Ever since electric traction was invented, measures have been needed to stop unwanted electric currents from interfering with signalling, telecoms, lineside power supplies and outside organisation interests. Writes Clive Kessell
The resultant protection is collectively known as immunisation, sometimes confused with medical activities using the same name. Ask the switchboard for the immunisation department and you will as likely get the doctor’s surgery!
The science of understanding electrical interference has been ongoing for many years, engaging the top brains in universities and research establishments. As new forms of rail electrification and traction emerge, so new technologies for S&T and power supply protection need to be devised. The interference is likely to have an adverse impact on many systems, often leading to disputes as to the method and cost of remedial action.
To better understand the latest situation, the IRSE ran a seminar recently entitled ‘Track, Traction and S&T – The Final Frontier’. It was chaired by David Bradley, a signal engineer with many years of experience in the subject, and it was an interesting day proving, if nothing else, that the word ‘final’ remains a pipedream.
The Dark Art
Ken Webb from Mott Macdonald used this title to describe the subject, noting that other names – ‘Arcane Discipline’, ‘Black Magic’ and ‘Not for After Dinner Speeches’ – had also been attributed. The main criteria are:
- Current management – getting amperes to and from the train;
- Frequency management – avoiding the creation of frequencies that can interfere with other systems, particularly for unsafe signalling conditions;
- Minimise coupling – eliminating or minimising the conductive, inductive or radiation linkage between the traction current and affected systems;
- EMC (Electro Magnetic Compatibility) design – ensuring that engineers responsible for minimising interference understand the established techniques and standards.
Safety and standards
Someone once said ‘the wonderful thing about standards is that there are so many to choose from!’ This was never more true than for electrical interference and immunisation, where a plethora of standards and instructions exist within Europe, the UK and Network Rail. Getting the responsible engineer to search all these out, realise how they relate to each other and then devise a practical compliance is a challenge in itself. The base standards were said to be:
- EC 61508 – setting out the functional safety of electrical and electronic equipment for safety related systems;
- IEC/TS 61000-2:2008 – describing the methodology for achieving the above;
- IET Guide on EMC functional safety.
These relate to the general control of electrical interference and are not railway specific. UK railway standards derive from the European EN 50121 series and RSSB Group Standards GE/RT8015 and GE/RT8270. However these will not guarantee immunisation safety. Three basic types of interference occur that can cause connection to other systems; conduction as a direct connection, induction as an electromagnetic coupling and radiation relating to high frequency radio waves.
All of these may occur on an electrified railway, be it 25kV AC overhead or both overhead and third-rail DC systems. Down the years, a portfolio of protective measures have been designed to limit induced voltages and currents to within prescribed limits. These can add considerable cost to an electrification scheme and thus the different engineering disciplines may argue as to their necessity.
In the UK, infrastructure owner Network Rail has responsibility for ensuring that railway systems are protected from electrical interference and that staff are not exposed to dangerously high voltages. To obtain optimum solutions, the company liaises with other European rail bodies to produce a suite of standards and instructions that give guidance on what can be expected and the remedial measures needed.
Maya Petkova, a principal engineer in the Asset Management Group, works with colleagues in Europe to both produce and update standards and to relate these to UK conditions. The resultant Railway Group and Network Rail standards set out the conditions and practices that must be met. This is not a static situation and the documents regularly need review and update to reflect emerging circumstances. Recent examples are:
- The growing impact of radio systems for rail control and communication including track mounted balises and train aerials;
- New types of track circuits (TCs) and axle counters;
- Transient effects of neutral sections on electrified lines;
- The impact of interoperability on both infrastructure and rolling stock;
- Pressure to use more COTS (Common off the Shelf) equipment which may have a lower EMC performance.Dominant in all of this is understanding how frequency management will be possible over the range from 0-1MHz to prevent unsafe operation of track circuits.
Dominant in all of this is understanding how frequency management will be possible over the range from 0-1MHz to prevent unsafe operation of track circuits.
The electrification engineer
So how do the different disciplines cope with the competing pressures of obtaining the most cost effective electrification configuration whilst keeping the railway safe?
For nearly 60 years, electrification on main lines has used the 25kV 50Hz AC overhead line system. The advantages are well known, since adopting the domestic mains frequency means avoiding expensive rectification and distribution networks. The current in the catenary wire does, however, produce an inducing electric field that affects lineside copper cables. It can be loosely regarded as a long single turn transformer and ways of preventing large induced voltages had to be found.
The most common solution is the use of booster transformers and return conductors. These transformers are mounted on the stanchions at roughly two kilometre intervals and inserted in series with both the catenary and the return conductor, the latter also being strung on the stanchions. Half way between transformers, the conductor is connected to the traction return rail. The effect is to ‘suck’ return current out of the rail into the return conductor that creates an opposing electric field to that of the catenary so as to balance each other out. The induced voltage limits into cables are 60V steady state and 430V under short circuit conditions, the latter only being present for a few milliseconds until the breakers shut off the traction supply.
