This paper was presented at a t the IRSE International Conference ASPECT99, London, September 1999
TRAIN DETECTION BY TRACK CIRCUIT THE EFFECT OF THE WHEEL/RAIL W HEEL/RAIL INTERFACE
R A Wood AEA Technology Technology Rail, UK SYNOPSIS
The wheel/rail interface interface plays a vital part in the correct functioning functioning of a track circuit. Unlike other track circuit components, it cannot easily be specified or accurately measured, as so many factors may affect it. Furthermore, the characteristics of this component can change unpredictably. Improvements made to track and rolling stock have had a profound (and sometimes detrimental) effect on the electrical characteristics characteristics of the wheel/rail wheel/rail interface. interface. At the same time, the track circuit circuit has grown from being an aid to the signalman into the primary safety system it has become today. This paper explores the factors which determine the electrical characteristics of the wheel/rail interface, and the remedial measures which have become necessary to cope with its deficiencies. Finally, an attempt is made to predict where problems may occur in the future. SYMBOLS AND ABBREVIATIONS
DMU S&T TCA
UIC ORE BRB 1.
Diesel Multiple Unit Signal an and Telecommunications Track Circuit Assister (Also known as a Track Circuit Actuator; TCA is a United Kingdom registered trade mark of AEA Technology plc) International Union of Railways Offic ffice e for for Rese Resea arch and Exper xperim ime ents of the UIC British Railways Board
HISTORICAL BACKGROUND
First Application of Track Circuits
The track circuit was originally developed in the steam age, as a secondary safety device for situations where a signalman might forget about the presence of a train. It was used as a backup backup to the operating procedures which, if followed correctly, rendered the track circuit unnecessary. Early track circuits did not have to contend with any electrification systems, and in most cases a simple dc track track circuit would would suffice. A low voltage voltage was used for the simple reasons that it was adequate for the job, and a higher voltage was wasteful in power. The only suitable power source was a heavy primary cell, which had to be physically replaced or refilled on site when discharged. The wheel/rail interface was, most of the time, a clean metal-to-metal metal-to-metal contact. Poor vehicle AEA Technology plc
dynamics meant that most of the time this was a wiping contact rather than just a rolling contact. Coupled with the heavy axle loads and rough wheel surfaces of tread-braked rolling stock running on jointed track, the wheel/rail interface presented a near-perfect electrical interface. Railhead Contamination
When track circuits were first introduced, each train had a heavy locomotive at its head and traffic was frequent. frequent. The railhead railhead was mostly free from rust and other forms of contamination, and track circuit operation was generally generally satisfactory. The properties of rust and leaves have not changed between 1909 and 1999, so why has track circuit performance deteriorated in the meantime? Measuring Track Circuit Performance
The signal engineer’s responsibility for commissioning and maintaining a track circuit generally went only as far as confirming that the shunt resistance required to drop the track circuit was above above a specified specified limit. In the UK, this has always been 0.5 Ω. This limit was presumably presumably based on early measurements of trains on a typical rail surface: the value used goes at least as far back as 1940 (1) (1).. That same reference also provides an early argument that the drop shunt is “...not a true measure of the capabilities of the track circuit unless the voltage between the rails is also given”. The drop shunt value is, in practice, a compromise between many factors, and determines the maximum length of track circuit for Page 1 of 8
a given ballast resistance. Other railway administrations use different values, for instance Dyer (2) quotes the US 0.06 Ω drop shunt, and some other railways used intermediate values. In the UK, a reduced value has been accepted for some track circuit types where the normal value cannot be achieved. The drop shunt test indicates that if the rail and wheel surfaces are clean, forming a good metal-tometal contact, then the track circuit will work. It does not prove that the track circuit will operate correctly when a train is present, since it does not test the condition of the wheel/rail interface.
conditions. A warm, humid climate will encourage rapid rust formation, which may be further encouraged by chemical contaminants present in the air. Although rust can also form on wheel surfaces, this is not a big problem because it will rapidly be removed by the normal movement of the wheel. Unless one alters the chemical composition of the rail surface or the atmosphere, little can be done about the initial rust formation. The effect on track circuit performance depends on the thickness of the rust film, and whether it has been rolled by the passage of wheels. A rust film which has been allowed to grow for a long period will exhibit insulating properties: a relatively high voltage may be required to break down this insulating film. This is the conventional insulation breakdown mechanism, where the insulating film is ultimately punctured at a microscopic level and becomes conductive. The rust film will be matt red-brown in appearance, and according to Graver and Kendal (3) may be typically up to 50 µm thick. Heavy (unrolled) Rust.
