Wind turbine blade surface damage detection by sound signal

The purpose of this study is to propose a method to detect the surface damage on wind turbine blades by using the sound signals. The noise of the blades is generated when the wind turbines are operated. In this work the sound signal of a normal wind turbine is measured first and the time-frequency spectrum is obtained by the short-time Fourier transform method. Then apply the time integration to obtain the marginal frequency spectrum and use the curve-fitting to obtain the noise referenced curve of normal condition. The summations of square error of normal and testing wind turbines are calculated and also the index is calculated. A threshold of the index value is determined based upon the comparisons of normal and damaged wind blades operating noise. This threshold is used to judge whether the wind blades are damaged. The results are verified by the photos of the wind turbine obtained during the repaired process.