An investigation into the performance of railway sleeper types and geogrid-reinforced ballast

Reliability and safety represent key features of any successful railway system and compromising either has undeniable ramifications. Railway industry practitioners are continually challenged to deliver reliability and performance improvements whilst facing ever increasing service demands. These improvements are typically achieved through track maintenance and renewal activities, which have to be balanced against a requirement to reduce whole life-cycle costs. Research focusing on the optimisation of railway track components, which can then be translated to field practice, presents a real opportunity to reduce the frequency of disruptive and often costly track maintenance activities and ultimately prolong the life span of a railway track. A laboratory study of track performance with particular emphasis on railway sleeper type and geogrid type as the main variables has been undertaken. The types of sleepers investigated were the concrete monoblock, twin-block, timber, plastic, and steel sleepers. The geogrids variants tested were the SSLA30 biaxial and TX130 triaxial geogrids with square and triangular apertures respectively. Testing undertaken involved the application of low frequency cyclic loads to railway sleeper sections and full-size sleepers installed on a 300 mm thick ballast with and without geogrid reinforcement. Bending tests, friction tests and hardness tests were initially performed to characterise the material and mechanical properties of the sleepers investigated. Preliminary cyclic tests were conducted with a Box Test apparatus and Composite Element Test (CET) apparatus to approximate field conditions. Full scale tests were subsequently performed with the Nottingham Railway Test Facility (RTF) which is designed to provide a closer representation of field conditions and simulate the passage of an axle load over three sleepers. The outer sleepers in the test facility provided the necessary boundary conditions for the middle sleeper, which was the primary focus of the tests performed. Measures of track performance included vertical track settlement, trackbed stiffness and formation pressure. Additional measurements were made of the differential deflection along the length of the middle sleeper to ascertain if sleeper bending occurred during the tests. Linear elastic and finite element analysis to determine the pressure on top of the subgrade and at the sleeper-ballast interface respectively were performed for idealised sleeper support conditions. The results of the numerical analysis were compared with the RTF pressure plate measurements and estimates of subgrade pressure calculated using empirically derived equations. The results showed that sleeper type influences the permanent settlement that develops in a railway track as well as the magnitude of transient live loads that is transmitted to an underlying subgrade. In line with the permanent settlement results, it is also apparent that trackbed stiffness is a function of sleeper bending stiffness. Measurements of formation pressure and resilient sleeper deflection revealed differences between sleeper types with regards to their ability to retain the as-built geometry of a trackbed, underlining the importance of the sleeper-ballast interface characteristics and sleeper bending stiffness. Traditionally used empirical equations for determining subgrade pressure were found to be conservative compared to subgrade pressures determined using linear elastic analysis and measurements of made of the same using pressure plates in the RTF. Finite element analysis to determine the pressure distributions at the base of different sleepers for a range of support conditions found the shape and magnitude of pressures determined to be consistent with the sleepers’ bending stiffnesses suggesting that sleeper properties should be an important consideration when predicting track performance. The use of the biaxial geogrid installed 100 mm above the base of the ballast reduced permanent settlement for all sleeper types without any significant bias towards any one sleeper type. Additionally, the use of the biaxial geogrid resulted in the delayed deterioration of sleeper support for all sleeper types. The application of the TX130 geogrid resulted in increased settlement and increased deterioration of the as-built trackbed geometry for all sleeper types owing to the grid aperture which proved unsuited to the standard Network Rail ballast gradation. It was proposed that a triaxial geogrid with a larger aperture may offer better results. It was also suggested that sleeper choice that includes consideration for the relative performance of sleeper types is possible for railway practice although it must be commensurate with the intended use of the track with due regard to cost and safety. The research concluded that the concrete monoblock sleeper, which is currently the prevalent sleeper type in the UK (with and without the biaxial geogrid), for the conditions simulated, presents the best opportunity to minimise the maintenance requirements of a railway track