Lifetime estimation of catenary components

Overhead line incidents and in particular fractures of the contact wire are at the origin of the most expensive railway incidents in terms of service interruption and train delays. Therefore, it is important to carry out preventive maintenance with a high safety coefficient. The SNCF (French Railways) aims at optimizing its maintenance by primarily estimating the lifespan distribution of various installed components. Information collected during catenary inspection is stored in a database. In an initial study, only the number of preventive and corrective replacements were used to estimate the lifetime distributions, as this type of data is easily available. A statistical method based on Markov chains and Markov jump processes was used for the estimation. Based on the encouraging results, it was decided to digitalize more precise data registered directly by the inspector on a paper support. These data ensure the identification of each installed component with the help of the pylon number. Several methods for lifetime data analysis are used to analyse these data. As the catenaries are not supervised continuously, methods accounting for censored data are required. The interval between two inspections cannot be neglected with respect to the lifespan of the catenary components. Eventually, the results of this study can be compared to the one obtained in the first analysis based on the synthesized data. In the future, these results will be used to find an optimal maintenance strategy: first, for the high-speed tracks in France and later for classic lines.