The management of track-bridge interaction using monitoring based calibration

Abstract

Continuous welded rail (CWR) has become an attractive solution in modern railway track to improve safety and rider comfort as well as decrease maintenance. CWR on railway bridges also has several advantages, however, additional rail stress can be expected and it is, therefore, important that the track-bridge interaction phenomenon is understood. Considerable longitudinal rail forces and displacements may develop in CWR track on long-span bridges. Any movement of the bridge deck induces a movement of the CWR track and additional rail stress. To reduce longitudinal stress in the rail, the length of the bridge can be reduced to a maximum expansion length that does not result in excessive bridge deck displacements, by altering the static arrangement of the bridge to relocate the “thermal fixed point”. An alternative method to reduce longitudinal rail forces and displacements on long-span bridges is to install a rail expansion joint (REJ). However, this is not an attractive solution as these devices cause local disturbance of the vertical track stiffness and track geometry which requires intensive maintenance.

The study first evaluates bridge temperature as an input parameter to the evaluation of the track-bridge interaction phenomenon. The results of the study confirm that the concept of an effective bridge temperature, supported by literature published by Emerson et al. (1976), is an effective means of defining the uniform bridge deck temperature and calculated effective bridge temperature results correlated well with bridge displacement measurements. Thereafter the study evaluates site data obtained to understand the mechanisms contributing to additional compressive stresses in the rail due to track-bridge interaction. This was compared to permissible levels of rail compressive stress which were taken from established codes of practice. The need for a rail expansion joint on the Majuba Rail Bridge was confirmed by numerical modelling and high compressive rail stresses resulting in lateral buckling can be expected in the absence of an REJ. The vast difference in recommended values in various literature, ranging from 30 m to 200 m, supports the need for further evaluation of track-bridge interaction due to total bridge expansion length. The numerical analysis also confirmed that alternatives to an REJ, such as ballasted track or Zero Longitudinal Restraints (ZLR’s) were ineffective on a bridge expansion length of 314 m. Finally, results from the numerical model demonstrate that additional compressive rail stress can be reduced to permissible values, as per UIC Code 774-3R (2001), by reducing the total bridge expansion length, by altering the static arrangement of the bridge, from 314 m to 120 m which compares well with results published by Esveld et al. (1995), Rhodes & Baxter (2016) and McManus et al. (2017).

The site investigation was undertaken on the newly constructed 68 km Majuba Rail Corridor to measure various parameters that influence track-bridge interaction. The structure is a continuous concrete bridge girder spanning 314 m over the Vaal River approximately 37 km south of Ermelo. Analysis of the bridge displacement, rail displacement and rail creep data indicate that although the rail is fastened directly to the bridge deck by means of elastic fastenings, a relative displacement of up to 6.5 mm, occurred between the bridge deck and rail. This is understood to be as a result of rail creep and “breathing length” associated with thermal loading. Approximately 78% of the displacement at the REJ is due to the high fixity Slab Track rail being pulled by the bridge and the remaining 22% is due to the ballasted rail track on the embankment “breathing” towards the discontinuity at the REJ. Rail creep for the high fixity switch blade accounted for less than 2% of the total rail creep. Rail creep is, therefore, primarily a function of “breathing length” from the ballasted track on the embankment for this bridge and REJ configuration. Midas Civil 2021 was used for numerical modelling. Displacement and axial stress results from a calibrated model compared well with the site data and published stress profile diagrams.