Font Size: a A A

Research On Dynamic Early Warning And Application Of Wind Turbine Gearbox Based On Multidimensional Monitoring

Posted on:2022-08-12Degree:DoctorType:Dissertation
Country:ChinaCandidate:Y J YangFull Text:PDF
GTID:1482306572972749Subject:Electrical engineering
Abstract/Summary:PDF Full Text Request
The gearbox is one of the key rotating components of the doubly-fed wind turbine.Once the gearbox is abnormal,it may cause the entire wind turbine to shut down.The gearbox failure will not only seriously affect the power generation of the wind turbine,but also will greatly increase the operation and maintenance costs of the wind farm.Wind turbine condition monitoring and fault early warning technology can optimize the maintenance mode and improve the safety and reliability of wind turbine operation.This article conducts condition monitoring and fault early warning research on wind turbine gearboxes.The paper mainly includes the following contents:(1)Aiming at the abnormal temperature monitoring of wind turbine gearboxes,A wind turbine gearbox process monitoring method based on Kernel principal component analysis(KPCA)and T~2 and square prediction error(SPE)statistics is proposed.This method can realize online monitoring and temperature warning of wind turbine gearboxes.Firstly,the principle of similarity between the same multi-dimensional feature points in the wind turbine group is used to construct the correlation coefficient matrix,so as to select a reasonable healthy wind turbine as a reference,and obtain the adaptive statistical control limit under normal operating conditions.Secondly,the moving window is used to dynamically adjust the data set of the KPCA model,which can perceive the time-varying characteristics of the system in time.Experiments show that this method can better realize the gearbox condition monitoring and early warning than the traditional KPCA monitoring model.The multi-dimensional parameters composed of the state variables of the supervisory control and data acquisition(SCADA)and the time domain characteristic indicators of the vibration signal of the condition monitoring system(CMS)can better adapt to changes in working conditions.(2)Based on the real data of the wind farm SCADA,the gearbox oil temperature prediction model based on multi-input improved ant lion optimization and support vector regression(M-IALO-SVR)is proposed.First,the correlation between the gearbox oil temperature and other state parameters of multiple healthy wind turbines is analyzed in different months,and the state parameters related to the gearbox oil temperature are reasonably selected.Secondly,to further analyze the performance of the gearbox oil temperature warning model based on M-IALO-SVR,the residual sequence is processed with a 95%confidence interval,and then the moving window statistical method is used to calculate the change trend of the residual mean and standard deviation.the test results show that when the gearbox is operating normally,the predicted accuracy of the gearbox oil temperature based on M-IALO-SVR is very high.When the gearbox operates abnormally,its temperature deviates from the normal range,and the statistical characteristics of the residuals also change.According to the trend of the residuals statistical characteristics,the abnormal state of the gearbox can be found in time.The feasibility of the wind turbine gearbox oil temperature warning model based on M-IALO-SVR is verified.(3)The non-stationary time-domain signals collected by the CMS system are resampled in the angle domain to obtain a stable angle-domain signal,and then the angle-domain signal is subjected to VMD decomposition and envelope order analysis.Experiments show that:First of all,when VMD decomposes various types of AM and FM simulation signals,the decomposition performance is better than Emprical Mode Decomposition(EMD)and Ensemble Empirical Mode Decomposition(EEMD).Secondly,VMD combined with Hilbert transform can more accurately diagnose gearbox faults,especially gearbox gear compound faults.Finally,the fault characteristics are more obvious in the high-speed operating mode than in the low-speed operating mode,and the fault characteristics are more obvious in amplitude demodulation than in frequency demodulation mode.(4)To realize the preventive intelligent operation and maintenance of the wind farm,a multi-dimensional intelligent monitoring system with the gearbox of the doubly-fed wind turbine as the main monitoring object is established.The system integrates SCADA system,CMS system,and gearbox endoscopic photos,and has functions such as data collection,transmission,processing,status monitoring,fault warning,fault diagnosis and performance evaluation.The above work is an active exploration and practice of the research on key technologies of intelligent operation and maintenance of wind turbines,and provides technical support for further optimization the preventive maintenance strategy of wind turbines.It has practical significance and academic value for improving the operational reliability of wind turbines and reducing the operation and maintenance costs of wind farms.
Keywords/Search Tags:Wind power gearbox, condition monitoring, fault warning, KPCA, SVR
PDF Full Text Request
Related items