Analysis of Early Crack Propagations in Rails

Abstract

Cracks in rails are a common cause of expensive repairs and sometimes dangerous failures. Mathematical models of the cracks can be used to investigate early crack propagation. One of these approaches uses Linear Elastic Fracture Mechanics (LEFM). In LEFM the stresses near the crack tip of cracks can be quantified by stress intensity factors (SIFs). These are related to the crack face displacements parallel and perpendicular to the crack surface. The main purpose of this study is to analyse how different factors will affect these displacements. Using contact mechanics, stresses between the two contacting bodies – wheel and rail –is quantified. To simulate crack deformation the eXtended Finite Element Method (XFEM) is used together with interaction properties managing the surface-to-surface contact of the crack faces. After building a numerical model, using CAE software, a linear elastic stress/strain analysis is performed. A Python script is employed in a post-processing extraction of node data and subsequent evaluation of crack face displacements. When studying the effect of variation of load position relative to the crack, the results showed that large displacements occur primarily when the maximum pressure load is before or after the initiated crack. Furthermore, when analysing the variation of the crack angle, it could be noticed that the crack is likely to propagate much faster when the crack is oriented transversally. Additionally, the influence of wheel–rail interfacial friction was studied and it was observed that it affected crack deformation differently depending on the crack inclination. A conclusion that could be drawn was that there are many trends that can be observed when varying different load and geometry parameters. The trends are influenced by many underlaying factors. The results presented in this study can hopefully be used as a basis on further studying of early crack propagation, and as a help to understand when the deformation of a crack will be the highest.