Available online at www.sciencedirect.

com

Transportation Research Part C 16 (2008) 515534 www.elsevier.com/locate/trc

Review

An overview of electromagnetic compatibility challenges in European Rail Trac Management System

Surajit Midya *, Rajeev Thottappillil
Division of Electricity, Angstrom Laboratory, Uppsala University, Uppsala, Box 534, S-75121, Sweden Received 15 May 2007; received in revised form 5 November 2007; accepted 6 November 2007

Abstract In Europe, the railway industry is rapidly getting transformed from traditional mode of public transportation to a very fast, more reliable, long distance and cross country operation. A new concept, called European Rail Trac Management System (ERTMS) is originated to make this transition smooth, reliable and compatible among dierent countries. Electromagnetic interference and compatibility (EMC) issues play a major role on the overall system design and performance of this. In this paper, an overview of the operational principles and major components of ERTMS and other modern railway systems are discussed in detail with an emphasis on possible EMC issues. Radiated and conducted interferences originated from dierent sources and their consequences on dierent subsystems and components are discussed and analyzed. 2007 Elsevier Ltd. All rights reserved.
Keywords: Rail transportation; Rail transportation communication; European Rail Trac Management System; Electromagnetic interference; Electromagnetic compatibility; Arc discharges

Contents 1. 2. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Railway signaling systems: a brief overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Traditional signaling using track circuits and axle counters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Brief of ERTMS system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Different levels of ERTMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . EMC challenges related to railways signaling and communication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Different railway standards related to EMI and EMC issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Interference with railway signaling systems and equipments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1. Interference with track circuits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2. Interference effects on axle counters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.3. Interference issues with the Eurobalise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.4. Interference with the leaky coaxial cable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 516 516 518 519 519 521 521 522 522 523 524 524

3.

Corresponding author. Tel.: +46 18 471 5801; fax: +46 18 471 5810. E-mail addresses: surajit.midya@angstrom.uu.se (S. Midya), rajeev.thottappillil@angstrom.uu.se (R. Thottappillil).

0968-090X/$ - see front matter 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.trc.2007.11.001

516

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

4.

3.2.5. Interference with the electromagnetic braking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interference problems due to pantograph arcing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1. Previous work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.2. Physics of sliding contact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.3. The situation in winter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.4. Investigations by OHL Ice Working Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.5. Radiated interference from pantograph. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Interference from power electronics and drives systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.

525 525 526 527 527 527 529 530 531 532 532

1. Introduction In the recent past, globalization on the European railroad transportation system had a great impact to the overall railway system and technology. Unlike the road transport or air transport, the technical specications of railway transport in Europe were dierent in dierent countries for various historical and political reasons. There were dierences in terms of track width, signaling system, traction feeding voltage and frequency, navigational standards, etc. So, the interoperability between dierent countries was dicult. With numerous mergers in the railway industry and the need to operate trains through dierent countries, a need for a railway system with cross country operability was building up. The ERTMS is the outcome of this. The main objective of this system as mentioned by de Tillere et al. (2003) was to set a standard for Europe in terms of onboard and trackside equipments like power feeding, signaling, communication, train control, complete certication process etc., to provide and improve the safety of the overall system for high speed trains with only few minutes headways between trains, enhancing the operational reliability and capability for carrying more trac with same infrastructure, reducing the maintenance cost, making it more ecient and enabling the cross country functionality for both freight and passenger trains. Although ERTMS had started in Europe at the beginning, the overall technology and concept is adopted by many other countries across the world because of its superiority over the conventional railway system. Electromagnetic noises, generated within the system or coming outside the system often hampers the overall system performance and sometimes create interference with the nearby civilian systems as well. Keeping in mind the huge size and complexity of the railway system, it is often dicult to identify the noise source and the coupling path by which it is aecting the victim. In this paper, operating principles of the critical components of the ERTMS are briey described from the EMC point of view to have a better understanding of the overall operation of the system, which will help in identifying the source, victim and the coupling path. Another important challenge of ERTMS and modern railway is to ensure the availability of the dierent radio based services because of high speed and more trac than the installed capacity. With increasing speed and capacity, it will get more and more challenging to ensure passage of all the required information through different subsystems and components within a very short time span. Section 2 describes dierent types of railway signaling systems and important components associated with it, an overview of ERTMS, its major components and dierent levels in brief. Section 3 provides information about major EMC challenges to railway signaling and communication both at component as well as overall system level. Section 3 also describes the major problems faced by the railway engineers associated with the related standards, major components, the detail analysis about the nature and possible consequence of the important problems related to EMC.

2. Railway signaling systems: a brief overview The main task of signaling system is basically to protect the train from collision with other trains, guide the train to maintain a desired speed, i.e., preventing it from undesired speeding or slowing, avoid derailing of the

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

517

train from the track changing positions till the connecting track gets ready. Pollack (1996) discussed about the evolution of the signaling system which can be divided in following ve major divisions:  Dark territory: Tracks have no visible signals and trains move within the territory following the time table.  Manual block signaling: A visible signal is displayed at the entrance of the block and hence, the cabin crew knows whether the train can proceed or not.

S Rail for return current (grounded)

I rail with insulated joints Length of the track circuit

Receiver Circuit

Track circuit
+

Signal feeding battery Red light circuit Toggle Switch Green light circuit Coil of track circuit relay

Resistor

Track circuit feeding battery (Transmitter)

Power supply circuit

S rail for return current (grounded)

Train

I rail with insulated joints

Length of a track circuit

Receiver circuit

Track circuit

Signal feeding battery Red light circuit Green light circuit Toggle Switch Coil of track circuit relay

Resistor

Track circuit feeding battery (Transmitter)

Fig. 1. Basic layout of the track circuit and corresponding (a) green and (b) red signal. (For interpretation of the references in color in this gure legend, the reader is referred to the web version of this article.)

