Track Circuits (rural, mainline, metro)
 

Hi all,
I'm new here! I am a University student studying electronic and electrical engineering. I am currently doing a project titled: Monitoring systems for Rail Networks. The task is to find a new innovative way of track signaling and locating trains.

I am having trouble finding out some specific information regarding current methods revolving around track circuits.

I would like to find out the number of track circuits employed around the UK and IF possible where these are located (i.e. rural, mainline, metro)

If anyone has any of these stats I would greatly appreciate this.

Thanks a lot!

Welcome Railproject.

What sources have you already tried?

I've tried quite a lot of websites, many of them not too good.

I did however find the following site to be good, having a lot of useful statistics I can use:
http://www.railway-technical.com/index.shtml

Most of my research has been through the internet. I have emailed Network Rail asking for any statistics they can share about trach circuits/signaling methods.

Can you suggest any useful sites?

Thanks!

Give up on the internet except to look for companies who might take pity on a poor student and help out.
Which are the signalling equipment supply and installing companies? e.g. Westinghouse, Plessy, GEC (must be more modern ones)

Which are the key Journals in S&T ?

This lot might help............http://www.irse.org/

The Institution of Mechanical Engineers has always been very helpful to me...even though I'm not a Member.

Speaking as a fool who has done two Research Degrees in my time.......eventually there is no alternative to getting out of the armchair and going to meet people, and librarians are generally embarissingly helpful and have spent hours helping me out. (except for the Library Dragon at the NRM :p )

I would imagine that the number would be astronomical. Then you will have to split it into ones under the wires as opposed to the more normal circuits, you may also like to check out those in wet tunnels?? The ones in Huddersfield tunnel on the right as you face it from the plaform had special dispensation to use a 4 ohm relay as opposed to the standerd which in those days was 8 ohm. This was because the ballast was always saturated and leakage occured across the sleepers etc.

We've been told on our NR briefing days that in future there will be a move away from track circuits towards axle counters on high speed lines, due to the signal sections being so long.

I've had personal experience of unreliable track circuits. Coming happily out of a cutting at 90 mph after travelling for over twenty miles on greens, the next signal which controlled a junction was red. There was no way I could stop in time, so that was a category B SPAD on my licence for evermore. But it was a relief at the time just to know that this was a system fault and that there wasn't a train making a conflicting move towards me as I'd feared.

An investigation into the signal 'going back' on me gave the reason as damp ballast in the signal section in advance.

I've often wondered about whether it would be possible to use digital signals to indicate the prescence of a train in section instead of having to tweak analogue circuits all the time.

The "obvious" way forward is to use GPS to locate every train in Britain to 1 metre accuracy in a giant computer, and do away with all trackside signals and circuiting. (You would need a transmitter at front and rear of the train so that if the tail end broke away it would be identified)

But would I have the confidence to hurtle at 125mph through busy junctions relying on a bunch of electrons in a supercomputer in Watford being right 100% of the time?

The trouble is that GPS was designed for use out in the open; like in the middle of the sea or in a desert. It constantly under-performs on the railway because of steel infrastructure, tunnels, and built up areas. The satellite antenna needs to see a large part of the sky in order to perform well, and this frequently doesn't happen on a train.

If you look at the 377 fleet, which use GPS in the Selective Door Opening process, they are constantly having problems. When a train arrives at a station where the on-board GPS doesn't have a satisfactory fix it will inhibit the door release. The train then sits there for an eternity while the the driver has to manually override the system and put up a manual release. It's farcical to have a safety system that needs to be bypassed.

I believe that one day we'll be extensively using a system like ERTMS with trackside beacons and in-cab signalling, but the investment will be so large that I wouldn't hold my breath.

ERTMS website...http://www.ertms.com/2007v2/what.html

Thanks for the info guys.

THe project brief our group has been given revolves around using the rails as a transmission line to send a signal down the line. When the signal hits the S/C (Train) it will reflect back to the source.

The idea is that we can then detect exactly where a train is within a section through monitoring the time taken for the reflected signal to be seen.

With this system we could also find out the speed of the trains and calculate time to enter next section etc....

A good description, and I agree that GPS isnt anywhere up to the failsafe reliability needed on a railway. But give the inevitable trend in all walks of like to huge centralised automated computer control I begine to imagine the Rail Network becoming like a giant model railway - with DCC control maybe :D rather than Block sections signalled from the trackside.

Whether that is the right way forward is another debate, of course.

