Collision and derailment near Salisbury (Fisherton Tunnel) 31/10/21

[Western Sunset]

What we know from photos:
1. 1L53 was 3-car unit 159102, with 52803 at the rear.
2. 1F30 was x2 2-car units; 158762 leading 158763.

The photo showing 159102 inside the tunnel alongside 158763, poses the question did 1F30 split between the two units whilst entering the tunnel, or was 158762 subsequently detached after the accident and moved forward?

What is also currently unclear is the part played (if any) by 1F27 travelling in the opposite direction, which had cleared the junction but stopped afterwards, though well clear of the line from Waterloo.

 

[w0033944]

The animation is wrong, both services would have used the same line (the down line) through the tunnel, the GWR first, then the SWR. Post #17 shows the idea..,

The latter. There is no bi-directional working. What appears to have happened is the front of the SWR train has run into the side of the GWR train on Salisbury Tunnel Junction. Thankfully the SWR train was deflected to the right, and momentum has carried it onwards whilst the remaining energy was dissipated. It is this deflection to the right which has made it appear to enter the tunnel "wrong line", in reality it was simply moving in a derailed state so could have ended up anywhere.

As has been posted before, it is very fortunate no vehicles seem to have collided or got trapped up against the tunnel portal. Some fractional differences in timing or position could have led to a very different outcome.

== Doublepost prevention - post automatically merged: 1 Nov 2021 ==



There is some awful stuff doing the rounds in the media at the moment. Sky are still reporting that the first train was derailed, and there is at least one anonymous "expert" who claims to be a Network Rail whistleblower who is alleging a serious failure with a mythical safety system which is supposed to "kick in". Probably just as well there's other stuff in the news at the moment or else this sort of nonsense would be getting more air-time.

Thanks both - this was confusing me as well. Sad to hear that one of the drivers appears to have suffered more serious injuries.

 

[357]

was 158762 subsequently detached after the accident and moved forward?

From a personal viewpoint having dealt with train evacuations I find this to be unlikely - but I wasn't there, and stand to be corrected.

What I would think is if it was moved it would only be far enough to let people climb down from the gangway of the other unit, if it was at such an angle to make walking through impossible.

First choice would be walk everyone through and then out a normal door. Climbing down from the gangway doors isn't easy and you have a coupler to argue with.

 

[Efini92]

We’re supposed to press the emergency call button and make a railway emergency call, when you press that button all signals in that area are supposed to go to danger, if you can’t get a call through for whatever reason it’s clips and dets and a nice walk.

I’ve heard numerous REC and not once has any signal gone to danger. I think it’s a railway myth.

 

[swt_passenger]

What is also currently unclear is the part played (if any) by 1F27 travelling in the opposite direction, which had cleared the junction but stopped afterwards, though well clear of the line from Waterloo.

Having read right through I believe 1F27 played no part in the actual incident, other than being stopped by a subsequent emergency call. However if it had passed only a little while later it could have been crossing the junction…

 

[AndrewE]

Localised low adhesion can develop very quickly, irrespective of time of day, and leaf fall related adhesion worsens with the passing of a number of trains over freshly fallen leaves (it is the compacted leaves that produce the very sticky black substance on the railhead that is the cause of low adhesion) until it reaches a low and then gradually improves as more trains pass over it.

it is also worth considering that if the train/trains in question were braking where preceding trains hadn't then any low adhesion for the previous trains wouldn't have been an issue, or if the train/trains in question were braking harder - exceeding the available co-efficient of friction of the contaminated rail - than preceding trains had braked - not exceeding the available co-efficient of friction - then again it wouldn't have been an issue.

It can even be something as prosaic as preceding trains running on a different part of the railhead due to differences in bogie/primary suspension/wheel arrangement and then the train/trains in question running a different, uncleaned, part of the railhead (as happened back in 2019 at the Norwich road level crossing near miss).

When I was involved we understood that it wasn't anything like black treacle that was the problem, but something that started out far "thinner" and (in the wet) was tracked out a long way down the line to the point where it was almost invisible to the naked eye but could still destroy adhesion.
A

 

[neonison]

Is there anywhere a picture of 158762, presumably some way west (down) of the incident?

 

[ChiefPlanner]

....or if it had to be singled, the provision of trap points at the end of the double line sections would have prevented the accident.

