Degree Name

Doctor of Philosophy


Faculty of Engineering


This thesis contains laboratory experimental results of strength, deformation and particle breakage of fresh and recycled ballast under monotonic and cyclic loadings, experimental evaluation of effectiveness of various geosynthetics in stabilising recycled ballast, and a new stress-strain constitutive model for ballast incorporating particle breakage. Ballast degrades progressively under heavy cyclic rail loadings, leading to deterioration of track substructure, rail alignment and train safety. Severely fouled ballast is often removed from the track and replaced by freshly quarried ballast, causing track maintenance very costly. Discarded ballast can be cleaned, sieved and recycled to track foundation. In this study, the shear strength and stiffness of both fresh and recycled ballast were investigated in a series of consolidated drained shearing tests using a large-scale triaxial apparatus. The degree of particle breakage was quantified by sieving the ballast specimens before and after each test and recording the changes in ballast gradation. The stress-strain, shear strength, stiffness and particle breakage results of recycled ballast were compared with fresh ballast. The crushing strength characteristics of fresh and recycled ballast grains were investigated in a series of single grain crushing tests on various particle sizes. A small track section comprising rail, sleeper, ballast, capping and subgrade was simulated in a large prismoidal triaxial apparatus in the laboratory. The settlement, lateral deformation and particle breakage behaviour of fresh and recycled ballast under field-simulated loading and boundary conditions were studied in a series of cyclic loading tests using the prismoidal triaxial rig. Three types of geosynthetics (geogrid, woven-geotextile and geocomposite) were used in this study to stabilise recycled ballast. The cyclic test results of recycled ballast stabilised with geosynthetics were compared with the fresh and recycled ballast without geosynthetics. In order to study the effect of saturation, the cyclic tests were conducted in both dry and wet conditions. Currently, there is a lack of appropriate stress-strain constitutive models for coarse aggregates like ballast, especially under cyclic loading incorporating particle breakage. The new constitutive model developed in this study incorporates the energy consumption due to particle breakage during shearing. A single non-linear function has been formulated to model particle breakage, and incorporated in the plastic flow rule assuming non-associated flow. The model has been developed based on the critical state framework and the concept of bounding surface plasticity. It captures the strainhardening, post-peak strain-softening, dilatancy and cyclic hardening features of ballast behaviour accurately. The model has been examined and verified against the experimental results. Finite element analyses using ABAQUS were also conducted to compare with the analytical model. This study clearly shows that the new constitutive model predicts the stress-strain, volume change and particle breakage of ballast to an acceptable accuracy for both monotonic and cyclic loadings.