The electrification engineer traditionally regards booster transformers as a necessary nuisance since they increase the impedance of the system thus losing potential power.
Even the 25kV system, however, is not capable of delivering sufficient power for high speed or very heavy load trains and so has been developed the 50kV auto transformer (A/T) system described at the seminar by David Hewings, the Network Rail electrification engineer. In this case, a transformer at the electrification feeder station which takes power from the grid delivers 50kV to an auto transformer with a centre tap connected to the return rail, one 25kV feeding the catenary and the other to a separate 25kV conductor mounted on the stanchions. The two feeds are in anti phase so help to cancel out the resultant electric fields. The ‘presentation’ to traction units remains 25kV and is thus unchanged. Because the system can deliver more power, the currents in the two feeds are higher, thus increasing the inductive effect.
In the same period, the means of controlling the system has been updated and the control of individual feeder stations by separately switched circuits has been replaced by a network known as a station bus system using IEDs (Intelligent Electronic Device). With more power comes a potentially higher fault current – up from 600A to 1200A – and thus much faster circuit breaker switches are needed to keep induced voltages within the prescribed limit. The package has led to GOOSE (General Operating Optical Sensing Equipment) that is based around a fibre optic LAN network, suitable for transmission of safety critical messages.
For many years, interference into telecom circuits represented a major problem. With long distance copper cable-based circuits, the longitudinal induced voltages could be very large, posing a risk to circuit performance, equipment damage and staff safety. The protection methods of booster transformers and additional cable screening were virtually mandatory, not only to prevent interference to railway circuits but also to British Telecom and other telecom providers’ systems which were routed close to an electrified railway.
This method is not perfect and such circuits may need to be further protected by either an aluminium screen on the cable(s) or the provision of a separate mutual screening conductor located in the cable route, these being earthed down to copper rods at 1km intervals so as to also carry an opposite phase current.
As well as high voltages, because the wires of a pair in a telecom cable were not always identical as to resistance and capacitance characteristics, the inducing field could introduce noise on to the circuit by what is called a transverse voltage. Under internationally agreed limits, this had to be restricted to 1mV, and required cable balancing techniques for this to be achieved.
Fortunately, the advent of fibre optic cables has been the solution to most problems and, for over twenty years, railways and public telecom operators have invested in this technology for long distance telecom links. It is not a complete solution since there remain a lot of trunk telecom copper cables in existence and copper tail cables are needed to connect from fibre ‘hubs’ to the end device. Keeping the length of the latter to broadly a 2km limit is good practice.
Telecom equipment at the lineside must be bonded to earth in a similar manner to signalling. Because some equipment is susceptible to high voltage transients caused either by electrification short circuits or lightning, the provision of fast transient earths at individual equipment cabinets protect against any high voltage ‘strike’.
The signalling engineer
Since signalling copper circuits are of relatively short length, induced longitudinal voltages are not the main problem. Of more concern is the potential interference to track circuits by conduction direct from the rails. This can be influenced by the type of track circuit deployed and whether it is based on single or double rail design. Success is measured by how well the TC operating frequency can be different from the 50Hz overhead line frequency and the harmonics that derive from this.
Unwanted frequencies can also be generated by the sophisticated types of traction units now in service and the objective is to prevent traction return currents from causing unsafe operation of track circuit equipment. The TI 21 track circuit used for many years is designed around frequency shift keying (fsk) with the frequencies chosen outside of the normal 50Hz harmonic range but recent tests with Class 90 locomotives have shown that the fsk principle is not good enough and that in certain circumstances, wrong side TC failures can occur. Thus digital encoding of track circuits, making each one unique, is the ongoing design preference.
Earthing and bonding remain a challenge and Peter Brown, the E&P design manager in Network Rail, described how a return screening conductor (RSC) connected periodically to the traction return rail can negate the need for individual earth rods at signalling lineside locations. Such an arrangement is ideal in areas using single rail track circuits or axle counters but impedance bonds will be needed if double rail TCs are in use. This arrangement has been chosen for the forthcoming Great Western main line electrification, subject to it passing a generic safety case approval.
The traction engineer
Every rolling stock designer is aware of the risk posed to signalling and other equipment by return currents and inappropriate frequencies generated by the train. Colin Place from Bombardier described the devices that are installed on Class 377 DC units in service on Southern and South East Trains.
Firstly, there is 1.5 tonnes per power car of filter equipment, intended to protect TCs from conducted electro-magnetic interference.
Also introduced when the trains were built was a Line Interference Monitor (LIM) based on Railtrack standards but experience in practice showed this to be so onerous that traction power was removed too often when the third rail experienced ice conditions and trains became stranded. The replacement is an Interference Current Monitoring Unit (ICMU) but even here when ice is present, drivers are trained to keep power on when coasting to ensure an arc keeps flowing to the third rail.
