Faster Failure Detection in Wind Turbine Drive-Trains
Machine learning is finding its way into wind energy. It can be beneficial in many aspects of the wind industry value chain, ranging from the planning phase of new farms to operational optimisation during their service life. For the latter it has big potential. A turbine has many sensors that allow detailed monitoring of its operation and this operational data can be used as input for machine learning strategies. By tailoring maintenance strategies to the information coming from anomaly detection based on monitoring algorithms maintenance can be optimised and turbine uptime improved. In particular by using already available SCADA sensor data, optimisation potential can be realised rapidly.
By Prof. Jan Helsen and Ing. Pieter Jan Jordaens, OWI-lab, Belgium
In order to keep increasing offshore wind energy’s market share it is necessary to further reduce the cost of electricity from this source. Figure 1 shows a typical cost breakdown for a wind energy project. Operation and maintenance (O&M) costs are an important cost driver, especially since these costs recur during the complete span of the project. Therefore, reducing O&M-related cost has a direct influence on the total cost of energy.
