Ørsted can now predict wind turbine generator failures at least three months in advance 70% of the time using a debris sensor and condition monitoring model.
The company said the system increased early detection from close to 0% before sensor installation to 70% in 2025 across 560 generators on one offshore wind turbine model.
Ørsted added that generator failures are costly and disruptive, often requiring jack-up vessels, specialist technicians and complex planning.
Engineers identified in 2021 that metallic debris builds up inside generator housings as stator wedges degrade, increasing the risk of short-circuiting and failure.
This led to the development and patenting of a debris sensor installed inside the generator housing to provide a measurable early warning signal.
The sensors have since been retrofitted across 560 wind turbines, with monitoring tailored to specific turbine models.
Using debris sensor data and other turbine inputs, Ørsted built a condition monitoring model that runs daily to detect abnormal debris or vibration patterns, according to a post on LinkedIn.
The model triggers inspection workflows, captures technician findings and feeds results into a continuous improvement loop.
“By 2025, we’d increased the percentage of generator failures we detect at least three months in advance from close to 0 % prior to the sensor installation to 70 %,” said Neil MacDougall, head of diagnostics & prognostics at Ørsted.
“In other words: we’d managed to go from reactive to predictive maintenance.”
Ørsted said the system has been running for over three years and required standardised data pipelines, structured inspection requests and clear prioritisation workflows to scale.
The company noted that predicting failures in advance enables more accurate forecasting of generator demand, supporting bulk procurement and lowering costs.
Ørsted added that the average predicted outage duration has decreased by 72% compared to unpredicted failures.
The company said the approach demonstrates how incremental innovation combining engineering insight and data analytics can improve reliability and reduce costs across offshore wind fleets.