Analysis of a typical railway turnout sleeper system using grillage beam analogy

Abstract

A simplified grillage beam analogy was performed to investigate the behaviour of railway turnout sleeper system with a low value of elastic modulus on different support moduli. This study aimed at determining an optimum modulus of elasticity for an emerging technology for railway turnout application - fibre composites sleeper. The finite element simulation suggests that the changes in modulus of elasticity of sleeper, E_sleeper and the sleeper support modulus, U_s have a significant influence on the behaviour of turnout sleepers. The increase in U_s from 10 to 40 MPa resulted in a 15% reduction in the bending moment while the increase in E_sleeper from 1 GPa to 10 GPa has resulted in almost 75% increase in the bending moment. The shear forces in turnout sleepers is not sensitive to both the changes of the E_sleeper and U_s while the sleeper with low E_sleeper tend to undergo greater settlement into the ballast. An E_sleeper of 4 GPa was found optimal for an alternative fibre composite turnout sleeper provided that the U_s is at least 20 MPa from the consideration of sleeper ballast pressure and maximum vertical deflection. It was established that the turnout sleeper has a maximum bending moment of 19 kN-m and a shear force of 158 kN under service conditions.

Introduction

Hardwood has been the preferred material for railway sleepers and maintenance work on existing timber sleeper track is continued to be provided by hardwoods [1]. In recent years, hardwood timber for railway sleepers is becoming more expensive, less available and is of inferior quality compared to the timber previously available. This problem has resulted in most railway industries searching for alternative materials for replacing timber sleepers. A review conducted by Manalo et al. [2] suggested that the advantages of hardwood timber sleeper can be simulated using fibre composite materials with the added advantages. Furthermore, fibre composites could be a more competitive sleeper material in specific application such as railway turnout as it has been increasingly difficult to get larger, longer and good quality hardwood timber. As the design of structures using fibre composite materials has been driven by the stiffness requirement rather than strength [3] and the cost of fibre composites are relatively higher than the traditional materials like timber, steel and concrete, it is important to ascertain the optimum stiffness of a fibre composite alternative suitable for turnout application. Such an investigation is very important to arrive at the best possible sleeper section that will satisfy both strength and serviceability requirements.

Turnout is a part of the railway where track crosses one another at an angle to divert a train from the original track [4]. Special sleepers laid on a turnout are called turnout sleepers [5]. A turnout consists of individual sleepers with varying lengths and fastening locations [6]. Because of the special nature of the turnout sleepers, their manufacturing procedure is different from that of the mainline sleepers which makes their maintenance more costly. The turnout sleepers are also produced with larger dimensions than the mainline sleepers to cope with the complex loadings due to the crossing of the train. It is important therefore to understand how the turnout sleepers respond to these forces to efficiently design an alternative sleeper from fibre composite materials. However, the complex structure of a railway turnout system makes the analysis of the behaviour of the turnout sleepers more complicated than the mainline sleepers.

Several researchers have analysed the railway track as a beam on elastic foundation and their results showed a very good agreement between the theoretical and the experimental results [7]. Kohoutek [8] analysed the railway sleeper as a longitudinal beam resting on an elastic foundation which is loaded by a pair of equivalent static load representing the train. In such a model, the contribution of the rail and the adjacent sleepers is represented by a distribution factor which is applied to the wheel load to determine the equivalent static load. This distribution factor is based on the type of rail gauge and the spacing of the sleepers [9]. The investigation conducted by Shahin [10] concluded that a 3-dimensional finite element analysis rather than a 2-dimensional simulation is a more accurate method to investigate the behaviour of a ballasted railway foundation, but the higher number of elements using this method greatly increased the computational effort. A 2-D beam model which further accounts for variation of subgrade within the length of individual sleeper was developed by Kohoutek and Campbell [11]. This model, which statically analyses the sleeper on elastic foundation, has the possibility to investigate different lengths, different ballast moduli or different parts of the sleepers with different sectional properties.

Shokreih and Rahmat [12] investigated the effects of Young's modulus on the response of railway sleepers as there are many materials being used for railway sleepers. In their work, sleepers were modelled as beams on Winkler's elastic foundation with a constant foundation modulus. The results showed that when the modulus of the beam is higher than that of the foundation, changing Young's modulus of the beam has little effects on the response of the sleepers but has considerable effects for lower modulus. Similarly, Shahu et al. [13] indicated that sleeper support modulus can change dramatically with track construction and this variation can have greater influence on the behaviour of sleepers. Further investigation conducted by Ticoalu [14] showed that using higher support modulus will create smaller rail seat bending moment on the turnout sleepers. These studies have shown that the analysis of beams on elastic foundation has been employed extensively and has been found to be appropriate for analysing railway structures. The results of these studies have also indicated that the bending rigidity and the sleeper support modulus directly influence the behaviour of railway sleepers. However, the finite element analyses of the abovementioned studies are implemented using only a single railway sleeper. The presence of at least two sets of continuous rails which connects the sleepers makes the inclusion of the entire turnout essential in the analysis. For this reason, the behaviour of turnout sleepers should be determined for a group of sleepers instead of a single sleeper, as the contribution of the neighbouring sleepers should be taken into account due to the joining effects of the rails.

In this study, a simple and rational structural model which considers the rail, sleeper, ballast, and subgrade in a railway turnout system is developed. The model also considers the effect of the adjacent sleepers on the behaviour of turnout sleepers through the rails secured to the sleepers. Subsequently, the response of the sleepers due to wheel load of a train passing in a railway turnout is investigated. The behaviour of sleepers with different moduli of elasticity and the influences of changes in the support modulus in the performance of turnout sleepers are analysed. Furthermore, the effect on the behaviour of timber turnout sleepers when one of the sleepers is replaced with a fibre composite sleeper to simulate the spot replacement maintenance strategy is investigated. The result of this parametric investigation could lead to an optimised section for an alternative composite sleeper in a railway turnout.