Micro-mechanical contact behaviour for railway ballast

Discrete or distinct element methods (DEM) are now using realistic particle shapes but consequently the correct contact mechanics are also now required. There is some existing work for the contact behaviour for sands but only a little on crushed rock materials (ballast and rockfill). Therefore, an inter-particle loading apparatus was designed and constructed to test ballast particles. It has a 3-axis control of forces and displacements, and is able to measure the displacements accurately with capacitive proximity transducers, and its high accuracy allows to resolve stiffness precisely. The stiffness of the apparatus was the key issue to make sure the compliance corrections were not large so careful calibrations and modifications were carried out throughout the development. The tested material in this research was granite, which is a typical railway ballast in the UK. The coefficient of interparticle friction (µ) was measured along with its evolution with cyclic loading and typically tended to stabilise at about 0.6-0.85 for dry samples. Continued sliding generated a large amount of powder, of which the mean size was 25µm. However, the roughness of the interface reduced. The µ values for the linear shearing before cyclic loading were lower than those for the first cycle (about 0.25-0.75). Water immersion has a clear effect on µ values but not the load level. Loading fresh contacts or loading old contacts made no difference to the µ values and there was no rate of sliding effect. Both normal and tangential load cycling gave hysteresis loops and the loop size increased with the load. The normal and tangential contact stiffnesses are highly non-linear and much softer than Hertz-Mindlin elastic solutions. Normal loading could be fitted with a Hertz model accounting for surface roughness but only using the whole particle size not the measured contact radius. However, these conclusions are only validated for granite and it would be interesting to see whether they are repeated for other ballast materials.