Automated assessment of wind turbine loading and dynamics using SCADA and vibration data

The effects of climate change are widely acknowledged within the scientific community. To limit these, the Paris Accords were signed in 2015. A core prospect of these accords is to limit CO2 emissions, for which it is needed to transition towards renewable energy sources. To enable their market penetration, they must be economically competitive with gray energy sources. Many renewable energy sources, such as wind, have known drastic reductions in their operational costs throughout the last decade. To further reduce the operating costs of wind farms, industry strives towards more optimized designs. These require detailed design insights in the behavior of wind turbines with respect to loading and dynamics. Such opportunities are presented through Industry 4.0, which has already pushed wind manufacturers towards more extensively instrumented turbines. Several challenges nevertheless still remain with respect to automated data processing, which this dissertation tackles.

First off, wind turbines are subject to variable and complex loading conditions, which are highly dependent on the farm location. A proper understanding in these is required to ensure that design lifetime requirements are met. Therefore, this dissertation develops a framework to automatically annotate different load cases based on SCADA data of the wind turbine. Based on this method, occurrence rates are obtained for different load cases. These are then compared with the ones used during the design, allowing to gain insights in the validity of the different design hypotheses.