Rail crash driver died after telling signalman train had passed

[R G NOW.]

I know I'm a bit old fashioned, but I find it surprising that the signaller doesn't have a very good idea of when a train would have likely passed the crossing or not. In the "bad old days" you did Train Entering Section at say 12.34, and the boxes at both ends would then know only too well that an express would be at that crossing at 12.38, a local at 12.40, etc. Someone phones you up at 12.36 and saying that the train has passed them would be spotted pronto.

Like in the above post I did. Why did this happen?. Surely the signalman got the red in section lights on the panel. But why did not the lights on the crossing stay on until the train had left that section. One would assume the bolts would be locked as well.

 

[DerekC]

Like in the above post I did. Why did this happen?. Surely the signalman got the red in section lights on the panel. But why did not the lights on the crossing stay on until the train had left that section. One would assume the bolts would be locked as well.

There are some very interesting questions around this incident - like the above. Normally we would be reading a comprehensive report from RAIB to find out what happened - so why isn't there one?

 

[R G NOW.]

On another thinking. I also noticed on youtube cab view videos that a lot of these user worked crossing have treadles that bring up a red light and lock the gates. That's what I imagine happens, or should happen anyway. I wonder if the driver of the land rover just opened the gates and there was no interlock or it was faulty.

 

[Taunton]

There are some very interesting questions around this incident - like the above. Normally we would be reading a comprehensive report from RAIB to find out what happened - so why isn't there one?

If it's like the Athelney, near Taunton, crossing accident report from a few years ago, the investigation team somehow fail to notice what is staring the rest of us in the face.

 

[Class 170101]

In this day and age you would think that all locos, EMUs and DMUs were fitted with GPS tracker though. Many buses do, and it's not expensive equipment.

Used for this purpose, though, it’d have to have next to no likelihood of giving any false indication of position, e.g. showing a last reported position any length of time after losing contact with the satellites and/or the monitoring system, which is more likely to happen with trains and the tunnels that they led through.

Wasn't GPS tracking installed on the Marks Tey to Sudbury branch after the crash there involving a lorry at a crossing near a sewage farm?

 

[tsr]

On another thinking. I also noticed on youtube cab view videos that a lot of these user worked crossing have treadles that bring up a red light and lock the gates. That's what I imagine happens, or should happen anyway. I wonder if the driver of the land rover just opened the gates and there was no interlock or it was faulty.

That’s really not how it works at many (or most) user-worked crossings. At many of them you have a phone on each side, and that’s it.

You do get miniature warning lights at busy or high-risk locations but they are by no means universal. Gate locks are, I believe, numerically fewer, but in any case they are also an optional feature depending on risks.

 

[R G NOW.]

That’s really not how it works at many (or most) user-worked crossings. At many of them you have a phone on each side, and that’s it.

You do get miniature warning lights at busy or high-risk locations but they are by no means universal. Gate locks are, I believe, numerically fewer, but in any case they are also an optional feature depending on risks.

You would of thought that all crossings would have to have some sort of locks or lights. especially now in these heath and safety times.

 

[R G NOW.]

On the line between oxford and Bicester village. They have done away with all crossings and installed bridges, but there is only one road crossing north of Bicester station.

 

[RichA]

Like in the above post I did. Why did this happen?. Surely the signalman got the red in section lights on the panel. But why did not the lights on the crossing stay on until the train had left that section. One would assume the bolts would be locked as well.

We have to remember that the track circuit indication covers a distance of 3.7 miles on the down line at Frampton, of which the crossing is roughly half way along within it, the crossing itself is just a standard user worked crossing (although it also has a footpath/bridleway element) without any form of warning lights, the deceased was the authorised user and key holder of the locked vehicle gates. He generally used the crossing twice a day and on the day of the incident was accompanied by another person who opened the gates whilst he obtained permission to cross, after that permission had been obtained the assistant heard a train approaching but was unable to get the attention of the vehicle driver to stop and the collision occurred.

