Prognostics of gear manufacturing errors for planetary gear systems based on power flow theory

Gear manufacturing errors are key parameters in planetary gear trains, which have effects on the load sharing, tooth stress and so on. Accurate estimation of manufacturing errors can help monitoring the conditions of planetary gear system. This study investigates the dynamic response sensitivity to model parameters for a nonlinear single-stage planetary gear set with coupled lateral and torsional motions. Power flow theory is introduced to assess the gear vibration and the parameter sensitivity. The response sensitivity equations are deduced with the direct method (DM). The influence of the rotating speed is considered in the sensitivity analysis. Then, the identifiability of the parameter estimation is investigated based on the sensitivity results. The Gauss-Newton method is applied to estimate the manufacturing errors. Gear meshing is a primary factor in gear vibration, so the sensitivities of its vibration power to the parameters are analysed in this paper. The estimated results are accurate when the collected data contain lower noise signal. The sensitivity and parameter estimation make it possible to provide support for the design and diagnosis of a planetary gear set.