518

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

 Automatic block signaling: This is based on the use of dierent types of track circuits to detect the train inside some particular block. Fig. 1a and b show a simple track circuit and its operating principle. This type of signaling does not require any manual intervention and the signal gets activated once the block occupancy is changed.  Centralized trac control: The occupancy of all the blocks are tracked from a central location and the signals are activated to ensure that maximum trac throughput is achieved for a given situation. Also, this ensures that proper priority is given for fast trains.  Cab signaling: Rails of the track are used as the transmitting medium for signaling. An onboard display inside the cab ensures the visibility of the signal irrespective of the atmospheric visibility condition. Following trains are immediately notied about the clearance and hence provide better track usage. In traditional xed block signaling, only one train can be present at one block and the physical location and length of the track is xed. Faster trains require longer distance to stop and hence, track circuit length should be long enough to provide proper safety. This reduces the capacity of the railway lines. Communication based signaling (CBS) is mainly based on a radio link, which is at the core for both voice and data communication. In CBS, apart from the exact location of the train, many other information like speed, health monitoring data, etc. are also transmitted to the central control unit. With the pinpoint location of the train known, railways are in a position to implement the moving block system. In moving block signaling, precise location of each train will be known continuously using various onboard, line side electronic equipments and communication systems and a safe distance (estimated from previous knowledge) will be maintained within which the train can stop itself safely. This headway distance is maintained continuously as the train moves through the communication with the central control center and this can reduce the headway distance signicantly as discussed briey by Pope and Harman (2007). This has greatly increased the throughput of a railway track without much increase of the infrastructure, which would have been expensive, time consuming and dicult to implement in many cities. Pollack (1996) mentioned that at Sau Paolo metro, with the existing infrastructure, the headway can be reduced to 66 seconds during busy hours using this moving bock techniques, maintaining the same operating speed. In the near future when train detection will be a new responsibility of the onboard computer and it will be displayed at the drivers panel. This requires the combination of the received train data with the status data, collected from the existing trackside equipments and the route information from the central control unit. Then the locomotive can provide control commands to the local wayside equipments. The train data will be continuously transmitted from the wayside control unit to the central control unit. Under normal operating conditions, all the information will be displayed in the locomotive and the wayside equipments will be automatically activated to their permissive condition. The existing signaling will then work as a backup system. In Automatic Train Control (ATC), high speed trains like Shinkansen or TGV which travels so fast that the driver almost has no time to acknowledge the trackside signals, the brakes of the train will be applied automatically depending on various conditions for safety reasons. ATC also controls the speed limit by checking the running speed and the permitted speed of the train and accordingly applies the brake and releases it once the speed is within the permitted limit. When the driver fails to apply brake or information from the trackside equipments cannot be retrieved, the Automatic Train Stop (ATS) system stops the train automatically. 2.1. Traditional signaling using track circuits and axle counters In a basic track circuit, there are three components, a transmitter, a receiver and a relay, connected to a triple pole switch, which in turn is connected to the red and green light signaling circuit as shown in Fig. 1a and b. The track circuit relay coil is energized when there is no train and the green light glows. Presence of a train shorts the two rails through the axle and the current drops in the track circuit relay coil, making it de-energized and making the red light glow as shown in Fig. 1b. The transmitter could be fed by audio frequency signals (usually between 1500 and 2600 Hz range) as in Indian railways, or by a near DC voltage in the range of few volts, typically 3 V, 9 V, etc. with 02 Hz as in countries like, Sweden, Denmark, Luxembourg, UK, etc., or fed by power frequency, i.e., AC with even harmonics as in countries like Switzerland, etc. Usually the electrical polarity of the feeding voltage is altered in each consecutive block to prevent a block

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

519

falsely powering the next block. The reason for choosing DC and even harmonics is the belief that in an AC railway these components cannot be present, which has been proved to be wrong later by Bormann (2003), Pollack (1996), Buhrkall (2005), Midya et al. (2007) and some other groups. Similar to the operation of track circuit, an axle counter detects the passing of a train in a block. These are installed at both the ends of each block with a counter head. As each axle passes the counting head at one end, the counter increments and when it passes the other end, the counter head starts decrement. When the net count is zero, the track is consider as empty. Usually its done by a centrally located computer, called evaluator. The information from the detection points of the train is transmitted by either telecommunication cable or wireless connection to the evaluator. It is mostly used in places like wet tunnels etc., where ordinary track circuits are not reliable, in tracks having electrically uninsulated steel sleepers which prevents the operation of the track circuit, and in longer sections of a track to save several track circuits. 2.2. Brief of ERTMS system As described by many researchers like Page (2001) and de Tillere et al. (2003), major subsystems and components of ERTMS are:  European Vital Computer (EVC), an onboard computer for controlling and managing all the information.  Signaling related technology, called European Train Control System (ETCS).  GSMR radio communication technology, between the vehicle and the central control station through the GSMR towers with 876915 MHz in uplink and 921960 MHz in downlink for voice and data communication.  Radio Block Center (RBC), a centralized signal encoder module.  Eurobalise active and passive transponders, which lay in between the tracks and send information to the onboard balise transmission module (BTM) at 4.234 MHz, once powered by 27.115 MHz microwave from it.  Euroloop, an inductive loop based localized communication system for semi-continuous mode of communication, where the line side signals get transmitted through a leaky coaxial cable laid across the track to an onboard antenna, similar to the onboard Eurobalise antenna. The uplink from track to train uses the spread spectrum modulated signal, spread over approximately 1.87.2 MHz. In UiC (2000), euroloop radio communication inside the tunnels at frequencies of 50100 MHz is briey discussed.

2.3. Dierent levels of ERTMS To ensure a smooth transition from previously used national railway technology to ERTMS, three levels are dened namely, levels 1, 2 and 3 as shown in Fig. 2ac, respectively. Most of the countries are at level 1 now and are getting transition to level 2. This transition is not exactly the same way everywhere:  Level 1, as shown in Fig. 2a, is a xed block spot transmission system (i.e., the driver machine interface is updated only when the train passes the signal). To avoid the replacement of the entire existing signaling system, it is overlaid on conventional railway signaling but designed to deliver ATC/ATS functionality and cab signaling. Information about infrastructure related data, viz., speed limits, balise linking distance, track information, i.e., gradient, etc. are programmed within the LEU, which determines the aspect of the signal to be displayed and passes a serial message to the Eurobalise containing these information. The drawback of spot transmission is that a train approaching a red signal will be forced to apply brake even if the aspect changes to green as the train comes closer. The mitigation of this problem, called in ll, is achieved either by placing additional balises on the approach of a signal, or euroloop is used for providing updated information, or GSMR is used for communication to allow the driver to maintain the speed by some modication in the ATP overspeed protection algorithms as mentioned by Bloomeld (2006). Trains are detected by track circuit and axle counters. Based on the received ETCS data, odometer data, safe speed prole and the breaking parameters, ETCS subsystem performs the ATP function. The information required to run the train will be displayed at the drivers panel.

520

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

Fig. 2. Overview of the ERTMS system at dierent levels (adopted from Page, 2001; de Tillere et al., 2003).

 Level 2, as shown in Fig. 2b is also a xed block but continuous transmission system. This uses the GSMR radio network to communicate with the wayside equipments and provides complete ll in in each block. Balises are still being used, mainly as position reference. Optical signaling is not required at this level and trains are detected in the same way like in level 1 by track circuits or axle counters. GSMR provides bi-directional transmission of all real time operational data regarding train, i.e., the train speed, location,

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

521

etc. Unlike the level 1, level 2 has the infrastructure data centrally within RBC and it uses the odometry data for the location of the train. With the information from the local interlocking blocks (IXL), RBC also checks the interlocking status and block occupancy ahead of each train and calculates the movement authorities. Then it transmits these information via the GSMR communication.  Level 3, as shown in Fig. 2c, is a moving block, continuous transmission cab signaling based on GSMR. All information between trains and wayside equipments, i.e., train position, train integrity, etc. are transmitted via GSMR. No optical signal, track circuits or axle counters are required. Balises are still used as position reference and to initialise and recalibrate the odometer in level 3 as well. RBC uses the train position and infrastructure status to decide on movement authorities.

