Prediction of rail damage using a combination of Shakedown Map and wheel-rail contact energy

Abstract

Rolling contact fatigue and wear are two key damage mechanisms that govern rail life. Although there are several different mechanisms affecting both their initiation and propagation, the trade-off between them is important and their accurate predictions can provide significant benefits when planning rail maintenance activities. Through integration with vehicle dynamics simulations, damage models based on the wheel-rail contact energy (Tγ) and Shakedown theory have often been used to predict damage. In this paper, the findings from previous studies were reviewed to identify their limitations. To assess the accuracy of the predictions, their input parameters were compared for sites with and without reported RCF defects from two lines on the London Underground network. The results indicated certain variations and hence, a new wear and RCF damage prediction method was developed using a combined Shakedown Map and Tγ approach. While the wear model predictions were validated by comparison with measured rail wear, the availability of field crack depth measurements enabled the validation of the new RCF crack depth prediction model. Reasonable predictions of crack depth and wear over consecutive intervals have been achieved on various sites which increases the confidence of the models to support future optimisation of maintenance planning.