Model-based compensation and uncertainty analysis of support bearing losses in a chain drive dynamometer

Abstract Efficiency testing of chain drives has received renewed attention in recent years from sectors including elite cycling, e-mobility and automotive engineering. This is because of a desire to maximise performance, reduce energy consumption and minimise wear rate at a component level. In particular, chain and lubricant manufacturers have developed test rigs that replicate transmission drives in order to measure the efficiency and wear behaviour of the chain and sprocket. This paper explores the impact that compensating parasitic losses from test rigs has on the measurement uncertainty. The analytical process of propagating measurement uncertainty for a dynamometer rig is presented, based on the use of modelled bearing friction losses in the rig and quoted datasheet values for the measurement transducers. Two commonly used bearing models are used to illustrate the process. It is found that the measurement uncertainty is dominated by the transducer uncertainties, particularly speed measurement and repeated experimental results corroborate this trend. However, experimental experience also suggests that the uncertainty derived using GUM Method B is conservative compared to the actual system. The research reported also highlights the importance of validating models of parasitic losses for measurement hardware.