Research And Application Of Wind Turbine Generator Condition Monitoring And Fault Diagnosis Method

As an important energy equipment,wind turbines are widely used in the world.But the high cost of equipment maintenance seriously affects the income of wind power enterprises.The application of condition monitoring and fault diagnosis system can effectively reduce the maintenance cost of equipment.However,for wind power,a new type of equipment manufacturing industry,the state monitoring theory is not yet mature,and the function of the state monitoring system is not perfect.Combining with the structure and operation characteristics of wind turbine,studying its fault mechanism,establishing fault vibration signal model,formulating appropriate condition monitoring strategy and selecting effective fault feature extraction method are the core issues to realize effective condition monitoring and fault diagnosis of wind turbine.Firstly,the working principle and basic structure of wind turbine are studied,focusing on the structure of gearbox of wind turbine.Typical faults of gears and bearings of wind turbine are summarized.According to the structure and operation characteristics of wind turbine,appropriate measuring points and sensor types are selected;fault characteristics,fault alarm level and state monitoring strategy are designed.Secondly,the fault mechanism of gears and rolling bearings is analyzed.Characteristic frequencies of gears and bearings are calculated.Spectrum characteristics of different faults of rolling bearings are studied through the analysis of fault simulation experimental data.Vibration signal models of local faults of solar,planetary and inner gear rings are established for planetary gear systems.The effectiveness of the local fault vibration signal model is verified by analyzing the data of fault simulation experiments.Thirdly,the application of envelope analysis in fault diagnosis of rolling bearings is studied.Aiming at the difficulty of setting envelope analysis band-pass filter,spectral kurtosis is used to determine the center frequency and bandwidth of band-pass filter.A fast spectral kurtosis method is proposed.The validity and practicability of spectral kurtosis method in rolling bearing fault diagnosis are verified by experimental data analysis.Finally,three examples of fault diagnosis for wind turbines are presented.