Data for: Deep learning-based aeroacoustic identification of trailing-edge cracks in wind turbine blades.
Description
This data repository contains the research data for the development of a deep learning framework for trailing-edge crack identification in wind turbine blades. The results are published in the corresponding paper: 10.1088/1742-6596/3224/6/062036 The three different features extracted from the raw aeroacoustic signals are collected in the subfolders: - STFT_64_channel: 64-channel spectra obtained from short-time Fourier transform (STFT), each channel representing a single microphone in a microphone array. - BF_map_data: 2D acoustic field maps obtained from beamforming method, showing the sound pressure level distribution in the airfoil plane. - spec_15_channel: 15-channel spectra obtained by integrating the beamforming maps within a specific region that emphasize the aeroacoustic signals close to the trailing edge.