Incident at Talerddig, Wales - 21/10/2024

[Towers]

There weren’t any undamaged coaches in the collision, forces would have gone through both coaches in the set.

An example of this being the Clapham crash, the 12VEP that was on the Basingstoke train had a car written off in the middle of the train.

Also worth noting that on another thread here recently it was reported that all three cars of the SWR class 159 involved in the Salisbury Tunnel collision have been cut up - I seem to recall there was some expectation that at least the seemingly unaffected rearmost coach might survive and see further service, but evidently not so.

 

[ainsworth74]

The RAIB have released their interim report into this accident:

Summary

1 At around 19:26 on Monday 21 October 2024, train reporting number 1J25, the 18:31 passenger service from Shrewsbury to Aberystwyth, collided with train reporting number 1S71, the 19:09 Machynlleth to Shrewsbury passenger service. Both services were operated by Transport for Wales Rail Limited (TfWRL).

2 The collision took place on Network Rail’s Cambrian line to the west of the passing loop located at Talerddig, Powys (figure 1). One passenger died and four other people were seriously injured. Neither train derailed in the collision, although significant damage was caused to the leading vehicles of both trains.

3 From each direction, the railway approaching Talerddig passing loop consists of a single line, with the loop allowing trains to pass each other. Train 1J25, travelling west, was due to stop in the loop to allow eastbound train 1S71 to pass it. Train 1J25 was braking as it approached and passed through the loop. Despite this it was unable to stop within the loop as intended. The train subsequently exited the loop, while still braking, and re-entered the single line, heading towards train 1S71. Train 1J25 travelled approximately 1,080 metres beyond its intended stopping point, before colliding with train 1S71.

4 At the time of the collision, train 1J25 was travelling at approximately 39 km/h (24 mph), while train 1S71 was travelling in the opposite direction at approximately 11 km/h (6 mph).

RAIB’s role and the context of this interim report

5 The Rail Accident Investigation Branch (RAIB) is responsible for conducting independent investigations into railway and tramway accidents in the UK. The purpose of its investigations is to improve safety, by establishing the causes of the accident, and making recommendations to reduce the likelihood of similar occurrences in the future or to mitigate their consequences.

6 RAIB does not apportion blame or liability, enforce the law or carry out prosecutions. RAIB investigations are independent of those undertaken by other public bodies, such as the Office of Rail and Road, and the railway industry itself. However, we will work alongside investigations being undertaken by others and may share certain physical evidence with them, where this is appropriate and allowed by the law.

7 A final report will be published by RAIB on completion of its investigation. All RAIB investigation reports are available on RAIB’s website. If found necessary, during its investigation, RAIB may also issue urgent safety advice to the rail industry.

8 This interim report is based on the evidence gathered and analysed to date and RAIB’s findings should be considered within this context. RAIB’s intended future actions for this ongoing investigation are described in paragraph 87. Topics relevant to the accident are likely to be covered in more detail within the final report following this work. This may also cause modifications to the interim findings contained in this report.

Interim report 01/2025: Collision between passenger trains near Talerddig

Collision between passenger trains near Talerddig, Powys, 21 October 2024.

www.gov.uk

 

[endecotp]

Interesting bit starts at para 71. The sand hoses were blocked with a “30 mm plug of organic material”, which they have used DNA testing to identify!

 

[ainsworth74]

One thought that occurs, and I presume the RAIB will investigate or at least perhaps provide an explanation for, is that this slide must have taken place over quite a period of time. We know that the train was in service braking for around 40 seconds before being put into emergency, it travelled the length of the loop (I think around 500m or so?), then we know it travelled another 900m before actually impacting the Shrewbury bound train. Whilst we don't know the exact speed profile (how fast was it going by the time emergency was kicked in, how much more speed did it lose after that point and the collision) it would seem that that must have taken maybe 60 seconds or so? At around 70kmh (well above the top bound for the impact speed) it would have taken 70 seconds to cover the 1,400m of the loop and the the point of impact.

