Wear 265 (2008) 1332–1341 Contents lists available at ScienceDirect Wear journal homepage: www.elsevier.com/locate/wear Modelling rail steel microstructure and its effect on crack initiation F.J. Franklin a,∗ , J.E. Garnham b , D.I. Fletcher a , C.L. Davis b , A. Kapoor a a Newcastle University, School of Mechanical & Systems Engineering, Newcastle upon Tyne NE1 7RU, UK b University of Birmingham, Department of Metallurgy and Materials, Birmingham B15 2TT, UK article info Article history: Accepted 18 March 2008 Available online 6 June 2008 Keywords: Rolling contact fatigue Crack initiation Ratcheting failure Computer simulation Pearlitic rail steel abstract The quality of rail steel has improved greatly in recent years and the material is more resistant to wear, plastic deformation and crack initiation; but track forces have also increased, and cracking of rails is a major concern. Different steel microstructures have different wear and rolling contact fatigue (RCF) behaviours when subject to cyclic, rolling–sliding, compressive contact. In order to capture the differences, it is nec- essary to model these at a microstructural level. This paper describes the development of microstructural models for incorporation into a wear and crack initiation model. A new mechanism is introduced which distinguishes RCF lives of pearlitic microstructures with different percentage volumes of pro-eutectoid ferrite. © 2008 F.J. Franklin. Published by Elsevier B.V. All rights reserved. 1. Introduction Loads on the rail have always led to an incremental process of plastic damage with each wheel pass, and the demands to carry faster and heavier trains exacerbates this problem. Depending on the applied forces, the depth of deformation can be anything from a few microns to 5–10mm. (The surface micro-roughness of wheels can cause extremely high contact stresses within about 50 m of the rail surface.) The strain (of about 10) which builds up is an order of magnitude greater than that found in ductile frac- ture tests at atmospheric pressure. Once the material’s ductility is exhausted, cracks initiate and, under certain conditions, may propagate deeper into the rail and potentially cause dangerous rail breaks. One control strategy is to grind the rails regularly to remove short cracks and to correct the railhead profile (which is affected by wear and plastic deformation). An engineering tool capable of pre- dicting the optimal grinding interval and grinding depth is under development for RSSB 1 to aid maintenance scheduling, and thereby offer considerable savings to the railway industry. The tool will pre- dict the time required for a crack to first initiate and then propagate to a significant length, taking into account the wear rate and the type of steel. A key input to this model is understanding of steel failure at the microstructural level. Most rails are pearlitic steel, but different grades are standard in different countries. Carbide-free ∗ Corresponding author. Tel.: +44 191 222 5681; fax: +44 191 222 8600. E-mail address: F.J.Franklin@ncl.ac.uk (F.J. Franklin). 1 The U.K.’s Rail Safety & Standards Board, http://www.rssb.co.uk/, Project T355: “Management and understanding of Rolling Contact Fatigue”. bainitic steels are also being introduced, and hard metallic coatings are being trialled. This paper discusses how microstructural observations and measured material properties of sections through used pearlitic rails and specimens from rolling–sliding, twin-disc tests (described in detail in a companion paper [1]) have been used to develop microstructure models for use in a computer simulation of wear and crack initiation. For the twin-disc tests, discs were machined from railhead and monobloc wheel rim transverse sections. Heat treatments were applied to some rail discs to investigate the effect of pro-eutectoid ferrite phase distributions and volume fractions. Tests were run to rolling contact fatigue (RCF) failure (defined below) and to percentages of fatigue lives. Analysis of strained material phases has included nano-hardness measurements [2]. The computer simulation [2–4] models the accumulation of plastic shear deformation (plastic ratcheting [5,6]) with repeated load cycles. The deforming material is modelled as a mesh of rectan- gular elements, and the material properties of each element can be selected to construct a representation of rail steel microstructure. For the simplest method described here, the computer simulation uses a simple regular microstructure of hexagons, representing the prior austenite grains with a grain boundary phase of pro-eutectoid ferrite and the interior of the grain being pearlite. In practice, how- ever, grain size and grain boundary thickness vary considerably from point to point within rail steel microstructure. To address this issue, a cellular automaton (CA) has been used to generate a more realistic microstructure. Examination of cracks forming in the heavily deformed near- surface pearlitic rail steel microstructure of sections of trafficked rails and twin-disc test specimens suggests that early crack prop- agation is along the most heavily deformed pro-eutectoid ferrite 0043-1648/$ – see front matter © 2008 F.J. Franklin. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.wear.2008.03.027