Gear Rattle Noise Prediction from Dynamic Simulation

Gear rattle noise is a common issue in manual gear transmissions and is often difficult to resolve. This paper discusses a methodology involving development of a simulation model for noise prediction and subsequent design of experiments (DOE) analysis to select optimal design parameters to reduce rattle noise. A one-dimensional torsional vibration simulation model for a tractor driveline was developed and was correlated with experimental measurements. This correlated model was used to calculate the torque variation between the gear pairs based on engine excitations. The standard deviation of this mesh torque was used as a metric and was correlated to noise ratings assigned by experts during experimental evaluation. Using this metric as the response variable, a DOE was conducted to determine the contributing factors and their influence on the rattle noise. Optimal design parameters were selected to achieve target value on the rattle metric. Physical prototype was built using these optimized parameters and was validated against noise ratings assigned by experts. It was found to successfully satisfy the subjective rattle criterion.