The TRAKATTK™ Model
A Cost-Effective Simulation Tool
There are a number of operational simulations available to examine train running, rail signalling and infrastructure issues. Some are very simple and unable to replicate the real world; others are extremely complicated and their use is only likely to be justified on the largest schemes. We believe that our model, based on the TRAKATTK™ software, enables us to have the best of both worlds - it is sophisticated enough to examine key issues on multi-user railways, without taking an inordinate amount of time to code up. This therefore enables the modelling of a wider range of options, within the typically-tight timescales needed in the modern railway environment.
Results are available (within weeks, not months) at the level of the individual train, with repeated runs of the simulation using different random latenesses of trains being used to analyse the impact of a timetable over an operating period. The visual simulation is useful for demonstration, particularly to those outside the railway, for whom theories of service bunching etc. may not be so comprehensively understood.
The model, as with other operational simulations, requires two types of inputs, relating respectively to the infrastructure and to the timetable to be operated.
Infrastructure data is normally taken from track plans (which of course need to include distance information, rather than being merely schematic) or engineering drawings, supplemented by line speed information. These enable routes to be set up, and conflicts to be identified. In addition, however, information is normally needed about the signalling system; for instance, the simulation can cope with approach control, but needs to be coded as to which signals this applies to.
The simulation also needs to be programmed to accept relevant stations or freight yards/sidings (whether on or off the screen) as timing points.
Timetable input is arranged through an editor with a 'cut and paste' or 'clipboard' function (see below). Trains are timed from the moment they enter the simulation; three columns represent (respectively) the name of the service, its stopping pattern (with times entered cumulatively), and this time of entry into the simulation. Such a format is invaluable in enabling hourly services to be entered easily, since only the entry time varies. Each service also has characteristics of train length and maximum speed.
Once the timetable has been set up, the programmer checks the requirements for auto-routeing, and then the simulation is available for use.
The model currently works from a DOS prompt (or equivalent) which gives access to the TRAKATTK™ environment. This is formed of a shell, which contains access to the range of available simulations (for operation) and to the timetabler (for editor). The timetabler contains access to various timetables for each of the available simulations.
Internal calculations reflect 30-second increments of real time. Display of these to the screen can be speeded up, to a maximum of 8 times real time.
Small delays are added to station stop times on a random basis, to reflect the numbers or actions of passengers boarding and alighting. This is important, to ensure that timetables reflect 'normal', rather than perfect, operating conditions.
The speed of the simulation naturally depends on the performance of the pc involved, but the program is not particularly intensive, and the 800% option enables a visual overview to be gained quickly. Batch operation (see below) occurs more slowly, and may take several hours to complete a range of analysis.
As with any model, results are obtained by examining the differences between (infrastructure and/or timetable options). The costs (capital and/or operating) need to be compared with the benefits (measured in terms of service reliability), especially where train operators have a performance regime with a railway infrastructure supplier.
The simulation can, of course, be watched to the screen. This is very valuable in demonstrating results to a non-technical audience, and does add considerably to one's understanding of how the railway in a particular area runs (or doesn't!).
However, every time a delay is caused (either by the random station stop time element, or by train movements conflicting), the results are written to a file, as are incidents of early running. The total delay during the simulation is also recorded to the screen, for those watching.
Because a vital part of timetable planning is to ensure that day-to-day variation can be accommodated, a batch facility is available. This enables the user to program a number of iterations (and of different simulations), leaving the pc to continue the calculations. Results can then be examined across a range of operating days, with the mean and maximum level of delay both being variables of particular interest. Data should be collected from a smaller time-period than that run, in order to avoid the period in which the simulation is being started.
TRAKATTK™ is able to provide a quick but realistic understanding of the operation of a rail system of a complete area, both visually and analytically. Its cost-effective nature means that results (including an infrastructure file and timetable variants) can be obtained for £5,000-£10,000, which is considerably less than for many of its competitors. Early uses of the model have included the Great Eastern main line between Shenfield and Colchester, and the rail network of North East Wales. For further information as to how the simulation could usefully be applied to your rail system, contact Dr Nigel G Harris or Mick Cordon at The Railway Consultancy Ltd.
Last updated: Sunday, 18th December 2016
© 1996-2017 The Railway Consultancy Ltd