Abstract Some challenging issues emerge for the health monitoring of composite wind turbine blades under the intrinsic noise of fatigue loading, including damage mode identification and singular signal detection. This work performs health monitoring of a 59.5-m-long composite wind turbine blade under fatigue loads by acoustic emission (AE) technique. First, the original AE waveform is acquired after wave attenuation calibration and sensor array arrangement. Second, a waveform-based feature extraction method is developed based on the wavelet packet decomposition (WPD) to capture the information contained in original AE signals, which covers all features for reconstructed signals in the frequency domain. Without the requirements for signal preprocessing, clustering analysis is conducted for damage mode identification and singular signal detection based on the extracted features. Third, two hyperparameters, including the scatter number and the selection of wavelet basis function, are demonstrated to show no effect on the results, indicating the robustness of the method. This method is proved to be effective and feasible for health condition monitoring of the blade.

中文翻译：

---

基于声发射的复合风电机叶片损伤模式识别与奇异信号检测

摘要 疲劳载荷固有噪声下复合材料风力机叶片的健康监测出现了一些具有挑战性的问题，包括损伤模式识别和奇异信号检测。这项工作通过声发射 (AE) 技术在疲劳载荷下对 59.5 米长的复合风力涡轮机叶片进行健康监测。首先，经过波衰减校准和传感器阵列排列后获得原始AE波形。其次，基于小波包分解 (WPD) 开发了一种基于波形的特征提取方法，以捕获原始 AE 信号中包含的信息，涵盖频域中重构信号的所有特征。没有信号预处理的要求，聚类分析基于提取的特征进行损伤模式识别和奇异信号检测。第三，两个超参数，包括散点数和小波基函数的选择，被证明对结果没有影响，表明该方法的鲁棒性。该方法被证明对叶片的健康状况监测是有效和可行的。