Research On The Fault Modeling And Monitoring For Variable Speed Constant Frequency Wind Turbines

As an inexhaustible renewable energy source,wind energy is widely used due to its irreplaceable advantages,such as no pollution and waste.Meanwhile,grid-connected wind power system,regarded as one of the most important production forms of renewable energy,possesses the favorable conditions for super-duper scale development.At present,Chinese wind power industries are developing from the form of extensive operation management to that of efficient and scientific cooperation,such that the effective methods for improving the power generation efficiency of a single wind turbine(WT)as well as optimizing the power generation performance of wind farms is the critical issues to be solved.Taking the large一scale variable speed constant frequency(VSCF)WT as the research object,this paper carried out the studies of the principle and characteristic of the faults for WT key components as well as the methods for faults modeling and monitoring by using the data collected with supervisory control and data acquisition(SCADA)system.As a result,the satisfactory power generation performance and its operating stability can be guaranteed.The online monitoring of WTs based on the data mining technology can also be realized.The performance of wind rotor which is one of the key components of wind turbine has great influence on the service life and operational stability of wind power generation system.In order to reduce the maintenance cost and enhance the reliability of fault diagnosis of WTs,the occurrence principles of the problems caused by mass and aerodynamic imbalances were discussed in detail.The models of wind rotor,transmission system,generator and pitch control system were all established while the corresponding simulation models under normal and imbalanced condition were developed via Matlab/Simulink software as well.The SCADA data measured in an actual wind power plant are utilized as simulation input by which the simulation studies of unbalanced faults were realized.Results shown that the power signal can generate shock excitation by the 1st frequency component of the rotor speed under imbalance fault.Additionally,the peak of the excitation would increase along with the enhancing fault severity.The related curves between the fault severity and the amplitude of 1st frequency were obtained via numerical fitting method.Thus,suggestions for troubleshooting were also given.For better completing wind energy prediction,condition monitoring and power control of WTs,a novel method for modeling the power curves of VSCF WTs which was composed of the advanced stochastic gradient boosting regression tree(SGBRT)and the partial mutural information(PMI)was proposed.The PMI algorithm was utilized to calculate the correlations between the external factors and the output power.After that,eight different parameters including wind speed,wind direction,pitch angle,yaw error,impeller speed,tip speed ratio,gearbox oil temperature and generator temperature are selected as the input variables for the modelling approach.Thus,the power curve model with multiple input variables was established by using SGBRT.Finally,comparative research between the power curve models with SGBRT and with typical IEC-12-1 model,differential evolution five-parameter model artificial neural network model and XGBoost were conducted.Simulation results indicated that the power curve obtained via SGBRT possesses the smallest maximum standardized prediction residual(6.56%)as well as the smallest forecasting errors which accounts only 2.09%of the rated power.The effectivenessand superiority of the proposed methods was verified and the proposed approach based on SGBRT can more accurately predict the wind turbine power output characteristics.