Collision between two trains at Llanbrynmair, in Powys, Wales - 21 October 2024

[whole diary]

[grahame]

Not seen anything saying if the coach's were removed by rail or road.

All of the major sources use the same words
https://www.bbc.co.uk/news/articles/cr7n5myvle0o
https://news.tfw.wales/news/cambrian-line-to-reopen-following-rail-accident
https://www.walesonline.co.uk/news/wales-news/wales-train-crash-update-line-30233213

Transport for Wales (TfW) is in the process of removing the two trains

As an educated guess, I would think that's because they or some of their readers might consider it insensitive or ghoulish to talk about how the remains are removed at a place where some were seriously injured and which perhaps triggered the passing of one person too.

[Mark A]

The line has reopened for the start of service this morning.

Mark

[grahame]

There is a good summary on Wikipedia including a track diagram and further detail not in the general press. It includes "The two trains were separated on 24 October and taken to Machynlleth by rail the next day" which answers a question asked here previously.

[infoman]

Thank you Grahame for the update.

[broadgage]

It seems to me that there is a small but real risk of a similar accident occurring in the future, unless mitigated by by very restrictive speed limits.
Sounds as though trap points would help. It appears that one train was driven correctly into the passing loop, where it should have been held at the signal protecting the exit from the loop onto the main line.
There has been no suggestion of any signalling failure, one may therefore presume that this signal was passed at danger due to poor adhesion.
Trap points, interlocked with the signal protecting the exit from the loop would prevent any repeat. Exit signal showing a proceed aspect=points in the normal position allowing train onto main line.
Exit signal at danger=trap points automatically set to divert the train into a sand drag.

[Witham Bobby]

It seems to me that there is a small but real risk of a similar accident occurring in the future, unless mitigated by by very restrictive speed limits.
Sounds as though trap points would help. It appears that one train was driven correctly into the passing loop, where it should have been held at the signal protecting the exit from the loop onto the main line.
There has been no suggestion of any signalling failure, one may therefore presume that this signal was passed at danger due to poor adhesion.
Trap points, interlocked with the signal protecting the exit from the loop would prevent any repeat. Exit signal showing a proceed aspect=points in the normal position allowing train onto main line.
Exit signal at danger=trap points automatically set to divert the train into a sand drag.

The Cambrian is operated on a unique ETRMS signalling system.  There are no signallers or signalboxes at the remote passing places. Train occupation of the route and speeds are controlled from Machynlleth.   The pointwork at each end of the passing loops is locally controlled.  There are no fixed running signals.   I believe (but stand ready to be corrected) that trains leaving the loop lines run through these trailing points as a matter of routine operation.  I don't see that trap points can be provided.  They are mostly absent now on the wider network at similar types of location.  Locally to me, Evesham West Junction and Ascott-under-Wychwood spring to mind as having no traps, although the junction points at these locations are operated from signalboxes.  The protection is provided by having an overlap ahead of the signal which gives access to the single line, in theory giving a safety margin before the fouling point of the junction is reached.

[grahame]

It seems to me that there is a small but real risk of a similar accident occurring in the future, unless mitigated by by very restrictive speed limits.
Sounds as though trap points would help. It appears that one train was driven correctly into the passing loop, where it should have been held at the signal protecting the exit from the loop onto the main line.
There has been no suggestion of any signalling failure, one may therefore presume that this signal was passed at danger due to poor adhesion.
Trap points, interlocked with the signal protecting the exit from the loop would prevent any repeat. Exit signal showing a proceed aspect=points in the normal position allowing train onto main line.
Exit signal at danger=trap points automatically set to divert the train into a sand drag.

I would have had some similar questions, but await reports rather than speculating too much.  I don't think that there ARE (real) signals there any more - the whole thing being under ERTMS^▸ (European Rail Traffic Management System.).  And, yes, it would appear that there are not real trap points either, or if there are they didn't trap when that might have been expected. Interesting from the reports that the train from Aberystwyth seems to have been stationary at the time of collision.

