Research For Fault Diagnosis Method And System Of Wind Turbine Gearbox

With the rapid development of wind power industry in the world,the global installation capacity of wind power generators has been increasingly growing.As the paramount component of a wind turbine transmission train,a gearbox has become the key maintenance object since gearbox fault rate is always the highest.However,wind turbines are often located in remote areas,which makes it very difficult to conduct maintenance and overhaul for wind turbines.Therefore,it is of great engineering significance to design and develop fault diagnosis system for wind turbine gearbox.The main research contents of this dissertation are listed as follows:1.The dissertation has studied basic structure of wind turbine and gearbox,analyzed main failure modes of gears and rolling bearings in gearboxes,summarized modulated characteristics of vibration signals of gears and rolling bearings and calculation methods for fault characteristic frequencies.2.Commonly used fault diagnosis methods,namely time-domain methods,frequency-domain methods and time-frequency analysis methods,have been summarized.Merits and limitations of aforementioned fault diagnosis methods have been analyzed.In addition,aforementioned fault diagnosis methods have been applied to actual fault diagnosis,and their effects have been analyzed,thereby providing theoretical basis for their application in the fault diagnosis system for wind turbine gearboxes.3.The basic theory of Adaptive Sparsest Time-Frequency Analysis(ASTFA)method has been investigated.The comparison result between ASTFA and Empirical Mode Decomposition(EMD)has proven the superiority of ASTFA method.Aimed at the problem of selecting initial phase function for Gaussian-Newton iteration in the ASTFA method,the dissertation has proposed the FA-ASTFA(ASTFA based on Firefly Algorithm,FA-ASTFA)method.The FA-ASFA method has well resolved the possible divergence problem of Gaussian-Newton iteration in the original ASTFA method.Simulation analysis results has manifested the effectiveness of FA-ASTFA method.Furthermore,FA-ASTFA method,has been adopted to process experimental signals of faulty gears,faulty inner race and outer race of rolling bearings.By comparing with ASTFA,FA-ASTFA could extract fault features of rolling bearings more effectively.Besides,FA-ASTFA has been used in life-cycle trend analysis of rolling bearings.Experimental results have indicated that FA-ASTFA could restrain trend fluctuation better,thusly smoothing the trend.4.The dissertation has analyzed requirements of the whole system and summarized necessary fundamental characteristics and functions of the system.Design and development of the system has been carried out in accordance with the order of selection and layout of sensors,design of data acquisition and pre-processing device,development of communication procedures,design of data base,development of client software for fault diagnosis system.Eventually,fault diagnosis of actual fault cases in gearboxes could be conducted by utilizing frequency-domain methods and time-frequency analysis methods in the system,thusly proving that the system has feasibility and practicability.