Near Miss At User Worked Crossing @ Bagillt - RAIB launch Investigation

[Y Ddraig Coch]

Another near miss at this user worked crossing for large vehicles to avoid a low bridge. The third incident in the last 4 years. All incidents had been granted permission by the signaller. This time from ROC in Cardiff rather than the local signal man as previously.

https://www.gov.uk/government/news/safety-incident-at-bagillt-user-worked-crossing

At around 11:56 hrs on Friday 17 August, a passenger train passed over a user worked level crossing, near Bagillt, North Wales, shortly after a user with a heavy good vehicle had completed using the crossing. The train, the 09:53 Manchester Piccadilly to Holyhead service, was travelling at around 75 mph (121 km/h). The private level crossing, which gives access to industrial premises, is only used by vehicles which are too tall to pass under a nearby bridge. The route over the crossing goes over two widely spaced tracks, and the crossing gates are 25 metres apart. The gates are kept locked. The crossing is fitted with telephones for users to contact the signaller and request permission to cross. To use this crossing, vehicle drivers must unlock and open both gates on foot, drive their vehicle over and reclose and lock both gates on foot.

The user requested permission to cross the railway with a ‘wagon’, and the signaller, based at the Wales Railway Operations Centre in Cardiff, granted it.

There had been similar previous occurrences at this crossing, on 31 October 2017 and 10 December 2014. At those times the level crossing was supervised by signallers based locally at Holywell Junction signal box. Our investigation will determine the sequence of events. It will also include consideration of:

• the actions of those involved
• the rules and guidance for signallers relating to factors which need to be considered when giving permission for the crossing to be used
• the information and guidance provided to users of private level crossings
• the management of safety risk at this crossing
• any relevant underlying management factors

Our investigation is independent of any investigation by the railway industry or by the industry’s regulator, the Office of Rail and Road.

We will publish our findings, including any recommendations to improve safety, at the conclusion of our investigation. This report will be available on our website.

 

[Chris M]

The report into this near miss was released last week.
https://www.gov.uk/government/news/...-irregularity-at-bagillt-user-worked-crossing

At around 11:57 hrs on Friday 17 August 2018, a passenger train passed over Bagillt user worked level crossing, Flintshire, shortly after a very large road vehicle had crossed. Railway signals had not been set to stop trains from approaching the crossing. A person assisting the vehicle driver, who was walking back over the crossing to close the gates behind the vehicle, was alarmed to see the approaching train and ran off the crossing.

The vehicle driver’s assistant had telephoned the signaller and obtained his permission before crossing the railway, but the signaller had not stopped trains approaching when a large vehicle needed to cross the railway, as required by the Rule Book. The user had not told the signaller that the vehicle was large, as required by a sign displayed at the crossing. The signaller did not ask questions to establish the size of the vehicle, and did not know that most people using this crossing did so with heavy goods vehicles, although some Network Rail staff were aware of this. Network Rail was unaware that this exceptionally heavy vehicle, subject to special requirements when on public roads, used the crossing regularly.

Underlying factors relate to Network Rail’s processes for risk management at this type of level crossing. These did not provide railway staff or road users with a coherent and consistent process for deciding when a vehicle should be treated as ‘large’, and did not provide an effective interface between signallers, crossing users and railway staff responsible for liaison with users and inspecting level crossings. An observation identifies further shortcomings in the information provided to signallers.

Recommendations
The report contains one recommendation addressed to Network Rail, seeking improvements in its management processes for user worked crossings with telephones.

The report also contains two further observations. One, relating to how signallers decide when it is safe for users to cross the railway at level crossings, provides evidence supporting the need for Network Rail to complete implementation of a previous RAIB recommendation. The other notes poor application of safety critical communication protocols in some training material.

The RAIB has identified five learning points. Four relate to dealing with requests to cross the railway at user worked level crossings. These cover clear communication about the characteristics of road vehicles needing to cross the railway, the circumstances when signal protection is needed, making allowance for differing train speeds when deciding when it is safe for users to cross and achieving safety critical communication standards when speaking with members of the public. The final learning point relates to correct use of safety critical communication protocols in training material.

Although a skim reading of that summary puts most of the blame on the signaller, the main failings are institutional and revolve around Network Rail not giving all relevant parties (including the signaller) the necessary information, in part because they hadn't collected it. It wasn't collected because the people doing the collecting didn't know they weren't getting all the relevant information, in part because they didn't know what information would be relevant.

