Chain drives are used in powertrains for the kinematic coupling of the cam shaft, the ancillary units and the balancing shafts with the crank shaft. Advantages of chain drives are their high load carrying capacity along with increased durability whilst simultaneously being maintenance-free. A crucial issue in the drive is the optimization in regard of friction, further improving efficiency, reducing exhaust emission and abrasive wear.
Modeling friction in drive systems requires precise description of the whole system dynamics. High-frequency oscillations occurring in the chain strands cause numerical problems in the friction computation. As a remedy, regularized friction curves are often used, being however not able to correctly determine all friction configurations and requiring a tradeoff between accuracy and computational efficiency. Another challenge is the sensitivity of the coefficient of friction to many factors among which are kinematic and kinetic quantities, lubricant, material and surface properties.
This contribution presents an approach for multi-body simulation of structure-variant chain drives, including bush, roller and silent chains. It describes the kinematic quantities of the bodies. Special focus is laid on modeling friction at its different points of origin. For the oscillation in the strands, a physically motivated elasto-plastic friction model is applied in the multi-body simulation and proven for its accuracy and computational efficiency. Also presented is an elastohydrodynamic model for computation of the friction coefficient, accounting for the before-mentioned factors.
The presented approach was implemented in an in-house Fortran-based simulation tool. The paper reviews results of the friction simulation pointing out the efficiency of the approach.