3. EMC challenges related to railways signaling and communication The beginning of EMC research on traction was to control the return current. There are many previous investigations and articles on both AC and DC railways on this by researcher like White (2007b), White (2007a), White (2006a), Hill and Cevik (1993), Hill (1994a), etc. Still, because of its complexity involved, there are many EMC related problems even today as reviewed by researchers like McCormack et al. (2006), White (2006b), Leeming et al. (1992), Hill (1997a) and Hill (1997b), while Bourne et al. (2003), Ogunsola (2003) and Ogunsola and Pomeroy (2003) etc. had described some practical management related issues dealing with railway EMC issues. Since railway is getting modernized dramatically over the last decade or so, its a challenge to keep the older existing services intact as well as starting newer systems and following the EMC standards, which are getting more and more strict in respect of unwanted electromagnetic emissions. There is a need for a proper understanding of the various issues related to EMC, both with the older systems as well as newer systems like ERTMS and their compatibilities with the non-railway surroundings and will be discussed in this section. Major sources of EMI in a railway system are:  Rolling stock: It includes propulsion drives, choppers/three phase inverter drives, circuit breakers, static converters, other facilities like, air conditioning unit, lighting, etc., electrical braking units, etc., corona discharges and breakdown of the onboard electrical equipments, arcing in the pantograph, etc.  Power supply unit: It includes substation equipments, switching, substation power electronics, load unbalancing in the three phase system, etc.  Environmental sources like, lightning, electrostatic discharges, geomagnetic induced current (GIC), etc.  Infrastructure: Improper earthing, trackside equipment failure, etc.  Track: Since it carries low frequency signaling information as well as traction return current, coupling with any other sources, e.g., nearby railway, DC components from dierent sources, etc. create interference. The EMI victims includes:       Signaling and communication equipments, including both power and communication cables, lines, etc. Radio communication. Traction power system network. Automatic control system. Trackside equipments. Utilities in the surroundings of the railway, e.g., TV, radio and other wireless communication.

Apart from direct eects of interference to the signaling, there are many indirect eects of the interference problems as well. 3.1. Dierent railway standards related to EMI and EMC issues Railway is a largely distributed complex network comprises of various types of moving and static electromagnetic noise sources and unfortunately compliance of the present standards do not give a guarantee of

522

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

trouble free satisfactory performance of the system. There are standards to deal with various EMC issues for both dierent railway applications and its eects on the surroundings like, CENELEC international standards EN50121:2000 (parts 15) specifying EMC for railway applications, code of practices by GM/RC 1500 between the railway and the neighborhood, CISPR publications by subcommittees A-I, various standards of dierent countries, etc. Also, there are other specications like, the Railway Industry Association species susceptibility levels of 20 V/m over a frequency band of 27500 MHz as mentioned by Allan et al. (1993) and Konefal et al. (2002), EN50121 insists to have a minimum distance of 10 m from the centerline of the outer track, which is often dicult to maintain in many cities and often not enough to keep the surroundings from being aected from the railway generated interferences, etc. It is reported that even when the train is 1 km away, it aects some applications at few hundreds of kHz (Bartlett et al., 1999). Because of the wide variations in the nature of the various interference sources like, pantograph arcing, power electronics and drives systems, etc., and lack of proper experimental and theoretical investigations, these standards are often found to be insucient to protect various applications both within the railway and outside civilian systems from being aected from interferences as often reported by many researchers and engineers. Hence, there is a clear need for proper investigations which can lead to more suitable standards. 3.2. Interference with railway signaling systems and equipments In general there could be two kinds of problems due to interference to the railway signaling and communication systems, viz., false occupancy, i.e., though there is no train, the signaling system will falsely show red light, or false unoccupancy, i.e., although there is a train within the track, the signaling system will show green light. False occupancy can cause delay in the operation of the railways, whereas false unoccupancy can cause collisions and accidents, which may cause disasters. 3.2.1. Interference with track circuits Since the rails carry both the return current of the power supply as well as the track circuit current, its very much vulnerable to receive interference from various sources. Hill (1990b) discussed about various safety and reliability features of synchronizable digital coding in railway track circuits and possible design to eliminate
Power supply circuit with DC component Arc

S rail for return current (grounded)

Train

I rail with insulated joints

Length of a track circuit

Receiver circuit
+

Track circuit

Resistor Signal feeding battery Red light circuit Green light circuit Toggle Switch Coil of track circuit relay

I int I int

Track circuit feeding battery (Transmitter) DC component of arcing current Signal lamp current Track circuit current Interference curent due to pantograph aring

Fig. 3. False signaling in track circuit due to DC component of the arcing.

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

523

interferences (Hill, 1990a). Interference to DC fed track circuits due to the DC components originated from pantograph arcing, DC railways running nearby, geomagnetically induced current, etc., are quite common and frequent in some countries. This type of interference to track circuits may result in false unoccupancy since the interfering current couples itself to the track circuits as shown in Fig. 3. For power frequency fed track circuits, false occupancy can occur if the track circuit relay gets de-energized by some interference current, antiphase to the normal track circuit. False unoccupancy can occur when the track circuit relay gets energized even when the train is still in the track circuit block. But it requires more current to energize the relay rather than de-energizing it and also the interfering current has to overcome the shunting eect of the train axles. Joos et al. (1998) mentioned that in general track circuits are more susceptible to false unoccupancy compared to false occupancy and hence causes delay in the train schedule. White (2006b) discussed about various interference sources of the audio frequency fed track circuits like, switching circuits of the modern converters/inverters, fundamental and multiples of the fundamental frequency of the chopper, four quadrant thyristor converter, inverter drive and the resonant frequencies of the electrical and mechanical systems, mechanical sliding of the pantograph and sometimes the power supply switching, etc. Hill (1997a) and Konefal et al. (2002) reported about some false signaling in North America in 1965 in the audio frequency fed track circuit because of interference from harmonics generated by the chopper unit. In case of a audio frequency (AF) fed track circuits, if unmodulated current of correct frequency gets induced in the track circuit, this may result in false occupancy. False unoccupancy occurs when the interfering signal of correct frequency and modulation rate gets coupled to the track circuit. But as Joos et al. (1998) mentioned, in this case also the shunting eect of the vehicle axles have to be overcome and hence track circuits are more susceptible to false unoccupancy compared to false occupancy. 3.2.2. Interference eects on axle counters Although in mainland Europe, axle counters are proved to be more reliable compared to track circuits, it sometimes fails because of power failures, electromagnetic interferences, etc. and may result in accidents, like the seven tunnel accident in UK in 1991. This was because of improper resetting of the counter, which sometimes require manual intervention. van Alphen (2004) and Bloomeld (2006) mentioned that axle counters can get aected by the magnetic elds close to the rails and also from the magnetic elds emitted from the train or

Z X

Balise antennas Noise region of interest

Fig. 4. Balise installation and possible region with potentially strong EMI capability (adopted from Pozzobon et al., 2003).