 

[Y Ddraig Coch]

The report into this near miss was released last week.
https://www.gov.uk/government/news/...-irregularity-at-bagillt-user-worked-crossing

Although a skim reading of that summary puts most of the blame on the signaller, the main failings are institutional and revolve around Network Rail not giving all relevant parties (including the signaller) the necessary information, in part because they hadn't collected it. It wasn't collected because the people doing the collecting didn't know they weren't getting all the relevant information, in part because they didn't know what information would be relevant.

This is where moving signalling to ROC's has a downside.

Local signallers would know there patch, know there crossings inside out and possibly be even able to see the crossing.

For all the benefits of ROC's lack of local knowledge is a down side. Hopefully NR will implement so more through checks and questions to be asked as a matter of course for UOC.

 

[ComUtoR]

Although a skim reading of that summary puts most of the blame on the signaller,

Please remember; the RAIB does not apportion blame.

The purpose of a Rail Accident Investigation Branch (RAIB) investigation is to improve railway safety by preventing future railway accidents or by mitigating their consequences. It is not the purpose of such an investigation to establish blame or liability. Accordingly, it is inappropriate that RAIB reports should be used to assign fault or blame, or determine liability, since neither the investigation nor the reporting process has been undertaken for that purpose.

 

[Chris M]

Please remember; the RAIB does not apportion blame.

Sorry, that was sloppy wording on my part.

 

[SilentGrade]

This is where moving signalling to ROC's has a downside.

What if said local signallers move into the ROC?

 

[Chris M]

What if said local signallers move into the ROC?

Even if they do, and retain all their local knowledge, then they will retire or move jobs at some point.
As the report makes clear, there needs to be a single document with all the relevant* knowledge on it, which is kept updated, that is available to signallers, crossing maintainers, crossing inspectors, people who liaise with authorised users, etc. and all of these people need to have input into deciding what is relevant.

 

[bobbyrail]

I am not a frequent user of user worked crossings but the rural crossings that i have used from memory have always stated that you must Stop, Obtain permission to cross, and Report to the signaller when you have safely crossed. Maybe it's just in the areas that i have travelled in but SOR has always been in my head. Did this crossing have any such rules in place.

 

[Randomer]

Yes, the report is very clear that the member of staff of the metal salvage yard phoned for and received permission to cross from the signaller...

 

[The BIgman1234]

The problem with these wonderful “ROCs” is that some clown who has never signalled a train in his/her life decides to triple the signallers workload . The usual line trotted out is that “you’ve got SARS or ARS” . Let me tell you having these systems is like having a trainee on with you , you have to watch it constantly.
Who knows why this incident occurred? Could have been the signallers workload , fatigue, misunderstanding of where the actual crossing user was .
But the signals protecting the crossing should have been replaced to danger and collared till the movement had been completed.

 

[Chris M]

The report makes it clear that fatigue was not an issue in this incident. The workload was mentioned, and passing comment made that it was potentially an issue, but it doesn't go into detail - presumably as doing so could not be justified in the circumstances (the signaller and their manager both said the workload was fairly typical, and there were other issues that were definitely contributory.
Yes, the signaller should have protected the movement with signals based on the actual nature of the movement, but the crossing attendant (an employee of the scrap metal dealer, not a railway employee) did not give the signaller the information required to know that. The crossing attendant didn't give that information because neither the signs nor the information the authorised user received from Network Rail made it clear that they needed to. Some bits of Network Rail knew that the crossing was primarily used by large and heavy vehicles but this information was not provided to the signaller, nor was the signaller aware that the formation was much wider than typical here. Additionally, the signallers were not given sufficient information to make a reliable and consistent judgement about the time a train would take to reach the crossing from any given point.
Based on the information available to the signaller, and their understanding of the nature of the movement, there was no need for signal protection according to either the rulebook or local instructions.

 

[MotCO]

Based on the information available to the signaller, and their understanding of the nature of the movement, there was no need for signal protection according to either the rulebook or local instructions.

Is there any downside with setting train signals to red until the second call from the railway crosser to confirm a safe crossing has been made? Is it that a train may be forced to slow down unnecessarily, but at least it would be safer?

 

[Chris M]

Is there any downside with setting train signals to red until the second call from the railway crosser to confirm a safe crossing has been made? Is it that a train may be forced to slow down unnecessarily, but at least it would be safer?

