Research On Method Of Vibration Trend Feature Extraction For Wind Turbine Transmission System

Becauce of the distribution of wind energy resource,wind turbine operation in the most remote places,poor working conditions and open environment.comparing to the general large rotating machinery,wind turbine is more prone to failure but not easy to repair.Especially,the transmission system components,such as main shaft and the gearbox,will shut down for a long time under variable loads,it's difficult to repaire and lead to high maintenance costs.Therefore,The condition monitoring of wind turbines is of great significance for preventing and reducing the faults of wind turbines,carrying out maintenance work in time and improving the generation efficiency of wind turbines.But for now,the trend analysis index of state monitoring still exist many problems.For example,the bearing vibration signal is weak,the trend of the feature extraction difficulty,trend analysis index fluctuates greatly under the influence of variable speed and there is lack of the component level monitoring indicators.According to the actual demand,this article takes the wind turbine transmission system as the monitoring object,and vibration analysis as state monitoring method,focuses on the extraction method of vibration characteristics of the key parts of wind turbine,Especially the gears and bearings of wind turbine transmission system.The main tasks of this article include:The vibration mechanism and vibration trend feature extraction strategy of wind turbine drive system is proposed.This article analysis the internal structure of a typical wind turbine transmission system,the main failure mechanism of the gear and bearing are described.According to the different fault characteristics of different components,the monitoring strategy of the key components of the wind turbine transmission system is proposed,including the extraction strategy of trend characteristics of different components and implementation strategy of the wind turbine transmission system condition monitoring system.A method of bearing signal separation and feature extraction for wind turbine drive system is proposed,which is the basis of bearing vibration trend feature extraction.For the reason of strong noise environment,nonlinear disturbance and high amplitude harmonic component of gear,the bearing'amplitude is too weak to be directly extracted from the spectrum.Because of the different fault mechanism and signal characteristics between gears and bearings,article using Resonance-based Sparse Signal Decompositio method to separates the bearing fault signal from the sensor data and then extraction the fault feature by envelope spectrum.Aiming at the problem that the frequency domain components of frequency spectrum overlap with each other in variable speed,it is difficult to extract the trend feature in frequency domain accurately.A method of frequency domain trend feature extraction based on order resampling is proposed.The wind turbine transmission system speed fluctuates greatly by the external wind speed.The order resampling is incorporated into different components of the trend feature index extraction process,weakening the influence of the speed change on the spectral analysis,and realizing the speed independent extraction of the trend feature index in the frequency domain.The mechanism of vibration of different key components are different.In order to extract the frequency trend characteristic index of different parts,it's extracts the edge band energy factor,sideband energy ratio as gear vibration trend characteristic index,and resonance sparse decomposition are combined to order envelope spectrum to extract bearing vibration trend characteristic index,thus has been implemented to the wind turbine transmission system state monitoring.It has developed a wind turbine transmission system state monitoring system and the trend of vibration characteristic index is integrated into the system for realizing the trend monitoring of large amounts of data,multi unit wind turbine transmission system of main components.