The Fault Diagnosis Research Of Wind Turbine

As a lot of wind turbines put into operation,the problems of fault diagnosis are taken more and more seriously. Bearing is an important component of the wind turbine,which may affect the performance of the whole machine direct. Therefore,the detection and diagnosis of bearing failure is very important. In this article,the bearing are taken as the research object.Taking a wind power operating company on-line monitoring and fault diagnosis of fan vibration signal as the background, the fan vibration signal used in the article is from the wind power operating company.Firstly,This article chooses the wavelet filtering to reduce the noise of original signal.Secondly,three methods are applied to extract the features:the method of fourteen of time characteristics and frequency characteristics,the method of wavelet packet energy,the method of intrinsic modal energy method. Finally,the characteristics are classified by the support vector machine(SVM).The main parameters affecting the performance of SVM are penalizing factor C and kernel function ?,parameters of support vector machine were optimized by particle swarm optimization algorithm.but the shortcoming of which is easy to fall into local optimum in the process of the optimal parameters. Therefore,in this paper,the particle swarm optimization(PSO) algorithm is improved,and the algorithm of particle swarm optimization is put forward based on simulated annealing(SAPSO).Labview calling the component implements generated by the M file.Thus,we can use the toolkit apart from the Matlab environment. Developing the model of fault diagnosis-SAPSO-SVM based on Labview,to classify the three types of rolling bearing vibration signals : the normal,he failure of outer ring,the failure of inner ring.Characteristic parameters are extracted by the above three methods to train the SVM,after that the three types of bearings are identified by SVM.According to the three kinds of results,the method of intrinsic energy is more effective to extract the fault feature. And the result based on the model of SAPSO-SVM and the result based on the model of PSO-SVM are compared,the conclusion of which is that the effect of the SAPSO-SVM is better.