Image: Thampapon Shutterstock

A predictive wind turbine vibration analysis system that controls 1,400 wind turbines, including 600 in Spain. The savings that can be achieved with a single detection can be more than 100,000 euros per machine on the cost of repairs. This analysis system can be implemented in any model of wind turbine.

Automation and machine learning are the key elements of this Endesa system, capable of processing up to 100,000 records per second.

Endesa is setting up a new predictive analysis system in its wind farms for the maintenance of wind turbines according to the vibratory behavior of this equipment, with savings in repair costs ranging from 15% to 95%.

The analysis system is carried out from the Enel Green Power diagnostic room in Spain, from its Renewable Energies department, which remotely supervises more than 1,400 wind turbines in Madrid, of which 600 belong to Endesa Spain and the rest to Enel Green Power. and are located in Mexico, Chile, Italy, Greece, among other countries.

Predictive analytics is the early detection of failures in major components of wind turbines, which enables the detection of failures which in some cases can be months in advance, allowing repairs to be planned for periods with lower costs, minimizing losses due to turbine interruptions and thus improving the efficiency of the entire system.

Automation and machine learning is machine learning, which is essential in this process, as the system supports processing up to 100,000 records per second.

Wind turbine components. Image: Andrea Crisante Shutterstock

Endesa stresses that the savings that can be achieved with a single correct detection, which can reach 100,000 euros in a single wind turbine, this in the case of faults in large wind turbine components, in which the repairs are considered as important corrections and involve operations. with cranes.

By having data from different turbine models, better predictive analysis can be performed in any wind farm in the world. Based on the know-how acquired, information is shared with manufacturers to make improvements to their models and an improvement in the learning curve is facilitated to make more and more precise predictions.

Endesa has implemented the analysis system in all wind turbines that have less than 5 years of operation, the “Condition Monitoring System” ensures the supervision and control of the condition of each machine.

More information: endesa.com