I wonder if there was a Railway Emergency Call made from the incident train during the slide? That would surely have caused the Shrewsbury train to stop (assuming it didn't also go into a slide!) and perhaps bought more time for the incident train to come to a stand without impacting it?

Perhaps an interesting one to see what transpired when the report comes out!

We have a bit more info about my questions from a few months ago about the radio communications from the incident train which makes for interesting reading:

44 After 6 seconds of full service braking, the on-board signalling system intervened with an emergency brake demand. The intended stopping point, at block marker MH1078, was now approximately 500 metres away.

45 The driver of train 1J25 then used the train’s GSM-R radio to call the signaller to report that the train was sliding and was probably going to pass the block marker. Approximately twenty seconds after starting the call, the train passed block marker MH1078 at approximately 54 km/h (34 mph). The area beyond block marker MH1078 is a designated low adhesion area (see paragraph 60).

46 At approximately 19:24, the call with the signaller ended, as the train ran through the points and exited Talerddig Up Loop at approximately 33 km/h (20 mph). The train then entered the descending gradient and, although the brakes remained fully applied, its speed increased.

47 The signaller then called the driver of train 1S71. The driver of train 1S71 was told by the signaller that train 1J25 had passed block marker MH1078 and was occupying the loop exit points. The signaller instructed the driver of train 1S71 to stop at block marker MH1081 (figure 7), and the driver mentioned that the train was struggling with low adhesion. At the end of the call, train 1S71 was travelling at 26 km/h (16 mph). The driver then shut off traction power, possibly in response to seeing the approaching headlights of train 1J25. Over the next 9 seconds, the driver of train 1S71 applied service braking and then the emergency brake. The collision occurred 4 seconds after the emergency brake was applied.

So the driver of 1J25 did use their radio to alert the signaller to the slide which allowed the signaller to phone up the driver of 1S71 to alert them which no doubt helped reduce the severity of the collision as it will have seen 1S71 slow down earlier than it might otherwise have done so with no warning. But I do still wonder why the driver of 1J25 didn't make a Railway Emergency Call instead? Feels like that would be a situation in which to hit the red button rather than phone up the signaller? That being said perhaps it wouldn't have made that much difference in the grand scheme seeing as the driver of 1S71 was reporting low adhesion, going into full service breaking immediately might have triggered it's own slide!

 

[renegademaster]

Interesting bit starts at para 71. The sand hoses were blocked with a “30 mm plug of organic material”, which they have used DNA testing to identify!

Likely trying to identify what type of leaves where blocking them up

 

[bramling]

We have a bit more info about my questions from a few months ago about the radio communications from the incident train which makes for interesting reading:



So the driver of 1J25 did use their radio to alert the signaller to the slide which allowed the signaller to phone up the driver of 1S71 to alert them which no doubt helped reduce the severity of the collision as it will have seen 1S71 slow down earlier than it might otherwise have done so with no warning. But I do still wonder why the driver of 1J25 didn't make a Railway Emergency Call instead? Feels like that would be a situation in which to hit the red button rather than phone up the signaller? That being said perhaps it wouldn't have made that much difference in the grand scheme seeing as the driver of 1S71 was reporting low adhesion, going into full service breaking immediately might have triggered it's own slide!

I suspect it will be concluded as an error that the second train wasn’t told to stop immediately, especially as it seems there was sufficient warning given that something was amiss. Wouldn’t have prevented the collision but would likely have avoided the driver being seriously injured.

 

[Krokodil]

That being said perhaps it wouldn't have made that much difference in the grand scheme seeing as the driver of 1S71 was reporting low adhesion, going into full service breaking immediately might have triggered it's own slide!

It wouldn't have slid for long uphill.

 

[SteveP29]

Sorry, I haven't had time to read the whole thread and I caught the report on the BBC website just now.

