Level crossing incident near Norwich new RAIB investigation

Taunton said:

Slowing down, not stopping.

Click to expand...

I think it would be quite easy at many crossings for a driver not to realise how close they had to be to get a good view down the line, and therefore how slow they had to go so they could stop short if they saw a train. When focused on getting that view, and slowing down more than anticipated, it would also be quite easy to forget to change into a lower gear. Then accelerating again could stall the engine, leaving the car rolling at a speed that wouldn't get it across the crossing. Or the driver might get focused on re-starting the engine and brake to a stop on the crossing to achieve that, then be unable to re-start due to some random problem or just getting flustered - particularly if the crossing sequence started just then too.

Llama said:

Actually Track Circuit Actuator is the rulebook name for that equipment, but that is a misleading name and the technical name is Track Circuit Assister.

Click to expand...

Unfortunately I only work to one rule book and glossary of terms. But if the techs call it something else, well C'est la vie

this seems to be where the footage came from

The RAIB have at last published their report into the above incident. A summary is posted below and the full report can be found here:

Summary​

On 24 November 2019, the barriers at Norwich Road level crossing, near New Rackheath, Norfolk, lifted as a passenger train from Norwich to Sheringham was approaching. Two road vehicles crossed the railway in front of the train, which reached the crossing less than half a second after the second road vehicle was clear.

The investigation found that there was contamination of the railhead in the area caused by leaf-fall and atmospheric conditions. This contamination had not been removed because there were no railhead treatment trains on the Norwich to Sheringham line at weekends. The narrow band on which trains’ wheels were running on the contaminated railhead, which was a consequence of the introduction of new trains, left the wheel-rail interface vulnerable to a poor electrical contact in the event of contamination. This caused the level crossing equipment to misinterpret the position of the train, and consequently it opened the crossing to road traffic while the train was closely approaching.

Recommendations​

RAIB has made three recommendations addressed to Network Rail regarding the planning of autumn railhead treatment, guidance on the introduction of new trains and the configuration control of signalling equipment. RAIB also identified two learning points concerning the investigation of incidents and the signalling design process.

Simon French, Chief Inspector of Rail Accidents said:​

“All too often the interaction between road users and the railway at level crossings leads to incidents and accidents. In many cases the actions of the road user are the immediate cause, but in this alarming event, deficiencies in the way the railway equipment operated placed two car drivers, and the people on a passenger train, in deadly danger through no fault of their own.

“Our investigation found that the installation at Norwich Road level crossing was a poor piece of engineering which had been in use for several years, and only luck had previously prevented an accident. A change in the type of train using the line exposed a weakness in the way that the crossing’s electronic control equipment was configured. This meant that the system lost sight of the approaching train and commanded the barriers to rise. Lessons from the trial of the crossing equipment in other locations, and from incidents involving the same equipment in service, had not been applied to the crossing at Norwich Road. It is important that the railway industry learns from this incident, and makes sure that it has effective processes in place to transfer such learning to where it is needed.”

Click to expand...

Another fantastic report by the RAIB. To think that on straight track, with the wheelsets in pristine condition, most stock only makes a 4mm wide contact is very surprising. I imagine close attention will be paid to all level crossings with similar detection methods in high leaf-fall areas going forward.

Roast Veg said:

Another fantastic report by the RAIB. To think that on straight track, with the wheelsets in pristine condition, most stock only makes a 4mm wide contact is very surprising. I imagine close attention will be paid to all level crossings with similar detection methods in high leaf-fall areas going forward.

Click to expand...

Its an interesting trade-off between passenger ride comfort and an engineering consideration. I imagine any rolling stock manufacturer / operator will be striving to minimise side-to-side movement and reduce flange contact etc, but the track engineer might see it differently?

(A possible illustration of the effect of separating the railway vertically - rolling stock operator and infrastructure maintainer objectives no longer align for the common good.)

Wilts Wanderer said:

Its an interesting trade-off between passenger ride comfort and an engineering consideration. I imagine any rolling stock manufacturer / operator will be striving to minimise side-to-side movement and reduce flange contact etc, but the track engineer might see it differently?

(A possible illustration of the effect of separating the railway vertically - rolling stock operator and infrastructure maintainer objectives no longer align for the common good.)

