Effect of geocell and geogrid inclusion on railway ballast stability

Heavy haul trains cause repeated cyclic loads on railway tracks creating progressive deterioration of the railway substructure resulting in frequent maintenance issues. The situation is exacerbated when the underlying subgrade is of soft clay with a relatively low unconfined compressive strength.

Geocells and geogrids are a potential alternative in providing reinforcement to the ballast and subballast, improving ballast stability and reducing stresses transmitted to the subgrade. In this study, a three-dimensional finite element (FE) model is developed within the explicit FE modelling framework using the FE program ABAQUS. The effects of geocell and geogrid reinforcement are studied over a range of practical conditions. A working train speed of 80 km/h was studied for N = 80,000 loading cycles where the train load varied according to a haversine load function with a frequency of 11 Hz (i.e., 11 cycles per second). Multibody contact was efficiently handled with realistic modelling of the rail, concrete sleepers and reinforcements. The model developed was validated against existing experimental results showing reasonable agreement. Further comparisons were performed with existing FE studies wherever possible. Models with and without reinforcement were studied to quantify the beneficial effect of the reinforcement.

It was shown that both reinforcing elements improved track resiliency by reducing track and subgrade settlement. Geocell reinforcement was found to reduce the track settlement by as much as 73% and subgrade settlement to an extent of 35%. In terms of track and subgrade stresses, the reduction was 71% and 29%, respectively. Under the same conditions, a single layer geogrid placed at the subballast-subgrade interface showed a reduction of track settlement of 45% and subgrade settlement reduction of 34%. Relative comparison with geocells and geogrids showed that geocells performed better than geogrids. While two layers of geogrids performed better than a single layer of geogrid reinforcement, it was less beneficial when compared to a single layer of geocell. This is due to the nature of confinement provided, while geogrids provide a planar reinforcement effect only, geocells provide a more three-dimensional confinement. Hence the increased cost of geocells over geogrids is justified. The lateral ballast spreading tendency is substantially reduced with geocell confinement. In general, with reinforcement the embankment was found to be more stable with less inward and outward movement of the ballast/subballast. Lateral spreading in the long term (at N = 80,000 cycles) was found to reduce by 24.6% at the subballast location and by 49% at the mid-height of the ballast.

The several performance improvements found were a strong function of the material strength properties. The reinforcement beneficial effects were more pronounced for a relatively weaker subgrade and ballast. For instance, a weaker ballast with a friction angle (β) of 45o showed a track settlement reduction of 32%, whereas a more competent ballast with a friction angle of 70o showed an improvement of ~13%. The potential for geocell or geogrid damage under repeated cyclic loadings were found to be low with strains within the reinforcing elements remaining within the elastic limit. The subgrade profile under cyclic wheel loading under various conditions are explored. Parametric studies are reported not limited to geocell stiffness and subgrade strength.