In this research project, involving KTH, Swedish Rail Administration and train operatingcompany SJ, timetable simulation is used on different areas in the Swedish railnetwork. The objective is to develop methods and strategies which can be usefulfor both long and short term timetable planning in the future.This study presents some of the result from simulations done on the Western mainin Sweden. Traffic is mixed with significant differences in average speeds. Thismakes the high speed passenger trains (X2000) sensitive for small delays and disturbancescaused by other, much slower trains.Simulation software RailSys was used to evaluate how increased and decreasedtime supplements for high speed trains affect punctuality. Also buffer times were increasedbetween X2000 and other trains. Delay distributions from previous projectswere implemented in order to model entry delays together with dwell and run timeextensions.Results show that increased buffer times can have a significant effect on punctuality.Some of the studied trains have a situation with dense traffic and high occurrenceof overtakings. This clearly increases average delays and contributes to a lowerpunctuality.

Suitable analysis methods are needed for evaluation of future timetable scenarios, both in short term operational planning and for strategic planning with a longer time horizon. One method is to use simulation software which makes it possible to model large networks. The Swedish Transport Administration (Trafikverket) is in a process where the aim is to start using simulation software RailSys as a tool for timetable planning. This will at first be applied for long term strategic planning with the possibility to also use it in operational planning further on.

The main focus in this paper is to estimate primary run time extensions from registered data. Ideally these should only represent primary causes, e.g. decreased vehicle performance, variation in driver behaviour or infrastructure malfunctions. These extensions are important in order to make simulations more realistic.

Different reduction levels of registered data are tested in order to estimate primary run time extensions. Registered data used are absolute values without distinction between primary and secondary causes. Calibration simulations are done on the Southern main line in Sweden where the mix of high and low speed trains is substantial.

Many lines in Sweden have mixed traffic, with both passenger and freight trains. In contrast to passenger trains, freight trains can deviate significantly from their timetable. This study evaluates the effect of modelling freight trains running ahead as well as behind schedule. In previous Swedish studies freight trains have been modelled as being on time or delayed. RailSys is used to simulate a section of the Southern Main Line and a fictive double-track line. Simulation results are compared to data from real train operation. Results show that freight trains can be modelled in a more realistic way. This can improve simulation analysis of freight operations.

Mixed traffic with large speed differences between fast and slower trains consumesmore capacity and makes the system sensitive to disruptions. This article focuses onadequate train configuration for increasing the top speed for express trains like theGreen Train and how that affects capacity on lines with heterogeneous traffic.Microscopic simulation of a future timetable selected by criteria revealed by analyticaltimetable analysis for a chosen structure of services combines the advantages of twomethods and makes it possible to reveal relevant characteristics for different alternatives.Punctual short stops through better train layout and skip-stop operation forregional trains are a few of the measures that are shown to have compensating effectsfor the increase in capacity utilisation and can reduce disruptions. Although it ispossible to reduce the perturbations by means of different measures, the basic problemwith mixing fast and slower trains on the same line still remains.

One way to model train operations and make predictions offuture outcome is to use simulation. Many lines and networksconnecting major cities have a high capacity utilization, meaningthat running additional trains leads to an even more strainedsituation and delays are likely to increase. The mix of averagetrain speeds is also related to capacity and delay propagation.Considering one line or several lines connected in a networka requested train traffic can consist of different train categoriesand departure frequencies. There are usually several possibletimetables satisfying this traffic demand. The infrastructure oftenimplies limitations on the type and volume of traffic that can behandled. Additionally constraints introduced by requests for regularintervals, minimum headways, passenger transfers betweentrains etc. can reduce the number of acceptable timetables.This paper presents an approach using combinatorial traininitiations and simulation to generate conflict-free timetables.These can then be simulated with random variations in departureand dwell times. This is implemented on a fictive single track linewith high speed passenger train traffic. The objective is to studyoutcome by varying allowance times and delays. Simulations arecarried out in RailSys, a software using synchronous simulationto model train traffic operations.

This paper aims at presenting a framework for obtaining and evaluating multiple timetablesin operational conditions on different infrastructure alternatives. Evaluating several differenttimetables, which share characteristics regarding departure frequencies, helps in gettinga more comprehensive picture of how scheduled and operational delays are affected byimprovements or reductions in the infrastructure.In order to flexibly create infrastructure alternatives a script is developed in which differentstation layouts can be defined and expanded into a railway line by connecting stationstogether. This is then compiled to files that RailSys can interpret, thereby facilitating in includingthe infrastructure as a variable in the analysis. Timetables are obtained by definingnominal train paths in a combinatorial approach and simulating these in RailSys in order toget conflict free timetable solutions.A case study is presented where the infrastructure, applied departure frequencies andsuperimposed stochastic delays are varied. Results from this give an indication of howmuch the scheduled and operational delays are improved when longer sections of a singletrack line are expanded to double track. The case study is fictive but the framework canbe used for evaluating real scenarios, e.g. where different infrastructure expansions areweighed against each other and only one of them can be realized.