A Railway’s safety philosophy is an essential factor when selecting an ATP/ATC system. Another important issue is the effect of the train control system on line throughput and network capacity.
Even assuming the existence of a well designed signal system, with signal spacing appropriate for line speed etc. the introduction of an ATP system usually requires headways to be increased, whilst still maintaining overlaps. This apparent contradiction stems from the fact that where previously an ideal driver may have been assumed in signal layouts, the introduction of ATP usually is the consequence of a more explicit risk evaluation concerning train and driver failure modes, combined with the requirement to prevent these risks causing accidents to a much higher degree of certainty. Secondly of course the introduction of technical systems will always introduce trade-offs between response times, availability, delays caused by spurious or erroneous interventions etc. Technology will almost always be less adaptive and flexible, especially in the case of vital systems, and limit the human operator’s degrees of freedom. On top of that, in the design, as in any technical system, (safety)margins are introduced and enforced.
ATP enforces line speeds and in most modern systems supervises brake curves. That limits a driver’s ability to make up time.
Defensive driving techniques (with or without ATP) involve. not entering the platform at maximum speed and coming to a stop within 1 m of the overlap or departure signal with screeching brakes etc. ATP enforces this.
Rolling stock engineers asked to specify guaranteed brake performance of their trains “to SIL 4 requirements” will incorporate their own safety margins. Engineers devising slip/slide algorithms required in position determination as part of brake curve supervision, will try to err on the safe side etc.
In this chapter we examine the capacity effects and a number of strategies to mitigate them or even improve line capacity using ATP/ATC systems. In practice their efficiency would depend on the proportion of the rolling stock being equipped with the required ATP/ATC capabilities.
In very simple intermittent ATP systems such as AWS or Indusi, the train-stop system only applies the emergency brake when the driver does not acknowledge a warning or passes a signal at danger. Therefore an overlap is required to allow a train to be stopped within the distance to the danger point that is being protected by that signal. In most of those systems a (standard) overlap is used, which for example is nominally 200 yards in the BR system. This overlap reduces line capacity, although of course that overlap would also have to be required and perhaps even have to be longer, on a non-ATP railway.
In German signalling systems, the overlap is only protected when the train is approaching the signal and is still at a considerable distance. Depending on time elapsed or position of the approaching train, the overlap can subsequently be released. This certainly improves the headway of trains, but on the other hand it increases the danger of an accident when the overlap has been released and the train is erroneously accelerating again.
Brake supervision and Release Speed
In ATP systems that check if the brakes are applied after receiving a “caution” or “prepare to stop” type of command from the trackside, the driver can no longer immediately react on an improvement of the signal aspect and release the brake or accelerate as he would have been able to do in the absence of ATP. To compensate for this some systems allow the driver to override, or release, this brake supervision after observing the signal aspect ahead improving. This of course has to be weighed against the risk of errors of perception etc as this override can easily become a reflex due to its repetitive nature especially on densely trafficked lines.
If supervision of braking curves is implemented, the brake can be released either when the speed is below a predefined value, or the driver has taken special a action, for example pressed a release button, or after the signalling information is updated by a new transmission from the trackside. This release mechanism is needed to avoid a deadlock. Without it the train, which is usually intended to stop short of the signal and hence the new transmission point, would not be able to resume driving once the signal ahead clears, as the train borne equipment has no knowledge of the improved signal aspect until the train can actually drive over the next transmission location and pick up the new information. The release speed can be a fixed speed or it can be a calculated one, depending on the train characteristics and the length of the overlap. With short overlaps, the release speed is quite low and the continuous supervision of the braking curve seriously restricts the throughput of high density lines.
The second option, the manual release of the brake or brake curve supervision by the driver, is introduces risks, as drivers can be misled by observing the wrong signal clearing, e.g. on the approaches to a station where many parallel tracks can exist and signals may be mounted on a common gantry. Although it is sometimes argued that recognising which is “his” signal is part of the driver’s route knowledge we should never not forget hat ATP/ATC systems are introduced to reduce the risks caused by human error and so dependence on human actions or perception should be avoided as much as possible in their design.
Modern intermittent ATP systems enforce maximum speed and provide full brake curve supervision. Besides transmitting signal information, balises act as location references and allow for compensation of long term errors in odometry (speed and location measurement). Long term errors caused by wheel wear etc. can calibrated in workshops or can be compensated for using the distance references the balises provide, but the short term error, introduced by slip and slide of motored and braked wheelsets, is difficult to detect and compensate for. Recent systems can use inertial navigation, Doppler radars etc. to reduce this error.
On higher density lines, to limit the capacity penalty, an intermittent ATP system may need to be provided with an early signal information update mechanism. This is usually described as providing “infill information”. Its purpose is to overcome the problem of a train being restricted to the braking curve imposed by a distant signal or similar, even after the signal it is approaching has cleared. It can be achieved either by using a supplementary semi-continuous data transmission or additional intermittent data transmission points in rear of the signals. Infill information can be transmitted by spot transmission systems, i.e. balises, or by (semi-) continuous transmission media. Examples of semi-continuous infill transmission are cable loops, leaky feeders, (local) radio etc.
