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

[Horizon22]

Wouldn't have appeared at my place, Drivers should get an email , but that's hit and miss too.

Communication to drivers is, frankly antiquated and needs root and branch change imho.


Which they aren't at weekends.

Late notices are still sent in bulk to remote locations and printed off next to the signing on points. RHTT can run over a LOT of separate routes indepdent of where TOC services run to-from. It would be incredibly difficult, if not impossible to advise every driver on every route affected about the potential for issues. The best place for it would be on a late notice (so it is a blanket piece of information) and for drivers to apply leaf fall driving policies, including notifying the signaller of reportable conditions. I wonder if the previous driver through SY31 reported any problems; how long had passed between 1L53 and the previous train?

 

[king_walnut]

The train half an hour earlier went over the route with no issues. I know the weather conditions were appalling that night but I'm fascinated as to what can make the railhead go from not even reportable to lethal in just half an hour.

 

[Moonshot]

The train half an hour earlier went over the route with no issues. I know the weather conditions were appalling that night but I'm fascinated as to what can make the railhead go from not even reportable to lethal in just half an hour.

Maybe it wasn't faced with red aspects and was coasting? Light drizzle is a nightmare ......as soon as it starts , conditions change very quickly

 

[Horizon22]

The train half an hour earlier went over the route with no issues. I know the weather conditions were appalling that night but I'm fascinated as to what can make the railhead go from not even reportable to lethal in just half an hour.

Interesting. I've seen plenty of reports whereby one driver has reportable rail conditions, then 15 minutes later another driver goes over the same section and reports that it is nothing that wouldn't be expected for the time of the year - or indeed there's any issue at all. Obviously drivers are human and have different interpretations of the condition of the track and in different circumstances (i.e the second driver in my example knows he might be approaching an issue and prepares accordingly or is on different aspects). Perhaps the reverse happened here, but that is just speculation. I'm sure the RAIB will look very closely at the prevailing weather conditions in the hours preceding the incident.

As an aside it is interesting that the two most recent serious incidents on the railway regarding passenger services (this and Carmont; Enfield Town we shall see if the newspaper reports are true) have both seemingly been directly due to extreme or strange changes in weather / weather conditions as opposed to any specific signal / track / crew failings. It's already known, but it's going to become a serious issue over the next 10-15 years if the climate continues to change as expected with wetter seasons and higher chances of downpours & flash flooding.

 

[30907]

Meanwhile the RAIB can look at why, in the prevailing leaf fall conditions, NR apparently cancelled the Treatment Train on both Saturday and Sunday.

Do we actually know this?

RTT shows two D800 workings on the Saturday and one on the Sunday afternoon - remarkably just before the accident - which look suspiciously like RHTT schedules (characteristics: random routes and reversals) over the down line and then the Laverstock Loop.
There was also a cancellation of a service which would have run to Yeovil and to and from Basingstoke and I wonder if that has caused someone to jump to conclusions.

BTW did the yellow warnings issued by the Met Office cover the Salisbury area? We know there was some decidedly unexpected weather on Sunday affecting Salisbury, the southern WCML and the southern ECML at least.

 

[king_walnut]

Maybe it wasn't faced with red aspects and was coasting? Light drizzle is a nightmare ......as soon as it starts , conditions change very quickly

Oh yeah definitely, I've had many squeaky bum times braking for stations just as drizzle has started.

Just, based on the cab ride video a few pages back and the new information today, that train slid for an absolutely ridiculous distance in absolutely pouring rain which generally doesn't have an impact on adhesion. I hope the report really details how the rails became that awful.

 

[Taunton]

The train half an hour earlier went over the route with no issues. I know the weather conditions were appalling that night but I'm fascinated as to what can make the railhead go from not even reportable to lethal in just half an hour.

But did the preceding train(s) have to pull up at the junction signal?

Going a long way back, I recall that the standard clearing point beyond a stop signal was a quarter mile. That's 440 yards, or 400 metres. Yet today's preliminary RAIB report says that the protecting signal for the junction is only 220 metres from the junction, at a key critical point. When was this relaxed, and what was the risk assessment of so doing. I wonder if before the Laverstock curve was built whether the signal was further back, but got pulled forward just to "fit in" with the layout.