 

[Mathew S]

True in principle, but railways have to meet safety standards that demand an extremely low level of unsafe failures. IEC SIL-4 level (which is usually the one quoted for signalling systems) demands a probability of dangerous failure lower than 1 in 100,000.000 per operating hour - putting it the other way round that's at least 99.999999% reliable. Even if you applied SIL-2 to this situation you would be looking for 99.9999% reliable.

In this situation "dangerous" would mean a user being told it was safe to cross when there was a train in the approach likely to arrive within the crossing time, including opening and closing gates etc. GPS (which I think was where the discussion started) just doesn't cut it on its own, even with augmentation. However there are some initiatives which look promising. One company I am aware of is combining GPS with image analysis from the forward facing camera and a map, so the train works out for itself (to a high level of integrity) where it is as it goes along, and could be asked (via data radio if only NR had implemented it with GSM-R) its exact location by the control centre. If it replies with a valid location, the time to arrival at the crossing could be calculated and a "safe to cross" (or not) message passed to the user. If the train doesn't reply you would be back to the current situation.

It was GPS, yes. And I agree, GPS alone is helpful but should never be relied on by itself for safety critical applications (not least because it can be switched off).
An augmented system like this was exactly the kind of thing I was advocating in favour of. No system can ever be perfect, but a system such as this would seem to be as near as you can get.

That said, I still say that the cost would be insane, and that the current system, used properly, is safe enough.

 

[Llanigraham]

There are some very interesting questions around this incident - like the above. Normally we would be reading a comprehensive report from RAIB to find out what happened - so why isn't there one?

I suspect simply because when the RAIB did their initial investigation they found that no-one on the signalling side did anything wrong. Certainly from reading the excellent report from the local paper at the inquest everything was done EXACTLY as per the Rule Book.

I said it before and I will say it again, there are thousands of these crossings in the UK used by thousands of people and there are VERY few incidents. Nearly every RAIB report on incidents at them have shown that it was the public at fault, generally by not following the instructions posted at the crossings.

And as for some of the ideas to "improve" (??) them, I really do not think some of you realise just how many of these types of crossings there are and how much it would cost to implement anything that would be infallible.

 

[Llanigraham]

On another thinking. I also noticed on youtube cab view videos that a lot of these user worked crossing have treadles that bring up a red light and lock the gates. That's what I imagine happens, or should happen anyway. I wonder if the driver of the land rover just opened the gates and there was no interlock or it was faulty.

Wrong. There are SOME that work that way, but it certainly isn't many.

 

[alxndr]

Like in the above post I did. Why did this happen?. Surely the signalman got the red in section lights on the panel. But why did not the lights on the crossing stay on until the train had left that section. One would assume the bolts would be locked as well.

I think you're slightly confused about the nature of this crossing. It's just a pair of gates and a phone. No lights or lockable bolts. The only way the vehicle driver can obtain information about whether it is safe to cross is by telephoning the signaller.

There are hundreds of such User Worked Crossings across the country, mainly for accessing properties.

 

[bramling]

Because the system works and there are very few accidents compared to the number of uses.
Read my reply at post #4 above.

It has always surprised me that there is any process which allows people to cross with anything less than signals either side collared at danger and nothing in section. However it does seem that this would be simply impractical on some routes with either long sections or a frequent service, especially when considered that the longer a person waits the more likely they are to take things into their own hands. It’s surprising there aren’t more accidents really.

 

[Mintona]

It has always surprised me that there is any process which allows people to cross with anything less than signals either side collared at danger and nothing in section. However it does seem that this would be simply impractical on some routes with either long sections or a frequent service, especially when considered that the longer a person waits the more likely they are to take things into their own hands. It’s surprising there aren’t more accidents really.

If you put a signal either side of the crossing to protect it you would end up with hundreds of extra signals being installed, some of them very close together. Further along this same line is a run of three farm crossings within a mile of each other, you could hardly put a signal in to protect each one, as they’d be too close together to provide adequate braking distance.

 

[alxndr]

We have to remember that the track circuit indication covers a distance of 3.7 miles on the down line at Frampton, of which the crossing is roughly half way along within it...