On older fleets where an ICMU does not exist, cruder monitoring systems were provided but proved less reliable than the on-board systems being protected. The latest thinking is that the next generation of trains is unlikely to be fitted with monitoring equipment since the desired limits are known to be exceeded many times daily but are of short duration. Protection systems to prevent over voltage, over current and device failure in the traction power electrics will be sufficient.
The power supply engineer
Providing a reliable power supply for S&T equipment at the trackside is vital; without power, nothing can function. Traditionally this has been via a 650V single-phase cable with a transformer provided at every lineside cabinet from which the local 110V AC and 50V DC power sources are derived. Later designs have adopted a 400V three-phase arrangement but both used a cable with a separate earth core.
In electrified areas, this earth was connected to the collective earthing system of the line and this could lead to high touch voltages if double earth fault conditions occurred.
A design to simplify and reduce the cost of lineside power supplies has recently emerged from Network Rail; known as a Class II supply this is based around an earth free system using only a two-core cable and its features were described at the seminar by Tahir Ayub from Network Rail’s Signalling Innovations Group. A full explanation of this power supply is contained within a separate article in this magazine.
Mixed Electrification Areas
The protective measures needed to protect vulnerable equipment in AC and DC electrification are quite different: in AC, earthing plays a vital role whereas DC systems should be kept earth free (or at least only a local equipment earth) so as to prevent high traction return currents using S&T cables as a return path. However, it is known that DC stray currents can exist well beyond the immediate rail area and there are many locations where the two systems are in close proximity, bringing with it the risk of equipment mis- operation.
Crossrail and Thameslink
Any new piece of electrified railway will raise fears as to whether interference will cause problems to existing co-located railways.
One such consideration is Crossrail – on its east-west route across London it will have two 400kV supply points at Pudding Mill Lane and Kensal Green and will use the 50kV A/T system delivering 67MVA from Stratford to Shenfield and Paddington to Maidenhead and 30MVA in the central section.
James Greatbanks from Bechtel described how an interference model has been produced looking at 16 engineering elements that might be affected. These included earth potentials, stray currents, electrolytic corrosion, resonance, rail voltage, screening factors and equipment loading. Modelling is based around short circuit conditions with and without a supplementary earth wire on the surface sections. The conclusion is that there will be negligible 50Hz flow between Crossrail and other railways.
The biggest risk is at joint stations and also at Heathrow’s baggage and airside terminals with their many electronic systems. London Underground continues to have reservations with the Hammersmith & City line being the most likely to be affected. Whilst modelling is much cheaper than live tests, no doubt confirmatory tests will be necessary once the line is built and equipped.
The north-south Thameslink project is an upgrade scheme to hugely increase the number of trains using the route. Martin Sigrist from Network Rail explained that it means converting the Midland main line electrification to the 50kV A/T system but retaining 25kV and booster transformers from St Pancras to Blackfriars and with the complication of having to change from overhead to third-rail electrification between Farringdon and Blackfriars, which will be dual voltage. The expected high DC currents raise the risk of these straying well into AC areas with risk to track circuit operation because of insulated rail joint (IRJ) arcing.
Early tests suggest 180 amps of stray current in the quiescent state (no trains) and 300+ amps with trains running, but measurements on the East Coast main line have shown only 0.5 amps detected there. The earthing and bonding plan for the central section is thus a complex issue with the multiplicity of cables, equipment cabinets and a fire main between Kings Cross and Farringdon all making possible parallel return paths for single rail track circuits and traction return. Contactors have been installed at Blackfriars to minimise the DC flow from southwards of this point, but many more tests and adjustments to the design will be needed as the project progresses.
At the end of the day, this seminar demonstrated that interference and immunisation is anything but a precise science. Whilst broad order protective measures have been established for 25kV lines over a number of years, the emergence of new types of rolling stock, computerised based signalling systems and more power from the overhead catenaries have needed improved methods of testing and monitoring.
In the past, engineers may have taken an over cautious approach and perhaps the adverse effects are not as bad as some would have us believe. Some technology advances – for example fibre optic cables and Class II power supplies – have made immunisation easier. The situation with DC electrification and particularly third rail is less well documented and adverse weather can vary traction current conditions thus impacting on interference levels.
The work to understand and protect against electrical interference will need to continue for many years yet.
The ongoing NW electrification project brings AC systems near to the Manchester Metro (tram) network. Recent tests carried out at three sites, Ordsall Lane, Castlefields and Eccles, were described by David Bradley to see if a problem would result. Four basic measurements were made at each site – two of voltage and two current – to determine any flows between the two rail operations. Both short circuit and steady state conditions were tested. Happily, stray currents did not penetrate as much as expected and thus it is concluded that track circuits will continue to function correctly.