524

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

Fig. 5. Radiated interference from the pantograph arcing, catenary wires and other radiating source and the interference from the traction drives to dierent components.

from the return current of the railway. Electromagnetic rail brakes can also create interference and false counting as discussed in detail in Section 3.2.5. 3.2.3. Interference issues with the Eurobalise The onboard antenna, called balise transmission module (BTM), is very sensitive to vertical magnetic eld. Sometimes even horizontal magnetic eld and electric eld components can cause interference too. From Figs. 4 and 5 it can be said that vertical H eld component may come from several sources like harmonics of the traction current, from the high frequency current components of the horizontally laid cables in the close proximity of the onboard antenna, current from the rails, etc. Although the microwave transmission is based on the vertical H component, sometimes horizontal components i.e., H x and H y , especially while tilting or cables with an inclination/ bend, can cause interference to this. Pozzobon et al. (2003) reported about this type of interferences and mentioned that the inuence of interfering E eld is not specied for the antenna module and needs further investigation. Suciently strong random bursts of short duration can cause wrong bit demodulation which may create diculties in decoding the balise telegram and may lead to operational failures as reported by Pozzobon et al. (2003) and many other practicing railway engineers. Both trackside and onboard balise antennas are subjected to analysis and tests for magnetic elds due to harmonics of the power system and transients. But, it has been observed and experimentally veried by many engineers and reported by Pozzobon et al. (2003) that often there are troubles in the balise detection module because of interference due to switching of the main circuit breaker, etc. Today, most of the traction drives (>90%) use resistive breaking to dissipate the energy while breaking/retardation as mentioned by Joos et al. (1998). The faster and sharp pulses originated from these variable speed electrical drives/power electronics modules and switching can cause interferences even in the GHz range as reported by many researchers like Stemmler (1993), Jahns and Blasko (2001), Joos et al. (1998), Chen (2000), Wisniewski et al. (2001), etc. Often many railway engineers and researchers report that the electrical connectors to the breaking unit can cause interference with the dierent onboard radio based services, balises, etc. as shown in Fig. 5. 3.2.4. Interference with the leaky coaxial cable The radiating leaky coaxial cable uses magnetic coupling between the wayside subsystem and the vehicle for transfer of information. Although there are standards and tests as mentioned in UiC (2000) on the loop trans-

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

525

mission module (LTM) receiver, onboard antenna and the coupling factor between them to check the interference levels, there are possibilities of interferences as discussed by Nakamura et al. (1996) and Pozzobon et al. (2003). This is particularly when the eld strength becomes too low or there are some obstacles in between etc. Also, there are possibilities of interference with vehicles other than the intendant one. 3.2.5. Interference with the electromagnetic braking In electromagnetics brakes, the breaking force comes from the electromagnetic attraction between the track and the brake plate. Generally a high current is applied to the coils of the strong electromagnets within the brake shoes, although in some cases permanent magnet is used. Instead of being economic and having several advantages like, regenerative braking, independent of the coecient of adhesion between the wheel and the rail, the main problem with electrical breaking was its less reliability because of the complex circuitry and strong interference problems with some electronic equipments located adjacent to the rail due to its strong magnetic eld and ferromagnetic material as discussed by Kroger et al. (1999) and Tolksdorf (1974). Today, many high speed trains like ICE train in Germany, Adtranz X2000 on Sweden, some trains operating in hill areas, etc. are tted with electromagnetic track brakes to be used for emergency purposes only. The strong magnetic eld causes interference problems with the electromagnetic axle counter and other similar sensors/devices. Also, when applied and hence lowered down to the rail, these brakes can mislead the axle counter, which may count it as wheel. This will stop the train automatically without any reason as discussed by Kroger et al. (1999). The heavy current required for the braking operation and the transient caused by its switching can saturate the rails and can create interference problems. 3.3. Interference problems due to pantograph arcing In the sliding contact between the contact wire and the pantograph, arcing is a common phenomena. This becomes more predominant with higher speed as observed with the test run of V150 TGV on 3rd April, 2007 in France as well as in many other high speed trains and in winter as observed in many cold countries. This arcing generate radiating electromagnetic waves in a wide band. Bartlett et al. (1999) reportedly measured EMI at UHF band at a distance of 0.12 km away from the train line. The major problems from the pantograph arcing can aect the railways in two ways, one, in conducted EMI to the traction power and signaling system and the other, radiated EMI to the wireless communication system. This radiation can cause interference with both the railway systems and non-railway systems in the near vicinity like, TV reception, radio communication, etc. as mentioned by Bartlett et al. (1999). This type of arcing in a sliding contact distorts the ideal sinusoidal waveform of the power supply and creates assymetry and hence a dierence in the voltage drop between the positive and negative half cycle of the AC waveform. This voltage drop is higher in the negative

Fig. 6. Propagation of DC current into the railway power feeding network with booster transformer (adopted from Kiessling et al., 2001).

526

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

half cycle compared to the positive half cycle of the power supply as reported by Bormann (2003), Bormann et al. (2007), Buhrkall (2005), etc. which results in a net DC EMF with the pantograph as positive electrode and the catenary as negative electrode as shown in Fig. 7. This kind of assymetry also generates even harmonics, interharmonics and subharmonics as described in IEEE Industry Applications Society (1981, 1992) standards and by many other researchers. This DC component and the harmonics propagate in the entire railway system including rail tracks, track circuits, other locomotives in the same track, substation supply transformer, vehicle transformer, etc. as shown in Fig. 6 and create false signaling as shown in Fig. 3, saturation of the transformers, corrosion problems, etc. Also, Buhrkall (2003b) reported that in some new models of trains like, OTU EMU, etc., the DC supervision system gets tripped, which results in delay and sometimes cancellation of train operation. The severity of this phenomenon depends on several parameters and in general, is predominant in winter. Brillante et al. (1998) reported that in the presence of multiple pantographs (0 series Shinkansen trains have 16, 12 and 6 pantographs parallel and the last few show clearly visible arcing because of the induced oscillation due to applied forces by the preceding pantographs), and some heavy freight trains with multiple traction engines, the situation is bit complex. Buhrkall (2005) mentioned that the following pantographs provide a return path to the DC component and harmonics and the magnitude of the DC generated from the arc in the following pantographs are lesser compared to the preceding ones as shown in Fig. 9. The magnitude of the DC voltage generated on the other pantographs depends on some other parameters like, number of following pantographs, the electrical load and some other circumstances like, removal of ice by the preceding pantographs, etc. The arc mechanisms of this type of sliding contact, both in the absence and presence of the ice and the physics behind this DC component generation is not well understood so far. In this subsection, a brief of the previous investigations, physics of sliding contact and the need for further investigations with the changing scenario due to ERTMS are discussed. 3.3.1. Previous work Some experimental investigations and numerical simulations on the pantograph arcing were done by Klapas et al. (1976), Buhrkall (2005), Galdi et al. (1998), Brillante et al. (1998), Tellini et al. (2001a,b) and Giannetti et al. (2001). Their investigations gave some characteristic information about the arcing, the variation in gap between the contact wire and pantograph, variation in the duration of attachments and detachments, interference level, etc. Because of the complexity of the problem and limitation of experimental facilities, most of these experimental works were limited to few hundred kHz to few MHz frequency band and they made some simplied assumptions to make the laboratory investigation and computational analysis possible at university and research laboratory levels. In the contexts of ERTMS and modern railways, some of these assumptions and simplications like, lower train speed by Klapas et al. (1976), lower frequency measurements by Brillante et al. (1998), only static attachment and detachment of the pantograph and contact wire instead of sliding contact by Tellini et al. (2001a,b), nature of the arc, which in reality is dierent than the test setup
Arcing at the pantograph

Arc generated DC current flow in the entire circuit

Fig. 7. The net DC component generated by the pantograph arcing and its propagation into the traction system (adopted from Buhrkall, 2005).