The only true indication of track circuit performance must be measured dynamically, during the passage of a train. This is never done during normal maintenance, but may be carried out by specialist staff after a wrong side failure has occurred due to suspected problems at the wheel/rail interface. Dynamic measurements on contaminated track have been carried out on numerous occasions, though reports on these tests are not in the public domain. The presentation and interpretation of the measurements are critical to the understanding of the track circuit performance. This type of measurement is fraught with problems, and it is often difficult to repeat a measurement with any consistency since any subsequent train movement will modify the railhead condition. 2.
CONTAMINATION OF THE INTERFACE
This paper discusses contamination of the “wheel/rail interface” rather than “railhead contamination”. This highlights the fact that contamination exists between two metallic surfaces: whether it adheres to the rail or wheel – or indeed to both – is immaterial as far as the track circuit is concerned. The characteristics of the different forms of contamination are very different. Contaminants may include rust, leaves, sand, brake materials, metal particles, grease, oil, and water. Some railways may have specific problems with other contaminants, for instance a railway administration in Australia indicated that it had short stretches contaminated with pine needles and occasionally crushed mealy bugs. Rust
Rust occurs naturally on the railhead, and ‘grows’ at varying rates depending on the climatic AEA Technology plc
When a number of wheels have rolled over a section of rusty track, the characteristics of the rust film change significantly. The number of wheels required to effect this change depends on many factors, and it is important to take into account the effect of the vehicle dynamics which are discussed later in this paper. The rust will be compressed by the wheels, causing a partial fusing of the rust layer. The rust film will take on a dark glassy appearance, and may be only 5 µm thick (3). From this point on, expert opinion as to what happens begins to diverge. Rolled Rust Film.
Early research carried out by ORE (4) identifies the rolled rust film as an indeterminate combination of linear (resistive) and semiconducting components, although a later ORE report (5) refers only to breakdown voltages, implying an insulating layer. The variance may be because the ORE reports do not clearly differentiate between rolled and unrolled rust. Short (6) reflects the BRB S&T philosophy of the early 1990s when he refers to the characteristics as that of a Zener diode. In practice, the actual characteristics of the film vary from point to point as the wheel moves on to a new contact patch, and all of the characteristics may be present in some form. The semiconducting phenomenon can be thought of as a bi-directional diode of indeterminate characteristics: the author does not consider that attempting to characterise the rust film as a particular type of diode is feasible.
Page 2 of 8
The mechanism for leaves becoming attached to the railhead is well described by Pearce and Watkins (7). Leaves are shed in the autumn, and fall onto the trackbed in large quantities. In dry weather, the turbulent air around a passing train lifts the leaves, which are then sucked underneath the train. Some of them are trapped by the wheels and squashed onto the railhead. The leaves eventually cover the whole of the railhead in a glassy black film, which exhibits a low adhesion coefficient when it becomes damp, and a high electrical insulation resistance when dry. For track circuit performance, leaf film is now a major problem. The three types of curve illustrated above show the typical range of characteristics (which may be present in any series/parallel combination). When the semiconducting characteristic ‘A’ is present, a small increase in applied voltage will cause a proportionally greater increase in current, and this can be very effective at only a few volts. The linear characteristic ‘B’ is purely resistive. The breakdown voltage of the insulation breakdown characteristic ‘C’ (greatly exaggerated on the graph) can be very low with a rolled rust film, perhaps down to one or two volts. It can be seen that if any non-linearity is present, increasing the voltage across the wheel/rail interface will greatly improve the track circuit shunting performance of the train. If the applied voltage is dc, the effect will be continuous. If the applied voltage is ac, the effect will be dependent on the instantaneous voltage. For a family of breakdown curves like ‘C’ for instance, a higher applied ac voltage will increase the proportion of the ac cycle over which the film is conducting. Electric trains are normally considered immune from poor track circuit shunting on rusty track because of the high voltage available at the wheel/rail interface (anything between 650 V and 25 kV). However, an electric train may only require a small current when coasting, and this only has to return through one rail (correct track circuit operation requires good contact with both rails). If other factors such as good vehicle dynamics come into play, track circuit performance may be inadequate. Leaves (and other vegetable matter)
Leaves were rarely a problem in steam days, for the simple reason that flammable vegetation at the side of the track was incompatible with machines that ejected burning cinders into the air.
