Study of variability in dynamic simulations of turnouts using the PGD method. Issue 3 (4th March 2023)

• Record Type: Journal Article
• Authors: Fan, D.; Sebès, M.; Qazi, A.; Pozzolini, C.; Bourgeois, E.
• Is Part Of: Vehicle system dynamics. Volume 61:Issue 3(2023)
• Page Start: 905
• Page End: 922
• Publication Date: 2023-03-04
• Subjects: Switch & crossing -- railway dynamics -- wheel/rail contact -- model order reduction -- proper generalised decomposition -- data-driven technique -- Monte Carlo; Motor vehicles -- Dynamics -- Periodicals; Electronic journals; 629.231
• DOI: 10.1080/00423114.2021.1977832
• Languages: English
• ISSNs: 0042-3114

Abstract

Abstract : The design and maintenance of turnouts implies considering a great number of external factors that have an effect on their structural integrity. The proper generalised decomposition (PGD) method is used in this paper to carry out a parametrised study of the vehicle-track interaction in a non-intrusive way. The reduced model is built using a defined domain for some of the influence parameters: the vehicle speed, the wheel/rail friction coefficient and the axle load. A major assumption of the PGD solution is its ability to express quantities of interest in a separated form. One of the goals is to show the relevance of separability in the field of railway dynamics. The example of a switch is chosen as a validation of method. The output of the PGD model is the dissipated energy per metre in the switch rail. The robustness of the reduced model is verified using the complete solution. Probabilistic simulations are carried out in order to explore the possibility of using the PGD solution as a numerical abacus in dynamic simulations of switches and crossings. These first tests show that the PGD solution is a promising reduction model method in railway dynamics.