The increasing number of installed wind turbines has led to a greater need for monitoring of their subcomponents. In particular, damages on rotor blades should be detected as early as possible, since they can cause long and hence expensive standstill times. In this work, a three-tier structural health monitoring framework is employed on the experimental data of a 34-m rotor blade for damage and ice detection. The structural health monitoring framework includes the functions of data normalization by clustering according to environmental and operational conditions, feature extraction, and hypothesis testing. In order to assess the framework and the methods applied with respect to ice detection, an ice accretion test was performed by gradually adding masses at the blade tip. First, a modal test by means of manual and impulse excitation was performed on the healthy blade and for all steps of the ice test. Subsequently, to induce damage, the blade was cyclically excited in edgewise direction for over 1 million cycles until failure occurred at the trailing edge. Finally, the initial modal test was repeated on the damaged blade. Modal parameters from system identification and further damage features, also called condition parameters, are presented and compared to each other. Results from the modal test show that structural changes due to damage at the trailing edge and added mass can be detected by changes in the condition parameters. Nevertheless, it is shown that some condition parameters exhibit higher sensitivity to damage than natural frequencies. Furthermore, a correlation between the amount of added mass and the changes in natural frequencies and some of the condition parameters is shown. For the analysis of the fatigue test, condition parameters were determined with and without prior data clustering according to the applied damage equivalent load, resulting in two realizations of the structural health monitoring framework. Results from the fatigue test show that the majority of condition parameters have good detection performance regarding structural change due to fatigue cracks and due to damage at the trailing edge for various confidence intervals. Finally, it is shown that the detection performance in the case of data clustering according to applied damage equivalent load is higher than without data clustering. This emphasizes the need of data normalization by clustering according to the environmental and operational conditions.
This article proposes the deployment of adaptive boosting (AdaBoost) for combining damage feature decisions and improving the detection accuracy of structural health monitoring algorithms. In structural health monitoring applications, damage-sensitive features are combined with classifiers to define decision boundaries and provide information about the structural state. Boosting algorithms combine multiple classifiers aiming at the improvement of their performance. In this study, AdaBoost is deployed on the realizations of a modular structural health monitoring framework, which consists of three tiers: data normalization based on environmental and operational conditions; extraction of damage features, also referred to as condition parameters; and hypothesis testing. Each condition parameter–hypothesis testing pair composes a classifier which is used in AdaBoost as a weak classifier. The integration of AdaBoost with the structural health monitoring framework is validated using experimental data of a 3-kW wind turbine located at the Los Alamos National Laboratory and data generated from a mechanical model of the same structure. The AdaBoost classifier is evaluated with respect to the error rate as well as the true positive and false positive rates, which are typically used in receiver operating characteristic curves. The AdaBoost classifier outperforms the framework classifiers in many cases, improving drastically the detection performance. However, it is shown that the boosting performance depends on the relative location of the condition parameter values on the condition parameter space. The overlaps between the condition parameter values to be combined are quantified using the Bhattacharyya coefficient, which provides a metric for assessing the boosting potential. Finally, omitting condition parameter values corresponding to specific environmental and operational conditions from the boosting process is proposed for obtaining optimum boosting results.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.