Role of Residual Stresses in Rail Squat Formation

Sarvesh Pal, William  Daniel and Andrej Atrens (University of Queensland), Vladimir Luzin (ANSTO-Bragg Institute) and Malcom Kerr (Rail Corp Australia)

Rolling-contact fatigue is fatigue caused by the alternating stresses associated with bodies in rolling contact. A squat, a major type of rolling-contact fatigue damage on some railways, is a surface or near-surface defect that generally occurs on the crown or ball of the rail head, in tangent and gently curved track.

Squats are easily identified visually, as they appear as dark spots or “bruises” on the running surface of the rails. Squats occur typically on a near-horizontal plane, with cracks propagating approximately 3–5 mm below the rail surface.

A research project has been initiated by the CRC for Rail Innovation to examine the formation of rail squats, identify formation mechanisms, and methods of prevent its formation. The project seeks to identify the effects of changes in operating conditions on squat formation and is supported by Australian Rail Track Corporation, QR National, and Railcorp NSW as industrial partners and academic institutions (University of Queensland, Monash University and Central Queensland University).

Many attempts have been made by researchers to find how a squat grows under a moving wheel-rail contact load. However, a squat’s life time from crack initiation point up to spalling off has not yet been studied. The local high stresses from a wheel rolling on a rail, exceeding the material yield point, can cause both squat-type fatigue cracking and can also produce other types of fatigue cracks such as head checking and gauge corner cracking.

The local normal contact stress of the wheel to rail contact can easily exceed the yield strength of the rail for a high axle load. The stresses induced in the rail are especially large for high torques applied during train acceleration and braking, which leads to the local plastic deformation of both wheel and rail.

The distribution of forces present in the contact zone during the interaction of the wheel and rail are explained by the Hertzian theory, which indicated that contact stressing leads to residual stresses (RS) created by the plastic deformation that builds up in the rail. The presence of tensile or compressive residual stresses may affect the crack growth rate and direction of a rolling contact fatigue crack.

Knowledge of the residual stresses gives insight into the likelihood of crack propagation (tensile RS) or crack arrest (compressive RS) in the rail matrix. The residual stresses are one of the most important initial conditions and will greatly affect the model of crack propagation leading to formation of a squat. Hence, in order to understand fundamental issues regarding rolling contact fatigue crack propagation and modelling the crack propagation behaviour it is necessary to measure the RS in the rail matrix.

Residual stress measurements have been carried out using Kowari residual stress neutron diffractometer (Figure 1). RS measurements were conducted on both a service rail and a virgin rail. The experiments were carried out using two types of setup, which are detailed below:

• High spatial resolution mode for accurate local stress measurements in the area of the squat (top surface) using a gauge volume of 0.5x0.5x10 mm³ periodically to a depth of 5 mm, where the largest dimension of the gauge volume was oriented parallel to the rail surface, while maintaining a 0.5 mm in-depth spatial resolution.

• Low-spatial-resolution mode for measuring global changes in the RS distribution on the transverse slices from the same rails as the high-resolution tests using a gauge volume of 3x3x3 mm3.

The high-spatial-resolution measurements of the service rail were intended to evaluate stresses near the surface where the wheel had made contact and found significant compressive stresses along the transverse direction. RS distributions below the surface in transverse and vertical directions were plotted for the virgin and service rails as a function of depth (Figure 2). The service rail showed a compressive stress of around -360MPa form the surface of the rail head to at least a depth of 5mm.

To develop understanding of the large-scale stress redistribution over the whole rail head measurements of the global RS distribution in the rail head were conducted on transverse slices of the rails in order to produce 2D RS distribution maps across whole rail head and to characterize RS acquired throughout the rail service (Figure 3).  Each of 255 mesh points were measured in three directions normal, N, vertical, V and transverse, T directions in order to resolve the stresses in the normal directions, to plot 2D stress component maps and to map d0 variation in the rail head.

In these measurements it was shown that large scale effects, such as the maximum of tensile RS between 20-30 mm below the rail contact surface covering a large area, develop and this is the result of the thermo-plastic history of the rail produced throughout the service lifespan of the rail. The residual stresses measured in this study are to be used in a finite-element model of the rail to simulate the effect of the residual stress state on crack initiation and propagation.

Figure 1. Rail slice samples mounted on the instrument sample table.
Figure 2. In-depth dependence of the transverse and vertical stress components for the virgin (V-Rail) and the sample after service (S-Rail).
Figure 3. Stress maps of the transverse and vertical stress components in the two rail conditions.