The BBC report seems to suggest that the trains should have passed each other via a passing loop, I get the impression that the trains were travelling in opposite directions, so if that's the case and apologies for my naivety, why were the points set so that both trains ended up on the same line through the loop?

 

[Dave W]

Sorry, I haven't had time to read the whole thread and I caught the report on the BBC website just now.

The BBC report seems to suggest that the trains should have passed each other via a passing loop, I get the impression that the trains were travelling in opposite directions, so if that's the case and apologies for my naivety, why were the points set so that both trains ended up on the same line through the loop?

They weren't. The collision was after the passing loop - the first train should have stopped in the loop to allow the other to pass. First train ran through the points.

Quite a damning preliminary inspection relating to the maintenance of the sanding equipment, I thought - although they hedged bets on if all the defects found were existing before the incident.

 

[edwin_m]

Yes, it does seem odd that the REC button wasn't used at that point. Might have given the other driver time to escape.

It seems the driver was only instructed to stop at the next block marker, hundreds of metres beyond the point of collision, instead of stopping immediately and leaving the cab.

It also seems the leading car of the westbound unit had four faults with the continuous sander, each of which would probably have prevented it working, though some might have been due to accident damage or have developed afterwards. It also had a second system to discharge emergency sand bottles if operated by the driver, which was apparently in full working order but not used.

 

[Wilts Wanderer]

It seems the driver was only instructed to stop at the next block marker, hundreds of metres beyond the point of collision, instead of stopping immediately and leaving the cab.

It also seems the leading car of the westbound unit had four faults with the continuous sander, each of which would probably have prevented it working, though some might have been due to accident damage or have developed afterwards. It also had a second system to discharge emergency sand bottles if operated by the driver, which was apparently in full working order but not used.

This is pertinent and it will be interesting to see what future reports say on this subject. I am reminded of a much worse head-on collision that occurred in the underground RER station at Paris Gare de Lyon in 1988 where the driver of an incoming train found he had no working service brakes while descending the 1:30 tunnel ramp and eventually collided with a fully loaded outbound train standing at the platform. It was later confirmed that the inbound train had a fully operational secondary braking system which the driver didn’t attempt to use, which would have quickly stopped the train. I wonder whether there is a parallel here, that a secondary sanding system was available but no attempt made to use it. Is there a common psychological factor perhaps relating to high-stress crisis situations?

 

[DDB]

Would the driver be aware no sand was being delivered? If they thought they were slipping despite sand maybe they didn't think there was any point adding more?

 

[millemille]

Would the driver be aware no sand was being delivered?

In short, no.

 

[voyagerdude220]

We have a bit more info about my questions from a few months ago about the radio communications from the incident train which makes for interesting reading:



So the driver of 1J25 did use their radio to alert the signaller to the slide which allowed the signaller to phone up the driver of 1S71 to alert them which no doubt helped reduce the severity of the collision as it will have seen 1S71 slow down earlier than it might otherwise have done so with no warning. But I do still wonder why the driver of 1J25 didn't make a Railway Emergency Call instead? Feels like that would be a situation in which to hit the red button rather than phone up the signaller? That being said perhaps it wouldn't have made that much difference in the grand scheme seeing as the driver of 1S71 was reporting low adhesion, going into full service breaking immediately might have triggered it's own slide!

I was wondering that myself. Surely if you're driving a train towards a red signal and you realise you're not certain that you will stop before the red/before you leave the loop, it's an emergency situation which warrants using the emergency call button on the in cab GSMR?

Also by pressing the button yourself it saves time relaying the message to the Signaller and waiting for the Signaller to stop the trains themself. It goes without saying that every second counted in the build up to this incident occuring.

Maybe Drivers need a refresher on when they should be pressing the emergency call button? I know it's stating the obvious, but as the Driver of 1J25 didn't make an emergency call despite them knowing their train was effectively out of control, I understand they had alot to deal with, but it's extremely quick and easy to press the button.

