Proposal of the rail profile selection method to reduce gauge corner cracking initiation by mitigating contact conditions of wheel and rail

Abstract

The gauge corner cracking (GCC) occurs on heat treated rails of the high rail in curved sections with a radius of 600 to 800m. In case of GCC propagates deeply, it may cause rail breakage. Therefore, it is very important to prevent the occurrence of GCC for the safety transportation of railways. In the previous research, the countermeasure method for suppressing the occurrence of GCC by applying worn profiles of rails to the high rail in curved sections due to the relief of contact pressure between wheel and rail. Since the wear development of rails is closely related to the contact conditions of wheels and rails, predicting of worn profiles of rail will be changed complexly due to various contact conditions. The aim of this study is to examine the cross-sectional rail profile that is the most effective in suppressing crack initiation for the high rail in curved sections with a radius of 600 to 800m where the occurrence of GCC is a concern. In the beginning of this study, a wear development analysis with multibody dynamics which was modeled in various radii at the appearance of GCC was conducted. Secondly, a wheel and rail contact analysis using a predicted rail worn profiles was conducted, and the occurrence of cracks was evaluated based on wheel and rail contact conditions. Finally, the optimum rail cross-sectional profile was searched by machine learning with neural network using the analysis results as a teacher data. In summary, the optimal rail cross-sectional profile with highly effective for the suppression of GCC initiation was determined and evaluated the suppression effect for crack initiation.

Keywords

• wear development
• fatigue index
• wheel/rail
• gauge coner cracking
• multibody dynamics
• archard
• machine learning
• neural network