Research On Fault Diagnosis Of Wind Turbine Bearing SVM Based On Permutation Entropy

With the development of the economy,people’s demand for energy is growing,and the excessive use of non-renewable resources such as coal will not only make people face the problem of resource depletion,but also bring serious environmental problems.Therefore,the clean energy represented by wind power has entered everyone’s vision and has been vigorously developed,bringing new solutions to resource problems and environmental problems,and with the installation of a large number of wind turbines,the situation that the unit is prone to failure has also brought challenges to the rapid development of wind power.Bearing is one of the important components of wind turbines,in the face of harsh environments,bearings have a good anti-interference ability is extremely important,when the failure occurs,quickly identify the fault parts,shorten the maintenance time,the stability of the operation of the unit has great significance.In this thesis,based on the fault characteristics of wind turbine bearings,combined with the advantages of support vector machine in fault diagnosis,a fault diagnosis model of wind turbine bearings based on optimized support vector machine is established.The main tasks are as follows:(1)For the vibration signal of the wind turbine bearing,the fault feature extraction is carried out by combining variational modal decomposition and improved multi-scale arrangement entropy.Firstly,the variational modal decomposition method is used for signal decomposition,and the parameters are set by the observation center frequency method,and then the signal is compared and analyzed with the empirical modal decomposition method;finally,the improved multi-scale arrangement entropy method is introduced,and the entropy value calculation is performed on each component of the variational modal decomposition after the parameters are determined by the control variable method,so as to better extract the fault characteristics.(2)The support vector machine has unique advantages in fault diagnosis,but its diagnostic accuracy and calculation speed are greatly affected by parameters.Through the comparative analysis of several common optimization algorithms,the cuckoo algorithm is selected to optimize its parameters,and the adaptive probability and step length are introduced to improve the shortcomings of the cuckoo algorithm,and finally the improved cuckoo algorithm is used to optimize the important parameters in the support vector machine,establish a variety of fault diagnosis models,and conduct simulation experimental performance tests,which verifies the effectiveness and superiority of the model constructed in this paper.