 

[bramling]

I was wondering that myself. Surely if you're driving a train towards a red signal and you realise you're not certain that you will stop before the red/before you leave the loop, it's an emergency situation which warrants using the emergency call button on the in cab GSMR?

Also by pressing the button yourself it saves time relaying the message to the Signaller and waiting for the Signaller to stop the trains themself. It goes without saying that every second counted in the build up to this incident occuring.

Maybe Drivers need a refresher on when they should be pressing the emergency call button? I know it's stating the obvious, but as the Driver of 1J25 didn't make an emergency call despite them knowing their train was effectively out of control, I understand they had alot to deal with, but it's extremely quick and easy to press the button.

I can imagine the driver would have been quite agitated whilst all this was going on, so the lack of an emergency call one can probably understand - and to be fair he seems to have communicated the relevant information to the signalling centre, and likewise shouting a warning to people in the front carriage probably reduced the severity of injuries.

I’m rather surprised the signalling system doesn’t seem to have intervened with the other train, as certainly other types of continuous ATP systems (such as used on metro applications) can detect an encroachment into a limit of movement authority and act accordingly.

It is somewhat bothersome that this incident seems to have happened quite easily on a route with the most sophisticated signalling and protection system going, and not too far down the road from Abermule of all places. Having been there a few years ago I kind of said to myself “this wouldn’t happen now”, yet here we are.

 

[edwin_m]

Would the driver be aware no sand was being delivered? If they thought they were slipping despite sand maybe they didn't think there was any point adding more?

In short, no.

They would certainly be aware that they were sliding by the WSP activity and lack of deceleration - been there done that during trial of said one-shot sander on a 159 back in 1996ish (not involved with the 158 fitment but I believe itw was the same design). But clearly there was nothing to tell the driver that the continuous sander was actuallly delivering sand. There could have been an indicator of system health that might have picked up the electrical faults, but it would be difficult to install anything that could alert to the reported blockage in the sand pipes.

The one-shot sander was used several times in those early years when no other system was available and proved highly effective in an emergency situation. I produced a report showing the almost instant return to normal deceleration when it was activated during a slide. But I wonder if, once the continuous sanders were fitted, sliding became so rare that a driver might not have had to use the one-shot for years so it wasn't foremost in their mind as something they could use in this situation. There are parallels with some air accidents where operating one of the many switches and buttons in the cockpit could have averted disaster - had the pilots remembered to do so while under extreme pressure.

 

[MotCO]

Presumably, in an emergency situation, your brain reacts to what you know, and that is probably the actions you take most frequently. I imagine a driver rarely uses the GSMR emergency button, a driver rarely uses the one-shot sander etc, so these were not used. However a driver does use the radio to contact the signaller and deploys the sander quite often, so these were the immediate actions he took.

Maybe there should be some refresher training of actions to take in an emergency or other stressful situation, and somehow try to undertake them in a stressful environment. A bit like learning to control a skid on the London Transport skid pan at Chiswick. Maybe a simulator could be used.

 

[zwk500]

Maybe there should be some refresher training of actions to take in an emergency or other stressful situation, and somehow try to undertake them in a stressful environment. A bit like learning to control a skid on the London Transport skid pan at Chiswick. Maybe a simulator could be used.

Does driver simulation training not already include emergency scenarios?

 

[MotCO]

Does driver simulation training not already include emergency scenarios?

In a stressful situation, or a relaxed classroom environment?

 

[sharpener]

Interesting bit starts at para 71. The sand hoses were blocked with a “30 mm plug of organic material”, which they have used DNA testing to identify!

... but no build up of sand behind the plug. Which might indicate the sander did not operate to release sand into the hoses in the first place? How much air pressure is involved and is this not sufficient to blow out a plug of debris?

