Based On Wavelet Packet And Higher-order Spectrum Of Wind Turbine Research Of On-line Monitoring Method

In recent years, wind power technology at home and abroad to get increasinglyrapid development, the rapid development of China’s wind energy development, hasbecome the world’s largest wind power market. Due to limitations of wind resourcedistribution, wind farms are located in areas far from the city, it is of greatsignificance to ensure the normal operation of wind power equipment and to avoidmajor economic losses that research and application of wind turbine on-linemonitoring system for real-time observation of each wind turbine components.Wind turbine on-line monitoring system is the multifunctional collection of onlinemonitoring diagnostic analysis system which compose signal acquisition, on-linemonitoring and signal analysis. This paper studies signal analysis, a large number ofvibration signal had been to extract collected which from online monitoring systeminstalled on many1.5MW direct-drive permanent magnet models wind farm. Usingsignal processing methods for analysis, focusing on the use of wavelet packet analysisand higher-order spectrum of the vibration signal processing, in order to accuratelyreflect the condition of information, early warning of abnormal conditions. It canaccurately determine the equipment malfunctions, make scientific maintenanceprogram, and enhance equipment reliability, improve operational efficiency.This paper mainly from the following three aspects of the research carried out:(1) Extract a lot of vibration signal data packets which collected from turbinemain bearings, Select two normal and two fault data, using wavelet analysis methodto compared, it can get the frequency range of the normal operation of the bearing toextract the signal characteristic frequency.(2) Select two normal and two fault data, using higher-order spectrum analysisprinciples by comparing, and it can get the frequency range of the normal operation ofthe bearing to extract the signal characteristic frequency. (3) Wavelet packet combining dual spectral analysis is using to process twonormal and two fault data, and compared to using a single method. Experimentalresults show that it’s to better eliminate the interference of high frequencycomponents, when the use of wavelet packet analysis combined with dual spectrum.And to avoid the lack of de-noising wavelet analysis method, get more good analysisresults.This paper analyzes the application of MATLAB simulation of vibration signals.Experimental results show that using double-spectral method combining waveletpacket for the turbine bearing vibration signal analysis has a good results. It providesa good basis for the online monitoring reflection of further information and earlywarning of abnormal conditions.