More robust switches through improved control of the switch rail

Rapport, 2022

Current switch rail control detectors are among the most traffic disturbing factors in the Swedish rail network. Figures from Trafikverket from 2017 shows that more than 1021 trains were delayed for more than 257 delay hours. The sensor is not an expensive component, but required sensor adjustments when maintaining the switch is time-consuming, costly and cause further traffic disturbance.

The aim of this feasibility study is to evaluate the possibility to enhance the reliability of switch rail control while retaining safety levels. The focus is on the case when foreign objects (e.g., ballast stones) are trapped between the switch rail and the stock rail. This will lead to a rail gauge reduction when the switch rail is closed by the drives. To indicate such a rail gauge reduction that may cause derailments, switch rail control sensors (TKKs) are used in Swedish switches. These control sensors indicate if the switch rail is sufficiently close to the stock rail but are, as mentioned, also major causes of traffic disruptions. It is therefore important to know if, and under which conditions the controls in fact add additional safety. This is the focus of the current feasibility study.

The investigation sets out with a detailed review of previous investigations into the use of switch rail controls and concludes that all Swedish investigations are based on a one-page report (M5745/87) from 1987. From a scientific perspective the conclusions of this report can neither be verified nor falsified, a fact that has been further established by studying all available reports and presentations that may provide insight into how conclusions in M5745/87 were achieved. In particular, the background to current regulations to prevent derailments related to narrow rail gauge are two ORE reports. This study shows that these reports are not applicable for the case of reduced rail gauge in switches.

The study has further studied deformation of ballast stones and loads from vehicles. Preliminary static calculations have been performed and indicate that a derailment cannot be achieved for the studied “worst normal case”. These conclusions must however be further ensured with simulations, and with tests in track. To this end, it has been ensured that simulations of dynamic switch negotiations can be performed. Also, a tentative test plan to validate analyses has been outlined.