Fault Diagnosis Of Wind Turbine Gearbox Based On Artificial Bee Colony Algorithm And Least Square Support Vector Machine

Energy crisis has affected people’s lives with the development of economic and thegrowth of population, and wind energy is widely used all over the world because it is themost abundant green energy, whose most favorable form is wind power. Wind power hasgood prospect for development as it is low cost and pollution-free. With the widelyapplication of wind turbines, fault diagnosis problem has gradually attracted people’sattention. In the wind turbine, the gearbox is the part with the highest failure rate, so thispaper studies the method of wind turbine gearbox fault diagnosis.The general fault diagnosis method is difficult to guarantee the performance ofdiagnosis because the fault sample of wind turbine gearbox is fewer. SVM is a type ofartificial intelligence methods for small sample, and is suitable for wind turbine gearboxfault diagnosis. LSSVM is the improvement of SVM, which has been improved both inthe running time and the rate of correct, so this paper selects LSSVM as the theoreticalbasis of fault diagnosis.The parameter of LSSVM has an important effect on LSSVM’s performance, so itshould be optimized in order to avoid the blindness of parameter selection. In this paper,Artificial Bee Colony Algorithm is improved with the reverse learning method andtournament selection strategy and put forward the BTABC, which is used to optimize theparameter of LSSVM. This paper uses BTABC-LSSVM method for wind turbinegearbox fault diagnosis.Finally, this paper gives a fault diagnosis model of wind turbine gearbox based onBTABC-LSSVM, and uses BTABC-LSSVM for the fault diagnosis. The analysis ofcomparative experiments shows that the method used in this paper has the betterrecognition rate in fault diagnosis compared with LSSVM with cross-validationparameter and wavelet neural network method. So it is practical in the field of windturbine gearbox fault diagnosis.