THE railways linking the vast iron-ore mines in the arctic north of Sweden with the ports of Luleå in Sweden and Narvik in Norway are a critical link in the supply chain for the country's mineral industry. The efficiency of these operations is therefore vital to the global competitiveness of Swedish iron-ore. Mining company LKAB has higher rail transport costs than competitors in Brazil and Australia, so it has launched an ambitious programme to develop the world's most efficient ore railway, with a programme of investment in higher capacity wagons, new locomotives, IT-aided planning and control, and improvements to infrastructure and terminals.
While greater efficiency is being achieved through fewer, heavier and longer trains, new driver assistance technology also has a part to play in bringing down operating costs by reducing energy consumption and optimising use of available capacity on the largely single-track line.
Following a successful trial, LKAB announced last August that it would roll out the Computer-Aided Train Operation (Cato) driver advisory system on all of its iron-ore trains.
Jointly developed by Transrail, Sweden, and Swedish infrastructure manager Trafikverket, and co-funded by LKAB, Cato uses radio communication to transmit operational parameters such as the operational timetable and train status in real-time between the control centre and the train. The advisories displayed on the driver interface in the cab enable the drivers to keep the operational timetable with very high precision while reducing energy consumption. These speed advisories are easy to follow, and can easily be adapted to suit heavy-haul operations in the extreme conditions encountered for much of the year on this railway.
During 2010, Cato was installed in two of LKAB's Bombardier Iore electric locomotives and was also integrated with the train dispatching system at Trafikverket's train control centre. Thereby, the plans and operational timetables made by the dispatchers could be used for optimal control of the trains.
The energy savings generated by any driver advisory system will largely depend on the traffic situation, line profiles, and type of trains on which the systems have been tested. Punctuality characteristics on the line and available running time compared with the fastest possible run, and utilisation of locomotives for additional tasks such as shunting will also have an impact.
The driving advisories calculated by Cato are based on the actual timetable that is valid in the current operational setting, including any traffic disruptions. Advisories are based on actual train characteristics, such as tractive/braking efforts, length, and weight. The advisories are updated whenever it is necessary, for example if the timetable is changed or if the driver has not followed previous advisories.
Preliminary results from the tests performed on LKAB trains showed energy savings of 20-25%. Figure 1 shows a typical example of energy measurements from point-to-point runs between two fixed stations (ie the train stops at both stations). Only runs with no intermediate stops are considered. The theoretical minimum energy use (as calculated by Cato) is also indicated and depends on the available running time. As the graph shows, there is significant variation in the energy consumption of trains operating without Cato.
Cato will always try to use all the available run-time so that energy consumption is minimised. Conversely, drivers without Cato tend to run faster than the available time, thus arriving unnecessarily early. It is easy for the driver to operate according to the advisories. The speed curves recommended are smooth and gentle, taking into account parameters such as application of traction power, braking and gradient. Figure 2 illustrates the deviation between Cato advisories and actual operation in metres per second. The diagram shows that 79% of the time, the deviation between advisory speed and actual operation was less than 1m per second. The fact that Cato is used in normal operations and not under specific test conditions, together with the irregular stopping patterns of the trains, means there are relatively few comparable observations for the Cato train for each pair of stopping points.
In any decision support system, it is important that data is correctly entered, but some data may be difficult to estimate correctly. Yet even with some data errors, the final advisories may be of good quality, showing that the system will tolerate some incorrect data.
Figure 3 illustrates the correlation between the error in the train's estimated running resistance and the energy consumption when Cato's speed advisory is followed. For example, the diagram shows that if there is a +20% error in the estimated running resistance, the deviation from the optimal energy consumption is only 0.6% for loaded trains and 0.4% in total for empty and loaded trains.
The calculation of advisory speed curves can easily be extended to take the current driving conditions into account, like reduced adhesion or poor weather conditions.
The use of target points is a key concept in Cato. A target point is an instruction to the train regarding the exact location and speed that the train should reach at a specific time. The target points are defined by the crucial points in the timetable. For example, target points can be created to make a loop on a single-track line operate more efficiently. The train due to stop in the loop receives a target point specifying the latest time it should reach the stop signal, while the passing (non-stopping) train gets a target point specifying the earliest (and latest) time it should reach the loop.