Click to expand...

I don't think that any sort of vertical integration would have prevented this situation any better. Network Rail still have complete judicial say over what operates on their infrastructure, and their specifications and standards drive rolling stock order details. At best, the communication between parties might have been better or more frequent, but from what I know of large organisations and their internal struggles, I'm inclined to think not...

All too often the interaction between road users and the railway at level crossings leads to incidents and accidents. In many cases the actions of the road user are the immediate cause, but in this alarming event, deficiencies in the way the railway equipment operated placed two car drivers, and the people on a passenger train, in deadly danger

Click to expand...

Bit like Athelney then ...

I never liked American level crossing predictors. There are many many thousands of them in the USA clearly, operating on, in some cases, really horrible track, but typically with big ass rolling stock grinding along and keeping the railhead clean. I've nonetheless seen some hair-raising videos of barriers not falling on very sparsely trafficked lines until (say) runaway wagons are almost on top of the crossing, or where they raised just as a train approached as in this case. I don't think vertical integration could have made significant difference; it is still an interface between different engineering departments and mainly a failing in S&T engineering. Note, this type of incident could also have happened if all the wheelsets of the old rolling stock had been replaced or ground back to 'as new' profile (although that wasn't likely at all in this case, as they were all about to be withdrawn or transferred away soon), so it really wasn't much to do with the introduction of new stock as such, although there were issues uncovered with that stock, particularly the delay in sanding activation which no doubt affected the emergency braking response effectiveness.

The inadequate 16-second manufacturer's default 'LOS' timeout was the primary problem clearly. Most traditionally controlled AHBCs will usually only open again after a warning sequence has been triggered following a subsequent series of 'strike-out' conditions involving both track circuits going clear AND operation and normalisation of a mechanical treadle switch very close to the crossing. There are some odd failure modes where the long failure timer could allow the barriers to open with a stationary train parked across the road, but never with a moving train approaching.

I noticed the very narrow bright metal band phenomenon on railheads of the ex LSWR mainline in the west country, where normally only the almost identical class 159 and 158 trains use the track east of Exmouth Junction. A lightweight track machine or a different class of unit along the route would have a high risk of disappearing completely from train detection due to its own narrow wheeltread contact patch being on a slightly different part of the railhead which, even without external contaminants, could be heavily rusted thus representing a higher resistance to track circuit current and more likely to register as a false clear. The modern TCB signalling on that route east of Pinhoe is axle counter based today however, which doesn't suffer the same problems, but some of the level crossings en route may still rely on local track circuits.

MarkyT said:

I never liked American level crossing predictors. There are many many thousands of them in the USA clearly, operating on, in some cases, really horrible track, but typically with big ass rolling stock grinding along and keeping the railhead clean. I've nonetheless seen some hair-raising videos of barriers not falling

Click to expand...

How I handle level crossings here was developed from my time when in the USA. It's compulsory for buses and trucks there to stop and check at every crossing, regardless of type. They always put on their 4-way flashers when doing so, and buses open the door to hear the train whistle better. So for me it's radio off, open window to listen, slow, and glance up and down. At night you would see the headlight. US crossings are also far better than here in clearing away any vegetation that restricts visibility for the vehicle driver. I've never experienced issues of US crossing non-operation, not all their trains are heavy axle load, in many mountain districts the train is preceded by a lightweight speeder or even a converted road pickup on rail wheels, which nevertheless always make the equipment operate.

Taunton said:

How I handle level crossings here was developed from my time when in the USA. It's compulsory for buses and trucks there to stop and check at every crossing, regardless of type. They always put on their 4-way flashers when doing so, and buses open the door to hear the train whistle better. So for me it's radio off, open window to listen, slow, and glance up and down. At night you would see the headlight. US crossings are also far better than here in clearing away any vegetation that restricts visibility for the vehicle driver. I've never experienced issues of US crossing non-operation, not all their trains are heavy axle load, in many mountain districts the train is preceded by a lightweight speeder or even a converted road pickup on rail wheels, which nevertheless always make the equipment operate.

Click to expand...

Thats astonishing, is it presumably to prove the route is safe for the train following behind? almost the modern equivalent of a man walking ahead with a red flag in the early days of road automobiles.