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

527

Fig. 8. Burning arcs attached to the rotating wheel (source: Bormann, 2003).

by Tellini et al. (2001a,b) and Klapas et al. (1976), etc., demands more realistic and detailed investigations to get a more clear picture of the interference levels at operational railway sites. 3.3.2. Physics of sliding contact In a sliding contact, usually the arc tends to attach itself to a particular point, called cathode spot or anode spot depending on the polarity, on the contact wire and then as the current collector moves, the arc gets elongated and nally, it is extinguished as shown in Fig. 8a and b. It attaches itself to a new spot again and continues the same way. Usually, when the speed is low, the arc may continue for several half cycles of the supply voltage before getting quenched, i.e., before current becomes zero. When the speed is high, the arc becomes quenched near the zero crossing of the applied voltage as observed by Bormann (2003) and Bormann et al. (2007). If the speed is very high, the arc will be quenched well before the zero crossing and then it may reignites again within the same half cycle depending on the gap length, voltage at that instant and train speed. The mechanical oscillation and the AC current waveform interacts in such a way that there is a very rapid transition from a highly conducting arc to a very low or almost a zero current. This superimposed nature of the two waveforms makes the situation worse from the EMI point of view by generating very sharp switching pulses. 3.3.3. The situation in winter Although arcing from pantograph occurs in all the seasons because of the particular nature of the sliding contact and vertical oscillation of the train, it becomes more critical and visible to the bare eyes in winter, especially below the freezing point. When temperature is above zero, usually there is a thin lm of water which makes a sliding contact smoother. This phenomena was investigated and reported by many scientists like Holm (1946), Shobert (1976, 1993), Slade (1999), etc. In winter this water lm gets frozen and hence the smoothness and self lubrication of the sliding contact is gone. Moreover, the ice layer on the contact wire acts as a dielectric layer. These two phenomena work together to make the sliding contact far away from being smooth and we see the clearly visible arc moves with the pantograph along the contact wire. For high speed trains, which generates visible arcing in normal weather, the situation becomes far worse in winter and this degrades the life and performance of the contact wire and the pantograph. 3.3.4. Investigations by OHL Ice Working Group As a part of a European Union project, an Over Head Line Ice Team (OHL Ice Group) was formed, consisting both academicians and industrial experts to investigate this arcing phenomena and its consequences. The group conducted three eld tests on four dierent types of trains at dierent test tracks with dierent supply in dierent environmental conditions. On December 2001, they conducted experiments at Luxembourg with BR185 locomotives and Z2000 EMU with 50 Hz supply at below freezing point and on March 2002,

528

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

Net DC current flow in the entire circuit DC current flow from preceeding unit to the following unit
Fig. 9. Flow of the DC current in the presence of multiple pantographs. Adopted from Bormann (2003) and Buhrkall (2005).

at Vasteras, Sweden with OTU EMU with both 16 2 and 50 Hz supply in articial iced condition and at Cerhe3 nice, Czech Republic, with Class 357 EMU in an ice simulated environment with 50 Hz supply. As reported by Buhrkall (2002), they observed that the DC component is independent of the RMS value of the fundamental component and the vehicle speed. They also measured the interference levels at both 16 2 Hz supply with 3 100 Hz track circuit feeding and 50 Hz supply with 77 Hz and 83 1 Hz track circuit feedings to investigate 3 on the possible eects. It was found that the level of the line current harmonics are proportional to the level of the DC. Buhrkall (2003b, 2005) reported that they also varied the load angle and tested both ACAC powered vehicles having four quadrant converters with AC motors and ACDC vehicles having phase angle controlled thyristor rectier with DC motor. They concluded that although both types generate almost same level of DC EMF, the rst type is advantageous since it runs at a slightly lagging phase angle whereas the second type runs at a lagging phase angle by nature. They also conducted some experimental investigations by running two dierent types of trains forming a multiple unit, i.e., two BR185 locomotives with AC motors, connected back to back and one Z2000 EMU with DC motors back to back at 25 kV 50 Hz supply. It was noticed that the net DC emitting from two vehicles in a multiple formation is lower than a single unit. This is probably because the rst pantograph removes the ice layer to certain extent and the following one acts as a return path to the previous one as described in the Fig. 9 since the generated DC EMF is lower in the following pantographs. They also noticed that the returning of the DC current via the second locomotive was around 50% compared to the rst one. The OHL Ice group conducted some laboratory experiments at Vasteras ABB facility to investigate further on this matter. Fig. 11 shows a simplied diagram of the experimental setup. They reported a possible DC voltage of upto 106 V depending on the voltage, the current, the distance between the pantograph and the conductor, the train speed (simulated experimentally by rotating the wheel) and the phase angle. Bormann (2003)

sinusoidal voltage input

Copper conductor Conducting plasma Burning plasma Conducting plasma Carbon collector

V V V

V V V

Fig. 10. The origin of the resultant DC current in the pantograph arcing because of dierences in potential drops in both the polarities.

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

529

Fig. 11. Experimental setup of the ICE team (source: Bormann, 2003; Buhrkall, 2003a).

and Bormann et al. (2007) discussed about the physics behind the origin of the DC component and discussed about the dierence in material properties to explain the dierence in the voltage drop between the positive and negative half cycle. Voltage drop in the plasma of the arc, as shown in Fig. 10 is higher when the pantograph side is positive and lower when the contact wire is positive. Thus a net DC component of the order of 518 V gets generated with static electrodes of dierent materials like that in the present case with the pantograph as positive as shown in Fig. 10. This leads to a current owing from the contact wire to the pantograph (Buhrkall, 2003b; Buhrkall, 2005; Bormann et al., 2007). The arc resistance and corresponding voltage drop depends on the electrode material, arc length, arc radius and the current. This phenomena requires further investigation to estimate the level of voltages with various test conditions and inuence of various parameters like electrode material, voltage, current, arc length, arc radius and the wheel speed. Their investigations with this experimental setup match with the results obtained during the site tests. Bormann et al. (2007) reported that it takes certain voltage to ignite an arc in a new spot. The ignition voltage of the arc depends on the electrode materials and the voltage and its polarity at that instant. The attachment and detachment process of the arc leaves burn marks on both the contact wire and the carbon and its aluminum frame as reported by Bormann et al. (2007) and this deteriorates the life and performance of the contact wire as well as the pantograph. They concluded that in general, the DC EMF increases with:  increase in speed of the rotating wheel,  increase in voltage and current,  increasing gap between the contact wire and the pantograph. When the speed is low, the DC EMF decreases with increasing phase angle, but when the speed is high, it increases with increasing phase angle (Bormann, 2003; Buhrkall, 2003b, 2005). 3.3.5. Radiated interference from pantograph As discussed in Section 2.2, ERTMS is very much dependent on the wireless communication and uses several frequency bands ranging from few hundreds of kHz to few GHz. Fig. 5 shows all the radiating sources which include the arc itself, the contact wire, the pantograph with its aluminum support, the connecting wire

