This work addresses the contribution of the wavelength composition of the spectrum of the rail support stiffness profile to the expected long-term settlement. To that aim, purely harmonic stiffness variations of different wavelength are studied. The frequency-domain model with a double periodicity level previously developed by the first and last authors is adopted to embed the stiffness profile in one of the periodicity layers. Additional resonance velocities at which the resonance frequency of the track system coincides with the support-passing frequency or its multiples are found. The susceptibility to degradation is assessed both by quantifying the mechanical energy dissipated in the substructure under a moving train axle within one wavelength of the support stiffness variation, and the work performed by the wheel-rail contact force. It is shown that shorter wavelengths and larger standard deviations of varying ballast/subgrade stiffness result in an increasing energy dissipation in the substructure, and increase the work performed by the wheel-rail contact force, therefore leading to a reduced lifetime of the track. The energetic quantities increase for lower mean values of the stiffness profile, confirming the proneness of tracks on soft soils to degradation. The influence of varying stiffness vanishes for wavelengths of approximately 16 times the sleeper span, which is equivalent to a track length of about 10 m. High railpad stiffness values result in increased energy dissipation but the influence is limited. In general, an increasing train velocity amplifies the rate of track degradation, with no stabilizing trend in the high-speed regime (300 km/h).

Systematic experimental investigations were conducted to study the microstructures and impact toughness of each heat affected zone (HAZ) formed during rail flash-butt welding. A high-strength carbide-free bainitic rail steel was subjected to different thermal simulation cycles to separately reproduce each HAZ subzone by tailoring the peak temperature (PT) with respect to 700, 850, 920, 1000 and 1350 °C, and hence to generate the corresponding microstructures by using Gleeble-3500 simulator. Results show that the HAZ subzones exhibit complicated microstructures depending on the PTs, and with increasing PT the dominant bainitic microstructure type evolves from polygonal bainitic ferrite (700 °C) to a mixture of fine bainitic ferrite and granular bainite (850–1000 °C), and finally to coarse bainitic ferrite and granular bainite (1350 °C). Impact tests demonstrate that the impact toughness initially increases significantly as the PT reaches 920 °C (i.e., fine-grained HAZ), beyond which the impact toughness starts to decrease. The fine-grained HAZ displays optimal impact toughness in HAZs, yet which is lower than the base metal. Moreover, the morphology and distribution of martensite-austenite (M-A) constituents is strongly dependent on the welding PT, and the high fraction blocky and coarse slender M-A constituents is considered to be detrimental for the impact toughness.

The vertical vibration and transmission characteristics of ballast are key factors that affect the dynamic stability of railway track structures and control the settlement of ballasted beds. Therefore, the following study was conducted to explore this topic. Firstly, through an impact hammer test on a ballast sensor with embedding chip, the vertical vibration data of the ballast was accurately measured. Therefore, the vertical vibration characteristics of a single ballast can be studied. Then, the vertical vibration characteristics at different positions in the stack were obtained by embedding ballast sensors into a ballasted stack. Finally, combined with field tests, a discrete element numerical model was established, then the vibration transmission speed and diffusion angle in a ballasted stack were calculated. The results of this study show that the damping ratio of ballast particles is less than 0.1, and the natural frequency is above 1000 Hz. The damping ratio and natural frequency of ballasts are greatly affected by their shape. The damping ratio of a ballasted stack is greater than that of ballast particles, and its natural frequency is lower. This indicates that the ballasted stack has the attributes of a soft material. The vertical acceleration transmission rate of ballasts is lower at frequencies below 257.94 Hz. This shows that the vibration suppression ability of the ballasted bed is better in the lower frequency range. As the depth increases, the vertical vibration transmission speed of the ballast gradually decreases, as does the accumulated external force. In the impact hammer test of a ballasted box, the average vertical vibration transmission speed was calculated to be 0.88 mm/μs, and the ballast vibration was transmitted downward at a diffusion angle of 35.32°–54.51° from the direction of gravity.

The continuous homogeneous rail constraint of embedded rail system (ERS) is realized by the encapsulation of rails with the elastic poured compound (EPC) which is a composite material. Previous treatment of EPC as linear elastic material was insufficient in the failure analysis of ERS. In this work, a hyperelastic model is developed to describe the mechanical properties of the EPC with engineering strain up to 150%. Physical tests of uniaxial tension, planar tension and quadruple shear are conducted. A 4-parameter Ogden model is determined by curve fitting and validated with a progressive validation strategy, and then is applied to the failure analysis of ERS. It is found that the material nonlinearity of EPC contributes noticeably to the decrease of the longitudinal stiffness of ERS. The 2^nd debonding is more probably caused by the failure of adhesive at the interface between EPC and rail rather than EPC itself.

Embedded rail system (ERS) is a new type of track structure with many advantages due to its continuous rail support. The rapid development of urban rail transit all over the world renders its application prospect broad. However, the cracks and debonds in ERS present a threat to the traffic safety and a possibility for high maintenance costs. In this work, a longitudinal pushing experiment was designed to explore the damage development process in ERS in order to help structural optimization and performance maintenance. The first order derivative of displacement-longitudinal force curve indicates that the damage process of ERS could be divided into three stages: linear elasticity, damage initiation and damage acceleration stages. The surface deformation of the elastic poured compound (EPC) was analyzed with the particle velocimetry and it is shown that the damage is possibly localized in a small EPC part. Statistics of the absolute displacements of a large number of interrogation areas show that their percentage distribution changes in agreement with the increment of rail displacement, which could be the basis for monitoring of EPC deformation in the breathing zone of continuous welded rail. The analysis of the deformation of EPC from side views, together with the qualitative analysis with finite element method, reveals that the large shear strain of rubber strip and the intense shear strain of EPC at rail foot are the main causes of damage initiation and growth in ERS under longitudinal force.