Application of Virtual Sensing to Offshore Wind Energy Converters with Jacket Substructures
1, Jan Häfele2, Alexandros Iliopoulos1, Wout Weijtjens1, Raimund Rolfes2
1Vrije Universiteit, Brussel, Belgium, 2Leibniz Universität, Hannover, Germany
Structural fatigue is supposed to be the main design driver for offshore wind turbines (OWT). Aiming at jacket substructures for OWT's, steel lattice constructions, fatigue critical welds can be found due to combined dynamic loading from wind, waves and power supply. Inspections are costly because of poor accessibility of the mostly under water located jacket welds. Moreover, equipping every weld with sensors is impracticable because of the vast number of welds. Therefore, Virtual Sensing estimates jacket fatigue using the accelerometers on the tower. This concept predicts current fatigue critical welds based on pointwise reproducing fatigue history and is already successfully applied to the welds of offshore wind turbines on monopiles substructures. The spatial estimation is performed by real-time tower measurements and extrapolation to any desired jacket location. Gathered information about fatigue-critical locations leads to a better assumption of structural health and supports maintenance measures. For example can the need of corrosion painting or inspections be backed or rejected. Overall virtual sensing will allow planning inspections more case-based, which is seen as an increase of efficiency and decrease of risk.
Structural fatigue is linked to the stress histories of welds. In order to perform a stress prediction for any location of a jacket, virtual sensing relies on structural modes and a set of sensors. Structural modes can be obtained from a computer model (FEM) or extracted from previous measurements, if available. Sensors for real-time measurements are only mounted on the tower to ease sensor maintenance and lower risk of failure. Together with the better dynamic properties of accelerometers, this results in estimating fatigue progress at the jacket from tower acceleration signals. The physical background of this method is built by Modal Decomposition and Expansion. MDE connects structural modes with measuring sensors in order to generate virtual sensors, which estimate fatigue history on a given joint.
The successful transfer of virtual sensing from monopiles to jackets within a proof of concept using FAST software (Aero-Servo-Hydro-Elastic simulation framework) showed already strengths and weaknesses of the method. All simulations were conducted considering a realistic variety of operational conditions, including different wind conditions and wind directions. Complementing the simulations, a sensitivity analysis for the different nodes of the substructure was performed. Results showed good matches for fatigue contributing frequencies as well as an accurate spatial estimation. Variation of wind direction and sensors give important references about geometric effects.
Virtual sensing is a method complementing maintenance of offshore wind turbines with jacket substructures by improving the prediction quality of structural health by means of real-time measurements. The method resolves concerns as sensor failure and environmental interactions to provide a robust fatigue estimation approach. It provides information about current fatigue progress on the jacket pointwise and without limitation to measured areas. The generated information can be used to evaluate the need of inspections and remaining lifetime, which is intended to help improving maintenance and cost planning.
• Basic concept of virtual sensing as a tool to reproduce the fatigue history on all locations of the jacket,
• Comments on challenges and important considerations for fatigue measurements and estimation in terms of jackets,
• Influence of different operational conditions and sensor setup on fatigue estimation accuracy,
• The role of virtual sensing in the entire work flow of fatigue monitoring.