Development of bainitic rail steels with potential resistance to rolling contact fatigue

Rails are a major capital and maintenance cost for railways in North America. While manufacturers produce clean steels with high quality, most rails made today retain the basic carbon–manganese chemistry of traditional pearlitic rails. This paper describes the development of a bainitic rail steel with potential additional resistance to rolling contact fatigue damage. It is shown that rails can be produced in bainitic steel without the need for complex heat treatments after rolling, and that bainitic rails can have higher hardness and fracture toughness than pearlitic rails. Although small‐ and full‐scale tests indicate that the wear performance of bainitic steel depends considerably on test conditions, the indication is that bainitic steel rails can have significantly better rolling contact fatigue performance compared to pearlitic rails. Reasons for the superior fatigue performance are not fully understood, although a number of hypotheses exist. A conclusion is that continued research would be useful to understand quantitatively the physics and metallurgy of wheel/rail contact.     

Statistical analysis on rolling contact fatigue in railroad axle bearing steel

High‐strength steels are widely used in high‐performance bearings utilized in most mechanical systems. However, there has been little statistical analysis regarding the fatigue failure behaviour of the material, where surface peeling resulted from contact fatigue during rolling is a significant life‐limiting mechanism. In this study, we examine the statistical behaviour of surface‐crack nucleation, propagation, and peeling in a high‐speed train axle bearing made of GCr15 steel by using a laboratory rolling‐contact equipment. We reveal that cyclic rolling‐contact leads to the formation of a hardness gradient in the outer ring of the bearing. The gradient layer is of several millimetres. The peeling rate could be as high as 28 μm per million cycles when the contact pressure is close to that applied in real service. Peeling‐induced cracking is dominantly transgranular. The incipient angle is about 23.2°, and its depth could be hundreds of micrometres. The findings reported here could be employed to assess the lifetime of bearings made of GCr15 steel and possible other engineering metals.     

Life prediction for rolling contact fatigue crack initiation of kiln wheels

This paper investigates the rolling contact fatigue life of kiln wheels with respect to the axis line deflection related with the applied supporting loads on wheels. Fatigue crack initiation criterions for elastic shakedown, plastic shakedown, and ratcheting material responses are applied to assess wheels responses with two sets of axial line deflection, one is measured in field and the other is optimal adjustment for the measured axial line deflection. The finite element simulations are performed by using the Bilinear material mode for nonlinear and kinematic hardening within ANSYS 11.0. By comparing life prediction from different criterions, it is showned that the low-cycle fatigue is the predominated failure. Results from different axial line deflection indicate that the optimum adjustment can greatly enhance the whole life of the supporting structure, that is useful for kiln adjustment and maintenance.     

Rolling contact fatigue cracking in rails subjected to in‐service loading

Rolling contact fatigue (RCF) is one of the most important failure mechanisms in rails with significant cost‐ and safety‐related implications on the operation of railway systems. In this work, a metallurgical analysis of RCF crack initiation and propagation, including geometrical characteristics of RCF cracks – length, depth from surface, angle of propagation and spacing between cracks, is presented. The role of proeutectoid ferrite in crack initiation has been studied. Analysis of the fracture surface of an RCF crack revealed a ductile initiation zone followed by a quasi‐cleavage crack propagation. Iron oxide formed in the interior of all cracks in rails exposed to stagnant water with implications to crack propagation rate because of crack closure effects. Sequential sectioning parallel to the rolling surface revealed that RCF cracks possess convoluted surfaces. The crack trace expands with depth from the rolling surface. Subsurface crack initiation has also been documented.     

On propagation of short rolling contact fatigue cracks

Recent accidents involving railway rails have aroused demand for improved and more efficient rail maintenance strategies to reduce the risk of unexpected rail fracture. Numerical tools can aid in generating maintenance strategies: this investigation deals with the numerical modelling and analysis of short crack growth in rails. Factors that influence the fatigue propagation of short surface‐breaking cracks (head checks) in rails are assessed. A proposed numerical procedure incorporates finite element (FE) calculations to predict short crack growth conditions for rolling contact fatigue (RCF) loading. A parameterised FE model for the rolling‐sliding contact of a cylinder on a semi‐infinite half space, with a short surface breaking crack, presented here, is used in linear‐elastic and elastic–plastic FE calculations of short crack propagation, together with fracture mechanics theory. The crack length and orientation, crack face friction, and coefficient of surface friction near the contact load are varied. The FE model is verified for five examples in the literature. Comparison of results from linear‐elastic and elastic–plastic FE calculations, shows that the former cannot describe short RCF crack behaviour properly, in particular 0.1–0.2 mm long (head check) cracks with a shallow angle; elastic–plastic analysis is required instead.     

