of the Datathon, Elewit formed a team with Predictland AI to tackle the challenge of obtaining insights from leakage current measurements in high-voltage insulators, a successful collaboration for the first semifinal of the 5th edition of the Datathon.

The challenge set was to characterise the behaviour of certain types of insulators in response to changing weather conditions. These insulators, exposed to a broad spectrum of environmental conditions, face risks due to the accumulation of dirt or corrosion, which could compromise their efficiency and safety. To tackle this challenge, leakage current measurements were taken in insulators to act as key indicators of their status. With a circuit of over 45,000 km of overhead lines in the transmission grid, optimisation of asset management is essential.

The results obtained during the test could allow different uses of artificial intelligence that have not yet been considered in this field to be explored:

• In the short-term, it would help achieve a higher degree of accuracy regarding the level of risk that a line poses depending on the weather.
• In the medium- and long-term, it would help with planning maintenance tasks for insulators.
• Lastly, to cover the entire life cycle of the line, it will help us to choose the most appropriate insulator for its design.

The analysis methods developed are not only applicable to the insulators analysed in this project, but they also offer a robust framework for addressing similar challenges in other types of insulators. For example, polymer insulators with a fibreglass core that come with additional challenges for measuring their health status, may benefit greatly from these innovative approaches.

This project emphasises the importance of technical experience and cutting-edge technology. Automated mass measurements in critical components of the electricity transmission grid may be exploited using artificial intelligence algorithms. This synergy between the human experience and advanced analytical capabilities represents the future of the industry, joining forces to achieve continuous optimisation and greater operational efficiency. Elewit and PredictLand have demonstrated how strategic collaboration can deliver innovative and sustainable results in this sector.