This has got me thinking. What you could do is provide the signaller with indications for "before crossing" and "after crossing" in situations like these where a crossing goes through a long track circuited section. To try to explain my completely speculative idea in simple terms:

A track circuit has a maximum length. Where the space between signals is longer than the maximum length of a track circuit it is divided up into different track circuits, so while the signaller simply gets in indication for 'AA' on this panel, on the ground this might actually be track circuits 'AA/1', 'AA/2', 'AA/3'... All of the track circuits are summated and this is indicated to the signaller, so that if any one of the track circuits becomes occupied, 'AA' on the panel shows occupied. This cuts down on the amount of information needed to be transmitted to the panel and because the signaller doesn't need the additional information in order to run their trains.

This does mean, however, that where a level crossing crosses a long section as far as they're concerned the train could be anywhere within that section. Their only options are to wait until the train has completely cleared the section or rely on the word of a member of the public saying that it has passed.

What could be done is to summate all of the track sections before and including the crossing and indicate those, and all of the track sections after the crossing, and indicate those. The signaller could then receive a phonecall asking for permission to cross, see that the section over the crossing was occupied, check an additional set of indications to see whether the train was approaching the crossing or if it had already passed, and permit or deny on that basis.

To illustrate an example:

A train has occupied track circuit 'AB/1'. This shows section 'AB' on the panel to be occupied, as is traditional. As one of the sections 1-3 are occupied the "approaching" indication is also lit. Once the train passes the crossing and occupies 'AB/4' or 'AB/5' the user calls back, the signaller confirms that the "approaching" indication as extinguished, the "beyond" indication has illuminated, and the user may cross. Admittedly, on a bi-directional line you would need to come up with a better name than approach and beyond.

I believe that doing this would be cheaper to install than a complete redesign as it wouldn't mean completely changing the track circuiting arrangements, but simply taking another output from each, summating them as necessary, and transmitting those to the panel. It still wouldn't be cheap, but would most likely be cheaper. The bulk of the cost would come from producing the initial design, testing, and accepting it into use. That cost would probably still make it prohibitive, and this would only provide a solution for track circuited lines. Axle counter sections have a theoretically unlimited length and thus there wouldn't be a direct equivalent possible.

The cheap option would be to have a Rulebook change and prevent permission being granted with a train in section, as is now the case for trackworkers line blockages, but realistically the public is too impatient for that would it would just make matters worse.

 

[Highlandspring]

Some signallers will use ‘unofficial’ indications of train position where possible (particularly in long Absolute Block sections) such as the operation of an AHB indicator or sounding of an annuciator but this practice is fraught with risk and it has led to many near misses in the past.

 

[6Gman]

You would of thought that all crossings would have to have some sort of locks or lights. especially now in these heath and safety times.

I would support that. So long as it's all paid for by the crossing users.

 

[4069]

If it's like the Athelney, near Taunton, crossing accident report from a few years ago, the investigation team somehow fail to notice what is staring the rest of us in the face.

Meaning what?

 

[bramling]

If you put a signal either side of the crossing to protect it you would end up with hundreds of extra signals being installed, some of them very close together. Further along this same line is a run of three farm crossings within a mile of each other, you could hardly put a signal in to protect each one, as they’d be too close together to provide adequate braking distance.

This is my point, it’s impractical, so we’d be down to holding the nearest signal in the rear at danger, which could be many miles away, or even if close still have an undesirable impact on capacity on a high-frequency service, or result in someone being asked to call back even when they might have seen the train pass and nothing else passes until they do eventually get permission.

However it still points towards a compromise “best of a bad job” setup which is far from ideal, and in many cases placing quite a burden of responsibility on the signalman, and effectively having a single point of failure in a safety-critical procedure. Naturally the busier the railway gets the more probability there is of a mistake occurring.

 

[R G NOW.]

I think NR are getting ready to resignal this line soon, and about time. The signalling in Gloucester and Cheltenham is now just over 50 years old. The signals need to be three aspect. On looking on a photo website. it appears that Gloucester has no control on that part of the crossing, by looking at that it is set from the Swindon workstation. I have reason to believe the telephone should have been wired to contact TVSC instead. I do not think that the signalman at Gloucester could have done anything to prevent it.