Not to mention other factors like train / signal box communications. Awful scenario as to SPAD occurred and signal man at Oxted powerless to do anything about it.

 

[CHAPS2034]

Self loading freight here.

I notice that 1F27, which passed Tunnel Junction towards Dean/Romsey just before the accident, was delayed for 130 minutes after passing Laverstoke South Junction.

Was this because everything was brought to a stand in the area after the accident (even though it was moving away from the accident location) or was it from another cause?

Any informed explanation gratefully received.

 

[WestRiding]

when you press that button all signals in that area are supposed to go to danger,

who told you that? either way, its not true.

 

[Annetts key]

We’re supposed to press the emergency call button and make a railway emergency call, when you press that button all signals in that area are supposed to go to danger, if you can’t get a call through for whatever reason it’s clips and dets and a nice walk.

With the type of signalling system in use at this location, there is no connection between the GSM-R system and the signalling system, so it is not possible for the GSM-R to put the signals to red.

Where? The only footage of flashing lights is of the SWR train.

On the BBC One BBC News at One, BBC Points West Lunchtime News, BBC One BBC News at Six and BBC Points West Evening News. The first two should now be available on BBC iPlayer.

 

[21C101]

I would suspect the signal before the signal protecting Tunnel Junction (ie the signal situated Andover side of the Laverstock Loop junction) and how the train reacted to whatever aspect that signal was or wasn't showing when said train passed it will be a line of enquiry.

 

[DNCharingX]

Media doing what media does.

Next it'll be that the driver was having a cuppa in the tunnel! (No disrespect intended to the driver.)

 

[neonison]

The point of impact of 1L53 was in the rear quarter of the penultimate coach of 1F30 which stopped about 30-35m west of the junction which suggests that the GWR train was moving when the collision occured and stopped only because of that collision.

 

[FrodshamJnct]

I would suspect the signal before the signal protecting Tunnel Junction (ie the signal situated Andover side of the Laverstock Loop junction) and how the train reacted to whatever aspect that signal was or wasn't showing when said train passed it will be a line of enquiry.

Weren’t there reports of the ‘signalling being knocked out by the derailment’? Or was that just media mis-quotes and speculation?

 

[swt_passenger]

Was this because everything was brought to a stand in the area after the accident (even though it was moving away from the accident location) or was it from another cause?

That was the explanation first posted last night and there’s been nothing to contradict it since.

 

[CHAPS2034]

That was the explanation first posted last night and there’s been nothing to contradict it since.

Thanks - must have missed that

In this instance, how large an area would be involved with a stop message?

I assume nothing is then allowed to move until cleared by the signaller once the circumstances have become clearer?

 

[ScotRail158725]

Is there anywhere a picture of 158762, presumably some way west (down) of the incident?

Its in the tunnel a few yards beyond the west end of 763

 

[Moonshot]

There appears to be some confusion on this thread as to what a GSMR emergency call actually does. When the big red emergency button is pressed, an immediate call is sent out to all other trains in the area to stop. The activator of the call is immediately put through to the controlling signaller....overriding any call the signaller may be on at the time. Control are also able to listen to the emergency call. What the big red button doesnt do is set signals to red automatically.....which would be impossible to do anyway if semaphore signals are involved.

 

[Llanigraham]

This thread is indeed illustrating that the sequence of events and factors that culminated in this incident could be somewhat complex and not what may first seem apparent.

I can remember seeing some of the paperwork produced by the RAIB for the Moreton-on-Lugg incident and that formed a VERY long timeline that was almost down to the second, and contained innumerable entries.

 

[21C101]

Weren’t there reports of the ‘signalling being knocked out by the derailment’? Or was that just media mis-quotes and speculation?

There were, from official sources, and there is no reason to doubt them.

However further reports today suggest that the derailment may have occured due to the trains colliding which, if true, makes the signalling being knocked out by the derailment irrelevant.

 

[Llanigraham]

I’ve heard numerous REC and not once has any signal gone to danger. I think it’s a railway myth.

Going back in time, but I once got 6 bells from the next Box down, and boy did those relevant levers get thrown back into the frame!

