Reliability Analysis Of Gear Transmission System Considering Failure Correlation

Gears are important devices for transmission,and their reliability has a great impact on the entire equipment system.Generally,gear tooth fracture and tooth surface failure are their main failure.In the gear transmission process,multiple factors will act on the entire gear equipment system,that is,the function of each failure mode will come from the same random variable,so there must exsit correlation between the various failure modes of the gear.Based on the above analysis,this paper considers the correlation between the modes when performing reliability analysis on gears.The main working of this paper are summarized as follows:(1)Using the ANSYS/LS-DYNA to build parametric model of the gear system for dynamics analysis.To verify the correctness of the model,comparing the maximum contact stress and the maximum bending stress obtained from the shaftless condition with the theoretical calculation value.A shaft model is established to compare the tooth surface contact stress and the root bending stress with no shaft,and analyze the influence of shaft deformation on the gear system.(2)Studies the reliability of gear single failure mode.For the problem that the single-time simulation of the gear system is too long.We improved the selection method and learning stop condition of AK-MCS active learning reliability algorithm,and proposed AK-MCS-K algorithm.The method combines the k-means algorithm to divide the points in the sample set U(x)<2 into k groups in each iteration,and then selects the best sample points according to the min(U(x))principle.The k sample points and the true response value are added to the test sample,and the Kriging model is updated until the relative error of the failure probability satisfies the requirement.The algorithm introduces parallel operation,which reduce the number of iterations and simulations under the premise of ensuring accuracy.(3)By analyzing the drawback of asymmetric system reliability model based on Archimedean Copula(AC)and Hierarchical Archimedean Copula(HAC)'s advantage over AC in describing the correlation between random variables,The method to confirm the structure of HAC from inner to outer is proposed.Kendall rank correlation coefficient matrix is introduced to establish the objective function and constraint of the best fit HAC.The optimization algorithm for seeking the best HAC out is proposed.Combine the failure mode reliability function with the established Copula correlation model to complete the reliability analysis of multiple failure modes of gear system.