For the Cato installation at LKAB, the target points are based on the detailed, operational and updated timetables made by the train dispatchers. As the timetable is updated, information is instantly communicated to the train so that driving instructions are always optimal.
Figure 4 illustrates the deviation between the target point time and actual arrival time. All target points suggest that a train should stop. In 56% of observations, the deviation from the target point was less than 20 seconds.
In only 6% of observations, the train arrived more than 10 seconds after the target point time. In 22% of observations, the train arrived more than 1 minute before the target point time. This is partly caused by the fact that it costs more to run more slowly. Cato also forbids excessively-slow speeds. Cato allows trains to arrive too early at stopping target points, but never too late. On the other hand, target points corresponding to non-stopping trains at loops forbid early arrivals.
Such precise timing of trains generates many positive benefits. The capacity utilisation of the railway can be increased because the train dispatchers can be more proactive in their planning, and unnecessary signal stops are avoided. Trains running in the same direction can avoid running into the signal shadow of the preceding train.
The architecture of Cato is shown in Figure 5. The solution consists of two parts, the Cato-TCC module at the traffic control centre and the Cato-Train onboard module, which communicate by digital radio. Cato-TCC is linked to the traffic dispatching system and sends running instructions, expressed as target points, to the train. As described above, the target points are interpreted from the operational timetable, which is constantly updated by the train dispatcher.
Communication with the traffic control centre allows assessment of the real-time traffic situation on a line, rather than just the planned timetable. This means trains run in accordance with the actual traffic situation. The optimal speed profile on a section of line will vary for different trains and depends on the available running time, which may vary from day to day. The optimal speed profile may in fact be very different from the speed profile normally chosen by the driver. Furthermore, it ensures that the train will arrive just in time.
Cato-TCC allows the traffic dispatcher to control the movement of the trains, and both dispatchers and drivers can trust that trains will arrive at the target points on time. Unnecessary braking and stops can be avoided. The train can also notify the dispatcher if it cannot reach the target point on time, which means the dispatchers can update the target points to make them attainable.
Cato-Train continuously calculates the optimal speed profile to ensure the train reaches its target points on time, with the running speed profile displayed as an advisory on the driver interface. In fact, any optimisation function can be applied, for example operation without regenerative braking, or minimum use of mechanical brakes.
The speed advisory can also take into account many aspects that are specifically relevant to extreme conditions and heavy-haul. For example, acceleration can take into account drawbar force on heavy trains, maximum traction and braking efforts can be reduced in situations with limited adhesion, and appropriate speeds can be maintained in curves.
Interface
The development of the driver interface has taken the following factors into consideration:
Safety At the very basic level the curves have to respect technical limits and tolerances set by infrastructure managers and vehicle technicians. For example, both permanent and temporary speed limits are considered in the driving advisories.
Ergonomics A pleasing layout designed with carefully selected colours and fonts, with careful consideration of the driver's working conditions. The Cato interface has been developed in cooperation with the department for Human-Computer Interaction at the University of Uppsala, Sweden.
Simplicity Experience shows that drivers can easily and accurately follow recommendations with little or no prior training, and
Comprehensiveness The system needs to present relevant and dynamic information about surrounding traffic and the line.
In conclusion, Cato has generated very positive results for LKAB, and it is now being rolled out across the Iore fleet in a SKr 50m ($US 7.2m) project which is expected to generate annual savings of nearly SKr 12m. The trains run as stipulated by the train dispatchers and with very high on-time performance. At the same time, the margins in the operational timetable are utilised for eco-efficient running, and energy consumption is therefore reduced by 20-25%. Furthermore, the communication between the dispatchers and the drivers is improved, as drivers are always up-to-date with current plans.
The implementation of Cato, or interoperable systems, is foreseen across the entire Swedish railway network. As the use of driver advisory technology becomes more widespread, we expect to see several suppliers offering the target point technology. It will therefore be essential to standardise these systems so trains from different operators can run smoothly over long distances, across several traffic control areas, supported by target point control. Within the European Union Railenergy project, the Energy Efficient Train Operation (EETROP) standard has been developed with the aim of ensuring interoperability between different suppliers' systems.
