Predictive maintenance plays a central role in the innovation and digitalization process of Global Power Generation's hydroelectric plants and has proved to be increasingly necessary to ensure greater efficiency.

In September 2018 was the launching of the PresAGHO (Predictive System and Analytics for Global Hydro Operation) project, focused on elaborating the world's first predictive diagnostics process to be integrated with maintenance processes, which Enel Green Power realized together with machinery manufacturers. The project has been based on a stream of activities leveraging O&M personnel's skills and field experience to develop distributed predictive diagnostics models for hydroelectric plants, by industrializing infrastructures and developing digital platforms to monitor and handle predictive alarms. The ambitious goal was to optimize performance and strengthen risk management, especially of the electromechanical parts in the plants. Initial results are already tangible: 1,800 predictive diagnostics models developed and applied in 320 plants and 80 O&M staff trained so far.

Another important achievement of the PresAGHO project is the e-maintenance platform, entirely developed with internal resources, where predictive models using Artificial Intelligence algorithms are defined and implemented. It allows for a timely diagnostic analysis of components to assess their status, optimizing maintenance plans and forecasting breakdowns by means of anomaly detection, from the first signs of potential malfunction. O&M now has a powerful tool for efficiently planning and executing inspections and maintenance, thereby significantly increasing safety and savings in terms of time and money.

Effective maintenance is possible thanks to the reinforcement of three pillars:

• The infrastructure that guarantees data availability;
• The commitment of O&M technicians as regards predictive diagnostics, spreading digital culture and creating a network of skills;
• The digital tool that predictively analyzes plant data and can be used by all O&M technicians to enhance their expertise.

In a few months of full operations, several anomalous situations have already been detected which would have caused production losses and machinery damage, if not handled in time.