A rolling contact fatigue crack driven by squeeze fluid film

ABSTRACT A new model of surface breaking rolling contact fatigue (RCF) crack driven by a coupled action of a squeeze oil film built up in the crack interior and a pressure exerted at the external contact interface was developed. The model can be applied to the ‘nominally dry’ contact couples with an occasional presence of liquid in the crack interior (wheel/rail contact) as well as to the elastohydrodynamic lubrication (EHL) conditions. In the first case, the contact load is a result of solid/solid interaction and can be determined by solving the FE contact problem, but the liquid contained in the crack interior forms a thin film between the crack faces changing their interaction into the type of liquid/solid. This liquid is being periodically squeezed under contact load and acts as a ‘squeeze film’ known from the lubrication theory. In the second case, the liquid (lubricating oil) is permanently present in the contact area and consequently in the vicinity of the crack mouth. This creates conditions for filling the crack with oil. Similarly as in the first case, the ‘squeeze oil film’ is built between the crack faces. The contact load in this case results from a liquid/solid interaction and can be approximated by the pressure and traction distributions obtained from the numerical solution of the elastohydrodynamic contact problem. In both cases the model can be used to determine the Linear Elastic Fracture Mechanics (LEFM) crack tip stress intensity histories during cyclic loading and consequently to predict the crack growth rate and direction. An example of applying the model to the EHL case is given to explain the mechanisms and phenomena leading to the crack front loading. The cycle of rolling a roller over the crack was numerically simulated to obtain the mixed mode (I and II) SIF histories. In the analysis, the EHD pressure and traction were determined through the full solution of the EHD line contact problem accounting for the presence of a crack, whilst the pressure in the crack was found with the use of the wedge shaped squeeze oil film (SOF) model. Possible effects of the mode I and mode II stress intensity cycles on crack growth rate and direction are discussed. The solution indicates high pressure in the neighbourhood of the crack tip, exerted on the crack faces by the squeeze oil film. This leads to the ranges of the mode I and mode II SIF variations, which are larger than for the ‘dry’ and ‘fluid entrapment’ models, and can be an explanation for the crack growth rate observed in practice     

The effect of contact load reduction on the fatigue life of pearlitic rail steel in lubricated rolling–sliding contact

Twin-disc contact simulation tests were carried out to investigate the influence of contact pressure variation on rail steel fatigue life. Both a colloidal suspension of molybdenum disulphide in an oil carrier fluid (similar to a commercial flange lubrication product) and water were used as lubricants. It was found that the reduction from 1500 to 900 MPa of the maximum Hertzian contact pressure (at which a molybdenum–disulphide-lubricated and previously worn rail sample was tested) extended the fatigue life of the rail steel by over five times. For water lubrication a similar reduction in contact pressure produced only a marginal increase in fatigue life. The results were found to be in qualitative agreement with the predictions of the newly developed Three Mechanism (TM) model of rolling contact fatigue, which is introduced here. This model combines the mechanisms of ratcheting and the fracture mechanics-based mechanisms of both shear stress- and tensile stress-driven, fluid-assisted, crack growth.     

Surface crack growth of silicon nitride bearings under rolling contact fatigue

Surface crack growth of silicone nitride ceramic bearings under rolling contact fatigue has been investigated from the viewpoints of contact stresses (ring crack model) and fluid pressure (wedge effect model). The mechanisms of these two models have been investigated independently; however, it was impossible to separate the effects of contact stresses and fluid pressure on surface crack growth. In this paper the effects of contact stresses (ring crack model) on surface crack growth are investigated. In the ring crack model the crack growth is caused by contact stresses around the circumference of the contact circle. The growth of surface cracks located inside and outside the contact track was observed in order to obtain data from which we could reexamine the ring crack model. The outside cracks under rolling contact fatigue were propagated by contact stresses alone and also the inside cracks grew as slowly as the outside cracks. We concluded that the cracks are propagated by the single effect of contact stresses. Preliminary observations of surface crack growth showed that the cracks were unaffected by wear and residual stresses.     

