Research On Sensor Fault Diagnosis Of Wind Power Generation System

The non-renewable and large-scale use of fossil energy has caused environmental pollution,global warming and other problems,prompting countries all over the world to develop new energy to achieve energy transformation.As a kind of clean and renewable energy,wind energy,with its short construction cycle,low environmental requirements,abundant reserves and high utilization rate,has been developing rapidly in countries all over the world.Wind turbines usually operate in harsh natural environments,such as the Gobi desert,savannah and sea surface.Due to the influence of alternating loads and environmental factors,their failure rate is relatively high.When a wind turbine fails,it is usually necessary to repair the system by shutting down for maintenance,which will increase the operating cost of the system and reduce its economic benefits.The sensor is used to detect and collect the information of the wind turbine,and then transmits the data to the controller.By analyzing and processing these data,the controller generates the input signal of the actuator to ensure the stable operation of the control system.If the sensor is faulty,it will generate a deviation signal,which will cause the controller to generate the wrong actuator input signal,so that the system can not run normally and stably,and may even lead to serious loss or shutdown.Therefore,for the stable operation of wind power system,the sensor fault diagnosis technology is very important.In this paper,the structure of permanent magnet direct drive wind power generation system is introduced,and the operating principle of the whole system and each part is expounded.The mathematical models of key parts of the system are established,including wind turbine,transmission system,permanent magnet synchronous motor and converter.The control strategies of total wind speed control,zero d axis current control and space vector modulation(SVPWM)in permanent magnet direct drive wind power generation system are introduced.The permanent magnet direct drive wind power generation system model is built in MATLAB/Simulink,which provides a foundation for the subsequent permanent magnet direct drive wind power system fault diagnosis simulation analysis.Secondly,taking the current sensor in the converter as the research object,the output signal is analyzed,the fault classification of the sensor signal is described,and the fault of the current sensor is described mathematically.Kendall correlation coefficient was used to evaluate the correlation between sensor faults and system parameters.The outliers in the fault sample data are determined by box diagram method.Fourier transform,principal component analysis and wavelet packet transform are used to denoise the fault sample data.Then,a fault diagnosis scheme based on particle swarm optimization(PSO)to optimize support vector machine(SVM)is proposed,and the principle of feature extraction based on fast Fourier transform(FFT),fast Fourier transform combined with principal component analysis(FFT-PCA)and wavelet packet analysis is described.The SVM fault diagnosis algorithm and PSO optimization algorithm were analyzed,and the SVM and PSO-SVM fault diagnosis models were established.In the Simulink model of permanent magnet direct drive wind power generation,the fault diagnosis simulation analysis of current sensor is carried out by using the above feature extraction method and fault diagnosis model.Finally,the simulation platform based on prototype controller is introduced,and the flow of wind power generation simulation platform is expounded by using fast control prototype,and the simulation verification is carried out.The sensor fault data is collected by the wind power generation experimental simulation platform,and the current sensor fault diagnosis experiment is analyzed by using the above feature extraction method and fault diagnosis model,so as to verify the validity of the current sensor fault diagnosis in this paper.