The report does discuss this.
A train may indeed be slowed down or brought to a stop unnecessarily. This is seen as undesirable so a larger gap ahead of the train is required, this can result in a significant wait on busy lines such as the one involved in this incident. The longer someone has to wait before being given permission to cross, the greater the likelihood of misuse (i.e. crossing without the signaller's permission), especially if there are gaps between trains passing over the crossing that are sufficiently long for a crossing user to cross safely (or to perceive the ability to do so based on their experiences of crossing roads) - this is obviously significantly less safe than the existing requirements.
For example if it takes a user 2 minutes to cross the railway and the nearest train is 5 minutes away, then it is clear that the user can be given permission to cross instantly and that they can do so without affecting the train service. However, due to signal positioning, etc. if signalling protection was required the trains could need to be a minimum of 10 minutes away in both directions to avoid unnecessary delay. Now think what happens if the typical interval between trains is 9 minutes. There would be multiple instances of gaps more than four times the necessary length, so the waiting user just gets fed up and after the fourth such interval (over half an hour after they arrived at the crossing) just crosses anyway, only to find that this gap was only 1½ minutes...

 

[The BIgman1234]

The report makes it clear that fatigue was not an issue in this incident. The workload was mentioned, and passing comment made that it was potentially an issue, but it doesn't go into detail - presumably as doing so could not be justified in the circumstances (the signaller and their manager both said the workload was fairly typical, and there were other issues that were definitely contributory.
Yes, the signaller should have protected the movement with signals based on the actual nature of the movement, but the crossing attendant (an employee of the scrap metal dealer, not a railway employee) did not give the signaller the information required to know that. The crossing attendant didn't give that information because neither the signs nor the information the authorised user received from Network Rail made it clear that they needed to. Some bits of Network Rail knew that the crossing was primarily used by large and heavy vehicles but this information was not provided to the signaller, nor was the signaller aware that the formation was much wider than typical here. Additionally, the signallers were not given sufficient information to make a reliable and consistent judgement about the time a train would take to reach the crossing from any given point.
Based on the information available to the signaller, and their understanding of the nature of the movement, there was no need for signal protection according to either the rulebook or local instructions.

Any crossing I have control over where the user request to use is protected by signals wether the rule book or box instructions say or not . It’s common sense and good practise . And those signals don’t get cleared until that user reports back in he/she is safely across .

 

[The BIgman1234]

Is there any downside with setting train signals to red until the second call from the railway crosser to confirm a safe crossing has been made? Is it that a train may be forced to slow down unnecessarily, but at least it would be safer?

In my box the signals are put to danger and if the user does not report back then it’s a caution for drivers till the first one reports the crossing is safe . Even if the user had said it will take 5 minutes and 10 have elapsed. It’s simple really , do you give a train a couple of minutes delay or run the risk of it hitting someone . If there’s one thing I’ve learned in the many years as a signaller and that’s the public are stupid when it comes to crossings !!!!

 

[bramling]

Any crossing I have control over where the user request to use is protected by signals wether the rule book or box instructions say or not . It’s common sense and good practise . And those signals don’t get cleared until that user reports back in he/she is safely across .

Personally I’ve got to say I feel much more comfortable with this. *Any* vehicle can break down on a crossing, and 3 minutes isn’t enough to inform the signaller and have signals put back to danger.

I know that AHBs operate on precisely the premise that vehicles aren’t protected, but they do tend to be better laid out than the average UWC such that the chance of getting stuck (for an average vehicle) is minimal.

There are two issues with taking the “always protect” line that I can see:
1) As it’s technically not quite what the Rule Book says, it presumably requires sympathetic local management to back up the signallers, especially if it results in delays that wouldn’t otherwise have occurred
2) It does potentially increase the chance of misuse. Having said that I think on balance I’d rather run that risk as the lesser of two evils.

 

[Sunset route]

In our ASC we always use the “always protect method” regardless of what’s crossing and the LOMs (yes there are two) have never complained about it.

 

[bramling]

In our ASC we always use the “always protect method” regardless of what’s crossing and the LOMs (yes there are two) have never complained about it.

Good to here there are managers about who are prepared to support their staff.

 

[The BIgman1234]

In our ASC we always use the “always protect method” regardless of what’s crossing and the LOMs (yes there are two) have never complained about it.

Same in our signalling centre . Why run the risk of a collision for the sake of a 2 minute delay .

 

[Llanigraham]

I had 3 UWC's at my Box, and we worked on the simple principle that if there was a train in Section, or we had been asked "Line Clear" then the user got told to wait. One of the crossings the (general) user could get stroppy sometimes if they had to wait too long, but tough it was their life that was on the line.
Our biggest problem one was a Footpath UWC where phones we added, where we used to ask people to wait. The number of times they never came back to us was large; we suspected that they ignored our instructions and crossed anyway.

