The Research Of Fault Diagnosis Method About Parallel Shafts Gearbox In Variable Condition

With the constantly developing of modern industry, Gearbox is playing an increasingly important role in petrochemical, aerospace, power, transportation and other fields. But the failure frequency of gearbox is high due to its structure and working environment, and the economic loss and security threat caused by the faults is relatively serious. Therefore, the research of fault diagnosis about gearbox is significant. As an example, the structure of gearbox in wind turbine is complex, which contains both fixed axis and planetary gear train. The location and operating method of gear and bearing in planetary train is different from fixed axis train, the path of vibration transmission is more complicated, and the fault response is much weaker. All of these increase the difficulty of fault diagnosis about gearbox.To solve problem above, this paper deduces vibration signal model of gear and bearing fault in planetary train, calculates fault characteristic frequency, and analyzes the characteristic of Fourier spectrum and Envelope-spectra about gear and bearing fault firstly. Then with the using of fault diagnosis test bench of wind turbine, Compact RIO and software based on LAB VIEW for data acquisition, the single gear fault and bearing fault in fixed axis train and planetary train, and a variety of compound faults under variable conditions are researched by Fourier spectrum and Envelope-spectra analysis. Then the actual cases analysis about wind turbine in Guangdong and Jilin are mentioned, improving the practical application of vibration research method such as Fourier spectrum and Envelope-spectra analysis for gearbox fault diagnosis in engineering field.During the running of wind turbine, the direction and speed of wind is always changing, the inertial force, elastic force and other alternating loads are also unstable. In this paper, the influences of time-varying speed, torque load and radial load on vibration characteristic identification is researched, it is for providing bases about the choosing of reasonable working condition for vibration feature extraction and the setting of alarm line in condition monitoring and the judging of vibration state.In addition, as a result of improper sensor location in condition monitoring field, the measurement is not accurate, data information is not comprehensive, and machine state analysis will not be correct. To deal with the problem, the method of characteristic value analysis in time and frequency domain is used to determine the relative ideal position among multiple vibration measuring points, it is for providing certain basis about the layout optimization of gearbox vibration measuring point in engineering.