Maybe it's accepted now for a 40mph limit. Yet the line is 90mph until shortly beforehand, and it seems to have been a single braking movement from high speed to the collision point. It's a high speed downhill ride westbound from the Wiltshire Downs into Salisbury, timetabled as such of course, and seems a bit minimalist to come from this to a short clearance before a key conflict point.

 

[quattromatt]

So can someone confirm if either SW or GW issued a notice to their drivers that weekend that the RHTT was not running over that route?

It should have been in the late notices.

 

[Link24]

Just to clear up a little misinformation going around, Saturday's RHTT was not cancelled. TOPS shows 3S82 running as planned 30/10/21 AM, treating the rails from both Dean direction and Andover direction throughout its shift. The Sunday RHTT however, does look like it was postponed to run in the evening instead of early afternoon.

 

[Grumpy Git]

The train half an hour earlier went over the route with no issues. I know the weather conditions were appalling that night but I'm fascinated as to what can make the railhead go from not even reportable to lethal in just half an hour.

Imagine you are driving a car over a stretch of road that has a surface temperature 1 degree lower than it was 30 minutes earlier. Normally no big deal, but if the temperature was previously 1C, then it suddenly becomes a very big problem for you.

Obviouisly freezing was not the case here, but the changes in conditions that can occur in a heartbeat should never be underestimated.

 

[rd749249]

Only with some signals and then only when a position light/calling on aspect is being shown (main aspect red, with two smaller diagonal white lights). When a main proceed aspect (green, double yellow or single yellow), is being shown, all the TPWS ‘loops’ are deactivated.

This is not quite correct. When signals are close, a TPWS OSS will arm on a signal yellow if speed is judged too high for the red that follows quite soon after. I have at least one example of this on the route that I drive.

Absolute nightmare for the driver. Very good to see SWR come out and say what they have said. I hope he has a speedy recovery and that SWR look after him.

 

[Watershed]

The standard overlap is 200 yards, and has been for several decades at least. Reduced overlaps are permissible subject to a risk assessment and indeed that's the case for SY31, which has an overlap of slightly less than 200 yards.

The distance to the point of conflict in this case was, however, just over 200 yards - so having a standard overlap (or even a 400 yard one) would not have prevented the collision.

 

[bengley]

This is not quite correct. When signals are close, a TPWS OSS will arm on a signal yellow if speed is judged too high for the red that follows quite soon after. I have at least one example of this on the route that I drive.

Absolute nightmare for the driver. Very good to see SWR come out and say what they have said. I hope he has a speedy recovery and that SWR look after him.

That's not armed for the yellow though, it'll just be a TPWS+ installation for the red.

 

[MarkyT]

This is not quite correct. When signals are close, a TPWS OSS will arm on a signal yellow if speed is judged too high for the red that follows quite soon after. I have at least one example of this on the route that I drive.

Absolute nightmare for the driver. Very good to see SWR come out and say what they have said. I hope he has a speedy recovery and that SWR look after him.

That's a special case though and I presumably in quite a high speed 4 aspect area. I think it's a feature of TPWS+ that additional OSS installations can actually be before you reach the previous signal. There are also fixed speed restriction OSS installations in some places that are permanently activated.

 

[Llama]

Going a long way back, I recall that the standard clearing point beyond a stop signal was a quarter mile. That's 440 yards, or 400 metres. Yet today's preliminary RAIB report says that the protecting signal for the junction is only 220 metres from the junction, at a key critical point. When was this relaxed, and what was the risk assessment of so doing. I wonder if before the Laverstock curve was built whether the signal was further back, but got pulled forward just to "fit in" with the layout.

Maybe it's accepted now for a 40mph limit. Yet the line is 90mph until shortly beforehand, and it seems to have been a single braking movement from high speed to the collision point. It's a high speed downhill ride westbound from the Wiltshire Downs into Salisbury, timetabled as such of course, and seems a bit minimalist to come from this to a short clearance before a key conflict point.