AEA Technology plc
Dry leaf film has a breakdown voltage of several hundred volts. It is impracticable to provide this voltage from any track circuit, and unsafe to provide a high voltage on the train specifically for this purpose. Electric trains have an in-built high voltage source, some of which will be present at the wheel/rail interface when leaf film is encountered. This is likely to break down the insulating film and enable the track circuit to work. Severe leaf film has however been known to cause track circuit failures under electric trains, for similar reasons to failures caused by rust film, which has been mentioned earlier. Unlike most other forms of contamination, leaf film can be transferred to wheel surfaces and cause the wheel surface as well as the railhead to be completely covered with an insulating film. In the UK, there have been instances where wheel contamination has had an effect at least 10 km after a train left the leaf-contaminated track. Sand
Apart from sand contamination blown from beaches and sand dunes, most sand is put on the track from the train, to improve adhesion. Traditional use of sand for adhesion improvement ensured that it was only used by a locomotive at the head of a heavy freight train. By the time the rear of the train had passed the sanded section of track, the sand would have been removed by the passage of many wheels. Modern freight locomotives are designed to lay sand whenever the traction system demands. This has led to a number of incidents where a locomotive running light has laid sufficient sand at low speed to apparently disappear from the track circuit. The sanding system can be switched off, but the layout of the cab controls does not normally lend itself to this being done while in motion. The problems of excessive sand were highlighted by a fatal collision between two trains near Sydney, Australia, in 1990. A preserved steam Page 3 of 8
train stalled on a gradient, and excessive use of sand combined with other factors caused the whole train to be insulated from the track circuit. A following electric train running at speed under clear signals struck the rear of the stalled train. Modern multiple unit trains in the UK (both diesel and electric) are being fitted with sanding gear for improved adhesion during braking, and some are now beginning to use sand in traction. Sand is applied ahead of the second bogie, so that the leading bogie will always be running on unsanded rail. Tests have shown that track circuit problems may be experienced when sanding at a fixed rate below about 15 km/h. For any modern sanding system, a key parameter is the rate at which sand is being laid. All current systems lay sand at a constant low rate regardless of train speed, whereas what is required (particularly at low speed) is a sanding rate which relates to the train speed. Other Contaminants
Sandite - a suspension of sand in a thixotropic gel - is applied under controlled conditions at locations which suffer from leaf contamination. Sandite is sometimes blamed for track circuit failures, however it is difficult to be certain whether it is really the cause, and the underlying leaf contamination may be the actual cause of the failure. In the context of “contamination”, composite brake shoe materials have been blamed for degrading track circuit performance. Although this may be true to a limited extent, the main culprit may not be the presence of a composition shoe but the absence of a cast iron shoe! Any contamination of the wheel/rail interface by metallic particles will generally improve the electrical contact. The main source of metallic particles at the wheel/rail interface is from traditional cast iron brake shoes. These improve wheel/rail contact for two main reasons. Firstly, they roughen the wheel surface, causing raised spots which will break through any other contamination - this will also happen to a lesser extent with composition blocks. This phenomenon is described by Armstrong (8). Secondly, small particles of metal will break away from the blocks and be crushed under the wheel. Grease and oil may contaminate the railhead: they are often deliberately applied to the wheel flanges or side of the rail to reduce friction. All railway administrations use lubricators to some extent, and excessive contamination can spread to the railhead. If excess oil or grease is present in AEA Technology plc
sufficient quantities to cause track circuit problems, it has normally already become known by the poor adhesion. Grease can also be present from overenthusiastic point greasing, and grease or similar slippery substances may be present from testing under low adhesion conditions (liquid detergent is sometimes used). In most cases track circuit operation is not affected. Not normally thought of as a contaminant, water on the railhead will improve track circuit performance significantly, particularly when performance has been degraded due to rust. Whenever tests have been carried out on rolled rust, there is a significant improvement when even a slight shower of rain falls. Once the rails dry out, track circuit performance reverts to its previous level. 3.