Wilts Wanderer said:

That's astonishing, is it presumably to prove the route is safe for the train following behind? almost the modern equivalent of a man walking ahead with a red flag in the early days of road automobiles.

Click to expand...

Always done it. It's to look out for rockfalls etc. They communicate by radio back with the main train about a mile behind.

Back to the Norfolk event.

Roast Veg said:

Another fantastic report by the RAIB. To think that on straight track, with the wheelsets in pristine condition, most stock only makes a 4mm wide contact is very surprising. I imagine close attention will be paid to all level crossings with similar detection methods in high leaf-fall areas going forward.

Click to expand...

The report says that similar crossings on Marston Vale had a longer detection timeout so were less prone to this problem, but for some reason the timeout was never adjusted on this route. All such crossings have now been adjusted and fitted with treadles which ensure the sequence is triggered even if the train is not detected. This is also required for other crossing types.

So it would appear enough has been done to prevent a repeat of this situation.

Wilts Wanderer said:

Its an interesting trade-off between passenger ride comfort and an engineering consideration. I imagine any rolling stock manufacturer / operator will be striving to minimise side-to-side movement and reduce flange contact etc, but the track engineer might see it differently?

(A possible illustration of the effect of separating the railway vertically - rolling stock operator and infrastructure maintainer objectives no longer align for the common good.)

Click to expand...

I recall visiting Heaton Lodge Junction in 1988 with the local p-way engineer who pointed out the same thing, at a time when nearly all trains on the route were nearly-new Sprinters. It's part of the reason these units were fitted with TCA soon after that.

Taunton said:

How I handle level crossings here was developed from my time when in the USA. It's compulsory for buses and trucks there to stop and check at every crossing, regardless of type. They always put on their 4-way flashers when doing so, and buses open the door to hear the train whistle better. So for me it's radio off, open window to listen, slow, and glance up and down. At night you would see the headlight.

Click to expand...

There is of course always the risk that if stopping just before a crossing, a driver gets into the wrong gear and stalls on the crossing itself. Less of an issue in America where most vehicles are automatics.

Taunton said:

How I handle level crossings here was developed from my time when in the USA. It's compulsory for buses and trucks there to stop and check at every crossing, regardless of type. They always put on their 4-way flashers when doing so, and buses open the door to hear the train whistle better. So for me it's radio off, open window to listen, slow, and glance up and down. At night you would see the headlight. US crossings are also far better than here in clearing away any vegetation that restricts visibility for the vehicle driver. I've never experienced issues of US crossing non-operation, not all their trains are heavy axle load, in many mountain districts the train is preceded by a lightweight speeder or even a converted road pickup on rail wheels, which nevertheless always make the equipment operate.

Click to expand...

There's some mention in the report about the comparatively low voltages and currents involved, and I'm sure the fairly old now tech has been refined to work very well over the years in all kinds of scenarios. I didn't mean to imply the tech was inadequate in any way. Some of these hair-raising incident videos occurred following manual resets under poorly controlled maintenance prodedures I think, but there could be issues with VERY rusty track after long periods of non-use. Crew also have to sound that special whistle code on approach and their horns are VERY loud. I assume there is signage on approach to each crossing to remind them. Most crossing accidents, in US, UK and no doubt everywhere else, are down to road user failings clearly.

Wilts Wanderer said:

Thats astonishing, is it presumably to prove the route is safe for the train following behind? almost the modern equivalent of a man walking ahead with a red flag in the early days of road automobiles.

Click to expand...

I guess if there's only a handful of trains a day on a very remote and challenging route, sending out an inspection speeder shortly before the first one is an effective strategy.

My antipathy towards the technology is mainly based on the fact that in the USA it was developed to offers a consistent warning time for road users based on a potentially very wide range of train approach speeds. Unusually extended road closure times at half barrier crossings are disliked by safety professionals because they encourage misbehaviour by road users who may be fooled into thinking the equipment has failed and take a chance to weave through after waiting for a while. On a passenger dominated line like Cromer today, the variability in approach speeds is tiny, hence a much simpler in principle fixed strike in point sensor system is perfectly adequate to ensure a consistent warning time. Network Rail have now effectively acknowledged this by putting in the fixed reinforcement treadle equipment, which must be sited for the fastest trains that are the TYPICAL trains at each of the predictor sites now. When a slow but very rare engineering train traverses the route now for example, it will strike in at exactly the same point and thus trigger a rather longer warning time before it arrives at the crossing. There is a statistically derived limit as to how many trains may exceed an upper limit of warning time for an AHBC in UK. As long as that is not exceeded, the crossings can still be legally compliant with their level crossing orders.

edwin_m said:

All such crossings have now been adjusted and fitted with treadles which ensure the sequence is triggered even if the train is not detected. This is also required for other crossing types.

