Structural Health Monitoring of Offshore Wind Turbine based on Online Data-driven Support Vector Machine

The structural health monitoring (SHM) of the offshore wind turbine based on data-driven is supposed to extract the numeral characteristics and classify the health condition from the data stream acquired from sensors. Traditional classification method like support vector machine (SVM) and clustering method cannot process data stream directly. In this paper, according to the features of data stream, the SHM system is designed to improve the effects of clustering method, and the health condition is classified online by time-domain and frequency-domain SVM classifiers based on data stream. The experiments are performed with the measured data of the vibration detection of the offshore turbine structures to evaluate the system. The experiment results show that the SHM system proposed in this paper can process the online vibration detection data stream and classify the health condition.