EFFECT OF ROLLING STOCK DESIGN
The development of rolling stock has always taken place in parallel with developments in the application of track circuits, but the signal engineer has had minimal impact on its design. When DMUs were first introduced into the UK, the BRB S&T department laid down criteria for the reliable operation of track circuits. Provided a vehicle had at least four axles, each with an axle load of at least 8 t, it was accepted that it would operate track circuits reliably. Specific rules and regulations were applied to vehicles which failed to meet these criteria. The second generation of “Sprinter” DMUs easily met these weight and axle load criteria, but incidences of track circuit failure were being identified with increasing regularity. With the introduction of the Class 158 DMU in 1989, the situation had deteriorated to such an extent that the unit could not enter service because of its poor track circuit performance. These units were actually slightly heavier than earlier DMUs despite having aluminium bodies (38 t instead of 32 – 36 t for a power car). The brakes were disc instead of tread. Armstrong (8) considers that the resulting lack of abrasion on the wheel surface would degrade the electrical performance of the wheel/rail interface, but this alone could not account for the tremendous change in performance. The main factor appeared to be the greatly improved suspension systems of modern bogies. Instead of bouncing, scraping, and sliding around the track, the wheel on a modern bogie runs straight and true along the rails, with a significantly reduced scrubbing action between wheel and rail. The bogie designer had achieved his ultimate aim of a smooth, quiet ride – but at what cost? Page 4 of 8
Interesting facts about track circuit performance began to emerge. If all services on a particular route were run by one class of DMU, substitution of another class for one service would cause the its track circuit performance to be much worse than the usual trains. It made little difference which way round the substitution was. The mechanism for this unusual behaviour is quite simple. The combination of excellent dynamics and smooth well-balanced wheels was causing the wheel to take up the same lateral position on the railhead when it traversed the same section of track at the same speed. Since the wheel normally only has a very small contact area, it was consistently running on the same narrow strip of rail. When a different class of vehicle came along, it would be running on a slightly different part of the railhead. An extreme case of this phenomenon was noted on Iarnród Éireann (Irish Rail) over twenty years ago, when the Class 071 6-axle diesel locomotives were introduced. When delivered, the locomotives failed to operate a number of track circuits when running light. On investigation, they were found to have been delivered with the wrong wheel profile. The wheels were thus running on a section of the railhead which had seen hardly any traffic since the rail had been laid, although the line itself had a frequent service. These locomotives weigh 99 t, so can hardly be classed as lightweight vehicles! As part of the UK remedial actions to cope with leaf-induced track circuit failure, the Class 158 was run as a hybrid set. This was effected by splitting the two cars and running each coupled with one car of another class. This had the immediate effect of increasing the surface area of the railhead which was ‘swept’ by the wheel tread, thus increasing the chance that one of the units would find a clean bit of rail to run on. It may also have caused the dynamics of the combined set to be degraded so that lateral wheel movement (and hence railhead abrasion) was increased. The lesson for signal engineers is not at all clear. Despite tests being carried out to quantify the effect of vehicle dynamics (8), it is not yet possible to predict with any accuracy what the track circuit performance of a particular vehicle will be. Neither is it possible to accurately predict how rapidly a new vehicle will remove rust or leaf film from the railhead. 4.