Click to expand...

Treadles have their own issues; they were the initiation of the Athelney (near Taunton) accident where a tamper was run wrong line over the crossing overnight and upset the treadle sequence. The accident train was the first service the following morning.

Taunton said:

Treadles have their own issues; they were the initiation of the Athelney (near Taunton) accident where a tamper was run wrong line over the crossing overnight and upset the treadle sequence. The accident train was the first service the following morning.

Click to expand...

R042014_140224_Athelney.pdf (publishing.service.gov.uk)

Athelney level crossing is configured such that the crossing closure sequence for an up train starts when the train operates a strike-in treadle located 1.1 miles (1.8 km) from the crossing (figure 3). Occupation of VY/2 track circuit, which is co-located with the treadle3 , will also initiate the closure sequence. This is described as ‘independent treadle and track circuit operation’. Later designs of automatic level crossing used ‘treadle reinforced track circuits’, in which the strike-in treadle operates the associated track circuit either by short-circuiting the track relay or by disconnecting the electrical supply to the track circuit.

Click to expand...

It is strongly implied, but on a quick skim doesn't appear to be actually stated, that the later design with treadle reinforced track circuits would not have this problem. There is a suggestion that other crossings on the Western with the independent operation were to be modified with directional treadles to prevent the same situation arising.

BBC news item with a significant inaccuracy - BBC News "
"The cars were crossing the tracks at New Rackheath on 24 November 2019 because the barriers had not come down." (my bolding).

Report is quite clear, barriers had lowered, but raised prematurely, I've reported it to BBC.

Norwich to Sheringham train in 'near miss' due to leaves on the line​

Published1 hour ago
Share

IMAGE COPYRIGHTGOOGLE
image captionThere were "heavy deposits of leaf debris" on the tracks near New Rackheath level crossing
A train almost hit two cars on a level crossing due to leaves on the line, an investigation in Norfolk found.
The cars were crossing the tracks at New Rackheath on 24 November 2019 because the barriers had not come down.
That Norwich to Sheringham train missed them by less than half a second.
A Rail Accident Investigation Branch (RAIB) report said barrier sensors were not working properly because leaves had not been cleared and "only luck" had prevented an accident previously. ......

Click to expand...

So the RAIB report backs up Greater Anglia's statements at the time that it wasn't an issue with the new trains. Perhaps those who were falling over themselves at the time to criticise GA and Stadler would like to eat some humble pie and admit that they were wrong?

I think it's astonishing that the system, having "lost" a train, will just default to assuming everything is fine after a set number of seconds. Trains don't just vanish. In my mind an LOS event should cause the system to failsafe, keep the barriers closed and alert a signaller or something.

MarkyT said:

My antipathy towards the technology is mainly based on the fact that in the USA it was developed to offers a consistent warning time for road users based on a potentially very wide range of train approach speeds. Unusually extended road closure times at half barrier crossings are disliked by safety professionals because they encourage misbehaviour by road users who may be fooled into thinking the equipment has failed and take a chance to weave through after waiting for a while. On a passenger dominated line like Cromer today, the variability in approach speeds is tiny, hence a much simpler in principle fixed strike in point sensor system is perfectly adequate to ensure a consistent warning time. Network Rail have now effectively acknowledged this by putting in the fixed reinforcement treadle equipment, which must be sited for the fastest trains that are the TYPICAL trains at each of the predictor sites now. When a slow but very rare engineering train traverses the route now for example, it will strike in at exactly the same point and thus trigger a rather longer warning time before it arrives at the crossing. There is a statistically derived limit as to how many trains may exceed an upper limit of warning time for an AHBC in UK. As long as that is not exceeded, the crossings can still be legally compliant with their level crossing orders.

