Squat Crack Propagation in Rails Induced by Rolling Contact Fatigue

Rolling contact fatigue (RCF) cracks in rails are critical concern for railway infrastructure, impacting safety, maintenance costs, and service life. RCF cracks initiate and propagate due to high stresses generated by cyclic wheel-rail contact. These cracks can form on the rail surface or within the subsurface of the rail head, exhibiting intricate patterns and enlarging under subsequent loads. The understanding and estimation of crack growth are paramount.

Squat cracks are widely recognized as a common type of RCF crack, which impact railway safety management due to their popularity and the significant increase in recent years. Large squats, which can cause depressions and spalling, lead to an elevated vertical dynamic load on wheelsets, accelerating deterioration of both the track and certain vehicle components. This deterioration is often accompanied by excessive noise and vibration, resulting in increased complaints against rail operators. Additionally, squats can compromise the effectiveness of ultrasonic rail testing by masking subsurface defects. The primary risk associated with squat defects is the potential for these cracks to propagate downward into parent rail along a transverse plane, which could ultimately result in a catastrophic rail fracture if left unaddressed.

Numerical studies, particularly those utilizing finite element models, have focused on stress intensity factors (SIFs) at the crack front to understand crack stability, growth trajectory, and propagation rate. Early 2D FEM studies indicated that smaller cracks tend to grow parallel to the rail surface, with observations consistent across several investigations showing that RCF cracks propagate at an acute angle before turning downward. However, these 2D models are limited as they can only address crack opening (Mode I) and shearing (Mode II) and often overestimate crack growth rates by neglecting three-dimensional constraints. Consequently, 3D FEM models have been developed to better capture the complex fracture mechanisms and Stress distribution. Despite these advancements, a systematic investigation of the sensitivity of wheel passage over pre-existing cracks and growth of twisted squat crack remains lacking, which is crucial for accurately predicting the associated damage under varying operational conditions.

To overcome the current limitations in numerical modelling of squat crack propagation, a two-stage simulation strategy was proposed to investigate the mechanisms of RCF-induced squat-type cracks in rails. It consisted of a dynamic FEM for wheel and rail interaction and a coupled static FEM/BEM for crack growth. The study found that there are two critical wheel positions for crack propagation during wheel approach and departure from the pre-existing crack. The different cracking mechanisms at these two critical wheel positions are correlated with the local physical phenomena. These include the traction at the wheel/rail interface, the relative distance between the contact patch and the crack, and the sliding at crack faces. Additionally, the twisting of the cracks from the transverse plane to the longitudinal plane at the railhead significantly reduces crack growth in mode I but enhances crack propagation in modes II and III.

The study also explored the effects of lateral wheel position variation, twist angles, crack face friction, and crack size on squat cracks. The results reveal that squats are significantly influenced by contact load variation, as increasing the distance between the crack center and contact center reduces stress intensity around the crack. Mode II emerges as the dominant growth mode, even when cracks extend beyond the contact region. Additionally, twist angles promote a shift in mode II and mode III depending on the angle. Importantly, when crack face friction is introduced, mode II remains the dominant growth mode, with a discernible reduction in both mode II and III.

The current work investigated actual squat crack growth from an early-stage tiny squat with a leading branch to the mature squat with both leading and trailing branches. The research highlights the three-dimensional propagation of squat cracks, occurring not only on the rail surface but also extending beneath it. The results show the formation of trail crack is due to the combined cracking modes of shearing and tearing and large opposite mode II shearing at the field corner of the crack. Maximum Shear Stress (MSS) criteria are valid before the squat crack grows to a certain deep. Mode I could assist in crack turning down at the mature growth stage.

The findings are expected to contribute for advanced maintenance technologies and improved rail material designs, ultimately contributing to safer and more efficient railway operations.