Research On Fault Diagnosis Of Wind Turbine Gearbox Based On Acoustic Signal

With the continuous increase of the installed capacity of wind turbines,how to detect the fault of wind turbines becomes more and more important.Gearbox which is an important component of a wind turbine usually operates in harmful conditions.As the result,the probability of a failure in a gearbox is relatively high.In recent years,acoustics has become a promising diagnostic method.Due to the characteristics of non-contact measurement and early fault sensitivity,acoustic signals have attracted wide attention in the field of mechanical fault diagnosis.In this thesis,based on the analysis of the main fault types of wind turbine gearboxes and their failure mechanisms,the corresponding diagnosis methods are identified.First of all,the principle of wavelet packet transform and Hilbert transform is analyzed,and the corresponding calculation program is compiled by using Lab VIEW and MATLAB software.Then,a transmission and acquisition system of acoustic signals acquiring from wind turbines as well as an experimental device for gear fault simulations are set up.Three main fault modes of gear faults have been simulated including chipped tooth 20%,chipped tooth 40%and tooth surface pitting corrosion.Acoustic signals from the different fault modes are respectively obtained and collected under fault conditions.Based on this acoustic signals,the typical fault features of acoustic signals are extracted by using a combination of wavelet packet transform and Hilbert transform.By comparing the analyzing results of the acoustic signals of the normal gearbox and the faulty gearbox,it is possible to accurately identify the fault characteristics of the three modes of faults.The analysis results show that the combination of wavelet packet transform method and the Hilbert transform method can be used to detect the acoustic signals for defining early the fault of gearboxes.