 

[cyclebytrain]

Having read the report, the question that stands out to me is why are we relying on a relatively untrained crossing user to use a telephone and describe their crossing needs and location accurately? All of these are liable to introduce errors, whereas cameras are both cheap and useful for other purposes and a grid of (e.g.) 4 buttons with pictures of lorry, van, car, sheep which the signaller can light either green or red is clear, unambiguous, faster than a call and is similar to a pelican crossing which users are likely to be more familiar with.

Is it just the cost of such systems, or are there other problems that I'm not seeing from 5 minutes of thinking about it?

 

[Llanigraham]

Problems:
Cost.
Power supply to the cameras.
Cabling back to Signal Box/Control Centre.
Remoteness.
Distance to the Box/SCC.
Local knowledge.
The sheer number of UWC's there are in the country. As said in the sections I controlled either side of my Box I only had 3. Now compare that with Machynlleth which has over 100!

And to be honest are we trying to find a solution to a problem that doesn't actually exist? There are thousands of UWC's across the country, used by thousands of users every day and the number of incidents, by percentage, is very low.

 

[bramling]

I had 3 UWC's at my Box, and we worked on the simple principle that if there was a train in Section, or we had been asked "Line Clear" then the user got told to wait. One of the crossings the (general) user could get stroppy sometimes if they had to wait too long, but tough it was their life that was on the line.
Our biggest problem one was a Footpath UWC where phones we added, where we used to ask people to wait. The number of times they never came back to us was large; we suspected that they ignored our instructions and crossed anyway.

At the end of the day misuse of a footpath crossing is unlikely to endanger the safety of rail staff / passengers. So whilst misuse isn't to be encouraged, that's their prerogative to take the risk. The prospect of a large road vehicle on a UWC is of far more concern. In all honesty it's always surprised me that the rules allow decisions to be taken based on time predictions and essentially relying on luck that nothing goes wrong while a vehicle is crossing (i.e. a breakdown). IMO it also places quite an unfair burden of liability on the signalman. I feel much more comfortable hearing that many places are taking the initiative to provide extra protection.

Having said all this, there are of course plenty of crossings where there is no requirement to phone the signalman at all, with the process essentially relying on sighting time. I'm even less comfortable with that.

 

[edwin_m]

Network Rail have tried to address some of the UWC risk with the Power Operated Gate Opening (POGO) system but to my way of thinking they made a basic error with this. The road vehicle driver has red and green lights showing whether it is safe to cross, and buttons to operate both sets of gates remotely, but the two are not interlocked and the onus is on the driver to check the lights are still green after opening the gates and before driving across. I imagine 99% of road drivers with no specialist knowledge would assume that if the system lets them open the gates it must be safe to cross.

 

[lineclear]

It must be nice to have a quiet line where every vehicle can be granted signal protection, but the reality is that doing so on busy lines would keep crossing users waiting for ten minutes or more. Keeping people waiting that long can and does result in misuse.

 

[tsr]

In all honesty it's always surprised me that the rules allow decisions to be taken based on time predictions and essentially relying on luck that nothing goes wrong while a vehicle is crossing (i.e. a breakdown). IMO it also places quite an unfair burden of liability on the signalman.

Indeed. It’s effectively “time interval” signalling, which was outlawed a long time ago when it comes to controlling the rest of the railway.

 

[Wilts Wanderer]

I personally find it astonishing this is done on a time-interval basis. Would you grant a possession on this basis without signal protection? Of course not, so why allow a vehicle across that could cause a serious collision. Or has everyone forgotten what happened at Hixon?

 

[Llanigraham]

It is quite obvious that some members here do not understand the job that signallers actually do in respect of UWC's, even after several of us have explained. I will repeat what I said above, are we trying to find a solution to a problem that doesn't actually exist? There are thousands of UWC's across the country, used by thousands of users every day and the number of incidents, by percentage, is very low.

 

[Llanigraham]

I personally find it astonishing this is done on a time-interval basis. Would you grant a possession on this basis without signal protection? Of course not, so why allow a vehicle across that could cause a serious collision. Or has everyone forgotten what happened at Hixon?

Hixon was not a UWC.

 

[edwin_m]

I personally find it astonishing this is done on a time-interval basis. Would you grant a possession on this basis without signal protection? Of course not, so why allow a vehicle across that could cause a serious collision. Or has everyone forgotten what happened at Hixon?

That's not the issue here. What happened in this case was that for various reasons the signaller wasn't aware that it was a particularly large and slow vehicle wanting to cross. Had they known that then they would have applied signal protection.

The suggestion upthread of using cameras seems a good one to me. If bandwidth is an issue they could just be still cameras trained on the approaches and on the crossing itself, and so the signaller could request a picture of the waiting vehicle and also confirm the crossing clear afterwards.