Study On Fault Diagnosis Method Of Wind Gear Box Based On On-line Monitoring Data

Under the the oil,coal and other resources shortage caused background of the energy crisis and ecological crisis,because of the resources abundance,sustainability,renewable,no pollution and other advantages,wind energy get fast development,attracts the common attention of society.With the annual growth of wind turbine installed capacity,the importance of maintenance services such as fault detection and maintenance are highlighted.As the wind turbine is installed on the high tower,the maintenance cost will be high once the wind turbine drive system fails.This paper studyed the fault detection methods of roller bearings and gears in fan gearboxes.The main tasks include the following aspects:(1)Author analyzes failure data on the earlier wind power development four countries,compares the collected failure data from wind farm in Zhangjiakou for consecutive two years,analyses the each parts of failure rate,we may discover the laws of potential,get the meaning of the gearbox fault detection,gets the analysis from the gearbox components affect the normal operation of the gear box failure rate.From the above data,the author gets the parts that the author needs to focus on in the gearbox.The existing methods of wind turbine gearbox fault detection are also analyzed,the advantages and disadvantages of each method are analyzed,and the general trend of various methods are prospected(2)Author has carried on the simple introduction about the role in the operation and the structure of the wind turbines and gear box,The author analyzes the cause and the vibration mechanism of the failure forms of gear and rolling bearing,the corresponding calculation formula is given for the corresponding vibration frequency of various fault forms,and the theoretical basis for the fault diagnosis of rolling bearing and gear is given.(3)Based on the wind turbine bearing early faults are faint and difficult to be found in time,author uses the variational mode decomposition method to deal with rolling bearing early fault signal,is expected to be from out of the original signal with low signal noise ratio contains abundant characteristic information of the signal components,in order to further suppress the noise in the signal component composition,realizes the accurate extraction of characteristic frequency,the double spectrum analysis methods as post-processing of VMD method,so as to put forward a kind of based on VMD and double spectrum of wind turbines bearing early faults diagnosis method,the simulation signal and the measured signal analysis results verify the effectiveness of the proposed method.At the same time,in view of the fan at the scene of the rolling bearing of the data,the author proposed the variational mode decomposition and the method of combining the Teager energy operator,a wind farm in zhangjiakou fault detection field data,the reliability of the results proved that the proposed method.(4)Based on the vibration of the gears contained the high frequency noise,the method of parameter choice usually directly given based on the experience of that in most of the research,and other gear detection problems,the author proposes a fusion of the EEMD and PCA fan gear signal noise reduction and performance degradation process evaluation method,whole life EEMD will gear vibration acceleration signal data is decomposed into multiple the IMF,using PCA compare the change trend in the whole life cycle of gear,to select the most can reflect the gear reconstruct signal part of the performance degradation of the IMF component for noise reduction,finally,PCA is used to reconstruct signal processing is put forward in continuous square prediction error signals in the whole life cycle of change trend,realization of data fusion and dimension reduction.