Identifying the key parameters that influence geogrid reinforcement of railway ballast

Abstract

A series of experiments are described involving the full-scale simulation of geogrid reinforcement for railway ballast, which allowed the key parameters influencing the reduction in vertical settlement (permanent deformation) under repeated loading to be studied. The results demonstrated that grid geometry, stiffness, rib cross-sectional shape and junction strength are all influential. The research data was applied as part of a wide ranging study to improve the effectiveness of ballast reinforcement and understanding of the fundamentals of grid/aggregate interaction.

Introduction

The maintenance costs for conventional ballasted railway track could be significantly reduced if the rate at which differential settlement develops under traffic loading could be decreased. This is because riding quality and safety deteriorate as permanent settlement accumulates. Consequently, practical techniques, which could reduce this, would be of great benefit to the industry. Walls and Galbreath (1987) reported that the periods between maintenance operations could be increased by as much as 12 times by the application of geogrid reinforcement to the ballast.

It was against this background that a major research project was undertaken at Nottingham to investigate the potential for wider and improved use of geogrid reinforcement of ballast to reduce rates of settlement in track. Following railway engineering practice, the term ‘settlement’ is used in this paper rather than ‘permanent deformation’, as used in highway engineering, to denote the irrecoverable vertical movement accumulated at the surface of the structure as a result of repeated loading.

Geogrids have been successfully applied for the reinforcement of compacted aggregate layers both in permanent pavements and for haul roads (Koerner, 1990), as well as in railway track ballast (Selig and Waters, 1994), for several years. Much of the background research has, however, either been empirical or has considered approaches based on large strain development (e.g. Hufenus et al., 2006; Moraci and Gioffrè, 2006), which is inappropriate for a high level railway track. In addition, much of the research has not addressed the fundamentals of interaction between grids and aggregates in an organised way appropriate to the repeated load, low strain conditions that apply. Consequently, the development of practical and reliable design methods and of the correct grid specifications for particular applications has not been as soundly based as is desirable to optimise performance and provide a background for the development of improved track design. It was considered that in addressing the problem of reinforcing railway ballast, which is an essentially single-sized aggregate, some progress could be made in understanding the fundamental mechanics of aggregate/geogrid interactions in a way which would be more difficult for a continuously graded material. In particular, it was considered that it should be possible to optimise the mechanical and geometric properties of the geogrid to maximise the reinforcing effect.

The research formed part of a multi-faceted study of railway ballast reinforcement that is summarised in Fig. 1. A particular feature of the work was that the experiments should be at full-scale, whereas similar studies elsewhere have used model scale tests (Raymond, 2002; Raymond and Ismail, 2003). One of the key objectives was to conduct both theoretical and experimental studies to improve the understanding of how grids interact with the aggregates in which they are placed when subjected to the stress and strain environment resulting from moving wheel loads on the track. The ‘Element Tests’ and ‘Computational Modelling’ aspects shown in Fig. 1 covered these activities. The theoretical and computational work has been reported by McDowell et al. (2006). It involved application of the Discrete Element Method (Cundall and Strack, 1979) for modelling of both the grid and the ballast, together with simple pull-out experiments to validate the results of the computations. This modelling was more detailed than that described by Wilson-Fahmy and Koerner (1993). One of the key findings was that the ratio between grid aperture size and nominal size of the aggregate should be 1.4. Consequently, for 50 mm ballast, the best aperture size should be 70 mm. This paper deals with the Element Tests.

The ‘Full-Scale Tests’ referred to in Fig. 1 involved the development and use of a laboratory full-scale Railway Test Facility, described by Brown et al. (2007), to compare the performance of track sections that were reinforced and unreinforced. The grid that was specified for these experiments emerged from the detailed study using Element Tests, involving a simpler test facility, that is the subject of this paper. The ‘Site Trial’ (Fig. 1) was conducted on a UK main railway line, from which, early performance data, reported by Sharpe et al. (2006), has allowed preliminary validation of the laboratory-based work.