Retained austenite stability on rolling contact fatigue performance of 8620 case-carburized steel

Varying levels of retained austenite (RA) were achieved through varying undercooling severity in uniformly treated case carburized 8620 steel. Specimens were characterized via XRD and EBSD techniques to determine RA volume fraction and material characteristics prior to rolling contact fatigue (RCF). Higher RA volume fractions did not lead to improvement in RCF lives. XRD measurements after RCF testing indicated that little RA decomposition had occurred during RCF. A continuum damage mechanics (CDM) finite element model (FEM) was then developed to investigate the effects of RA stability on RCF. The results obtained from the CDM FEM captured similar behavior observed in the experimental results. Utilizing the CDM FEM, a parametric study was undertaken to examine the effects of RA quantity, RA stability, and applied pressure on RCF performance. The study demonstrates that the energy requirements to transform the RA phase are critical to RCF performance.     

A single parameter to evaluate stress state in rail head for rolling contact fatigue analysis

Based on the Smith‐Watson‐Topper (SWT) method, a phenomenological approach for multiaxial fatigue analysis, the maximum SWT parameter is proposed as a single parameter to evaluate the stress state in the rail head for assessing the fatigue integrity of the structure. A numerical procedure to calculate the maximum SWT parameter from a finite element analysis is presented and applied in a case study, where the stress and strain fields due to wheel/rail rolling contact are obtained from a three‐dimensional finite element simulation with the steady‐state transport analysis technique. The capability of the SWT method to predict fatigue crack initiation in the rail head is confirmed in the case study. Analogous to von Mises stress for strength analysis, the maximum SWT parameter can be applied to evaluate the fatigue loading state not only in rail head due to rolling contact fatigue but also in a generic structure subjected to a cyclic loading.     

The development of a crack propagation model for railway wheels and rails

Rolling contact fatigue (RCF) and wear of railway wheels and rails are the main phenomena that affect their maintenance costs. When crack propagation and wear rates can be predicted, maintenance planning can be optimised, and cost‐effective measures can be developed. Several RCF models exist, but none which can be used in combination with vehicle dynamics simulations and can predict the actual crack depth. This study shows the development of a crack propagation model that can be applied for both railway wheels and rails. Two unknown material parameters in the model were calibrated against crack measurements in a curve on the Dutch railways over a period of 5 years. Two different RCF models were used to calculate the stress magnitudes for the propagation model. The propagation model can be used in combination with vehicle‐track dynamics simulations and shows promise in predicting the actual crack depth and/or surface length. Further research is needed to determine the model's validity for other operational conditions.     

On the Ekberg, Kabo and Andersson calculation of the Dang Van high cycle fatigue limit for rolling contact fatigue

Recently, various methods have been proposed to assess the risk of rolling contact fatigue failure by Ekberg, Kabo and Andersson, and in particular, the Dang Van multiaxial fatigue criterion has been suggested in a simple approximate formulation. In this note, it is found that the approximation implied can be very significant; the calculation is improved and corrected, and focused on the study of plane problems but for a complete range of possible friction coefficients. It is found that predicted fatigue limit could be much higher than that under standard uniaxial tension/compression for ‘hard materials’ than for ‘ductile materials.’ This is in qualitative agreement, for example, with gears' design standards, but in quantitative terms, particularly for frictionless condition, the predicted limit seems possibly too high, indicating the need for careful comparison with experimental results. Some comments are devoted to the interplay of shakedown and fatigue.     

Investigation of rolling contact crack initiation in bainitic and pearlitic rail steels

The stress–strain history and the crack initiation lives of bainitic and head‐hardened pearlitic rail steels were determined under rolling contact loading by implementing the semi‐analytical Jiang–Sehitoglu rolling contact model that incorporates both ratchetting and multiaxial fatigue damage. The calculations revealed that the bainitic steel withstands higher loads than the pearlitic steel at low shear tractions, however; both materials behave in an increasingly similar manner as the shear tractions increase. Furthermore, maximum damage occurs in both steels when ratchetting and fatigue damage coincide on the surface. In addition to shedding light on the rolling contact fatigue (RCF) performance of bainitic and pearlitic rail steels, the current work also establishes a methodology for the realistic prediction of crack initiation under RCF.     

