Wind Turbine Blade Fault Prediction Based on EEMD-HHT and CBCNN

Wind turbine blades are one of the most critical components of a wind turbine system. Due to their prolonged operation and exposure to harsh environments, they are prone to faults such as cracks and icing. Therefore, fault diagnosis of wind turbine blades is essential. This paper proposes an algorithm based on EEMD-HHT and convolutional attention mechanisms within a convolutional neural network to predict fault types. First, Ensemble Empirical Mode Decomposition (EEMD) is employed to effectively suppress mode mixing and extract a series of Intrinsic Mode Functions (IMFs), which are then subjected to Hilbert transform to obtain the Hilbert spectrum. Subsequently, a convolutional attention mechanism is utilized to guide the focus on the Hilbert spectrum by introducing attention modules in different layers of the convolutional neural network, allowing the model to automatically adjust its focus on different regions of the Hilbert spectrum according to task requirements. Finally, the Hilbert spectral feature maps, with enhanced sensitivity to important features, are used as input to train the convolutional neural network (CNN) model for fault type classification, achieving fault type prediction.