A method to separate primary and secondary train delays in past and future timetables using macroscopic simulation

Abstract

Punctuality is a key factor in railway operations and is affected by both primary and secondary delays to differing degrees. Being able to separate these two types of delays is very important when simulating operations, and when conducting punctuality improvement efforts. However, it is not easy to estimate the relative proportions of primary versus secondary delays using historical data. In this paper, we demonstrate a method that uses repeated runs of a macroscopic simulation tool to estimate what share of delays has been primary or secondary. Using the Swedish region of Skåne as a case study, we estimate that about 36% of delays in 2019 were primary, leaving 64% as secondary. We further show that in order for operations to reach the targeted level of punctuality, 95% instead of the observed 87%, primary delays would have had to be cut by half. Using a draft timetable for 2025, we also simulate what the punctuality would be given different assumptions of primary delays. Assuming the same level of primary delays in 2025 as in 2019, we estimate that the punctuality would drop by a further 5 percentage points due to increased density of operations. In order to reach the punctuality target of 95% in 2025, primary delays would instead need to be reduced by two-thirds. At the request of the infrastructure manager, we also show the predicted geographical distribution of secondary delays in this future timetable. Our results highlight the need for drastic delay reduction measures to reach set targets.

Keywords

• Railway
• Simulation
• Delays
• Punctuality