Numerical assessment of the loading of rolling contact fatigue cracks close to rail surface irregularities

Rolling contact fatigue damage of railway rails in the form of squats, characterised by local depressions and cracks located at the rail surface, has been linked to the occurrence of local rail surface irregularities. This study concerns rolling contact fatigue cracks in the vicinity of fairly smooth surface irregularities, here denoted dimples. The influence of factors such as dimple geometry, cluster effects, and crack size is evaluated. To this end, dynamic vehicle–track simulations featuring realistic wheel and rail profiles are employed to characterise the dynamic impact during a wheel passage. The contact load in the vicinity of the dimples is then mapped onto a 3D finite element model of a rail section containing a crack in the rail head. The crack loading is finally quantified by multimodal stress intensity factors. The analyses establish that also shallow dimples might have a significant impact on the crack loading. This effect is increased for larger or multiple irregularities but decreases as the crack grows.     

Effects of sliding on rolling contact fatigue of railway wheels

This paper proposes a numerical approach based on a steady‐state algorithm to predict the rolling contact fatigue crack initiation in railway wheels in practical conditions. This work suggests taking into account the cyclic hardening of the wheel's material and one of its originality is to conduct a complete numerical approach whatever the loading level. The main stages are the characterization and modelling of the material behaviour, the determination of the stress–strain fields using a numerical steady‐state method and the application of a high cycle fatigue criterion. Computations were made with the Abaqus FE commercial software. Three cases are studied: rolling with or without sliding and skating. The numerical results give several types of mechanical responses: elastic or plastic shakedown. Otherwise, the results show that the location where the shear stress is maximal is not the same as where the risk of crack is the highest.     

Analysis of contact fatigue crack growth using twin-disc tests and numerical evaluations

The contact fatigue damages on the rail surface, such as head check, squats are one of the growing problems. Fracture of rail can be prevented by removing the crack before it reaches the critical length. Therefore, the crack growth rate needs to be estimated precisely according to the conditions of the track and load. In this study, we have investigated the crack growth behavior on rail surface by using the twin-disc tests and the finite element analysis. We have verified the relationship between the crack growth rate and the variety of parameters as cracks grow from the initiation stage.     

A numerical approach to subsurface crack propagation assessment in rolling contact

Subsurface crack mode II propagation parallel to the contact surface is a damage mechanism leading to dramatic failure in many components subjected to cyclic loading. A weight function (WF) was elaborated for calculating the applied mode II stress intensity factor (SIF) of a crack in a two‐dimensional half‐space in plane strain condition, for crack completely closed and frictionless contact between the crack faces. With respect to other methods, the WF allows faster SIF calculation, thus being suitable for simulation of many repeated load cycles and fatigue crack propagation. The WF was applied for simulating a case of rolling contact experiments found in the literature, and good agreement between experimental and numerical results was obtained, showing the effectiveness of the WF method in damage tolerant design.     

Flaking failures originating from microholes of bearings under rolling contact fatigue

Rolling contact fatigue tests were carried out using plates with microholes (diameter was about 100 μm and depth was about 140 μm) under three different loads (maximum values of Hertzian stress were about 3250, 3550 and 3840 MPa, respectively) and the surface cracks initiating from those holes were observed. It was found that there is a threshold value of maximum Hertzian stress whether surface cracks originate from microholes or not, and its value is between 3250 and 3550 MPa. However, flaking failures occurred even when the stress values were lower than the threshold value. In order to investigate the relation between the flaking failures and the cracks, sectional observations of the subsurface cracks were made before and after the surface layer separations. From these observations, it was found that the subsurface cracks caused the flaking failures even when the maximum value of Hertzian stress was lower than the threshold value of surface crack initiation.     

Numerical simulation of fatigue crack growth

A numerical simulation of fatigue crack growth which uses currently available crack tip stress and strain fields is described. The essential features of the numerical model are the concepts of damage accumulation cycle by cycle and repeated re-initiation at the tip of the growing crack. The failure criteria employed are a combination of a failure condition and a critical distance over which this condition must be achieved. This critical distance, the material size parameter, has a magnitude which depends on the failure mechanism.

The use of the model to illustrate the effects of stress ratio and environmental effects is described and the ability of the model to predict the onset of bursts of crack growth due to static failure mechanisms is demonstrated. The phenomenon of self-arresting cracks is also displayed.

Material characteristics are included in the model and comparisons with experimental data are presented for a C-Mn steel used in the fabrication of offshore structures.