Effect of multi-parameters interaction on transmission gear rattle based on RBF neural network

This paper proposes a method to analyse multi-parameters interaction on transmission gear rattle. Firstly, a simulation model of manual transmission was established, and the angular velocity of each loose gear as well as the mesh force were obtained. Then the loose gear angular velocity was measured on a manual transmission gearbox to verify the model. The derivative of gear mesh force was taken as the rattle index (jerk index), and was calculated using forward difference method. A radial basis function (RBF) neural network was applied to map the relationships between the selected input parameters and rattle index. The results show that gear backlash has the largest influence on gear rattle, followed by the inertia of the loose gear, the speed fluctuation and the drag torque. This study can be easily extended to other types of transmission systems to control the gear noise and improve sound quality.