Measuring and Modeling the Mechanical Properties of Bicycle Tires

Abstract

It has been shown that tire deformation can play an important role in the stability and handling of a bicycle. It is expected that an accurate understanding of tire behavior is necessary for correct understanding of rider behavior and that correct understanding of rider behavior is necessary for optimizing bicycle design. That certainly has been the case for motorcycles. Several instances of published bicycle tire stiffness data exist, but they seldom agree with each other, do not all measure the same properties, and often are missing key pieces of test configuration data, such as tire size, rim width, inflation pressure, or vertical load. In this project, three different test devices specific to bicycle tires were developed. Measured bicycle tire stiffnesses are presented for 14 different tires ranging from 22 to 50 mm wide, at inflation pressures from 2 to 11 bar (29 to 160 psi), under vertical loads from 304 to 731 Newtons (68 to 164 lb); for a total of about 120 different parameter combinations. Normalized cornering stiffness was found to vary from below 0.15 to over 0.35, which is ±40% from the average, and normalized camber stiffness varies from below 0.0075 to over 0.015, which is ±33% from the average. Based on numerical simulations, this is more than sufficient to influence bicycle stability and handling. Tires approach the camber stiffness necessary, without slip, for the net ground reaction force to be in the plane of the wheel, obeying the so-called tangent rule, but most tires with most inflation pressures and under most loads presented here fall below that. A numerical model, based on an analysis developed by Rotta for slender toroidal tire cross sections in contact with the ground, was also developed to provide insight into how the tires generated the forces they do and attempt to predict them from simpler measurements. Although actual values generated by the model do not exactly match measured values, the trends in contact patch size and lateral stiffness values generated do correspond well with measured data as parameters vary, such as inflation pressure, vertical load, and rim width.