Research On Fault Feature Extraction And Identification Method Of Large Scale Wind Turbine Gear Transmission System

In view of the characteristics of the faults in the gear transmission system of large wind turbine,the fault feature extraction method and the fault pattern recognition method were studied.The composition and working mechanism of wind turbine were described,and the gear transmission system structure and the common faults were mainly introduced,then the calculation formula of wind turbine fault characteristics frequency of gear transmission system were given,the characteristics of the vibration signal of the gear vibration and the causes of failure were described briefly.The time domain statistical indexes of original vibration signal that was collected were calculated.According to that the variance of time domain statistical index can indicate the discrete degree of different states,and the index of fault characteristic element in time domain statistical index is pointed out.The frequency domain characteristics under each fault state are analyzed by means of amplitude spectrum and spectrum analysis.By the analysis of time and frequency domain characteristics of the fault gear transmission system for wind turbine vibration signal,it could help us to understand the causes of the fault characteristics.The study provides the basis and guidance for the selection of fault feature extraction.In the light of the advantages and shortcomings of EMD was used to wind turbine gear transmission system fault diagnosis the simulated signal was used to prove that the EEMD method can reduce the modal aliasing effect,and put forward a reasonable method to determine the set of two main parameters in EEMD.The correlation coefficient method was used to select the IMF component,then the energy ratio of screened significant IMF component energy and total energy was calculated,the energy ratio as the fault feature element was used to structure fault feature vector.According to the algorithm and defects of the existing grey correlation,the improved grey similarity correlation algorithm was proposed.The improved grey similarity correlation algorithm was used the classification of fault identification of gear transmission system of wind turbine.The experiment proved its effectiveness via the compare with multi class support vector machine method.The results showed that the accuracy of the gray similarity correlation degree algorithm is better,higher real-time performance.