Add us to your taskbar by dragging this icon RGI logo to the bottom of your screen.

Close

Software tools ease planning headaches

02 Jan 2009

SOFTWARE: A range of approaches is available for planners seeking to model timetables and assess the capacity of a railway. Three case studies from Europe illustrate the variations between them.

Wolf-Dietrich Geitz, Partner, Railistics GmbH

For decades, timetabling and capacity allocation was a traditional process undertaken in-house by Europe’s national railway companies. However, the open-access policies imposed by recent European legislation have forced the sector to rethink relations between infrastructure managers and the incumbent and external train operators.

A handful of scientists have spent several years developing a theoretical basis to define and allocate capacity on a network and to create time-tables which work in the ‘real world’ by identifying and eliminating conflicts. IT companies have transformed these theoretical basics into powerful tools to simulate timetables and operations from simple branch lines to complex networks. Nevertheless, not every problem facing a planner can be solved by an exact formula or ‘off-the-shelf’ software. An understanding of the mathematical background of a particular software suite, its capabilities and limitations, and sufficient data are all required before a project can begin. Equally, knowledge of the operational railway is also needed to interpret the results correctly.

Specialist software development has led planners to consider the standards and definitions required to calculate timetables, define infrastructure and quantify the difficult issue of capacity. UIC has finally succeeded in developing an initial set of guidelines on capacity, based on research at the Technical University of Aachen and other institutions.

This is the first time that a standard definition of capacity has been established which is applicable to a wide range of route and network characteristics. UIC 406 is based around the occupation period of a section of line, defined by a number of parameters including obvious factors such as line speed or acceleration and braking curves, but also more complex variables such as the approach time to red signals and the constraints imposed by particular signalling technologies. These ‘occupation periods’ can then be added together to give an indication of the available capacity along a particular route, depending on the type of operation required.

Potential benefits

Large network simulations and path management tools have been applied by several major European railways for macro-level analyses of network capacity. However, this article will examine the other end of the spectrum to show how IT tools and definitions such as UIC 406 can offer an answer to case-specific operational questions.

Our experience suggests UIC’s approach to quantifying capacity can have practical use in the open-access environment. Three examples will illustrate how a small investment in simulation and operational analysis can allow focused investment at a time of ever-scarcer budgets. In particular, they show how the demands of operators and infrastructure managers might be balanced, even in potentially complex scenarios involving several parties.

Case studies

1. Connecting a new production site to a regional line

The situation. A factory has a siding served by a 50 km single-track route. The factory receives four pairs of freight trains each night. Planned expansion would require eight pairs of trains distributed across the day. The line carries a regular-interval hourly passenger service. Three stations have crossing loops, of which two are needed to maintain the passenger service; the third is used to regulate late-running trains. This is one of two loops long enough to accommodate a freight train.

The infrastructure manager wishes to remove the third loop to reduce costs, believing there would be sufficient capacity over a 24 h period for the extra freight services. These could only operate at night as the passenger operator does not want to change the pattern of the regular-interval timetable. The infrastructure manager, freight operator and factory cannot agree on the actual capacity needed, or on the operational changes required to fit the reduced infrastructure.

Shifting all the freight services to night operation would require further investment in loading facilities and sidings at the factory, which would make its expansion uneconomic.

The methodology. We decided to base our analysis on a tool which works on given timetables and not on a statistical approach, since the planned infrastructure changes are based on existing operating patterns. We developed a matrix of infrastructure and freight operating scenarios, including some minor adjustments to the passenger timetable which could be acceptable to operator and passengers, and variations in the factory’s logistics programme which we saw as manageable and commercially-viable. Simulations were undertaken with a simple tool which shares the UIC 406 methodology and was appropriate for this relatively straightforward calculation.

Outcome. The results showed that operational and commercial demands could only be met by keeping the third loop, or installing an additional loop at one of the two other stations. A compromise fulfilling the stakeholders’ commercial requirements, albeit with a reduction in some operational quality, was found by lengthening the other passing loop to accommodate freight trains.

Benefits. The combination of a tool based on real timetables with commercial calculations showed that the implementation of the infrastructure manager’s original plan would have been detrimental to either the factory or the passenger operator, and possibly to all three parties. This approach also proved that the addition of irregular services to a regular-interval timetable could not be simulated properly with a purely statistical method.

