Optimization of the high-frequency torsional vibration of vehicle driveline systems using genetic algorithms

Vehicle drivelines with manual transmissions are exposed to different dynamic engine torques under driving conditions. Engine torque can dramatically vary with throttle demand from coast to drive condition and, conversely, with throttle release from drive to coast. Abrupt application or release of throttle in slow moving traffic or rapid engagement of the clutch can be followed by an audible response, referred to in industry as the clonk noise. This paper presents a complete dynamic model of a vehicle driveline for the optimization of high-frequency torsional vibration by the distributed-lumped (hybrid) modelling technique (DLMT). The model used is first validated against experimental tests. Parameter sensitivity studies have been carried out using the model to identify the important components affecting clonk. Three key parameters have been chosen from the parameter study. To optimize these key factors, genetic algorithms (GAs) have been used in this multi-parameter optimization problem. The GAs show significant reduction in the driveline noise, vibration and harshness (NVH).