Vision-Based Modal Analysis of a Wind Turbine Tower with Variable Cross Section

The wind turbine tower (WTT) is the structure’s supporting pole and the pylon accounts for 15–30% of the total cost. As WTT is huge in construction, traditional contact measurements such as acceleration and strain -requires a large number of sensors, which is costly and complex for applying such sensing system to monitoring such structures. With the development of computer vision and high-speed camera technologies, more attention has been paid to the visual measurement of vibration. In order to solve the vibration model of wind turbine tower under the coupling of ice and snow accumulation and variable wind loads, this paper aims to evaluate the vision-based method by exploring the fundamental techniques involved in its two critical steps. The first is the FEA numerical model as the theoretical basis to analysis the mechanism of different masses on a cantilever beam with variable sections. The second is Gaussian fitting method, aiming at obtaining the vibration process with high accuracy through edge detection. On this basis, FFT transformation is carried out to obtain the natural frequency of the beam. Finally, a small cantilever beam with variable cross section and mass at tip is tested to verify the techniques. The natural frequencies obtained from visual measurement are compared to that of traditional acceleration measurement and theoretical calculation model, proving the techniques of interest is sufficiently good.