Schadensdiagnoseverfahren für die Zustandsüberwachung von Offshore-Windenergieanlagen

Vibration-based damage diagnosis is a promising approach for monitoring different structural and machine components of offshore wind energy plants. The aim of continuous monitoring is to reduce the inspection intervals, which especially offshore, are very time and cost consuming. At the same time an automatic damage diagnosis system provides the operator with necessary and reliable information about the plant's state, which allows the initiation of maintenance actions also if the plant is unreachable due to bad weather condition.
This work is focused on the development of new methods which are included in one monitoring system performing damage diagnosis for different component of offshore wind energy plants, such as: tower, foundation and drive train. The development is based on the extension and combination of different signal analysis approaches in structural dynamics, pattern recognition and information theory.
On this base a structural health monitoring concept dealing with measured data was developed and implemented. This includes structural damage detection algorithms in the presence of changing environmental and operational conditions, model-based damage localization by means of modal data and sensor fault identification. For condition monitoring purposes different approaches for signal analysis of measured time data belonging to the drive train components are extended and implemented. The applied algorithms combine different frequency domain approaches. These are used to detect damages in rolling bearings and planetary gears in an earlier stage compared to the established condition monitoring algorithms.
The developed and applied methods in structural health monitoring are validated by means of simulation models, measured signals from different laboratory structures and measured data from a long time monitoring of the 5MW prototype wind energy plant built by Areva-Multibrid. The condition monitoring methods are validated by means of measured signals from different wind energy plants.