Data powers wind turbine predictive maintenance
The failure of wind turbine components can lead to weeks of costly downtime, repairs, and production losses. For many machines, faults can show up early as anomalies in behavior measured by sensors, enabling quick correction. For machines as complex as turbines the challenges are greater, not least because the "normal" behavior against which changes are measured varies, as no two turbines are the same. Proxima Solutions, part of BKW Group, worked with CSEM to optimize its maintenance processes in this uniquely challenging sector.
We can now make the best use of data and predictive diagnostics to increase the energy production capabilities of wind farms. Given our experience with CSEM, we would like go further, and apply the same approach to other technologies.
Working with Proxima's data scientists, CSEM developed a series of artificial intelligence algorithms. These algorithms provide early, high-accuracy problem detection, diagnosis, and prioritization by "learning" the usual behavior of each turbine, and matching this to the expert coding knowledge of BKW’s field technicians and wind engineers. The system is now in use at almost all of BKW's European wind farms, where it anticipated around 40 anomalous behaviors in its first 18 months. The core technology could also be applied to solar panels and batteries.
In 2023, DNV acquires AI-enabled wind software-as-a-service provider Proxima Solutions to strengthen its GreenPowerMonitor business.