Railway Substructure System Based on Asphalt

Over the past few decades, the rail industry has been aiming to provide an infrastructure capable of accommodating faster, heavier and more frequent trains along with lower tolerance for maintenance-related delays and no compromise on travel safety. Under these conditions, conventional ballasted tracks are approaching their performance limits – mainly due to ballast breakage and fouling that undermine track stability and require ever more frequent maintenance. This state of affairs has led to the development of slab stack technology – a ballastless track infrastructure based on Portland cement concrete. The slab track solution was able to increase track stability, reduce maintenance frequency, and essentially provide for all above-mentioned requirements. However, the initial construction costs are very high, and maintenance activities (when necessary) are both expensive and time-consuming. Slab track technology is also known to amplify noise levels.
In this context, the current study focused on another type of ballastless track technology - based on asphalt concrete. While asphalt concrete is widely used for the construction of various transport infrastructures, its use within the rail industry has been very limited with only few reported implementations of ballastless asphalt tracks - mainly in tunnels. Although the solution seems workable and promising, data from the existing tracks are not available (at least not publically) and related scientific literature on the idea is very limited. Therefore, the overall aim of this work was to contribute to the understanding of the behaviour of ballastless asphalt tracks – with emphasis on mechanical responses to train-like loading. This goal was pursued by combining a full-scale experimental investigation with numerical and analytical model developments.
Experimental investigation was carried out by constructing and mechanically interrogating a full-scale mockup. Part of the mockup was built inside a steel container; it was further constructed, instrumented, and tested in a laboratory environment. It consisted of three concrete sleepers resting on asphalt concrete layer underlain by an unbound granular layer, followed by a mat to represent subbase and subgrade. The mechanical interrogation was carried out under different types of vertical loads such as ramp, pulse, and sinusoidal. Moreover, passages of a train with maximum axle loads of 200 kN and speeds of up to 200 km/h were simulated by sequentially loading the three sleepers. Different types of mechanical response were measured in the mockup, such as vertical stresses, horizontal strains, vertical accelerations, and relative displacements between components. It was observed that the ballastless asphalt track exhibited time-dependent and non-linear behaviour. In general, the measured responses within the asphalt track were of very low magnitude, indicating little or no damage under the applied loads.
A three-dimensional finite element model of the mockup was developed, including all major track components and trackbed layers. The non-linearity of the unbound granular layer was considered through stress-dependent elasticity. The time-temperature dependency of asphalt layer was incorporated using linear viscoelasticity theory assuming thermorheological simplicity. The properties of these two layers were obtained from separate laboratory element tests. Implicit dynamic analysis was carried out under simulated train passages, and calculated model responses were validated by the ability to predict response measurements in mockup.
While the experimental investigation and the numerical modelling focused on analyzing responses due to vertical loads, attention was also given in this study to the evaluation of track responses due to longitudinal loads. This was deemed essential - considering that sleepers in ballastless asphalt tracks rest on the top of the pavement, often without any physical anchorage or side support (unlike traditional ballasted tracks where sleepers have horizontal support from crib and shoulder ballast). In this context, a two-dimensional model was developed within an analytical framework to analyze track responses under longitudinal loads induced by train braking. All equations were purposefully developed in closed-form to serve as an easily implementable first-order engineering tool. The model clearly demonstrated that if longitudinal train loads were only intercepted by means of slipper-asphalt friction then heavier sleeper types would be needed to ensure longitudinal track stability under train braking events.
Though this study focused primarily on analyzing mechanical responses in ballastless asphalt tracks, some initial research effort was expended on developing an approach that can be useful for design. Specifically, work was done on a priori estimation of track modulus - a basic engineering parameter needed in new track design, as well as in condition evaluation of existing tracks. The concept of track modulus is closely associated with the idealization, commonly employed for railway tracks, of rails as infinite beams resting on a continuous spring support. A semi-analytical method was suggested for calculating track modulus based on elasticity solutions, and a parametric investigation was carried out to demonstrate the sensitivity to different input parameters. By means of the elastic-viscoelastic correspondence principle this work can be extended to address time-dependent track support. Thus, the track modulus concept can be extended to ballastless asphalt tracks - further promoting industry acceptance of the idea.