OTHER FACTORS TO BE CONSIDERED
Track Design
The introduction of continuous welded rail has AEA Technology plc
given trains a much smoother ride, but the reduced discontinuity at rail joints is likely to have an adverse effect on track circuit performance. If a section of jointed track is contaminated with rust or leaf film, as each wheelset passes a joint there will be a very high impact load as the wheel runs onto the next rail. This will cause the wheel/rail interface to have a very low resistance as it will break through any surface film and effect a pure metal-to-metal contact. The period of contact may be quite short, but the vibration set up by this “hammering” at the joint may well affect the contact at other wheels along the same rail. This phenomenon has never been formally tested. However, some results from track circuit tests on rusty rails show a cyclic reduction in residual voltage of about 40% corresponding to 60 foot (rail length) intervals – at only 15 km/h. Fortunately the track over which a wrong side track circuit failure is most critical – point and crossing work – also has the most discontinuities in the rail surface. This will tend to mitigate against any track circuit failures. Track Formation
Track formation below the rails has a similar effect to train suspension above. If the formation is hard and unyielding, such as a concrete bridge deck, track circuit performance will improve, with a measurable reduction in residual voltage particularly on rolled rust. If the formation is soft and spongy, such as over marshy land, this could conceivably tip the balance between acceptable and unacceptable track circuit performance. Track Circuit Timing and Sequencing
Track circuit timing does not directly affect the wheel/rail contact, but can mitigate (or mask – depending on the point of view) the effect of poor contact. If leaf contamination is intermittent, or if the track is jointed, a track circuit with very fast drop-away and very slow pick-up characteristics is quite likely to function correctly in adverse conditions, even if the rail voltage is low. Many audio frequency track circuits exhibit significant differential occupied/clear response times which can help maintain a track circuit in the occupied state. As an example, the Adtranz TI21 will drop in 50 ms, but takes 2 s to pick – a 40:1 differential. Evans (9) describes a method by which track circuit sequencing can be used to guard against intermittent track circuit operation caused by railhead contamination. In the UK, additional checks on track circuit sequences are applied Page 5 of 8
where automatic route setting is in use.
passed for a pre-determined time.
Level Crossing Design
Rolled Rust
Level crossings must be singled out for special attention. Firstly, they are more common on lightly used routes. Secondly, any delay in detecting a train on the approach will be much more noticeable than in any other location, and will have a greater impact. It is no coincidence that on two railways known to the author, poor track circuit performance was first identified as a serious problem at a level crossing. In one case there was a near miss, and in the other a collision with a road vehicle, both within a few days of the introduction of new rolling stock.
Conduction through rolled rust requires the voltage across the wheel/rail interface to be raised, but by a much smaller amount than for heavy rust. This enables the voltage source to be safely provided either from the trackside or from the train.
In the UK, automatic crossings are generally initiated by a track circuit backed up by a treadle at the strike-in point: this is not the case in many countries. 5.
SOLUTIONS FOR POOR CONTACT
Heavy (Unrolled) Rust
Where a section of track regularly has no traffic over it for days at a time, special precautions must be taken. If the section involved is short (such as a crossover or the end of a terminal platform), a narrow strip of stainless steel weld can be deposited on the surface of the rail. This effectively changes the chemical composition of the running surface, and prevents it from rusting. Rust will still form on the railhead, but the stainless steel will provide a rust-free electrical connection for the track circuit. This method has its limitations, in that the rough rail surface will cause poor riding, and the weld deposit will wear rapidly under heavy traffic. For longer track sections, a high voltage track circuit may be provided. The voltage required to break down the rust film will depend on the thickness of the rust film, and may vary from a few volts to several tens of volts. In all but the most severe cases, a track circuit which provides 80 100 V across the rails will be sufficient to ensure satisfactory performance. It is normal for track circuits providing voltages of this order to generate impulses rather than a steady voltage: this considerably reduces the total power requirement, and reduces the electric shock hazard.
The most logical source for this additional voltage is the track circuit itself. Increasing the rail-to-rail voltage at any point on the track circuit will cause disproportionately higher current to flow between the wheels, and the track circuit will ‘see’ a low resistance. It is also necessary for the track circuit to be able to maintain a reasonably high voltage (ORE (10) suggests 2 to 2.5 V) when the track circuit is occupied. A simple dc track circuit can be improved in this way by introducing a resistance in series with the relay, and making a corresponding adjustment at the feed end. This increases the rail-to-rail voltage and desensitises the relay by a corresponding amount, so for a purely resistive shunt there is no change. The wheel/rail interface however is now operating further along the V/I curve so on rolled rust there will be an improvement in shunting performance. This type of improvement is not new (1,2). Track–based improvement.
