Large Wind Turbine Yaw System Inverter Open Circuit Fault Diagnosis

The yaw system is one of the key actuators for fast and accurate wind alignment and capturing maximum wind energy from wind turbines.The frequent start-stop,long-term alternating thermal stress and extreme temperature differences of the yaw inverter lead to a high failure rate of its power components.As the yaw converter operates at high altitude,the vibration noise is easily covered up by the aerodynamic noise when it fails,and the power device failure will not affect the yaw alignment within a short period of time,in addition,it is not easy to detect the encapsulated power device during regular inspection at high altitude,making the yaw converter failure has strong concealment characteristics,if not detected early it will cause secondary converter failure and even cause damage to the yaw motor.This paper carries out research on the yaw inverter open circuit fault diagnosis method with the yaw inverter as the research object.The main research work is as follows:(1)The composition structure,working principle and operation process of the yaw system are introduced,and the topology structure and principle of the yaw inverter and the causes of inverter open-circuit fault failure are analysed;a three-closed-loop yaw system simulation model containing current-speed-position is built by applying parameters such as the actual yaw pinion radius and yaw large gear radius,the validity of the simulation model was verified by comparing it with the actual data collected by Supervisory Control And Data Acquisition(SCADA)system of a wind farm.(2)A Gramian Angular Difference Field(GADF)image coding based yaw inverter open-circuit fault diagnosis method is proposed to address the problem of insufficient fault feature extraction capability of the single-dimensional data input Convolutional Neural Network(CNN)structure during model training.Firstly,the current vector phase during open-circuit faults is briefly analysed,and the current vector phase is proposed as the fault feature quantity.Secondly,considering the high and nearly constant three-phase current amplitude and frequency at the early stage of yaw execution,and the fast rotation speed of the current vector trajectory,it is determined that the current data in the vicinity of the yaw speed is collected as the open-circuit fault monitoring data,and then the current vector phase is obtained.Finally,considering the problem that the CNN structure with single-dimensional data input has insufficient fault feature extraction capability during model training,the one-dimensional current vector phase signal is encoded into a two-dimensional image by GADF image,and an effective fault diagnosis model is obtained with fewer fault data training samples.Comparative analysis reveals that the two-dimensional image encoding enhances the open-circuit fault features,contains more adequate fault information,and achieves model accuracy approximating that of the one-dimensional feature volume with fewer training samples;the test set input classification model,the results illustrate that diagnosis of single and two power device open-circuit faults can be effectively achieved;in addition,it is found by the validation set input model generated in a noisy interference environment that the fault diagnosis results are not The proposed method has good anti-interference capability as it is not affected by the yaw angle and mechanical torque variations.(3)The off-line training of CNN requires a large number of fault samples and a long training time,which affects the wide application of the yaw inverter open-circuit fault diagnosis method based on GADF image coding in practice.To solve the above problems,a yaw inverter open-circuit fault diagnosis method based on current vector modulus and phase interval is proposed.Firstly,considering the yaw operation characteristics,the inverter output three-phase current data from the yaw start to the maximum yaw speed is selected,and the current vector modulus value and phase information are obtained by Clark transform;Thirdly,the standard deviation of the current vector modulus value before and after the open-circuit fault is obtained,and a threshold value is set to identify the open-circuit fault;Secondly,based on the current vector phase interval characteristics,a characteristic quantity is constructed to achieve the fault power device;Finally,the effectiveness and immunity of the proposed method is verified by simulation under two cases of noise interference and variable torque,and the feasibility and effectiveness of the proposed method is also verified by RT-LAB semi-physical simulation.The above research can provide technical support for the condition maintenance of the yaw system,and has important engineering application value for improving the safety and reliability of wind turbines and improving the economic efficiency of wind farms.