 

[crablab]

I'm a little surprised that an emergency brake application doesn't trigger the emergency sander, especially if the train is in a slide.

 

[saismee]

I'm a little surprised that an emergency brake application doesn't trigger the emergency sander, especially if the train is in a slide.

Sounds like a reasonable failsafe to me... sliding + emergency braking should trigger the one-shot sander automatically. It saves the driver from being distracted with yet another control, and likely could've prevented this.

 

[Annetts key]

I’m rather surprised the signalling system doesn’t seem to have intervened with the other train, as certainly other types of continuous ATP systems (such as used on metro applications) can detect an encroachment into a limit of movement authority and act accordingly.

Is that a function of the interlocking, rather than that of the ATP or ETCS/ERTMS or similar system? Certainly with conventional signalling systems it's the function of the interlocking, where specified.

However, to avoid problems caused by "right side" signalling equipment train detection failures, over run protection may not be specified as a requirement.

But regardless, without more details, all we can do is speculate.

 

[bramling]

Is that a function of the interlocking, rather than that of the ATP or ETCS/ERTMS or similar system? Certainly with conventional signalling systems it's the function of the interlocking, where specified.

However, to avoid problems caused by "right side" signalling equipment train detection failures, over run protection may not be specified as a requirement.

But regardless, without more details, all we can do is speculate.

It isn’t really speculation as clearly the system didn’t stop the second train. But one would think that a reasonably sophisticated system would identify that a train’s signalled route has been violated. Even a conventional signalling system will revert signals back to danger - as indeed happened at Ladbroke Grove, though sadly too late to make much difference but it did reduce the speed of the HST a bit.

 

[Snow1964]

I'm a little surprised that an emergency brake application doesn't trigger the emergency sander, especially if the train is in a slide.

So am I, in the same way an emergency (very heavy) brake application in a car also switches on hazard lights.

It would seem logical to have also triggered the back up one shot emergency sander without requiring someone to manually remember (when they have lots of other things happening) that it is available.

 

[millemille]

... but it would be difficult to install anything that could alert to the reported blockage in the sand pipes.

There is a sensor available, off-the-shelf, from Knorr-Bremse that is installed in the sand hose to detect sand flow, or lack of, and there are thousands in use in mainland Europe for over a decade.

== Doublepost prevention - post automatically merged: 1 May 2025 ==

I'm a little surprised that an emergency brake application doesn't trigger the emergency sander, especially if the train is in a slide.

The issue with one-shot sanders is that once they are operated the train is out of service until the bottles can be replaced. And there are very few classes of rolling stock (passenger, freight or OTM) left in the UK using one-shots so getting the bottles replenished after they've been operated is a challenge.

 

[Snow1964]

The issue with one-shot sanders is that once they are operated the train is out of service until the bottles can be replaced. And there are very few classes of rolling stock (passenger, freight or OTM) left in the UK using one-shots so getting the bottles replenished after they've been operated is a challenge.

Presumably rather less of a challenge than sorting out and trying to repair the consequences of it not being used when it leads to a crash.

I am guessing it would be in same maintenance category as when other emergency equipment (eg a fire extinguisher) has been discharged. Needs replacing (or swapping) with refilled one.

Ultimately there is no point in having emergency equipment if restoring it when (it rarely) gets used causes a problem. That's not way to think of it.

 

[sheeldz]

Would the driver be aware no sand was being delivered? If they thought they were slipping despite sand maybe they didn't think there was any point adding more?

This is a human factors issue in this case, because the brain works heavily on what is expected, and in this case the "correct" action was being undertaken - sanding was applied - but the wrong result was observed. What would have happened in all likelihood is the thought process of "I pressed the button, sanding is being delivered, and we're not stopping".

This is called an unrevealed failure, and is an interesting issue to consider - how would this error be revealed, how can you tell the driver? You'd think holistically that the driver would realise that the sanding was not being applied to the wheels, because it was not braking, but in the moment it is likely the driver would've assumed the sanding was being applied and simply wasn't effective, and as such more sand wouldn't do anything.

