Application of Impact-Driven Vibrations to Estimate Axial Stress in Continuous Welded Rails

Track segments welded together form a continuous welded rail (CWR). Although CWRs are widely used, they are prone to buckling in warm seasons. To prevent rail buckles, accurate estimates of the axial stress and rail neutral temperature (i.e., the temperature at which the axial stress is zero) are needed. This study proposes a noninvasive method to determine the axial load in CWRs using numeric models and field test data. A general finite element model under varying boundary conditions and axial stresses was formulated. The model was then validated experimentally by testing a real track in the field. During the experiment, the rail was subjected to the impact of an instrumented hammer, and the triggered vibrations were recorded with conventional accelerometers. These vibration frequencies were compared with numerical predictions to estimate the neutral temperature of the rail using a machine learning algorithm. The estimates showed good agreement with the measurements conducted by a third independent party that used a very cumbersome approach based on strain gages.