Abstract
The railway system is a complex dynamic system, characterized by numerous non-linearities and imperfectly known random environmental variables. As a result, its approval, monitoring and maintenance are still largely based on observation, expertise and semi-empirical approaches.
However, the joint introduction of simulation and data analysis is opening up new prospects for optimizing system operation and reliability. This presentation will illustrate, through several applications, how physical models of varying levels of complexity and experimental measurements can mutually enrich each other, whether for model calibration, the introduction of uncertainties, the physical interpretation of data or the resolution of optimization problems.