The first function of infill information is to allow a train standing before a red signal to start up again after the signal clears, without having to creep up to the balise at that signal before its ATP information can be updated and the brake curve / release speed supervision “ended”.
The second function of infill information is to prevent trains approaching a signal just before it clears, to be or stay caught under the brake curve. A well designed signal system and timetable tries to avoid trains being checked by signals and in simple terms, a driver would hope to see signals improving from a caution to a proceed aspect just before he reaches it. Experienced drivers will try to regulate train speed to achieve that. Any disturbance will result in trains being checked by brake curve supervision and employing infill information will help mitigate this effect. Experience gained with the Dutch ATBNG system, as well as in the UK ATP pilots on GWML en Chiltern lines indicate that infill information will be required at around 65% of the equipped signals. Where and over which length infill information should be available varies with traffic patterns and is timetable dependent. This is especially true if infill balises are used. Usually infill information over distances of 300 – 500 m is sufficient.
The obvious benefit of continuous ATP/ATC systems is that they can be thought of as continuously providing infill information. They offer a way to implement ATP / ATC on a well designed railway whilst minimising capacity sacrifices. An added benefit is that they also provide a means of informing a train driver of a signal reversion or revocation.
Continuous systems largely reduce the ATP/ATC capacity penalties, and in some cases they even can improve headways. This is mainly due to the quasi instantaneous update of signal aspects information and speed/brake curves to be supervised they offer.
Examples where the introduction of continuous ATP / ATC systems has had a positive effect on capacity, are usually related to retrofitting existing lines, where signal layout etc. is sub optimal for modern rolling stock. On some German lines for example, special train categories are allowed to run at higher speeds than allowed by the original distance between distant and main signals. In this mode of operation, these trains equipped with “better braking”, when running under full LZB supervision, the in-cab signalling takes precedence over the line side signals.
Raising Line Speed
In most cases raising maximum line speed involves introducing multi-aspect signalling, increasing block length and / or increasing the distance between distant and main signals. If it can be shown that modern rolling stock, e.g. through additional track brakes, has a guaranteed brake performance for those higher speeds still commensurate with the existing signal spacing, that cost can be avoided. Trains equipped with a continuous ATP / ATC system, usually employing in cab signalling and full speed and brake curve supervision can then be allowed to travel at the higher maximum speed under the supervision of such systems.
Examples of this strategy can be found in Germany where line speeds were raised to 200 km/h for certain train classes when they are under full LZB supervision. In that case the in-cab signal takes precedence over the lineside signals.
In a similar manner, in-cab signals can be used to provide a longer “electrical sight” and maximum speeds can be raised without requiring multi-aspect block signals to be installed.
Both strategies can also be implemented using an ERTMS level 2 system.
It has to be noted that some railway administrations / safety authorities do not approve of the use of in-cab signals taking precedence over lineside signals, as it involves different interpretations of signal aspects depending on train type. The obvious hazard is that a driver would mistakenly assume the higher line speed was allowed on a non-equipped train.
Increasing Block Density
The advantages of “mixed-signalling” systems can be taken even further if we want to increase line capacity by sub dividing blocks on a line to reduce headways, either because the original block length was too long for the required line speed and traffic demand has increased, or as before, if rolling stock development has delivered better braking capabilities.
In a conventional ATP / ATC systems this still involves installing additional train detection equipment in the subdivided blocks. In ERTMS level 2 (and similar systems) instead of erecting real signals, we can now use “virtual signals” at the (new) block boundaries. These block boundaries are usually marked by a balise and a sign and the virtual signals show their aspects, or rather imdicate their movement authorities through the in-cab signal.
In ERTMS level 3, if and when that implementation of the moving block principle becomes available, it might even be possible to dispense with the additional balises and / or leave out the additional train detection equipment. Examples of this can be found in Germany on the so called “Neubaustecken” implementing a concept known as CIR-ELKE.
For headway improvement, the moving block principle can be applied. In moving block systems, each train reports the position of its train end to a central location/ coordination unit, which issues and transmits a movement authority to each train in the system, using the reported position of the end of the preceding train as a target point. The moving block principle requires several features:
• Centralised track side unit
• Accurate positioning of trains
• Train end supervision
• Bi-directional vital data transmission
The moving block principle is frequently applied in urban mass transit systems, where low speeds and short headways combined with uniform train characteristics make it very effective. In mainline traffic, the advantage of the moving block is mainly related to cost, as the railway can save on investment and maintenance for train detection systems, as these no longer need to be track based. However, these systems require the end of the train position to be reported accurately and safely. Trains can and do break occasionally and therefore the separation of even a single car or wagon from a train must be detected and train end position reporting must take account of this immediately. The problem of “train integrity proving” as this requirement is often referred to, has not been solved satisfactorily for loco-hauled trains and especially for freight trains.