Increasing the length of wind turbine blades for maximum energy capture leads to larger loads and forces acting on the blades. In particular, alternate bending due to gravity or nonuniform wind profiles leads to increased loads and imminent fatigue. Therefore, blade monitoring in operation is needed to optimise turbine settings and, consequently, to reduce alternate bending. In our approach, an acceleration model was used to analyse periodically occurring deviations from uniform bending. By using hierarchical clustering, significant bending patterns could be extracted and patterns were analysed with regard to reference data. In a simulation of alternate bending effects, various effects were successfully represented by different bending patterns. A real data experiment with accelerometers mounted at the blade tip of turbine blades demonstrated a clear relation between the rotation frequency and the resulting bending patterns. Additionally, the markedness of bending shapes could be used to assess the amount of alternate bending of the blade in both simulations and experiment.s The results demonstrate that model-based bending shapes provide a strong indication for alternate bending and, consequently, can be used to optimise turbine settings.

中文翻译：

---

移动加速度计的技巧：使用3D加速度计和基于模型的弯曲形状来监视风力涡轮机叶片的弯曲。

为了最大程度地捕获能量而增加风力涡轮机叶片的长度会导致更大的载荷和作用在叶片上的力。特别是由于重力或不均匀的风廓线而导致的交替弯曲会导致载荷增加和即将出现疲劳。因此，需要在运行中对叶片进行监测，以优化涡轮设置，从而减少交替弯曲。在我们的方法中，加速度模型用于分析周期性发生的均匀弯曲偏差。通过使用分层聚类，可以提取重要的弯曲模式，并针对参考数据分析模式。在替代弯曲效果的模拟中，通过不同的弯曲模式成功代表了各种效果。使用安装在涡轮机叶片尖端的加速度计进行的真实数据实验表明，旋转频率与产生的弯曲模式之间存在明确的关系。此外，弯曲形状的标记可用于在模拟和实验中评估叶片的交替弯曲量。s结果表明，基于模型的弯曲形状为交替弯曲提供了有力的指示，因此，可用于优化涡轮设置。