Edit to add - Witham Bobby posted while I was writing the above .. which I have chosen to leave even though it broadly duplicates what he says

[Witham Bobby]

I would have had some similar questions, but await reports rather than speculating too much.  I don't think that there ARE (real) signals there any more - the whole thing being under ERTMS^▸ (European Rail Traffic Management System.).  And, yes, it would appear that there are not real trap points either, or if there are they didn't trap when that might have been expected. Interesting from the reports that the train from Aberystwyth seems to have been stationary at the time of collision.

With the ERTMS radio link, the driver of the train from Shrewsbury will have been able to warn the Machynlleth signaller, who could have instructed the driver of the train heading towards the Tallerdig loop to stop.  That is speculation, but reasonable

Short sketchy outline of the system here: https://ertms.freightliner.co.uk/news/a-trip-on-the-cambrian/

[broadgage]

I would have had some similar questions, but await reports rather than speculating too much.  I don't think that there ARE (real) signals there any more - the whole thing being under ERTMS^▸ (European Rail Traffic Management System.).  And, yes, it would appear that there are not real trap points either, or if there are they didn't trap when that might have been expected. Interesting from the reports that the train from Aberystwyth seems to have been stationary at the time of collision.

Edit to add - Witham Bobby posted while I was writing the above .. which I have chosen to leave even though it broadly duplicates what he says

Surely the remote control, rather than a local signal box, does not prevent the use of trap points.

[ChrisB]

And the eastbound train then collides with the likely derailed westbound service? Which would cause most damage/injury?

[AMLAG]

I am reassured when on the Cornish main line that where the double line goes to single
line at Largin East with a 1 in 57 falling gradient towards Bodmin Road, there has historically, for
obvious good reasons, been a trap point and long sand drag.

[grahame]

I am reassured when on the Cornish main line that where the double line goes to single
line at Largin East with a 1 in 57 falling gradient towards Bodmin Road, there has historically, for
obvious good reasons, been a trap point and long sand drag.

Trap points (and sand drags) were very much more "necessary" in the days before all trains had brakes throughout but they remain of significant use in many situations today.  I seem to remember a turbo on its way into Paddington from the sidings not all that long ago that overshot a signal and ran off the line.   It did (and these things do) still do damage, but that damage was lessened.

Fewer trains run away or overrun these days, and they probably don't get as far when they do.  So I'm not shocked at the reduction in the number of trap points.  However, Salisbury has shown up (and Talerddig probably has) that overruns are not eliminated and that on the occasions that trains run through there can be significant damage done if they overrun so far that they come into contact with another train.  It will be very interesting to see what the RAIB^▸ (Rail Accident Investigation Branch) have to say, if anything, about trap points, drags, safety zones, etc.

[ChrisB]

Salisbury wouldn't have benefitted from trap points as the train joining the through route was signalled in front of the 'straight on' train - that did the sliding into the train from the left. So the only place a trap would take that 'straight on' train would be into the tunnel arch, if indeed you could even get a trap point there

[grahame]

From the BBC^» (British Broadcasting Corporation - home page) https://www.bbc.co.uk/news/articles/cde7dwkpz5xo

Both trains involved in a fatal crash in mid Wales were moving at the time of the collision, investigators have revealed.

David Tudor Evans, 66, died after two trains crashed near the village of Llanbrynmair in Powys on 21 October.

It was initially said that a service from Shrewsbury to Aberystwyth hit a stationary train which was travelling from Machynlleth to Shrewsbury.

Now, the Rail Accidents and Investigations Branch (RAIB^▸ (Rail Accident Investigation Branch)) has said both trains were moving at the time of impact and there was conflicting evidence about how fast they were going.

On Tuesday, the RAIB issued an update on the investigation, adding that an automated system that helps train wheels grip tracks failed.

It said initial analysis indicated the 1J25 service from Shrewsbury to Aberystwyth was travelling between 15mph (24 km/h) and 24mph (39 km/h).

The 1S71service from Machynlleth to Shrewsbury was travelling at about 6mph (10 km/h) in the opposite direction.

The RAIB said it inspected the automatic sanding system, which automatically sprays sand via hoses when a wheel slide is detected during braking, in a bid to generate more friction.

It said the inspection of the system fitted to train 1J25 showed these hoses on the leading vehicle of the train "were blocked and apparently unable to discharge sand".