Research On Fault Diagnosis Method For Wind Turbine Doubly Fed Induction Motor Based On Fuzzy Recognition

Doubly-fed induction motor is an important part of the wind turbine to achieve constant frequency during variable wind speed, the normal operation of the doubly-fed induction motor is related to the safety of the wind turbine. The harsh environment and changing operating conditions of wind farm results in frequent generator failure, and wind farms are generally remote location, if the device is damaged, it takes long time to prepare replacement parts. If found early failure of the generator, the specific failure mode diagnosis and timely implementation of the adjustment operation mode and maintenance measures can shorten repair time and lower maintenance costs.Domestic and foreign scholars have done a lot of research work on the condition monitoring and fault diagnosis of double-fed induction motor, and with the in-depth study, some research has been productized and applied to wind farms. Because of insufficient research depth, the function of the products is not perfect, with frequent false alarm, mainly due to the existing system is mainly set threshold alarms, watching data trends, and do not link the monitoring parameters with the operating conditions, and the threshold is set not reasonable. And the results of these diagnostic systems give only a preliminary diagnosis, do not target a specific failure mode, which is bad for the implementation of the follow-up maintenance program.In this paper, on the basis of theoretical research of fuzzy recognition, propose a multiple fuzzy recognition fault diagnosis process of wind turbine doubly-fed induction motor. Under the guidance of the process, study the structural characteristics and working principle of double-fed induction motor to determine the typical failure modes, including rotor unbalance, rotor misalignment, bearing failure, stator windings inter-turn short circuit, rotor windings inter-turn short circuit. Set up the dynamics model of typical failure modes for different fault failure mechanism analysis, do research on multivariate fault symptoms when the failure occurred, combined with fault tree analysis, get the reasons of the failures and the corresponding maintenance, and establish failure knowledge base for each failure.Considering the wind turbine variable operating condition, for vibration signals, study the feature extraction method in the time domain and feature extraction method in frequency domain based on angular domain resampling; For SCADA (Supervisory Control And Data Acquisition) parameters, based on identification of operating conditions and the Gaussian distribution model to extract the fault feature of SCADA parameters, select the temperature parameters and electrical parameters related to the operating parameters to designate operating range, according to a Gaussian distribution characteristics of interval data to delineate the various operational parameters of the threshold range. Considering the multiple fault symptoms of failure modes, combined with fuzzy recognition method to study the integration of multiple fault symptoms to avoid one-sidedness and inaccuracy of single parameter identification to reach an accurate system fault identification purposes. After detecting a fault, adjust operation mode and arrange reasonable maintenance schedule.Based on fault feature extraction methods, fault pattern recognition and fault knowledge, combined with the actual situation, develop the generator condition monitoring and fault diagnosis system, as an module of "wind turbine online condition monitoring and fault diagnosis system ", to apply to the actual site.