Predicting Atmospheric Icing and its Impact on Overhead Contact Wires

Sammanfattning

Ice accretion on overhead contact lines poses a significant challenge in the dimensioning of the railway system. The accreted ice will impose an additional vertical load from its self-weight and affect the wind loads by changing the wires cross-sectional area and shape. The current dimensioning methods, used by Trafikverket, often rely on simplified approximations and constant values when calculating the effects of wind and ice loads. While the approach has not led to any known issues, more accurate calculations are desired to not needlessly overdimension the system. This thesis aims to, with the use of Computational Fluid Dynamics (CFD), evaluate the current way of calculating ice and wind loads. This is done by simulating in-cloud icing, more specifically rime ice, in the software ANSYS FENSAP-ICE followed by flow simulations in ANSYS FLUENT for the drag coefficients. A parametric study during the ice accretion simulations showed how different meteorological parameters will have a slight effect on the final ice shape. Thereafter the fluid simulations in FLUENT were conducted using both Unsteady Reynolds-Averaged Navier-Stokes (URANS) and Large Eddy Simulation (LES) to determine the drag coefficients for both non-iced and iced contact wires. The results were validated against experimental wind tunnel testing. The results showed how the drag coefficients of the iced contact wires have a dependency on the meteorological parameters under which the ice was accreted. Additionally the flow simulations showed how the current assumptions, for both the clean and iced contact wire, may underestimate wind loads due to an underestimation of the drag coefficient. Furthurmore, the fluid simulations showed a strong dependency of the simulation setup, especially with regards to the chosen turbulence model, where LES provided substantially better force predictions. From the gathered ice geometries and drag coefficients, as well as empirical formulas and correlations, a dynamic ice accretion model is presented. While the model contains rough approximations for geometric shapes and meteorological conditions, it demonstrates the benefit of using a dynamic model, as opposed to the today’s static assumptions, during dimensioning of the overhead contact line system.