Where a track circuit has an ac supply the effect is similar, although it is the peak (rather than the rms) voltage which is significant. Rail impedance is higher, and the voltage drop along the rails (which is largely insignificant for a dc track circuit) must be taken into account. Many types of ac track circuit have a relatively high voltage at the feed end, reducing to as little as a quarter of a volt at the receiver end. For most types of track circuit, the critical parameter of rail voltage at drop-away is not clearly specified, which makes comparison between different types in this context rather difficult. Short (6) explains in some detail the rapidly varying voltage signature which is exhibited by trains running over rolled rust, and a differential time response can make a track circuit operate on the short ‘troughs’ of low voltage. A similar but less pronounced effect can be achieved by using follower relays or interlocking timings with a differential time response, although in most cases they were not provided for this purpose. A more radical means of improving conductivity of a rolled rust film at the wheel/rail interface is to provide the additional voltage through the wheels of the train. Many interesting developments have used this principle. It can however cause departmental Train-based improvement.
If neither of these solutions can be applied, then either the rust film must be prevented from forming (by running more trains over the track), or the operating procedures must assume that the track circuit will fail to detect a train if no traffic has AEA Technology plc
Page 6 of 8
headaches in many railway administrations, because it requires “signalling” equipment to be fitted on a rail vehicle - who pays for it and who maintains it? Around 1958, the Budd Company developed a novel means of providing this voltage for a number of single unit railcars running on the Boston and Maine Railroad in the USA (2). This system used a 150 V 400 Hz motor alternator to induce a voltage into the leading axle of the railcar. A pickup coil on the trailing axle detected this signal, and provided an indication to the driver that the system was working. A major drawback of this system was that the two wheelsets had to be electrically insulated from the vehicle body. In 1980, BR Research (now AEA Technology Rail) produced in conjunction with Leyland Bus a prototype design for a lightweight 2-axle railcar. An ac voltage was directly injected between the two axles using axle-end brushes, and a pick-up coil on one axle was used to confirm the system was operational. Although the prototype vehicles successfully ran with this system in the UK, Northern Ireland, and in the USA, development was eventually abandoned in 1984. At this time, the production vehicles were built as 2-car sets (known as “Pacers”, now classes 140-144) and were being fitted with cast iron brake shoes. It was considered that track circuit performance would be adequate without any assistance. In 1985 when British Rail began building fleets of modern “Sprinter” DMUs, some problems with track circuit operation were noted. Development of an inductively coupled Track Circuit Assister was begun. When the disc-braked Class 158 DMU was developed, this was found to be very poor at operating track circuits, and development of the new device was stepped up. Ultimately the device, now known as the Track Circuit Assister or Track Circuit Actuator (TCA), was fitted to the whole fleet of Sprinters. It has also been successfully installed on many railcars and track maintenance vehicles around the world. The TCA consists of a single turn copper conductor which acts as the primary of a transformer. The secondary is formed of the axles, wheels and rails which surround it, and into which an ac voltage at 165 kHz is induced. The general arrangement is illustrated below. The induced voltage presents itself across any of the four wheel/rail interfaces at the extremities of the aerial circuit, and provides a minimum voltage of at least 5 V rms at the wheel/rail interface.
AEA Technology plc
Davies (11) describes how the TCA has now been further developed by AEA Technology Rail. The Mk II TCA has been fitted to the DMU fleet of Iarnród Éireann and is being fitted to new DMUs being supplied to UK train operators. TCAs are also being actively considered for other new and existing rolling stock. Leaves