530

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

from the pantograph to the traction drives, etc. Because of the complex geometry and position of the radiating structures as shown in Fig. 5, its often very dicult to nd the most suitable location for the vehicle antennas and dierent wireless based sensors and services. In 1999, an investigation was conducted mainly from Banverket (The Swedish Rail Administration) at Kiruna (north of Sweden) on the eects of pantograph arcing in the presence of ice with the railway radio communication services. Signicant level of EMI in the range of 60 MHz1 GHz were measured with a train speed of 100 kmph. The measurement antennas were placed at the front inside the second vestibule, having a glass front window. But the measured data was bit noisy and inuenced by test condition. 3.4. Interference from power electronics and drives systems DC motors and the low frequency (e.g., 16 2/3) AC motors dominated the traction industry for decades because of its over load capability and loadtorque characteristics and are still being used widely as discussed by Hill (1994a,b), Steimel (1996a,b) and Stemmler (1993). Since the second half of the last century developments in rugged semiconductor devices, high performance AC machine control algorithms along with high speed digital processors to implement these and developments in material science leading to the development of low cost neodymiumiron (NdFe) permanent magnets are having a signicant impact on the traction drives system. AC adjustable-frequency drives started to dominate traction system worldwide, both for urban light rails and intercity heavy rails. France has pioneered in the use of high voltage power frequency supply for the traction system in 1950s. Because of numerous advantages this is quickly adopted in many other countries. Another very important advantage of AC drives over DC is that for similar power and torque, AC motors are compact with lesser weight and physical dimensions. Although synchronous machines were selected for TGV in at least two versions in 1993, squirrel cage induction motors came up in the dominant position for new generation of traction vehicles. Once the thyristors came into the market in 60s followed by line commutated controlled rectiers, tap changers in the AC supplied vehicles were replaced and collector motors with DC supply ruled the traction industry even with its drawbacks like, poor power factor, limited performance, high losses and maintenance. Turn on thyristors, with forced commutation were used to built step down choppers for controlling the collector motors with controllable pulsed DC voltage. But this technique was used only for lighter vehicles due to its limitations. Forced commutated inverters were used for feeding three phase induction motors with variable voltage and frequency, i.e., V control. Once gate turn o thyristors (GTO) came into market in 80s, the expenf sive, voluminous and complicated turn on thyristors are gone and inverter fed three phase induction motors started becoming a preferred choice. However, GTO traction inverters have their limitations in terms of switching frequencies. For PWM switching frequency, its limited to 500 Hz or less and require complicated gate circuitry which reduces its reliability and increases the switching losses. Developments of high voltage IGBTs like, 3.3 kV1.5 kA and 6.5 kV0.6 kA, with higher switching frequencies in the range of 1500 Hz, simpler gate drive and higher eciency made it preferred for newer drive units. Although DC drives will continue to evolve in some sectors, three phase voltage source inverters with six switch topology for variable speed induction motor drive applications is becoming a preferred choice. This brings in many issues to be addressed properly, like input power quality, eects of transients and fast switching and related interference issues, etc. di Modern IGBTs evolved to such an extent that it requires a minimum of snubber circuit to limit the high dt dV and dt , or in some drives applications even without any snubber circuit at all. But the output of these snubberless inverters may contain switching rates more than 1000 A/ls and 10,000 V/ls, which may result in undesired and sometimes harmful consequences. High dV switching rates can interact with the inverter output dt cables and the machine windings because of large transient voltages across the outermost turns of the stator windings, which can cause catastrophic failures of the stator winding insulation as investigated by many researchers like Kerkman et al. (1997), von Jouanne and Enjeti (1997), Aoki et al. (1999), Ogasawara and Akagi (1995), Jahns and Blasko (2001), etc. Longer cables make this phenomena frequent and hence some machine manufacturers provide better insulation levels, at least for the outermost turns of the winding. High dV switching rates can generate unbalanced charge build up in the parasitic capacitance coupling between the dt machine stator, rotor, inverter switches and the ground. These accumulated charges can discharge through the bearing in the absence of any other discharge path and can eventually cause bearing failure by creating serious

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

531

spitting of the bearing balls as experienced and reported by Jahns and Blasko (2001), Erdman et al. (1996), Busse et al. (1995), Bhattacharya et al. (1999), Link (1999) and Macdonald and Gray (1999). Manufacturers came up with solutions like insulated bearings and grounded rotor shaft bearing, etc. Another problem with di the high dt and dV is that it increases the level of conducted EMI in the drive input lines and the ground. There dt are some preventive measures that can signicantly reduce this common mode EMI, like, adding common mode inductors in both input and output lines together with small capacitors between each DC link bus and the grounding which will prevent the high frequency currents from reaching the utility grid or machines, improved PWM switching algorithm, etc. Dierential mode EMI can be mitigated by using active or passive lters on the inverter output lines as reported by Jahns and Blasko (2001). But, this additional ler is always bulky because of the high power rating and sometimes not a feasible solution. Most of the industrial drives (>90%), including the traction drives systems do not return the power back to the utilities till date, and the power dissipation path often creates noises with the line side equipments, balises, etc. as shown in Fig. 5. Interestingly, Joos et al. (1998) showed that AC propulsion produces more interferences than DC propulsion because of the requirements of variable frequency operation of AC traction system. Onboard power electronics, i.e., power converters/inverters are one among the major contributors of the interference in a wide frequency band. A major drawback of the previous analysis of the interference caused by the power electronics and drives was that it used lumped circuit modeling, which does not demonstrate results beyond few MHz. Whereas, Chen (2000) and Lombardi (1994) investigated and found that measurements and dierent types of modeling show interferences at higher frequencies. To avoid the complexity of the problem, which is actually a very vast one, some researchers like Chen (2000) divided the entire system into smaller subsystems and then analyzed the interference, which gives a much more clear picture of the actual sources causing interference. Wisniewski et al. (2001) reported that in Sweden, SEMKO tested six standard inverters with ve of them had higher EMI radiation than the permissible limit in the European market. Between 1997 and 2000 they performed several tests on dierent commercially available variable speed drives and measured considerably high radiation in 30300 MHz by using a biconical antenna and from 200 MHz to 1 GHz by using a log-periodic antenna on the actual installation site. Chen (2000) reported that at frequency ranges of few hundreds of MHz, the power stage of the inverter contributes more to the noise. Based on the electrical layout of the system, these HF current components get coupled to dierent systems and subsystems. It is observed that IGBT imposes signicant capacitive loading beyond few MHz to the signal source. Several researchers like Chen (2000) reported that the noise gets distributed through several independent paths like, through cables to the traction motor, through the motor windings stray capacitance to the chassis of the train, through the stray capacitance of the switching devices with the chassis and in many other possible ways. Wisniewski et al. (2001) concluded that the level of the interference is high enough to cause bit-error rate of the digital communication systems. This shows that there is a potential risk to aect a wide variety of radio services including digital communication services, mobile phone service, navigation system, commercial broadcast, radio and paging services. The level of interference on the victim receiver depends on several parameters like, the frequency of interest, distance from the drive system, cable length, type of cable being used, conguration of the site and many other parameters. 4. Conclusion Railway signaling and communication is quite sensitive to both conducted and radiated electromagnetic interference from various sources and have potential risks of false signaling. Old believes like DC or even harmonics cannot be present in an AC fed railway are proved to be wrong and hence, the track circuit feeding has to be designed properly. Also, these interference problems deteriorate the health of dierent components and reduces the estimated life time and reliable operation of the railway. Both theoretical investigations and experimental observations found that all wireless and radio based applications related to railways are quite vulnerable to these electromagnetic noises generated from various sources. ERTMS is very much dependent on the radio based services for several applications and proper caution has to be taken in various forms to protect these services from possible EMI. Thus, it is very important to investigate the level of interference and possible coupling mechanisms from all these sources to the radio based services to make the ERTMS and other modern railway system more robust and reliable. Proper experimental and theoretical analysis of the noise sources will