That's been the case for decades. Clearing point is an absolute block term and yes the normal standard is still 440yds. The normal overlap distance has been 220 yards on TCB installations for decades now, obviously there are variances to this.

This is not quite correct. When signals are close, a TPWS OSS will arm on a signal yellow if speed is judged too high for the red that follows quite soon after. I have at least one example of this on the route that I drive.

Absolute nightmare for the driver. Very good to see SWR come out and say what they have said. I hope he has a speedy recovery and that SWR look after him.

Sounds like TPWS+. If you could get hold of the signalbox special instructions for the location and check the TPWS appendix, or maybe a yellow peril, that'd clear it up which signal the OSS is related to but I'd say it's almost certainly going to be the red, not the yellow.

 

[MarkyT]

The standard overlap is 200 yards, and has been for several decades at least. Reduced overlaps are permissible subject to a risk assessment and indeed that's the case for SY31, which has an overlap of slightly less than 200 yards.

The distance to the point of conflict in this case was, however, just over 200 yards - so having a standard overlap (or even a 400 yard one) would not have prevented the collision.

Measures approx 160m from signal to clearance point at the diamond crossing, which is the first conflict point at the junction in question. The clearance point at the converging turnout where the collision took place is about a further 30m further on, so about 190m from SY31. 160m is usually considered an acceptable overlap at an approach speed of 50mph (subject to risk assessment). the 200yd spec became 200m rather than 183m some years ago and I think may have been extended to 220m now for new schemes by default.

 

[cogload]

But did the preceding train(s) have to pull up at the junction signal?

Going a long way back, I recall that the standard clearing point beyond a stop signal was a quarter mile. That's 440 yards, or 400 metres. Yet today's preliminary RAIB report says that the protecting signal for the junction is only 220 metres from the junction, at a key critical point. When was this relaxed, and what was the risk assessment of so doing. I wonder if before the Laverstock curve was built whether the signal was further back, but got pulled forward just to "fit in" with the layout.

Maybe it's accepted now for a 40mph limit. Yet the line is 90mph until shortly beforehand, and it seems to have been a single braking movement from high speed to the collision point. It's a high speed downhill ride westbound from the Wiltshire Downs into Salisbury, timetabled as such of course, and seems a bit minimalist to come from this to a short clearance before a key conflict point.

The layout was resignalled in 1981, the same time as the curve was put in. The PSR kicks in at Laverstock North, not on the approach to SY31.

"A single braking movement"? What does that mean?

 

[alxndr]

This is not quite correct. When signals are close, a TPWS OSS will arm on a signal yellow if speed is judged too high for the red that follows quite soon after. I have at least one example of this on the route that I drive.

That's the first I've ever heard of that, although not saying some oddities don't exist as they definitely do!

 

[Railsigns]

The normal overlap distance has been 220 yards on TCB installations for decades now, obviously there are variances to this.

The standard TCB overlap length is 180 metres (197 yards) and has been for more than twenty years.

 

[30907]

Just to clear up a little misinformation going around, Saturday's RHTT was not cancelled. TOPS shows 3S82 running as planned 30/10/21 AM, treating the rails from both Dean direction and Andover direction throughout its shift. The Sunday RHTT however, does look like it was postponed to run in the evening instead of early afternoon.

Thanks for the inside confirmation of my post - am I right that the similar late Sunday evening, which actually ran early not long before the SW train, was also a RHTT?

EDIT: I understand from another forum that the working I referred to
https://www.realtimetrains.co.uk/service/gb-nr:B31707/2021-10-31/detailed
was an engineering train. Apologies if anyone was misled.

 

[Crossover]

One detail which had been unsettling me, was that from the outset it appeared the driver had remained in his cab throughout, and this now appears to have been confirmed. Serious respect for that.

Seconded!

 

[bramling]

The train half an hour earlier went over the route with no issues. I know the weather conditions were appalling that night but I'm fascinated as to what can make the railhead go from not even reportable to lethal in just half an hour.