EDIT: There is a placement switch on the up side towards London, but not one from Swindon as it looks like the route before that is set from Swindon.

 

[The Nomad]

With regard to GPS signals, surely the answer is that it shows when the position signal last refreshed - either as a timer or, say, a red outline if it's been more than a period of time.

Thus in this scenario, the signaller would know almost exactly where it was or say that it's not safe to cross presently until the train is located exactly.

 

[High Dyke]

You would of thought that all crossings would have to have some sort of locks or lights. especially now in these heath and safety times.

You may be right, but that may come back down to cost of implementation. However until recently it was identified that there are a number of User Worked Crossings (UWC) that signallers aren't aware of. Following a miscommunication regarding the correct identification of a level crossing near Barnetby (North Lincolnshire), which luckily didn't result in an incident, it was found there are numerous crossings on the rail network that are not identified on signal box diagrams or in the signal box instructions - though this should be changed by now, or work in progress.

 

[ForTheLoveOf]

You may be right, but that may come back down to cost of implementation. However until recently it was identified that there are a number of User Worked Crossings (UWC) that signallers aren't aware of. Following a miscommunication regarding the correct identification of a level crossing near Barnetby (North Lincolnshire), which luckily didn't result in an incident, it was found there are numerous crossings on the rail network that are not identified on signal box diagrams or in the signal box instructions - though this should be changed by now, or work in progress.

Are you saying that these weren't on the Sectional Appendix or that the signallers didn't have a copy of the Sectional Appendix in the box?

 

[Alanko]

In this day and age you would think that all locos, EMUs and DMUs were fitted with GPS tracker though. Many buses do, and it's not expensive equipment.

From a layman's perspective, similar questions are asked about aviation; why either some fairly mainstream piece of helpful technology isn't commonly employed, or why some antiquated technology is still heavily relied on.

I think the challenge would be phasing GPS into other working practices and standards and training up vast numbers of staff so that it doesn't all go wrong when somebody gets too reliant on what a dot on the map is telling them.

In this era of 'The Internet of Things' GPS in rail vehicles feels entirely possible, so there must be a compelling reason for why it isn't done.

 

[Llanigraham]

With regard to GPS signals, surely the answer is that it shows when the position signal last refreshed - either as a timer or, say, a red outline if it's been more than a period of time.

Thus in this scenario, the signaller would know almost exactly where it was or say that it's not safe to cross presently until the train is located exactly.

As has been said previously, there are too many signalboxes that would need equipping with a display, plus there is the astronomical cost of providing such a service, plus GPS is not dependable.

 

[alxndr]

I think NR are getting ready to resignal this line soon, and about time. The signalling in Gloucester and Cheltenham is now just over 50 years old. The signals need to be three aspect. On looking on a photo website. it appears that Gloucester has no control on that part of the crossing, by looking at that it is set from the Swindon workstation. I have reason to believe the telephone should have been wired to contact TVSC instead. I do not think that the signalman at Gloucester could have done anything to prevent it.

EDIT: There is a placement switch on the up side towards London, but not one from Swindon as it looks like the route before that is set from Swindon.

Presuming you're talking about Frampton crossing, TVSC takes control some 4 miles away. They do not control that area, receive indications for it, nor can they send trains past their last signal without Gloucester's input. Gloucester is certainly the correct signaller for a user at that crossing to be speaking with.

 

[sefton]

I suspect that from their initial investigation they found that the signalled had done nothing wrong and this was just an unfortunate accident.

Even if someone is following the rule book, it doesn't mean that improvements cannot be made to prevent someone else dying.

It is rather poor form to continue to operate a system that you know kills people without even considering whether anything could be changed to prevent or reduce the possibility of the same thing happening again.

 

[philthetube]

I would support that. So long as it's all paid for by the crossing users.

Not the crossing users fault that someone else wants to make access to their property more expensive.

 

[talltim]

Even if someone is following the rule book, it doesn't mean that improvements cannot be made to prevent someone else dying.

It is rather poor form to continue to operate a system that you know kills people without even considering whether anything could be changed to prevent or reduce the possibility of the same thing happening again.

And yet we do in so many other areas, where safety is already lower. Roads are a prime example.