Post-monitoring evolution: mathematical model for neutral temperature development in continuous welded rail

Abstract

The long-term development of neutral temperature in continuous welded rail (CWR) was investigated using measurement systems equipped with self-compensated strain gauges in quarter bridge configuration. Field measurements on railway lines were conducted over a two-year period following the installation of the CWR. In tracks with lower traffic intensity and stiffness, the CWR neutral temperature could easily reach 40 degrees C during summer but dropped below 30 degrees C in winter. In contrast, in stiffer tracks with higher traffic intensity and heavier loads, the long-term deviation of the CWR neutral temperature remained within significantly narrower limits (closer to the CWR installation temperature - 23 degrees C). Regardless of track stiffness or traffic intensity, the trends in CWR neutral temperature development were found to be similar for the left and right rails within the same profile of monitored section. The proposed machine learning (ML)-enhanced, and regression-based mathematical models accurately represent the long-term development of neutral temperature in CWR.