A Particle Swarm Optimization Tool for Decoupling Automotive Powertrain Torque Roll Axis 2014-01-1687

A typical powertrain mount design process starts with performing the system calculations to determine optimum mount parameters, viz. position, orientation and stiffness values to meet the desired NVH targets. Therefore, a 6 degrees of freedom lumped parameter system of powertrain and mounts is modelled in Matlab®. The approach is to decouple the torque roll axis mode from the remaining five rigid body modes so that the response to the torque pulses is predominantly ‘oscillations about Torque Roll Axis’. This is achieved by optimizing the above mount parameters within specified constraints so that ‘Rotation about the torque roll axis’ is one of the natural modes of vibration. The tool developed here uses ‘Particle Swarm Optimization(PSO) algorithm’ because of its ease of implementation and better convergence to the solution. The algorithm is programmed in TK solver®. Further, for the given torque input, the harmonic response of the powertrain mounted on optimized mounts, is evaluated for comparison with NVH targets.

This paper presents the design criteria to develop the objective function for optimization of automotive powertrain mounts. A detailed procedure of determining problem constraints from specified NVH targets and system interfaces is explained. The optimization is done for powertrain mounts of a 625cc single cylinder Load carrier Diesel engine and a 1.2L three cylinder passenger vehicle Diesel engine. The results obtained from the developed tool are compared with NVH simulation results of baseline mount configuration defined by the supplier.