Development of a tram track degradation prediction model based on the acceleration data

Abstract

Although vibration is considered as one of the important factors in passenger ride comfort, yet it has not been applied for predicting tram track degradation in tram network. Rail track degradation prediction models form an essential part of the rail infrastructure maintenance management systems. Vibration can be measured by acceleration signals. The acceleration signal is derived from the movement of railway vehicles on rail structure. In this study, vehicle acceleration data along with other track structural parameters have been used to predict tram track degradation index which can be considered as a representative of tram track quality. The index used in this study has been developed based on a mixture of tram track geometry deviations of several years. Three types of machine learning models have been employed for creating the prediction models. In this study, Melbourne tram network data have been applied for developing as well as predicting the degradation index. Based on the evaluation results, the proposed random forest regression model made more accurate predictions on track degradation compared to other developed models. The results of this study can help tram track managers to deploy cost-effective maintenance strategies by applying vehicle acceleration data in their decision-making processes.