532

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

give a precise knowledge of the EMI levels that various radio based services will be facing. Also the related standards should be modied to consider the higher level of interference, which was not properly observed because of lack of proper experimental and simulation facilities. Once the ERTMS will be at level 2 and level 3, the higher speed and lower gap between the trains will demand more reliable and robust signaling and communication without any potential threat from electromagnetic interference. Also, the consequences of intentionally radiated EMI has to be investigated properly to protect the ERTMS system from possible threats. Acknowledgements The authors would like to thank Banverket (The Swedish Rail Administration) and Bombardier Transportation for providing funding for this work, Dr. Nelson Theethayi of Uppsala University and Mr. Stuart Shirran of Bombardier Transportation for their valuable feedback and Dr. Dierk Bormann of ABB Corporate Research and Dr. Thorsten Schutte of Rejlers AB. for the valuable discussions and feedback. References
Allan, J., Chan, W., Shaoz, Z., Mellitt, B., 1993. Low frequency and radio frequency electromagnetic compatibility for rapid transit railways. In: Proceedings of the Fifth European Conference on Power Electronics and Applications, Brighton, UK, September, pp. 106111. Aoki, N., Satoh, K., Nabae, A., 1999. Damping circuit to suppress motor terminal overvoltage and ringing in PWM inverter-fed AC motor drive systems with long motor leads. IEEE Trans. Ind. Applicat. 35, 10141020. Bartlett, E.J., Vaughan, M., Moore, P.J., 1999. Investigation into electromagnetic emissions from power system arcs. In: Proceedings of the 1999 International Conference and Exhibition on Electromagnetic Compatibility, EMC York, York, UK, July, pp. 4752. Bhattacharya, S., Resta, L., Divan, D.M., Novotny, D.W., 1999. Experimental comparison of motor bearing currents with PWM hard and soft switched voltage source inverters 14, 552562. Bloomeld, R., 2006. Fundamentals of European rail trac management system ERTMS. In: Proceedings of the 11th IET Professional Development Course on Railway Signalling and Control Systems, June, pp. 165184. Bormann, D., 2003. Origin of DC component in railway line current upon operation with iced-over overhead line: theory, experimental results and conclusions. ABB Internal Report of the project: Investigation of DC Current because of Ice on Contact wire, Ref. No: SECRC/PT/TR-2003/038, May. Bormann, D., Midya, S., Thottappillil, R., 2007. DC components in pantograph arcing: mechanisms and inuence of various parameters. In: Proceedings of 18th International Zurich Symposium on Electromagnetic Compatibility. Munich, Germany, September, pp. 369372. Bourne, A., Bradshaw, J., Lalies, A., Price, P., 2003. Managing electromagnetic compatibility on the london underground system. In: Proceedings of the IEE EMC Assurance in a Railway Environment, September, pp. 111. Brillante, S., Ferrari, P., Pozzobon, P., 1998. Modelling of electromagnetic emission from catenary-pantograph pliding contact. In: Proceedings of the International Conference on Computer Aided Design, Manufacture and Operation in The Railway and Other Advanced Mass Transit Systems, Lisbon, Portugal, September, pp. 881890. Buhrkall, L., 2002. Ice on the Overhead Line, AC Railway Systems: Analysis of the Ice Recordings. The ESC User Group OHL Ice Working Group. August. Buhrkall, L., 2003a. Ice on the overhead line of AC electried railways. Summary Report From the OHL Ice Team, November. Buhrkall, L., 2003b. Ice on the Overhead Line of AC Electried Railways: Conclusions from the OHL Ice Team. The OHL Ice Working Group, June. Buhrkall, L., 2005. DC components due to ice on the overhead contact wire of AC electried railways. Electrische Bahnen 103 (8), 380389. Busse, D., Erdman, J., Kerkman, R.J., Schlegel, D., Skibinski, G., 1995. Modeling and damping of high-frequency leakage currents in PWM inverter-fed AC motor drive systems. In: Proceedings of the 21st IEEE International Conference on Industrial Electronics, Control, and Instrumentation (IECON), vol. 1, Orlando, FL, USA, November, pp. 698705. Chen, C., 2000. Characterizing the generation and coupling mechanisms of electromagnetic interference noise from an electric vehicle traction drive up to microwave frequencies. In: Proceedings of the Fifteenth Annual IEEE Applied Power Electronics Conference and Exposition, APEC, vol. 2, New Orleans, LA, USA, February, pp. 11701176. de Tillere, D.G., Emery, D., Curchod, A., 2003. Managing systemic innovations in rail systems: the case of ertms technology. In: Proceedings of the 3rd Swiss Transport Research Conference, Monte Verita, Switzerland, March, pp. 121. Erdman, J., Kerkman, R., Schlegel, D., Skibinski, G., 1996. Eect of PWM inverters on AC motor bearing currents and shaft voltages. IEEE Trans. Ind. Applicat. 32, 250259. Galdi, V., Ippolito, L., Piccolo, A., 1998. Arcing in AC railways: a mathematical approach. In: Proceedings of the International Conference on Computer Aided Design, Manufacture and Operation in The Railway and Other Advanced Mass Transit Systems, Lisbon, Portugal, September, pp. 891904. Giannetti, R., Macucci, M., Tellini, B., 2001. EMI measurements in line-pantograph contact discontinuity in railway transportation systems. In: Proceedings of the 11th International Symposium on Trends in Electrical Measurement and Instrumentation, IMEKO TC-4, Lisbon, Portugal, September, pp. 2590.

