< Back to previous page

Project

System Level Simulation of Drivetrains by Advanced Gear Contact Techniques

In recent years, the engineering and manufacturing industry has been reshaped by technological forces and the pursuit of sustainable development. The rise of computing power has enhanced the role of numerical modelling and simulation to the extent that they represent an everyday solution for engineering problems. Numerical modelling facilitates design validation in an early stage and enables testing different configurations in the virtual environment. This results in shorter lead times and more cost-effective solutions. At the same time, the increasing environmental awareness demands for more stringent requirements. Lightweight designs and lighter materials help engineers to significantly reduce the ecological impact of modern machines, without compromising functionality
or product performance. Nevertheless, lightweight structures have led to new challenges in developing numerical methods capable of accurately and efficiently describing their dynamic behaviour. An example is given by the evolution of gear transmission designs, where lightweight structures as well as tailored solutions to minimize the emitted noise are being developed. These geometries invalidate the assumption in foundation of the state of the art analytical methods used for modelling the contact dynamics of the meshing gears. Analytical modelling strategies fail to account for the wide range of manufacturing parameters currently used in transmission design. Their mathematical formulation cannot cope with arbitrary gear geometry such as gears with holes in the body or modified tooth flank shapes. More sophisticated methods such as Finite Element (FE) methods or Flexible Multi-Body (FMB) descriptions suffice for the required accuracy but lead to prohibitive computational time especially for dynamic
analyses. The high cost of these methods is mainly related to the large number of degrees of freedom of the underlying models - since finely refined FE meshes are required in the proximity of the time-varying contact location to correctly capture the non-linear contact stresses, as well as the cost of imposing the contact constraints. 

The focus of this dissertation is on providing solutions which allow simulating dynamic gear contact problems for novel (lightweight) designs efficiently and without sacrificing accuracy. Three methods are proposed. Firstly, a method that exploits Model Order Reduction (MOR) strategies for contact problems is developed. The novelty of the method consists in the combination of a semi-analytic contact model with MOR schemes that reduce the underlying linear FE model.

Secondly, the above-mentioned MOR scheme is extended by parametrizing the reduction space. 

Lastly, a numerical-based description of the gear meshing stiffness is developed. This method starts from a series of static non-linear FE simulations reduced by means of the above-mentioned parametric MOR scheme. The static simulations are used for assembling a look-up table to interpolate during the dynamic contact problem.

The results obtained with the different methods have been validated against experimental data, showing an excellent capability of capturing quantitative relevant behaviour such as e.g. static transmission error for different torque levels. Moreover, the capabilities of the proposed methods are compared though the numerical simulation of different gear dynamic contact problems. The discussion of the results allows to identify which method proves to be more suitable for modelling a certain transmission phenomenon. In this last comparison results obtained using Siemens LMS Virtual.Lab Motion have been included, since the novel gear modelling strategy implemented in the software has been co-developed by the author of this manuscript and is partially based on the proposed MOR scheme. In conclusion, the research summarized in this manuscript has not only provided novel numerical modelling strategies for gear contact simulation, but also has paved the way for further developments in the field of lubricated contact problems, suggested at the end of this work as future research tracks.

Date:8 Dec 2014 →  28 Jan 2019
Keywords:contact mechanics, Multibody dynamics, model order reduction
Disciplines:Control systems, robotics and automation, Design theories and methods, Mechatronics and robotics, Computer theory
Project type:PhD project