 

2. Connecting a freight yard to the main line

The situation. An industrial complex which includes a busy intermodal terminal is connected to a double-track electrified main line by an 8 km single-track branch. The branch is part of the national network while the terminal and plant sidings are private. The junction between the two includes a yard with six roads to store trains or empty wagons. While regular freight is in decline, intermodal traffic is booming, and an expansion of the intermodal terminal has become an urgent priority. The owners of the terminal are ready to invest but require sufficient capacity within the yard and along the branch. Around 70 trains/day are currently using the branch line, but a further 30 intermodal trains need to be accommodated to justify the capital investment.

The infrastructure manager believes that the capacity of the branch, yard and main line is sufficient for the maximum number of trains proposed. However, the investors in the terminal want to be sure that there is sufficient capacity even if the additional trains are not evenly distributed throughout the day. A capacity shortage would lead to discussions between the various parties to decide who might ultimately have to pay for more track.

The methodology. This case required a combination of fixed and stochastic approaches. While passenger services on the main line were fixed for the next decade, the timetable of most existing freight trains was variable within a certain timeframe, and times of the extra trains were not available. The terminal processes and capacity were also fixed. We decided to use a tool with fixed timetables, apply a stochastical curve determined by logistics parameters and terminal capacity, and add the extra services to the line with concrete timetables. The trains’ likely characteristics were then factored into the calculation, and the capacity of the terminal sidings extrapolated from the resulting timetable.

To include flexibility for existing freight services, each train was given a maximum deviation, again determined by logistics parameters. We then applied the UIC 406 methodology to determine a final capacity.

Results. The analysis showed that the existing infrastructure was sufficient to accommodate all the extra trains if some flexibility of peak-hour timetables were permitted, accompanied by a small reduction in punctuality. The owners of the terminal and the surrounding complex both decided that these constraints were acceptable, and the investment has started. Rail freight volumes will continue to grow without the need for a potentially harmful dispute over investment in additional infrastructure.

 

3. Integration of a new passenger service

The situation. Regional passenger services around a mid-sized city have been increased using some existing, some re-opened and some new infrastructure. Passenger numbers are climbing, and communities without a rail service are increasingly demanding their own link.

The existing network includes two busy main lines which must accommodate a growing number of long-distance passenger and freight trains, while inter-regional services are also growing. In addition, a new residential area and economic development zone are to be built with their own connections.

The regional network handles around 600 trains/day over 90 km of double-track electrified main line, plus 10 km of electrified and 20 km of non-electrified single track. The new line is around 10 km long and connects to the existing network at three points. One major station, three busy junctions and a freight yard have been included in the analysis. The track and stations across the network are owned by five different companies.

The regional transport authority requested a basic hourly service on the new line, with scope to run more services at peak times. However, some trains will be expected to join or divide with others during their journey, adding complexity.

The problem. The regional transport authority needed to have a detailed timetable and operational structure to allow the new line to be integrated into the network. It wished to know about potential operational difficulties, and to be prepared for discussions with two different infrastructure owners. It also needed to know about potential capacity bottlenecks arising from the new services, and the investment required to alleviate them. The analysis also had to find a balance between investment in rolling stock, infrastructure and stations, as well as establishing minimum journey times on key routes.

The methodology. A full scale simulation might have appeared most promising at first sight. But budget and time were limited, not all the stakeholders involved were familiar with simulation, and infrastructure data for parts of the network were limited.

We decided to see if a fixed timetable approach could be undertaken in such a complex case. We undertook some pre-analysis and comparison with similarly-complex studies, which suggested that a fixed timetable approach was feasible. However, it appears we may have come close to the limits of the complexity that this approach can handle, although this still gave a much more comprehensive analysis than appeared likely at the outset.

Results. The analysis proved that the line can be integrated into the existing network. To ensure a high degree of punctuality, a scenario involving the doubling of a short section on one of the single lines would help to meet the reliability requirements at the least cost. An exchange of rolling stock between two of the existing lines would give improved acceleration to boost capacity at a key bottleneck.

Benefits. The transport authority and infrastructure managers agreed that the plans could be realised with a minimum of investment. Additional benefits such as more efficient use of rolling stock could even reduce operating costs. The most important result of the study from our point of view was a genuine understanding of the transparent methodology, results and recommendations. This has allowed future decisions to be based on facts, and provided an opportunity to improve the cohesion of a network served by more than 10 partners.

Decision tools

The decision over which planning tool to use has to be made carefully on a case-by-case basis to deliver results with the desired content and quality.

Our own results have shown that the methodologies described above generate realistic results with comparatively little effort. Results are usually accepted by stakeholders if the methodology is transparent and statistical, and the stochastical methods are scientifically proven. But above all, operational and logistical expertise is essential to allow an informed decision between the modelling alternatives.