Three-Dimensional Microstructural Modelling of Crack Initiation in Rail Steel

Named Investigators: Prof Ajay Kapoor (University of Newcastle); Dr Claire Davis (University of Birmingham)
Researchers: Dr David Fletcher, Dr Francis Franklin (University of Newcastle); Dr John Garnham (University of Birmingham)
Industry Collaboration and Mentors: Andrew Lezala (Metronet); Phil Heyes (HSL); Network Rail; Corus Rail Technology; AEA Technology Rail

Background:
Rolling-sliding, cyclic contact of wheel and rail progressively alters the microstructure of the contacting steels, eventually leading to micro-scale crack initiation, wear and macro-scale crack growth in the railhead. Relating the microstructural changes to subsequent wear and cracking is being accomplished through modelling at three spatial scales: (i) bulk material (ii) multi-grain and (iii) sub-grain. The models incorporate detailed information from metallurgical examinations of used rails and tested rail material. The initial 2-dimensional models representing the rail material are being further developed into 3-dimensional models. Modelling is taking account of thermal effects, and traffic patterns to which the rails are exposed.

Brief Summary of Research Methods:
The research carried out within this project is a combination of very detailed microstructural observations and characterisation, novel methods for three-dimensional crack analysis and advanced computer modelling. Specifically the experimental work has involved: metallurgical analysis (optical and scanning electron metallography; macro-, micro- and nano-hardness testing; chemical analysis); heat treatment; twin-disc rolling-sliding contact testing; X-ray computed tomography; focussed ion beam microscopy. The modelling work is using Dyna3d explicit finite element code, and in-house developed C/C++ code for materials modelling and summation of damage under variable rail traffic.

Original Project Scope and Objectives (taken from project A6 proposal):
The main aim of the project is to predict rail life based on knowledge of particular rail steel microstructures. This will enable rail grades to be selected appropriately for specific routes, according to traffic / track geometry, in order to minimise maintenance costs. The objectives are:

To understand and model the three-dimensional nature of the rail steel microstructure and the implications for crack initiation and growth.
To identify critical factors limiting rail life and suggest rail management to minimise their impact.
To provide guidance to infrastructure owners and rail manufacturers on desirable rail microstructures for minimising crack initiation and wear.

Work Done:
Following the work done within project A2 a further series of twin disc rolling-sliding tests have been carried out on rail material (as received and heat treated to generate a range of microstructures with different pro-eutectoid (PE) ferrite fractions) at varying fractions of rail life. These tests have generated samples with rolling contact fatigue cracks at different lengths corresponding to initiation events within one prior austenite grain to cracks of length equivalent to several grains.

These samples are enabling the complex interaction between the initiating and propagating small cracks and the 3D microstructure to be investigated, for example using X-ray tomography, FIB / SEM analysis (shown below) and serial sectioning. The materials modelling work is concentrating on developing the ‘dynarat’ computer simulation tool. In the simulation, the material properties of each element can be selected to construct a representation of rail steel microstructure, initially as a hexagonal pattern requiring input from the microstructural characterisation work described above. This is a reasonable approximation of rail steel microstructure for the wear model and for indicating the probable depth of initiating cracks.

However, to understand how cracks initiate and begin to grow in real steel microstructures, the regular and 2D nature of the hexagonal microstructure is not appropriate since barriers to crack growth are not represented. Hence 3D microstructures are being generated using 3D cellular automaton (which uses random processes to ‘grow’ grains) or Voronoi polyhedra. The Voronoi method has the advantage that grains can be represented by their surfaces, making manipulation and graphical representation easier and reducing computation time. The surfaces are also the interfaces where crack initiation is likeliest to occur (confirmed by experimental studies), so identification of potential crack paths through the microstructure (following plastic deformation as a result of ratcheting) is relatively easy and the probability of initiation of relatively long cracks can be assessed.

Of course, representation of pearlitic rail steel as ‘pearlite’ grains with (prior-austenite) grain boundaries is overly simplistic. At the sub-grain level, standard pearlitic rail steel is a 3D composite of ferrite and cementite phases, plus non-metallic inclusions. Failure is dependent on how stress and strain are accommodated by these phases, inclusions and the bonds between them. This is being investigated using an elastic-plastic large strain deformation explicit finite element study of the microstructure, which will provide property information at the micron level, with a detail impractical to measure experimentally. For suggesting rail management strategies to minimise the impact of crack initiation the focus is on rail grinding. Software written in C is being developed to rapidly sum and quantify the interacting processes of crack initiation, propagation and wear damage over periods of years from multiple types of rail traffic. Through this model the effect of rail grinding interventions can be examined to find optimum strategies for removal of damage.

Work Remaining – and proposed resource allocation; person months and dates of deployment: At Birmingham further experimental work is proposed to examine the small cracks in relation to their 3D interaction with the microstructure using the techniques mentioned above. At Newcastle, work will focus on modelling of crack initiation and early growth focusing on predicting the rail management required to minimise impact of this damage. Software will be developed for modelling grains as Voronoi polyhedra.

Outputs:

Poster from RRUK Workshop 2008
Presentation from RRUK Workshop 2008
J.E. Garnham and C.L.Davis, Chapter on ‘Rail Materials’ in Wheel/Rail Interface Handbook edited by R. Lewis and U. Olofsson, to be published by Woodhead publishing in 2008
F.J. Franklin, J.E. Garnham, D.I. Fletcher, C.L.Davis and A. Kapoor, Chapter on ‘Modelling damage in rails’ in Wheel/Rail Interface Handbook edited by R. Lewis and U. Olofsson, to be published by Woodhead publishing in 2008
D.I. Fletcher, F.J. Franklin and A. Kapoor, Chapter on ‘Rail wear and fatigue’ in Wheel/Rail Interface Handbook edited by R. Lewis and U. Olofsson, to be published by Woodhead publishing in 2008
J.E. Garnham and C.L. Davis, “The role of deformed rail microstructure on rolling contact fatigue initiation”, Accepted by Wear
F.J. Franklin, J.E. Garnham, D.I. Fletcher, C.L.Davis and A. Kapoor. “Modelling rail steel microstructure and its effect on crack initiation”, Accepted by Wear
J.E. Garnham, F.J. Franklin, D.I. Fletcher, A. Kapoor and C.L. Davis. ‘Predicting the life of steel rails’, IMechE Jnl Part F: J. Rail and Rapid Transit special issue, 2007, Vol 221, No 1 pp45-58
F.J. Franklin and A. Kapoor. ‘Modelling wear and crack initiation in rails’, IMechE Jnl Part F: J. Rail and Rapid Transit special issue, 2007, Vol 221, No 1 pp23-33
D.I. Fletcher, F.J. Franklin, J.E. Garnham, E. Muyupa, M. Papaelias, C.L. Davis, A. Kapoor, M. Widiyarta and G. Vasić, “Three-Dimensional Microstructural Modelling Of Wear, Crack Initiation and Growth In Rail Steel” World Congress on Rail Research, WCRR 2008, Seoul Korea
P. Hyde, D.I. Fletcher, A. Kapoor, S. Richardson, “Full-Scale Testing to Investigate the Effect of Rail Head Treatments of Differing pH on Railway Leaf Films” World Congress on Rail Research, WCRR 2008, Seoul Korea