An unified framework for studying gear dynamics through model reduction techniques

Abstract

Contribution of gear, shaft and bearing stiffnesses to the gear dynamics is an important aspect that can affect the gearbox behavior. Nowadays, it is possible to simulate the gear mating with a very accurate numerical approach, but with the drawback of very huge computational effort. Gear designers in first design steps need instruments able to give good description of gear dynamics without spending much time in high performance numerical simulations. They need an easy instrument for making decisions in a short time. In this paper, an unified framework for studying gear dynamics is proposed. The gears and mechanical elements are modeled in a very simple way, reducing as much as possible the number of dofs, but maintaining the highest fidelity and with different complexity levels. This is reached using MoGeSeC master dofs selection and building a reduced parametric LTI model of each element. The gear mating model is the interface and it is handled with different nonlinear Maxwell models. By using block-oriented approach, it is possible to have a library of increasing complexity elements that depict the gear dynamics and share the same platform, so it is possible to increase the model complexity simply changing a block in the simulation framework and to compare the different approaches.