I lost count of the number of cat B Spads I've had. I even had one last week when a signal went to red against me as I was approaching it because of a power surge at the signalbox controlling it. Its the cat A ones you gotta worry about. :)

I remember when the old chestnut of leaves on the line first made an appearance in the newspapers. I'm pretty sure it was something to do with the Pacer trains that had just been introduced. When I first learned the 143's, they had some sort of a track circuit booster fitted to them because the engineers figured they'd have trouble activating track circuits because the vehicles were so light and only had four wheels per unit. Unfortunately the booster system proved to be unreliable and units were disappearing from the signallers panels because of the leaves which is the reason why there were so many problems.

You are correct about the fact that axle counting is the way that Network Rail wants to move foreward. As far as I am aware, the only section of track that actually uses this system so far is between Toton and Clay Cross Junction near Chesterfield. This system is used in conjunction with a new radio system thats being introduced called IVRS (Interim Voice Radio System) which will replace the currently out of date NRN system and also be used instead of Track Circuit clips which will now become obsolete with the axle counters being used.

I am one of the few drivers on the network who has been issued with my own personal IVRS unit because I sign that particular section of track that the system is in place on. Being issued with my own personal unit is only a temporary thing because all trains will eventually have the units fitted as standard. Some have already been fitted to some Class 66's.

Here's a small report about the new masts being put up and how some people aren't very happy about them.

http://www.mastsanity.org/index.php?...=100&Itemid=44

I appreciate that this is said with an element of fun, but this is a subject that I've given quite a bit of thought to over the years, so I'll jump in and have a rant. Remote control of trains certainly works on 'closed' systems like LUL and DLR, but these are pretty simplistic arrangements and were conceived as shiny new installations. The mainline railway on the other hand still hasn't received enough money/attention just to keep it turning over in reasonable order.

Firstly of course there's unbelievable investment required for the automation itself, but much more significantly is the fact that our railway has been operating under "make do and mend" since 1945. Railstaff wrestle on a daily basis with a system which is permanently knackered. Being on the spot enables signallers/drivers/station staff/etc to step in and prevent the house of cards collapsing. Layman understandably view the railway as a large trainset, whereas it's actually unbelievably complex when lift you the lid and look at what goes on.

I think you may be right about automation and remote control - it would make financial sense to do this, but it's gonna take a while. I've got 19 years service still to go and I doubt very much whether I'll be out of a drivers job before I go (at least I b****y well hope so ;)).

HTH,
Foggy

Yup, amen to that :mad:

I've heard of this, the legendary leaf fall 'stealth trains' :D

I've seen the IVRS units when I rode in the cab of a mate, but I didn't twig what the technology was behind them. The link mentions GSM-R masts. Does this mean that the IVRS you are using is GSM-R technology, or is it just that the people who wrote the link are having a good old rant at everything ?

Cheers,
Foggy.

I'm no expert but yeah it does seem that they are using GSM-R technology. I can think of quite a few places where the masts are going up. There's one on the up side of the main line at York Yard North is (where the junction for Clifton Loop was) for instance, and another at Thirsk. I can make a note of them when I see them if you like.

Rant accepted ! My point is a bit more subtle. I am interested in failsafe and "robust" systems. Another word is "redundant", ie if any one part of a system fails, there are unused systems that will take over. The Internet is a example of a robust operating sytem.

Yet there are people and big multinational companies with commercial motives for centralising and controlling everything in the face of practical experience which says that what is being attempted is impossible e.g the NHS computer system.

So while Railproject is trying to solve the technical problem of bouncing signaling data up and down rails or catenary wires, the bigger question is whether this solution fits into a centralised or decentralised contol system.

Well thankyou :D
OK, that's a very interesting view. With regard to redundancy I remember reading somewhere that in Japan their points have two sets of motors, so that if one motor stalls or fails it is clutched out, leaving the other one to continue. Diagnostic data is sent to the technicians for immediate attention.

There was in the 50s talk of an experimental single tracking the service to Hull using a loop system for passing. This was to mainly operated by wireless control. I never heard any more than the rumours at York but the way it was described it sounded feasable for lightly trafficted lines.

Thanks for the input from so many people, athough I would like to know what any of your thoughts are regarding the concept of our 'solution' (measuring time taken for signal to be reflected from train short-circuit)

As a group we are only about to start the design stage of this project.

Any thoughts as to what are going to be major issues regarding this method of detection?

No probelms with reliable track circuits - that can only be a good thing and I wish you well with the project.

