Design and validation of a methodology for wind energy structures health monitoring

The objective of Structural Health Monitoring (SHM) is the verification of the state or the health of the structures in order to ensure their proper performance and save on maintenance costs. The SHM system combines a sensor network attached to the structure with continuous monitoring and specific, proprietary algorithms. Different benefits are derived from the implementation of SHM, some of them are: knowledge about the behavior of the structure under different loads and different environmental changes, knowledge of the current state in order to verify the integrity of the structure and determine whether a structure can work properly or whether it needs to be maintained or replaced and, therefore, reduce maintenance costs. The paradigm of damage detection can be tackled as a pattern recognition problem (comparison between the data collected from the structure without damages and the current structure in order to determine if there are any changes). There are lots of techniques that can handle the problem. In this work, accelerometer data is used to develop statistical data driven approaches for the detection of damages in structures. As the methodology is designed for wind turbines, only the output data is used to detect damage; the excitation of the wind turbine is provided by the wind itself or by the sea waves, being those unknown and unpredictable. The damage detection strategy is not only based on the comparison of many data. A complete methodology for damage detection based on pattern recognition has been designed for this work. It handles structural data, selects the proper data for detecting damage and besides, considers the Environmental and Operational Conditions (EOC) in which the structure is operating. The damage detection methodology should always be accessed only if there is a way to probe that the sensors are correctly working. For this reason, it is very important to have a methodology that checks whether the sensors are healthy. In this work a method to detect the damaged sensors has been also implemented and embedded into the damage detection methodology.