Efficient human force transmission tailoring chainrings to a specific cyclist

Abstract

The bicycle chainring is extremely efficient (~98%; Spicer et al. (2001)) although the rider interface is subject to noticeable losses. These losses are partly due to the “one size fits all” nature of the circular chainring. A brief visual study of any group of competitive cyclists will show variations in technique and motion, combined with the inherent inter subject variability with relation to muscle make up and activation, this would suggest the ability to maximise the output from any one rider is likely to be held back in some respect. Creating a chainring specifically tailored to a rider could improve efficiency and increase output for a given level of effort. Analysis and modification of the chainring shape has appeared before in the literature, though the analysis and modification was purely theoretical, and any other chainring shape analysis has used non rider specific non-circular chainrings.

Experimental torque data for varying power and cadence is collected, for a competitive cyclist, using a stationary bicycle fitted with instrumented cranks mounted on a trainer, which enables power output to be controlled at varying cadences. Experimental data from the instrumented cranks shows the torque profiles to be asymmetric in nature. This indicates that calibration of the model using experimental data must be carried out. The data also shows a large dependency on the cadence for the “purity” of the torque profile with smoother profiles being produced closer to the optimum cadence. Motion capture data is also collected; this is used to drive the musculoskeletal model used to predict the muscle activity which is occurring for the given motion. With torque data collected a torque model is built based on local cadence. The chain ring shape is then manipulated so as to affect better muscle efficiency in powering the bike. Manipulation is carried out via the motion capture and torque data provided to the musculoskeletal model. Comparisons are made between these results and both circular and production non-circular chainrings.

Two optimised shapes are presented; one a chainring based on a Fourier series expansion, and the other based on an offset ellipse. These shapes both give a reduction in maximum muscle activity of approximately 15% in the musculoskeletal model. Discussion of the specific normalised muscle forces are given, with limitations and ideas for possible future work being given.

More information

Identifiers

Catalogue record

Export record