But as future user (driver) of this system I'm afraid I feel fairly luddite about real time speed information. Given the current situation of delay attribution that already causes a Spanish Inquisition over who is to 'blame', I can only see shedloads of grief being caused a system that could be used to display running speeds.

The concept is fine, but the problem is what is done with the information. Every driver regulates his train differently according to his/her assessment of weather, adhesion conditions, braking force, etc. As is so common in British Industry, the data for delays is handled by clerks who wouldn't know the difference between a fishplate and a fish supper. They have the authority to make hell for people at the front end, and they consistently do. It can only put addtional pressure on signallers and drivers, and give rise to thousands more accountants.

(NB, this isn't paranoia about drivers having overspeed events, as trains have carried data recorders for years).

Well not surprisingly you've been very coy and not told us anything technical about what you will be doing. If I were to assume that you might be 'firing' very high frequency signals along the rails and measuring the 'echo' time, I would suggest that you might find alot of nasty harmonics generated by the traction packages and electric motors on the trains themselves (since even modern diesel locos use electric motors). However DSP is dead clever nowadays, so you can probably filter out the noise once you know what the nasties look like.

Hope this helps,
Foggy

As a group we do not need to 'worry' too much as to what will be done with this information. The project is set up as though we have been given the task of increasing accuracy of current technology. We have also been given guidelines that suggest using the method I wrote about earlier (sending pulse....receiving reflection....)

We have decided as a group that we need to focus on a particular type of track, as they do vary quite a lot, in terms of electrification, characteristics etc.... From what I've read and looked up it seems that most mainline railway lines are powered by 25kV AC 50Hz. Am I correct in this?

As you pointed out, we would need to know the 'noise' generated by the traction systems/electric motors and also the noise generated by the electrification system. Once these are known we should theoretically be able to filter them out. I've been searching forthis info but I honestly have no idea where to start! There is just so much information around. Any pointers....

I have also been trying to find the characteristics (generally) of a railway line in terms of impedance and admittance (resistance, inductance, conductance and capacitance). With these figures we should be able to work out the 'best' frequency to send down the line.

Hmm, I'm not sure I'd say most. Someone on here may be able to tell you just what proportion of route miles in the country are OHLE/3rd Rail/Not Electrified.

3rd rail 750v DC accounts for nearly all of the South Eastern Region, Southern Region, and South Western region.

Sure, and it's not just noise. You would be looking at some hefty traction currents in the running rails. Several hundred amps per commuter train, considerably more for locos and intercity stock.

Why not speak to the guys currently doing the job; Network Rail's Signalling & Telecom engineers ?

I have absolutely no idea what 1 km of steel rail looks like electrically, but from the postings earlier on this thread you can see that the characteristics do change significantly, particularly due to dampness in the ballast.

HTH,
Foggy

I'm not an engineer but I know model railways sometimes use infra red beams to detect trains passing signals but how that could work in real life is anyones guess, badly I suspect.

The question is how you would be 100% sure that it was a train that was tripping the beams and not a bird, wandering badger or a couple of kids waving branches in front of the detectors.

But it would have to be sensitive enough that every time a real train passed it would trip the beam.

Well there already is a system for getting instantaneous indications of a train's presence - treadles. You don't get false readings with these as they require the flange of a train to operate them. (AFAIAA these tend only to be used where track circuits can't be such as in the middle of complex junctions, or to trigger events like occupational crossing warnings).

thats true, you don't get many of them on model railways, so it didn't cross my mind!

Good thread this one,
I have like a few others have tried the national database for rail signalling and track circuitry, it appears that because the system was inherited by railtrack, from the former British Railways, no actual plan or scematic, is held on record anywhere except at the former regional depots of the big four companies, lol. seems funny that an infrastructure essential like this could be left so low on the list of essential paperwork to be centralised ?
I would like to know the answer to the questions above though , it seems we may all be safer if this matter was brought to the attention of the people responsible for our safety whilst on the rails surely ?
however
things are bypassed daily so i won,t hold my breath
best regards Derby

Quick question:
Impedance bonds. Are they basically a piece of equipment which provide a path for traction currents to pass through, but still keep track circuits electrically isolated?

I just want to make sure I understand them correctly.

Basically yes. They're known coloquially round here (SE England) as spiders, and we're told that the postion of the spider does not always mark the end of the track circuit !!

As usual Wikipedia has something to say here... http://en.wikipedia.org/wiki/Track_circuit

But remember that Wiki tends to tell things like the Americans do it ;)

HTH,
Foggy

So, the project has now been broken into several areas.

