Dynamic Analysis of an Automotive Power Transfer unit : Towards prediction of TE and housing vibrations

Sammanfattning: This work describes the use of Multi-Body Simulation (MBS) to create a virtual prototype of a geared drive called Power transfer unit (PTU). PTU is a subsystem of the all-wheel drive driveline responsible for transfer of power between front and rear axles in an Automobile. The objective of the developing the prototype is to simulate the dynamic behavior of the PTU. Focus is on predicting the gear transmission error(TE) and gearbox housing vibration level. A Hypoid gear set, bearings, tubular shaft and housing are the major components in the PTU. This work is carried out at GKN Automotive which specializes in development of Automotive All wheel drive systems. When developing such geared systems one important characteristic analyzed is the noise and vibration it generates. And for companies like GKN it is desirable to predict these characteristics as early as possible for two reasons, to avoid late design changes and to speed up the product development cycle. To achieve this, a validated virtual model which is computationally efficient is desired. The methodology followed contains of two facets, development of the MBS model and validation of the developed model with physical testing. An integrated MBS-FEM approach is used, an FE modal reduction technique is used to create flexible components with which a virtual prototype is built and simulated in an MBS tool MSC ADAMS c . Gear contact and bearings are defined using an analytical approach which considers the nonlinear stiffness and damping. A dynamic analysis and system level modal analysis is performed to predict the TE, housing vibrations and PTU modal parameters. Experimental modal analysis and physical testing on test rig are performed to measure the actual values of the above predicted outputs. Parameters like damping, contact stiffness of the model are then tuned to achieve correlation. When comparing test and prediction, close correlation is seen in the TE and for housing vibration a similar trend is observed with some deviations. Predicted TE is heavily dependent on gear contact parameters. On the modal parameter comparison, a correlation of five modes and mode shapes below 2500Hz is seen which shows the validity of the MBS model. Parameter studies are performed to study the effect of bearing damping and preload on housing vibrations and TE. It is observed that an optimum value of preload and damping is essential to avoid unnecessary vibrations. In conclusion, the model with some fine tuning of damping parameters can be used for virtual noise and vibration analysis of the PTU.