Principle

The driving pattern, e.i. the speed over time diagram, has a considerable influence on the energy consumed by a train on a given trip. For given restrictions (timetable, stops, speed restrictions on the way and installed traction power) a shortest time driving strategy can be determined, which is basically given by

• Full acceleration up to maximum speed given either by speed limit or by maximum traction power
• Speed holding at maximum speed until train has to start braking
• Braking at the latest possible point in order to come to a stop when reaching the station

Time tables usually include a recovery time added to the minimal running time to allow for short delays. This recovery time is normally between 5 and 12% of the minimal running time.

This recovery time allows to apply different driving strategies which save energy in comparison with the shortest time driving strategy. Depending on their experience and skill, many drivers have always used timetable buffers to apply a more energy efficient driving style. Today, driving advice systems (DAS) exist that calculate (and continuously update) the optimum driving strategy much more exactly than any driver could. They are based on train positioning (GPS, Galileo or other), train, track and timetable data as well as algorithms to calculate driving recommendations.

The situation in suburban networks

In suburban service, good saving effects can be achieved by coasting only. In contrast to driving advice systems in main line operation, a continuous calculation of the optimal train speed on the remaining trip to the next station is not reasonable, given the short average distance between stations in local networks. Driving advice systems for local networks can therefore be limited to giving coasting recommendations, telling the driver when to shut off the traction motors in order to reach the station on time. Usually a suburban DAS only foresees one coasting phase directly before the next stop but no intermediate coasting on the way (as may occur for DAS in main line operation).

Components of a DAS

A DAS requires essentially the following on-board components:

• Storage medium storing all the relevant data for an individual trip (infrastructure data, vehicle data, time table)
• Information system monitoring driving time and train position
• Computer unit using the above data to determine driving strategies and display them to the driver

Several issues have to be resolved in a satisfying manner in order for DAS to be operable, reliable and safe:

Train positioning

Train position is essential for the calculation of driving recommendations. For driving advice systems on main lines a number of methods for train positioning exist. In suburban networks the wheel impulse counters (odometers) are usually sufficient. They count the number of revolutions of several driven axles. In order to correct errors caused by wheel slippage the position is updated at regular stops in stations. This method is very accurate in local operation due to frequent updating in stations.

Data supply

A DAS requires different classes of data to be updated in different time intervals:

• Permanent data: Vehicle data
• Long-term data: Track data base (to be updated annually)
• Mid-term data: Time table
• Short-term data: Data on temporary low-speed sections (to be updated daily or even in real-time), future options may include actual weather and track conditions as well.

The Siemens Metromiser

Siemens and the Technical University Berlin have developed the Metromiser, a driving advice system for for light-rail, suburban and metro systems. The algorithms were partly taken from earlier projects at University of South Australia at Adelaide. The Metromiser consists of two components: an off board timetable optimiser (TTO) and an on-board unit (OBU):

• The timetable optimiser is an off-board based software program (running on Windows NT) checking the energy efficiency of timetables. Using basic data (acceleration, rolling behaviour of the train, topology, passenger flows etc) it draws up a new energy-optimised timetable fitting in with the existing running schedule of the railway network.
• The on-board unit converts the data into driving recommendations such as: “make full use of permitted speed”, “coast” or “brake”. The OBU has learning capacity: the specific rolling qualities of the individual vehicle (which may differ even within one vehicle class) are recorded and used for the calculation of further recommendations.

All data are stored on-board on a flash disk. They are automatically updated via a W-LAN at several stations.

Punctuality

In-service testing in different systems have shown clear improvements in punctuality.

Wear

Coasting instead of braking reduces the wear of the brakes especially if electric braking is not possible.

Passenger comfort

Small improvements in passenger comfort can be achieved through smoother driving. In addition, early train arrival often leading to trains waiting outside stations (a situation hardly comprehensible for passengers) is ruled out by DAS.

Acceptance by drivers and management

Some drivers will be reluctant to go by the recommendations of a DAS for several reasons:

• If drivers experience train delays caused by insufficient or wrong infrastructure information in the DAS (e.g. temporary low-speed sections), he will no longer follow the system’s recommendations.
• Some drivers may experience DAS as a step towards automatic train control and therefore as a step towards losing their jobs.
• If drivers are not intensively familiarised with the DAS, many will not understand the system and will take the recommendations as irrelevant additional information.

Some decision makers think that acceptance by drivers puts too much uncertainty to the effectiveness and thus payback of the system.

Delayed trains

The effectiveness of the system is reduced if train punctuality in the network is low, since the DAS has no effect in this case.

Success factors concerning the network

• A pre-condition for the use of DAS is the existence of time buffers in the timetable. The saving effect of DAS increases with increasing time buffers.
• The DAS can only take into account predictable stops or decelerations (stops in stations, slow-speed sections etc.). If unexpected stops at red signals due to traffic situation occur, driving advice systems are uneffective and may even lead to delays. DAS are therefore especially well-suited to for systems with homogeneous service, e.g. suburban systems with few crossings and junctions.

Winning drivers’ acceptance

• Special training programmes have to be given to drivers
• Communication of the new system should focus on the purely advisory character of the system and stress that the system supports drivers to fulfil their core task being punctuality.
• Human-machine interface has to be optimised and the information presented in a clear and functional manner.

Integration into existing ICT infrastructure

When developing or introducing a DAS, synergy effects with existing hardware platforms as well as positioning and data supply systems should be sought in order to reduce costs and overall system complexity. Furthermore, DAS should be as far as possible platform-independent. The Metromiser was developed for a platform which is no longer produced by Siemens. Therefore the system has to be modified now to be compatible with a new platform currently introduced by Siemens.

Comparing a DAS guided driving strategy with a shortest time driving strategy one gets the theoretical saving potential. For typical recovery times offered by the time table it lies between 20 and over 30 %.

The real saving potential is however considerably lower, for three reasons:

• Even without DAS drivers do not pursue a shortest time driving strategy, but use a more energy efficient driving style according to their experience, skill and motivation. If there are no delays this may cut energy consumption by 20%.
• Due to delays the DAS often cannot be used. This may well be the case for about 50% of the trips.
• Some drivers may ignore the advice due to the purely advisory character of the system:.

Taking these factors into account, the remaining potential in suburban transport lies between 5 and 20%. In some networks, the potential is even higher: The two-year test on the suburban lines in Adelaide yielded savings between 12 and 30% depending on the line. Siemens claims that the energy savings achieved with the Metromiser are an average 10% in existing timetables and with an additional timetable optimisation even an average 15%.

Hardware

The price of the on-board units will show certain, but no major scale effects, if DAS become a wide-spread application.

Software development

Major scale effects in mid-term.