Research On Vibration Signal Of Wind Turbine Tower Based On

With the depletion of traditional energy and environmental problem becoming more and more serious, renewable energy which is low-carbon and clean has heightened its position in global energy strategy. Wind energy has shown well perspective of exploitation and utilization at present and it is mainly used to make electricity. As the load-bearing device of wind power generator, wind power tower directly affects the life of wind power generator.The accidents of wind power generator are frequent in recent years, which have caused serious economic losses of enterprises. However, the health examination of wind power tower is regularly and manually in practical applications and we cannot find malfunction of wind power tower as soon as possible in this way. Therefore, this paper researches the vibration signal of wind power tower that in normal use to extract its vibrating feature, in which way we can provide reference for subsequent checks of wind power tower.Focuses on vibration signal of wind power tower, this paper investigate as following:(1)According to the non-stationary of tower vibration signal, this paper study several time-frequency analysis methods which are commonly used and analyze their limitations, such as short-time Fourier transform, Winger-Ville distribution and wavelet transform;(2)Empirical Mode Decomposition (EMD) is emphasized. EMD is a new time-frequency analysis method, which is suitable to process nonlinear and non-stationery signal. In this method, waves in different scale or trend of original signal are decomposed step by step, in which way the original signal is decomposed to a sum of some intrinsic mode functions which are in different scale features. The characteristics of original signal could be extracted through the research of intrinsic mode functions;(3)Aiming at the noise signals which are produced during collecting vibration signals of wind power tower, this paper put forward a signal de-noising method based on KPCA. Through the phase space reconstruction, this method expands one-dimensional vibration signals of wind power tower as the multi-dimension vector which includes tower vibration signal and noise. And then extracts kernel principal component by KPCA to achieve de-noising signal;(4)Aiming at the problem of end effect in EMD algorithm, this paper describes it specifically by simulation and use mirror extension method to handle the end effect. The simulation results show that mirror extension method can restrain the end effects of EMD effectively. This method is used to pick up characteristics of vibration signal after eliminating noise.This paper studies the vibration signal of wind power tower which is in normal use and eliminates its noise within the method based on KPCA. After that EMD method is used to decompose the de-noising vibration signal. Eventually characteristics are collected and it can be used as a reference for Subsequent checks of wind power tower.