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

533

Hill, R.J., 1990a. Optimal construction of synchronizable coding for railway track circuit data transmission. IEEE Trans. Veh. Technol. 39 (4), 390399. Hill, R.J., 1990b. Safety and reliability of synchronizable digital coding in railway track-circuits. IEEE Trans. Rel. 39 (5), 581591. Hill, R.J., 1994a. Electric railway traction. I. Electric traction and DC traction motor drives. IEE Power Eng. J. 8 (1), 4756. Hill, R.J., 1994b. Electric railway traction. II. Traction drives with three-phase induction motors. IEE Power Eng. J. 8 (3), 143152. Hill, R.J., 1997a. Electric railway traction. VI. Electromagnetic compatibility disturbance-sources and equipment susceptibility. IEE Power Eng. J. 11 (1), 3139. Hill, R.J., 1997b. Electric railway traction. VII. Electromagnetic interference in traction systems. IEE Power Eng. J. 11 (6), 259266. Hill, R.J., Cevik, I.H., 1993. On-line simulation of voltage regulation in autotransformer-fed AC electric railroad traction networks. IEEE Trans. Veh. Technol. 42 (3), 365372. Holm, R., 1946. Electrical Contacts. Almqvist and Wiksells, Uppsala, Sweden. IEEE Industry Applications Society, 1981. IEEE guide for harmonic control and reactive compensation of static power converters. IEEE Industry Applications Society, 1992. IEEE recommended practices and requirements for harmonic control in electrical power systems. Jahns, T.M., Blasko, V., 2001. Recent advances in power electronics technology for industrial and traction machine drives. Proc. IEEE 89 (6), 963975. Joos, G., Kapila, R., Friem, R., 1998. Electromagnetic interference issues in the specication of AC and DC propulsion systems for light rail vehicles. In: Proceedings of the 1998 ASME/IEEE Joint Railroad Conference, Philadelphia, PA, USA, April, pp. 4147. Kerkman, R., Leggate, D., Skibinski, G., 1997. Interaction of drive modulation and cable parameters on AC motor transients. IEEE Trans. Ind. Applicat. 33, 722731. Kiessling, F., Puschmann, R., Schmieder, A., 2001. Contact Lines of Electric Railways: Planning, Design and Implementation. Publicis Corporate Publishing, Munich Erlangen, Germany. Klapas, D., Hackam, R., Benson, F.A., 1976. Electric arc power collection FPR high-speed trains. Proc. IEEE 64 (12), 16991715. Konefal, T., Pearce, D.A.J., Marshman, C.A., McCormack, L.M., 2002. Potential electromagnetic interference to radio services from railways. Final Report, March. <http://www.yorkemc.co.uk/research/railways/>. Kroger, U., Grautstuck, H., Wirth, W., 1999. Electromagnetic or permanent-magnetic rail brake, January. Leeming, D.A., Allan, J., Mellitt, B., 1992. Electromagnetic compatibility on rapid transit railways-a systems approach. In: Proceedings of the Eighth International Conference on Electromagnetic Compatibility, Edinburgh, UK, September, pp. 227234. Link, P.J., 1999. Minimizing electric gearing currents in SSD systems. IEEE Ind. Appl. Mag. 5 (4), 5566. Lombardi, E.J., 1994. Engineering tests performed on the x2000 and ice high speed trainsets. In: Proceedings of the 1994 ASME/IEEE Joint (in Conjunction with Area 1994 Annual Technical Conference) Railroad Conference, Chicago, IL, USA, March, pp. 1321. Macdonald, D., Gray, W., 1999. PWM drive related bearing failures. IEEE Ind. Appl. Mag. 5 (4), 4147. McCormack, L.M., Seller, S., White, R.D., Hutchison, R., Hooper, P.W., 2006. Railway electrical systems integration practical application of the v cycle for electromagnetic compatibility (EMC). In: Proceedings of the Institution of Engineering and Technology Seminar on EMC in Railways, Birmingham, UK, pp. 119126. Midya, S., Schutte, T., Thottappillil, R., 2007. Consequences of DC components in AC railway and its elimination. In: Proceedings of Railway Engineering-2007, London, UK, July. Nakamura, M., Tsunomachi, H., Fukui, R., 1996. Road vehicle communication system for vehicle control using leaky coaxial cable. IEEE Commun. Mag. 34 (10), 8489. Ogasawara, S., Akagi, H., 1995. Modeling and damping of high-frequency leakage currents in PWM inverter-fed AC motor drive systems. In: Proceedings of the 13th IAS Annual Meeting Industry Applications Conference, vol. 1, Orlando, FL, USA, October, pp. 2936. Ogunsola, A., 2003. EMC assurance and management. In: Proceedings of the IEE EMC Assurance in a Railway Environment, September, pp. 18. Ogunsola, A., Pomeroy, S., 2003. EMC assurance and safety critical apparatus in a railway environment. In: Proceedings of the IEEE International Symposium on Electromagnetic Compatibility, vol. 1, Hertfordshire, UK, May, pp. 429432. Page, C.R., 2001. ERTMS the global signalling solution. In: Proceedings of the Rail Solutions Asia, March, pp. 0120. Pollack, M.W., 1996. Communications-based signaling: advanced capability for mainline railroads. In: Proceedings of the 1996 IEEE 46th Vehicular Technology Conference, Atlanta, GA, USA, pp. 15991603. Pope, M.E., Harman, I., 2007. Vision and potential for future signalling strategies. UIC Signalling Panel of Experts, Ref. No: Signalling Strategy v1 0 DraftPdC, January. Pozzobon, P., Amendolara, A., Vittorini, B., Henning, U., Schmid, R., Shirran, S., Ahlstedt, S., 2003. Electromagnetic compatibility of advanced rail transport signalling. In: Proceedings of the World Congress on Railway Research (WCRR), Edinburgh, UK, October, pp. 12501263. Shobert, E.I., 1976. Carbon, graphite and contacts. IEEE Trans. Parts, Hybrids, Packag. PHP-12 (6), 6274. Shobert, E.I., 1993. Sliding electrical contacts. In: Proceedings of the 39th IEEE Holm Conference on Electric Contacts, Pittsburgh, PA, USA, September, pp. 123134. Slade, P.G., 1999. Electrical Contacts: Principles and Applications. Marcel Dekker, New York, NY, USA. Steimel, A., 1996a. Electric railway traction in Europe. IEEE Ind. Appl. Mag. 2 (5), 617. Steimel, A., 1996b. Electric railway traction in Europe. A survey of the state-of-the-art. In: Proceedings of the IEEE International Symposium on Industrial Electronics, ISIE, vol. 1, Warsaw, Poland, June, pp. 4048. Stemmler, H., 1993. Power electronics in electric traction applications. In: Proceedings of the 19th International Conference on Industrial Electronics, Control and Instrumentation, vol. 2, Maui, Hawaii, USA, November, pp. 707713.

534

S. Midya, R. Thottappillil / Transportation Research Part C 16 (2008) 515534

Tellini, B., Macucci, M., Giannetti, R., Antonacci, G.A., 2001a. Conducted and radiated interference measurements in the linepantograph system. IEEE Trans. Instrum. Meas. 50 (6), 16611664. Tellini, B., Macucci, M., Giannetti, R., Antonacci, G.A., 2001b. Line-pantograph EMI in railway systems. IEEE Instrum. Meas. Mag. 4 (4), 1013. Tolksdorf, G., 1974. Electromagnetic or permanent-magnetic rail brake, April. UiC, 2000. FFFS for euroloop subystem. <http://www.uic.asso.fr/>. van Alphen, G.-J., 2004. Electromagnetic compatibility between rolling stock and rail-infrastructure encouraging European interoperability. Proposal to EU Specic Targeted Research Research Project RAILCOM, April. von Jouanne, A., Enjeti, P.N., 1997. Design considerations for an inverter output lter to mitigate the eects of long motor leads in ASD applications. IEEE Trans. Ind. Applicat. 33, 11381145. White, R.D., 2006a. AC/DC railway electrication and protection. In: Proceedings of the Ninth IET Professional Development Course on Electric Traction Systems IET Seminar Digest, vol. 2006, Manchester, UK, November. White, R.D., 2006b. Electrication traction and signalling compatibility. In: Proceedings of the 11th IET Professional Development Course on Railway Signalling and Control Systems, June, pp. 165184. White, R.D., 2007a. AC 25 kV 50 Hz electrication supply design. In: Proceedings of the Third IET Professional Development Course on Railway Electrication Infrastructure and Systems, Birmingham, UK, May, pp. 81114. White, R.D., 2007b. DC electrication supply system design. In: Proceedings of the Third IET Professional Development Course on Railway Electrication Infrastructure and Systems, Birmingham, UK, May, pp. 3562. Wisniewski, M., Kubichek, R., Pierre, J., 2001. EMI emissions up to 1 GHz from adjustable speed drives. In: Proceedings of the 27th Annual Conference of the IEEE Industrial Electronics Society, IECON, vol. 1, Denver, USA, December, pp. 113118.