Influence of dry friction in the dynamic response of accessory belt drive systems.

Abstract

Dry friction plays an important role in the performance and the dynamic response of flat belt accessory drives. Dry friction in the pulley slip arc allows the pulley to transmit power to its attached accessory. In automotive front end accessory drives (FEAD's), dry friction in the tensioner arm is used to dampen the system's dynamic response to crankshaft pulley excitation. The influence of dry friction is not just limited to these beneficial effects. As a direct result of dry friction, during engine operation stick-slip tensioner arm motions may occur which lead to the phenomena of sub- and super-harmonic accessory pulley motions and system secondary resonances. Furthermore, as the belt slides against the pulley in the slip arc, power losses occur which degrade the system's efficiency. Belt drive designers do not have adequate models to evaluate frictional effects such as belt slip, power loss, and friction dominated belt dynamics. The commonly used (non-analytic) Coulomb Law for dry friction makes models difficult to develop, and to evaluate analytically and numerically. The objectives of this research are to develop a fundamental understanding of frictional effects in belt drive systems, to develop models incorporating dry friction, and to develop efficient solution methods for evaluating these models. First, models of the entire FEAD are developed using a Coulomb law representation of the tensioner dry friction. Employing both numerical integration and an approximate, analytical technique (incremental harmonic balance), stick-slip tensioner motions, sub- and super-harmonic accessory response, and secondary resonances are predicted when the FEAD is excited by crankshaft harmonic motion. Next, a study on dry friction in the belt/pulley interface reveals that a single dimensionless parameter $\Omega$ governs the belt response to harmonic tension fluctuations. Models of the interface are developed for the cases of small and large convection (i.e. belt speed) and are exercised to predict the power loss occurring at the interface.