Research On Fault Prediction Technology For Key Components Of Wind Turbines

Fault prediction technology can effectively identify the potential fault information and avoid the vicious equipment accident,high maintenance cost and high power loss.It is the key to realize the transformation of post-maintenance to predictive maintenance.Therefore,this paper systematically analyzes the historical operation data,online monitoring data and test data of wind turbines and Extracts the fault characteristic parameters which are hidden in the multi-source data.It launches the research on fault prediction technology for key components of wind turbines.It can provide technical support for intelligent operation and maintenance.Firstly,to improve the quality of the data source,the missing,invalid and distorted data which is called "bad data" needs to be identified and reconstructed.Then,it establishes a improved nonlinear state observation mode in order to launch the fault prediction of the gear box based on the SCADA data.It presents a new method of wind turbine gearbox bearing fault prediction thinking about sample optimization.It launches the signal analysis of fault resonance frequency and proposes a gear box fault prediction method based on HHT.Based on the D-S evidence theory,the two kinds of prediction results will be dealed with information fusion.It will Get more scientific,more realistic prediction results.Secondly,according to the field test,it launches the power performance test and the yaw accuracy test on yaw system.It can Effectively identify the potential failure information of the yaw system and present a new idea to repair the yaw fault and optimize the yaw dead zone.Finally,based on the historical data,it uses the application of multi-attribute comprehensive evaluation and presents a new method for evaluating the state of wind turbines based on grey entropy AHP and TOPSIS method.It can sort the wind turbines to determine the range of wind turbines with poor performance,and which were ragarded as “suspect turbines”,so as to achieve the purpose of fault prediction in the angle of the whole turbines.Compared with the existing methods,the results show that the proposed method in this paper is more accurate,more practical and effective.It can provide reference for the relevant research.