Research And Implementation Of Wind Turbine Gearbox Condition Monitoring And Fault Diagnosis System

Along with the development of economy, global energy crisis andenvironmental crisis have become increasingly prominent. With itsadvantages of low cost, cleanness, safety and renewability, most countriespay attention to the wind turbines. So in recent years, wind power hasobtained rapid development. However, with the large-scale running ofwind turbines, fan faults and sudden accidents are frequent. The normalproduction of electricity is seriously affected. In order to guarantee the safeand reliable operation of wind turbines and reduce the occurrence of faults,the research and development of on-line monitoring and fault diagnosissystem for wind turbines have very important significance.Gear box is a high incidence component of wind turbine faults. Underthis background, this paper mainly studies related problem and effectivemethods of the wind turbine gear box fault diagnosis. The generalalgorithms of fault diagnosis for wind turbine gear box have theshortcomings of low diagnosis accuracy, long operation and large amountof calculation. In order to solve the problems of wind turbine gearbox faultdiagnosis, this paper intends to use intelligent fault diagnosis technologythat the improved quantum genetic optimizes the least squares supportvector machine. Meanwhile, a system that based on virtual instrumenttechnology is designed to complete the implementation of diagnosticmethod and the construction of the monitoring and faults diagnosis.The main research contents of this paper are as follows: Combiningwith the operation principle of wind turbine and characteristic ofmechanical structure, the internal common faults of wind turbinegearboxes and the characteristics of the faults vibration signals are studied. This provides a theoretical basis for the development of the system;Wavelet demising and wavelet packet decomposition method are describedin detail. The common fault characteristics of wind turbine gearbox areintroduced. The time domain and frequency domain fault characteristicswere mainly analyzed. Because of the shortcomings of QGA algorithm,MPQGA algorithm was proposed and it was applied to the parametersoptimization of LSSVM. MPQGA-LSSVM model was set up. UCI datasample showed that the improved algorithm was more efficiency andaccuracy. Finally it was applied to the fault diagnosis of wind turbinegearbox and good results were achieved; Then the overall structural ofwind turbine gearbox condition monitoring and fault diagnosis system wasdesigned, the system hardware was selected, the software developmentplatform based on LabVIEW was determined; Based on the wind turbinesimulation gearbox test rig of laboratory, a wind turbine gearbox system ofcondition monitoring and fault diagnosis was developed. With theapplication of virtual instrument technology, the use of LabVIEWprogramming language, the combination of fault diagnosis methodsdescribed above, and the insertion of MPQGA-LSSVM model in theinterface, the acquisition of gearbox vibration signal, remote transmission,data processing, data storage, offline analysis, fault diagnosis, databasemanagement, and other functions were constructed. Offline analysisincluded time and frequency domain analysis and fault feature extractionmodule, the wavelet noise reduction module, wavelet decomposition andenergy feature extraction module, power spectrum and Hilbert analysismodule and other modules.