I decided to take the section we have called Filtering.

Basically the idea behind this sub-system is that it will filter the received signal of any noise and be left wit the reflected pulse which can then be used to determine train position.

My first obective is to identify all the possible areas were noise coild be created which could be imparted into the rails (therefore into our square wave (with highliy skewed duty cycle))

Below is a short list of what I think could be the main causes for noise within rails:
- Traction noise due to rectification
- Electrical equipment onboard train such as air conditioning
- power converters or other on-board equipment
- Mobile Phones?
- Overhead power lines in close proximity to rails

The traction noise is probably going to be my biggest source of noise right? As this is only a project and not an actual design I am thinking of limiting my outlook to only newer trains with IGBT controlled induction motors.

What I need to find out is what this noise looks like (electrically) so that I can start thinking of different ways to filter these out from my wanted signal.

If anyone has any information regarding these sources of noise and possible ways of filtering the noise then I would appreciate any input.

Thanks!

Perhaps you should approach Bombardier to find out what sort of electrical products are passed into the running rails by a modern train. IIRC they make the 375/376/377 series EMUs that are taking over the South East at the moment.

Link to them at - http://www.railway-technology.com/co...ier_transport/


Also since what you are doing is mildly similar to radar with aircraft, it might pay you to have a look at what sort of software/DSP is already developed in that field for distinguishing between the 'return' and the 'clutter'.

I would rule out mobile phones, but there is an awful lot of other intermitent airbourne RF noise from other transmitters. Emergency services radio transmitters are very powerful for example.

What about broad spectrum RF noise from sparking? either from OHLE, or from good old 750V third rail.



*****


Possbly a bit off topic for a undergrad project, but the "lateral thinking" apprach would say "why fight the noise?" Why not [electronically] look for a distinctive traction motor noise as the signal you are trying to detect, rather than eliminate

Hello, and "bon courage" for your project!

The current track-circuit systems vary according to the status of the line: electrified or not, 750 V DC 3rd rail (ex Network SouthEast and various metros); 1500 V DC usually overhead (France, Netherlands); 3kV DC (Belgium...);15 kV 16,66 Hz (Germany, Austria and Switzerland) and 25 kV 50 Hz (England, France...).

Usually, over non electrified lines, or DC electrified lines, 50Hz track circuit was used.

With AC electrification, systems have evolved to higher frequencies, of course off the harmonics of the traction current. More recent systems use periodic impulsions of high frequency (something like three impulsions at ~150Hz every second), process which proved to be efficient short sections of rail in dirty environment (leaves, oil...) especially in station areas.

Coming back to your solution, you mentionned a number of external sources of noise. DC usually does not create noise, provided that it is properly filtered in the substation, but DC trains are now equiped with thyristors which "cut" periodically the current in order to deliver the adequate power to the motors. The frequency they cut is around 400 Hz and is perfectly audible on recent trains (jubilee line for instance). Your system will have to avoid these frequencies.

Last question: if you process the signal on the reception side, how long will the process take, and how long is the free distance that it will represent ?

Good luck! (In case your project is not yet done, which I -now- doubt, according to the date of the former posts!!)

Hi SDX, thanks for some intersting info there.

In the UK, the last couple of generations of 750 V DC mainline EMUs have used AC motors. This started with 465/466 Networkers in the 1980s , and has been continued with the more recent 375/376/377/378 series. The main benefit of this is that the motors can be used for dynamic (regenerative) braking. This not only saves on brake disc usage and is a more 'powerful' brake, but also keeps the Green lobby happy as it enables some of the train's kinetic energy to put back into the Traction Current supply, or even the National Grid theoretically.

Because the motor frequency varies as a function of train speed, some units like 465/2 and 466 (ie, the MetCam units) have a device called an Interference Monitor, which will actually cut off traction power if the train's return current contains a frequency near to 50Hz into the running rails - which I'm told is to prevent damage to the ac track circuit.

And it's not just EMUs; Diesel locos are undergoing similar development. The ubiquitous Class 66 loco was conceived with DC traction motors, but the new Genesis Project hybrid loco (powered byboth Diesel & AC) which GE is currently developing for Freightliner is to have AC traction motors.

Hello, and thank you for your comments. As far as I know, there are in France some areas where using "récupération" (i.e. producing and sending it back to the over-head line) is forbidden due to incompatibility with track-circuit. So, either they use the rehostatic braking (dissipating energy through resistors) or use engines without this feature!
The idea is, on the long term, to modify track-circuit in order to protect them from these effects.