Click to expand...

I completely agree with you that predictors on the line to Cromer don't make a great deal of sense, but the fitting of a reinforcement treadle doesn't undermine their functionality. Para 137 of the RAIB report:

All HXP3 predictor crossings have been fitted with pairs of reinforcement treadles. The new pairs of treadles are fitted at the equivalent of the strike-in point, the point at which the minimum warning time for the road user is achieved with the fastest train. The treadles will force the crossing sequence to start if the predictor has not registered the train while it was running between the limit of approach and the treadles. If the HXP3 predictor has registered the train before it reaches the treadles but not started the sequence due to the train’s speed, the treadles will have no effect.

Click to expand...

Sorry for the ignorance - but what exactly is a treadle - a bit like a tripcock but doesn't stop the train of course?

Mcq said:

Sorry for the ignorance - but what exactly is a treadle - a bit like a tripcock but doesn't stop the train of course?

Click to expand...

Wiki is your friend:


In railway signalling, a treadle is a mechanical or electrical device that detects that a train axle has passed a particular location. They are used where a track circuit requires reinforcing with additional information about a train's location, such as around an automatic level crossing, or in an annunciator circuit, that sounds a warning a train has passed an exact point. They also serve as a critical backup in the case of track circuit failure. The important difference between a treadle and a track circuit is that while a track circuit detects a train over a distance as long as several kilometres, a treadle provides detection at a single fixed location.

They need a fair bit of maintenance and on busy lines wear out quickly. I know of one crossing (near a very complex junction) that cost over £50k a year just replacing the damn things.

jfowkes said:

I think it's astonishing that the system, having "lost" a train, will just default to assuming everything is fine after a set number of seconds. Trains don't just vanish. In my mind an LOS event should cause the system to failsafe, keep the barriers closed and alert a signaller or something.

Click to expand...

The timer helps to assure that a momentary track circuit failure, or maintenance activities such as those contributing to the Athelney accident I linked above (where crossing didn't have the relevant feature) don't cause the crossing to close and stay closed indefinitely waiting for the non-existent train to pass. This could lead to road users zigzagging the barriers and being hit by a train when it eventually arrives. The report notes the standard timer value is 120s - if a train can disappear off a track circuit for that length of time then it's highly likely to suffer a disastrous derailment or collision next time it goes through a junction.

And WHY was level Crossing Predictor technology (which implicitly resulted in the omission of track circuit re-enforcement controls (treadles)) thought necessary in the first place in the UK - and by what organisation?

moggie said:

And WHY was level Crossing Predictor technology (which implicitly resulted in the omission of track circuit re-enforcement controls (treadles)) thought necessary in the first place in the UK - and by what organisation?

Click to expand...

I believe that the reason was to reduce the chance of road users misusing level crossings which they believed had failed because the barriers had been down for "longer than normal" without a train passing.

A predictor system detects the speed of approaching trains so that the time interval between barriers being lowered and a train arriving is similar for all trains, irrespective of their speed.

As to organisation I'm unsure - probably Network Rail or its predecessors?

jfowkes said:

I think it's astonishing that the system, having "lost" a train, will just default to assuming everything is fine after a set number of seconds. Trains don't just vanish. In my mind an LOS event should cause the system to failsafe, keep the barriers closed and alert a signaller or something.

Click to expand...

Quite! I work with electrical fail-safe systems and I can scarcely believe the way this equipment operates.

jfowkes said:

I think it's astonishing that the system, having "lost" a train, will just default to assuming everything is fine after a set number of seconds. Trains don't just vanish. In my mind an LOS event should cause the system to failsafe, keep the barriers closed and alert a signaller or something.

Click to expand...

Grumpy Git said:

Quite! I work with electrical fail-safe systems and I can scarcely believe the way this equipment operates.

Click to expand...

Agree. Although I'm not particularly surprised given what I've witnessed first hand and heard from colleagues about American industrial equipment safety standards. I AM surprised that the due diligence process when selecting this equipment didn't pick up this rather obvious design flaw.
I also don't really understand why this predictor equipment was installed at this location. It's a 55mph line speed, with the occasional freight to/from North Walsham. It's not like there's ever likely to be such a disparity in the time the train takes to arrive after the barriers go down that people decide to start driving round them.