Since it is not feasible to provide sufficient voltage to break down the insulating leaf film, a number of options must be considered. In order to prevent leaves arriving at the railhead, the trees can be cut down. Unfortunately if this is done the most likely tree to grow back first (and at a rapid rate) is the Sycamore, which has one of the largest leaves in the UK and will make the problem worse. A better solution is to carefully manage the vegetation so that slow-growing small-leafed trees and bushes predominate. Removal of the leaf film is not simple. The passage of a number of trains and a lot of rain will eventually remove the film (depending on the vehicle dynamics). Railways all over the world have tried water jetting, wire brushing, and scraping, all of which can be successful to varying degrees but which require frequent treatment. Sand or Sandite is laid primarily to improve adhesion. In addition to the formation of a high adhesion layer on top of the leaf film, abrasion of the leaf film will occur, which will to some extent accelerate its removal. The leaf film can be punctured by metallic particles. In order to prevent a Sandite layer causing track circuit problems, stainless steel shot is added: this becomes embedded in the leaf film, forming an electrical bridge between the wheel surface and the railhead. When the rail surface is wet, the presence of water may also improve track circuit performance. If all else fails, the train can be detected by some means which does not rely on a track circuit. Treadles may be used to back up track circuits on the approach to level crossings where timing is particularly critical. If a train is fitted with a TCA, a Page 7 of 8
TCAID (TCA Interference Detector) can be used to detect the presence of the TCA at a range of about 100 m. The TCAID is a battery-powered device which detects the signal emanating from a TCA as it propagates along the rails. Having detected the TCA it short circuits the rails and uses the track circuit merely as a means of communicating with the interlocking. Sand
In 1933, a number of railway engineers in the UK met to discuss the problem of track circuit failures due to excessive sanding. The solution was to fit de-sanding gear to large numbers of locomotives, which used a jet of hot water to remove the sand from the railhead after the passage of the driving wheels. This gave the additional advantage that the sand would be available to improve adhesion, but would not remain to cause any additional drag to the wagons being hauled. The equipment remained in service well into the 1940s. The main problem appeared to be gravity sanding, which could lay sand at an excessive rate: modern compressed air sanders eject much less sand. The use of electrically conductive particles in sand has been discussed from time to time, but no serious trials have taken place. This may be a suitable solution if the problem persists. 6. CONCLUSIONS
The introduction of modern rolling stock with improved suspension systems can seriously affect the electrical performance of the wheel/rail interface, leading to an increased number of wrong-side track circuit failures. In the UK and elsewhere, the introduction of the TCA has restored an acceptable level of performance to low voltage track circuits, but the seasonal problem of leaf contamination remains an issue. It is likely that similar problems will be encountered by rail operators all over the world. We are still a long way off being able to predict with any certainty the track circuit performance of a train before it has been delivered. Despite numerous reports on the subject, the electrical performance of the wheel/rail interface is not well understood: tests and measurements are expensive, time consuming, and often produce indeterminate or conflicting results. The signal engineer must assume the best types of suspension will be used, and respond accordingly. The ultimate solution may be delivered by the move towards transmission-based signalling systems, which dispense with the need for track AEA Technology plc
circuits. In the meantime, signal engineers must work with their rolling stock and civil engineering colleagues to ensure safety is not compromised. REFERENCES
1.
Tattersall AE et al, 1940, “Railway Signalling and Communications – Installation & Maintenance”, The St Margaret’s Technical Press Limited, London, UK
2.
Dyer TK, 1981, “Lightweight Vehicle Track Shunting”, US Department of Transportation, Federal Railroad Administration, Washington DC, USA
3.
Graver GG and Kendal K, 1971, “Failure of Track Circuits due to Rail Contamination”, BRB Chemical Research Division, Derby, UK
4.
ORE, 1957, Interim Report No. 1, “Question A4: Shunting Sensitivity of Track Circuits – Interim report on tests carried out on the electric insulation of an oxide layer on rails”, Utrecht, Netherlands
5.
ORE, 1962, Interim Report No. 7, “Question A4: Shunting Sensitivity of Track Circuits – Electrical contact between rail and rolling wheel in the case of oxidised surfaces; Impedance of wheel-sets as a function of frequency”, Utrecht, Netherlands
6.
Short RC, 1991, “Irregular Operation of Track Circuits by Modern Diesel Multiple Units”, ASPECT 91 International Conference on Railway Control, 492-502
7.
Pearce TG and Watkins DJ, 1987, “Adhesion and Leaves: a Review of the Problem and Potential Solutions”, British Rail Research, Derby, UK
8.
Armstrong DS, 1991, “Effects of Creep Force and Wheel Roughness on the Operation of Low Voltage dc Track Circuits”, British Rail Research, Derby, UK
9.
Evans NJ, 1995, “Leaf Fall Protection Circuits”, ASPECT 95 International Conference on Advanced Railway Control, Session 5, 28-35
10.
ORE, 1962, Final Report, “Question A4: Shunting Sensitivity of Track Circuits”, Utrecht, Netherlands
11.
Davies P, 1997, “TCA: the Train-Borne Signalling Solution”, IRSE Seminar, Birmingham, “Signalling Improvements for the New Railway” Page 8 of 8