 

[Horizon22]

There appears to be some confusion on this thread as to what a GSMR emergency call actually does. When the big red emergency button is pressed, an immediate call is sent out to all other trains in the area to stop. The activator of the call is immediately put through to the controlling signaller....overriding any call the signaller may be on at the time. Control are also able to listen to the emergency call. What the big red button doesnt do is set signals to red automatically.....which would be impossible to do anyway if semaphore signals are involved.

Indeed. There could theoretically be a scenario that leads to an incident whereby an emergency call was made which was too late for an issue to be prevented. In this circumstance though that isn't necessarily relevant as the signal to the rear should have been protecting the GWR train anyway. It should of course fail-safe and if black (blackout), the driver would treat it as red.

 

[Dan G]

Its in the tunnel a few yards beyond the west end of 763

How do you know this? Have you seen it?

I don't believe for a moment that couplings are so weak that the drag of a derailed car would open it up when the units are travelling at no more than 20 mph.

Equally, I'm surprised the lead unit was deliberately moved after the accident, though that seems the more likely explanation.

I wouldn't mind a little more evidence that 158762 is actually there beyond "take my word for it".

 

[357]

I don't believe for a moment that couplings are so weak that the drag of a derailed car would open it up when the units are travelling at no more than 20 mph.

They come apart surprisingly easily when twisted beyond a certain amount of degrees.

 

[scotraildriver]

How do you know this? Have you seen it?

I don't believe for a moment that couplings are so weak that the drag of a derailed car would open it up when the units are travelling at no more than 20 mph.

Equally, I'm surprised the lead unit was deliberately moved after the accident, though that seems the more likely explanation.

I wouldn't mind a little more evidence that 158762 is actually there beyond "take my word for it".

If you search the BBC reports from last night there is one showing it sticking out the other end of the tunnel with its hazard lights on.

 

[kingqueen]

If you were evacuating, you wouldn’t take the time to carefully make up the gangway, secure all the doors away etc, they’d just be flung open and get out.

I often wonder what would happen to me, or other wheelchair users, in such circumstances. Particularly where (as in this case) there's no chance of parallel transfer into another train.

 

[Llama]

Thanks - must have missed that

In this instance, how large an area would be involved with a stop message?

I assume nothing is then allowed to move until cleared by the signaller once the circumstances have become clearer?

The GSMR 'cell' which encompasses about 5km IIRC, and any adjoining GSMR cell.

From my GSMR training which was with an ex-VWC area manager, one of the places with pretty much the best value for a Railway Emergency Call (REC) would be Crewe station as it would then affect six adjoining GSMR cells (the lines to Chester, WCML North, Manchester, the Potteries, WCML south and the Marches).

On receipt of a GSMR REC a driver must stop their train immediately whilst listening to the call, press the 'ST' acknowledgement button on the GSMR when stopped and then when the REC is terminated any drivers who know they are definitely not affected by the incident may proceed. They are supposed to proceed at caution and expect the next signal to be at its most restrictive in case the controlling signaller has had to use an 'all signals to danger' device - the signaller will have enough on his plate without half a dozen drivers then ringing in to say that they have had an irregular aspect sequence. Equally, the signaller would also be overloaded if every single train in the area had to contact the signaller for permission to proceed, even the ones clearly not affected.

Drivers of any trains who are or may possibly be affected by the incident must not move until the controlling signaller has given them each specific permission to proceed along with any necessary instructions.

 

[dingdinger]

I think last night for quite a while after the incident RTT (as an example in post #70) was still showing the SWR train as a 9 car. This was subsequently changed, I think someone said there’d been alterations during the day that hadn’t filtered through to open data sources...

It was definitely a 3 car as I was on it until Basingstoke. Originally the service was shown as cancelled but as there were crew to work it and a unit available it ran.

 

[357]

I often wonder what would happen to me, or other wheelchair users, in such circumstances. Particularly where (as in this case) there's no chance of parallel transfer into another train.

The emergency services have various ways to get wheelchair uses out of trains, with various senses of urgency.

However, the greater good is high on the agenda and providing it is safe you could be on the train for some time.

But rest assured, as soon as it is safe and possible you will be helped from the train.

In the very worst case, it's likely to be other passengers helping while the driver is laying emergency protection (a hike 1.25 miles up the track to lay dets and flag down the next train for assistance)