It’s not a design flaw. In fact it’s a good idea that the barriers raise after a certain time as it stops dangerous complacency with road users and keeps road traffic from being blocked for hours. The issue is the timing of the barrier release...16 seconds might be ok if linespeed is 5mph, but at the actual speed of this line then 2-3 minutes would have prevented or at least drastically reduced this risk.

perhaps part of the reason these predictors were installed on this line was due to the close proximity of crossings to stations and wanting somekind of standard?

jfowkes said:

I think it's astonishing that the system, having "lost" a train, will just default to assuming everything is fine after a set number of seconds. Trains don't just vanish. In my mind an LOS event should cause the system to failsafe, keep the barriers closed and alert a signaller or something.

Click to expand...

66701GBRF said:

It’s not a design flaw. In fact it’s a good idea that the barriers raise after a certain time as it stops dangerous complacency with road users and keeps road traffic from being blocked for hours. The issue is the timing of the barrier release...16 seconds might be ok if linespeed is 5mph, but at the actual speed of this line then 2-3 minutes would have prevented or at least drastically reduced this risk.

perhaps part of the reason these predictors were installed on this line was due to the close proximity of crossings to stations and wanting somekind of standard?

Click to expand...

That default reset time of 16 seconds is the problem and perhaps the concept of a full reset entirely. Thinking about this further, I'm pretty sure traditional relay circuits for AHBCs DO NOT reset in this way at all, although I don't have any real-world example drawings to examine and confirm this, unfortunately. I believe failure to 'strike-out' should cause the barriers to stay down and red lights to continue flashing indefinitely. That's clearly undesirable from a road user safety perspective, but with no crossing clear check possible at this type of crossing it is the only failsafe option. The only realistic approach to minimising such events is to design very high reliability into the systems.

Total power loss will cause the barriers to fall or stay down, but a 12-hour standby battery is provided to practically eliminate that likelihood, and mains power failure is immediately alarmed to the box so a tech can be called out. The 120 second figure quoted in the report is the standard time after which the power off/failure alarm is sent to the box if the circuitry fails to normalise for this time after activation, but the road remains closed. The equipment doesn't reset and reopen it. If the signaller knows the alarm is because of an unusually slow train taking a long time to clear, they can safely ignore it but if the alarm comes out of the blue, they must caution any approaching trains. There are many sensors incorporated in the circuits, and they monitor various sequences of events in normal operation, any of which, if interrupted, could plausibly send the alarm, including events such a road vehicle demolishing the barriers, actuator machine cabinets, road lights, or miscreants cutting and stealing cables!

A controlled or replaceable auto signal must be located within 5 minutes running time to protect an AHBC I recall. Where signals are within the strike-in, they are connected to the crossing, not to interlock them as such, but to enforce a clearance delay so minimum road warning time is ensured should a train be stopped at the signal for any reason. Selectable stopping/non-stopping controls may are incorporated where routine stopping of some trains is expected at or near such a signal, at a station platform for example. Signals are provided especially and solely for this purpose, where necessary.

It is LOCALLY MONITORED crossings of various types, with drivers' red/white flashing lights (for approaching rail traffic) that normally reopen to road traffic after failure timeout; I think in these cases the barrier machines are biassed to rise under total power failure, so the barriers are effectively forced down by the controllers and actuators, unlike AHBCs where they are designed to fall under gravity.

66701GBRF said:

It’s not a design flaw. In fact it’s a good idea that the barriers raise after a certain time as it stops dangerous complacency with road users and keeps road traffic from being blocked for hours. The issue is the timing of the barrier release...16 seconds might be ok if linespeed is 5mph, but at the actual speed of this line then 2-3 minutes would have prevented or at least drastically reduced this risk.

Click to expand...

I remember in the storm of 25 January 1991 I was in business offices overlooking Drayton Lane crossing, east of Chichester. It's CCTV full barriers now but it was AHB then. The barriers failed down with lights flashing for most of the day; rail service had been suspended. Presumably there was no equipment on the crossing to deal with whatever had failed. When we went to leave I got out, and from mid-crossing on a dead straight line handsignalled us (and another) across.