Research On Fault Diagnosis Of Gearbox In Wind Turbine

At present, the environmental pollution and energy problem has become an increasingly important issue. Wind power has been taken as one of the priority development energy sources. The increase of wind power capacity and complicated system structure will not only cause power outage, but also raise serious accidents when the set is at fault.Since the electrical system fault signal has a great signal and noise ratio (SNR), conventional methods can be well used to diagnose faults. The mechanical system of wind turbine is a complex, nonlinear and time-varying system. Its fault vibration signals are non-stationary signals under the influence of wind. Conventional methods cannot extract the fault feature accurately. The research on advanced fault diagnosis techniques can effectively shorten the time for repairs, save money for maintenance.Sparse signal Decomposition Method Based on Multi-scale Chirplet is a new signal processing method which is proved to be more effective than conventional time-frequency analysis methods. The method of generalized demodulation can transform the non-stationary signals into the stationary signals. After the implementation of generalized demodulation, the signals satisfy the standard of stationary demand in FFT analysis. Based on the equal angle-interval sampling technique, the envelope order analysis is able to analyze vibration signals of the wind turbine.The present thesis, funded by project"Research on Condition Monitoring and Fault Diagnosis Technology of Large Wind Turbine"(Project's Serial Number: 2009AA04Z414) supported by The National High Technology Research and Development Program of China and by project"Sparse Signal Decomposition Based on Multi-scale Chirplet and Its Application to Mechanical Fault Diagnosis"(Project's Serial Number: 50875078) supported by National Natural Science Foundation of China, does some theoretical researches and experimental analysis about Sparse signal Decomposition Method Based on Multi-scale Chirplet.The main research work and its results include:(1) According to the structure of wind turbine and the environmental characteristic of wind field, fault mechanism of wind turbine's gearbox is researched. Based on the feature of wind turbine with variable speed, an experimental system is set up. The analysis shows that the key to the fault diagnosis of the wind turbine's gearbox lies in feature extraction and recognition of the modulated signals.(2) Due to the problem of extracting instantaneous frequency from vibration signals of the wind turbine's gearbox, Sparse Signal Decomposition Based on Multi-scale Chirplet was introduced in the analysis of vibration signals of the wind turbine's gearbox. The principle of Sparse Signal Decomposition Based on Multi-scale Chirplet was briefly introduced in this thesis. The simulation and experimental results prove that the method has strong noise immunity and good time frequency gathering property, it can effectively analyze signals whose instantaneous frequency change continuously. Especially, it is suitable for analyzing vibration signals of the wind turbine's gearbox with variable speed.(3) In signal processing method based on generalized demodulation, it is difficult to extract phase function of multi-component non-stationary signals with time-frequency distribution consisting of curves. Sparse Signal Decomposition Based on Multi-scale Chirplet was introduced in generalized demodulation, which effectively solves the problem mentioned above. Simulation and application examples prove that non-stationary signals can be transferred into stationary signals by carrying out general decomposition using the phase function extracted by Sparse signal Decomposition Method Based on Multi-scale Chirplet. The method is suitable to process non-stationary signals and diagnose gearbox faults with variable speed.(4) Due to the problem of speed extraction and the inadequacies in precision of traditional speed extraction methods based on instantaneous frequency estimation in order tracking, the Sparse Signal Decomposition Based on Multi-scale Chirplet is used to estimate rotational speeds in the thesis, then order tracking is carried out on envelope signals of the wind turbine's gearbox. Simulation and application examples demonstrate that the method can accurately extract rotational speed signals, avoiding the installation of speed measuring equipment in order tracking analysis and saveing costs.The method of Sparse Signal Decomposition Based on Multi-scale Chirplet has manifest effectiveness in decomposing the multi-component non-stationary signals with time-frequency distribution consisting of curves. On the basis of this method, generalized demodulation and envelope order analysis can be used to analyze vibration signals of the wind turbine's gearbox effectively.