I deal with this kind of engineering interlocks in my day job and I am going to steal this as a great example of it. Would be interesting to brainstorm an engineered solution.

In engineering parlance, this error would straddle the line between a slip or lapse, and a Mistake. A mistake is where an action, done in good faith, doesn't provide the desired correct outcome, or a slip or lapse, where the correct procedure undertaken by highly trained personnel just doesn't quite do it right. I'd lean on slip/lapse.

For completeness, there is a third - a violation, where it's deliberately incorrect but not maliciously so.

 

[edwin_m]

I'm a little surprised that an emergency brake application doesn't trigger the emergency sander, especially if the train is in a slide.

Sounds like a reasonable failsafe to me... sliding + emergency braking should trigger the one-shot sander automatically. It saves the driver from being distracted with yet another control, and likely could've prevented this.

The one-shot sander was a retrofit and therefore aimed to be separate from all existing circuits except for taking power. While something else triggering the emergency brake is relatively straightforward, getting the emergency brake to initiate some other action involves tapping into safety-vital circuits and puts the whole thing into a different league of complexity.

Presumably rather less of a challenge than sorting out and trying to repair the consequences of it not being used when it leads to a crash.

I am guessing it would be in same maintenance category as when other emergency equipment (eg a fire extinguisher) has been discharged. Needs replacing (or swapping) with refilled one.

Dealing with the consequences of an accident is indeed much more serious, but usually also much less frequent than dealing with the consequences of unnecessary activation of a system when most of the time it would have made no difference to the outcome. So the balance of cost of the two depends on a whole range of factors.

I would hope that the depots supporting units fitted with the one-shot would have some stock of spare cylinders - they are fire extinguisher bottles filled with sand instead of whatever they usually contain. But this is a 30-year-old design and it may be that the particular type of bottle is now obsolescent.

== Doublepost prevention - post automatically merged: 1 May 2025 ==

This is called an unrevealed failure, and is an interesting issue to consider - how would this error be revealed, how can you tell the driver?

Where there is a potential unrevealed failure (which doesn't become apparent until the equipment needs to do something) then the solution is usually some kind of diagnostic or test that hopefully reveals it before time. In this case, according to the report, the sander is tested daily by pressing a button on the underframe and observing sand being discharged. This would have revealed the pipe blockage but not the electrical faults as it bypassed those circuits. As I think I pointed out some way back on this thread, the ultimate test would be to activate the whole end to end function of the system via an external input to the unit that controls it. Also, the report notes that the test couldn't be done on the day of the accident because the train was stabled in a platform where the button wasn't accessible.

 

[millemille]

Presumably rather less of a challenge than sorting out and trying to repair the consequences of it not being used when it leads to a crash.

I am guessing it would be in same maintenance category as when other emergency equipment (eg a fire extinguisher) has been discharged. Needs replacing (or swapping) with refilled one.

Ultimately there is no point in having emergency equipment if restoring it when (it rarely) gets used causes a problem. That's not way to think of it.

I speak from first hand experience, having been involved in several projects replacing with one-shot sanders with refillable automatic or manual sanders, that vehicles not being able to enter service, or being taken out of service, due to one-shot operation is much more of an issue to operators than the theoretical consequence of it not being used when needed.

But I would doubt very much that this would have been a factor in whatever reasoning, if any, the driver went through to explain why the one-shot was not operated.

As others have said, when subject to extreme stress the cognitive reasoning and training can go out of the window. I've looked at enough On Train Data Recorder (OTDR) downloads in post accident investigations to know that driver actions can often make no sense at all when viewed in the abstract while sitting at a desk in an office. But put yourself in the cab of a train, in the dark, not stopping and knowing you are on a single track line and somewhere in front of you is a train coming up the same line towards you.