Experimental and numerical analysis of automotive gearbox rattle noise

Abstract

The aim of this work is to characterize the rattle noise of automotive gearboxes, resulting from impacts between toothed wheels of unselected gear ratios. These stereo-mechanical impacts are modeled by a coefficient of restitution which describes damping associated with the elastic deformation during the impact and the squeezing of the lubricant film. The dynamic response of loose gear firstly depends on the design parameters and the engine operating conditions. The unknown parameters are the drag torque and the coefficient of restitution. They are identified experimentally through implementation of optical encoders in an actual automotive gearbox and the operation of a specific test bench which replicates the automotive power train. Models of the different drag torque sources are validated from analysis of the free damped response of the drivelines. The coefficient of restitution and its probability density function are measured from experiments under stationary operating conditions. A non linear model is built. The dynamic response of loose gear depends on the dimensionless backlash, the coefficient of restitution and a dimensionless parameter proposed to describe the rattle excitation level. Experiments under controlled excitation are performed to validate the assumptions, to confirm the ability of the parameter proposed to describe the rattle noise threshold, and to characterize the dynamic response. The non linear model predictions are fitted with the drag torque and coefficient of restitution previously identified. They are compared with measurements to demonstrate the ability of the model to predict gear rattle for any loose gear, any gearbox and any operating condition.

Prediction of airborne radiated noise from lightly loaded lubricated meshing gear teeth

“…These have been mostly experimental, often involving determination of coefficients of restitution to describe the effect of lubricant damping through its squeeze film motion [11], as well as any hysteretic elastic deformation of the impacting solid surfaces. Using a torsional vibration model, gear rattle noise was calculated for a 5-speed gearbox, employing the main design parameters and use of various empirical formulae [12].…”

Non-intrusive rattle noise detection in non-stationary conditions by an angle/time cyclostationary approach

“…The relative velocity between the pinions of the 4th unloaded meshing gear is estimated with encoders B and C. A previous work [6] used a similar instrumentation to study automotive rattle noise under stationary operating conditions. In our work the interest is focused on detection in non-stationary conditions.…”

“…Some authors investigate phenomenological approaches [5,6] while others build global models [7,8]. The experimental analysis of rattle noise requires test benches able to reproduce engine acyclism by means of a universal joint for example [9].…”

“…Many studies are focused on idle gear noise (gearbox on neutral) and use electro-dynamic translation exciters [10,11]. Kadmiri et al [6] exploit a dedicated test bench replicating excitations from an automotive powertrain. They use two optical encoders in order to measure the Instantaneous Angular Speed (IAS) of driving and loose gears under stationary operating conditions.…”

Multiobjective Pareto Optimal Design of a Clutch System

“…Some of these studies focused on neutral gear rattle [34][35], some of them focused on backlash effects [15,[36][37]. Barthod et al studied gear rattle phenomenon to understand the rattle threshold and rattle noise evolution in relation with excitations and mechanical gearbox parameters [38].…”