Optimal Design And Reliability Consideration Of Mainshaft Bearing For Wind Turbine

Mainshaft bearing for wind turbine works with restrict working conditions all year round. The forces supported by mainshaft bearing are complicated, change with the wind changed and produce vibration. Mainshaft bearing supports parts of axial loadings, besides radial loadings and is installed at the height of meters. It is inconvenient to lift and replace, and the cost is great high. Thus, good lubrication, sealed, high reliability and life for mainshaft bearing will be required. However, bearing life has a relationship with structural design, whose premise is that defining bearing principal parameters. So, principal parameters are designed in order to improve bearing life and reliability. It can further promote the design level of bearing for wind turbine and shorten the gap between developed countries and our country by studying optimal design and reliability of principal parameters.The rated dynamic loading of bearing 24176CA/W33 is regarded as the objective function based on the failture characters of mainshaft bearings for wind turbine. Optimal values of principal parameters can be attained by using MATLAB optimized toolbox. Then, the structures of inner ring, outer ring and cage are designed on the basis of principal parameters, which make structural design reasonable and meet demands. After that, bearing load distribution, contact stress, distortion etc. are analyzed, which improve bearing life and make it meet operating requirements.Geometric and physical meanings of shape and scale parameters are studied detailedly on the basis of two-parameter weibull distribution theory. Shape and scale parameters of weibull distribution for different sample capacities are calculated by adopting maximum likelihood method, which is one of parameter estimation methods studied in this paper. Then, the same set of fail data for bearing life are analyzed by using maximum likelihood method, bootstrap method and grey bootstrap method respectively. By compared, estimated values of bearing life calculated by bootstrap method is smaller than that of maximum likelihood method and the result calculated by grey bootstrap method is close to maximum likelihood method while failure number and sample capacity have the same number 10 and 9, namely r = n=10and 9, it is smaller than maximum likelihood method when r = n<9. Eventually, the changes in shape of weibull distribution curve and reliability functional curve for the same set of data are analyzed with poor information methods.