It isn’t really a direct comparison. Did the previous train call at Grateley, in which case it could have been travelling slower? Was it a 3-car or 6-car? Was the signal “on” for that train? Had it just started raining? Had the wind picked up in the intervening half hour and blown a load of leaves down? Was sanding gear operational?

Lots of variables, which no doubt the RAIB will look at.

 

[reddragon]

Is there a fund to raise money for this driver perhaps set up by his colleagues?

 

[Peter Sarf]

Yes I’d agree with that too. I know that some of my colleagues on here have put an awful lot of work into sifting out things that haven’t added anything of value to this discussion over the last few days and what has been left has hopefully been relevant and informative.
This isn’t a section that I personally monitor, but I know for a fact that it’s taken a fair amount of work by the other staff members to keep things on an even keel and I’m sure that we all appreciate what they’ve done here even though it’s subtle enough that we might not realise what goes on behind the scenes always.

Lest we forget the moderators.

The missing posts can now be found in various spin-off threads, such as:

• Signal spacing and where 4 aspects are needed
• Can fireworks cause poor railhead conditions?
• Network Rail's vegetation problem
• West of England line current service provision
• Do heavy locos cope with poor railhead conditions better than lightweight units?
• Was the Cowden crash due to cost cutting?
• Unnerving experience on 1803 from Paddington
• Do class 158's/159's have forward facing cameras?

If we missed posts, feel free to use the report button

Thankyou to all the moderators for your efforts to keep this very rapidly growing thread tidy.

One of the images from within the tunnel shows a lump of the rear corner of the 158 has been torn off and is against where the 159 driver's side cab bulkhead should be - the small window and the signage for the cycle area can be clearly seen. That's clearly quite some damage to the 158.

Photos below; not mine.

(The first image shows a general view inside the tunnel, with the fronts of both trains alongside one another. The second picture is a zoomed version of the first, showing a section of class 158 bodyshell resting inside the damaged cab of the class 159).

Your powers of recognition impress me. Gives an indication of the state of the inner end (and side) of the leading coach of the second 158.

This time of year can be a nightmare to drive. Ordinarily I’m straight on the brakes for DY with at least a 50% reduction but in this weather, it’s more. And if it does go into a slide, well that’s my worst nightmare. Fortunately it’s only happened to me at a very low speed and it was contained, but it does shake you up. Being the first unit behind a RHTT isn’t a pleasant experience either.

I sincerely hope that the driver is ok in body and mind after this.

The slide would be an ordeal. The drivers worst fears then coming true is beyond my comprehension. Let us not forget the nature of the job a train driver does - it is not like driving your own car. It requires a lot of learning, responsibility and I assume a cautious mentality.

One detail which had been unsettling me, was that from the outset it appeared the driver had remained in his cab throughout, and this now appears to have been confirmed. Serious respect for that.

Yes. The driver must have had a reasonable amount of time to abandon his station - but did not. As I said before I hope driver has a good recovery and no mental baggage from the ordeal.

This may seem a basic question and apologies if it’s already been asked but surely signal SY31 being at red meant there was at least a single yellow before this signal, if not a double yellow proceeding that depending on signalling around that area?

My understanding is the driver slowed at the yellow and realised the train was slipping at that stage. The train was slipping, despite the drivers efforts, before the red signal came into view.

The train half an hour earlier went over the route with no issues. I know the weather conditions were appalling that night but I'm fascinated as to what can make the railhead go from not even reportable to lethal in just half an hour.

Conditions can easily change in 30 minutes and the weather that night indicates it was likely to have changed. Also if the preceding train had not needed to brake then nothing would be discovered anyway.

 

[HowardGWR]

Reading this second RAIB report, may we assume that NR already have an order in for extra chainsaws?

 

[Thumper1127]

Just to clear up a little misinformation going around, Saturday's RHTT was not cancelled. TOPS shows 3S82 running as planned 30/10/21 AM, treating the rails from both Dean direction and Andover direction throughout its shift. The Sunday RHTT however, does look like it was postponed to run in the evening instead of early afternoon.

You are right and I apologise to all and fess up that I posted the original saying that one wasn’t activated (the Saturday one) and one was cancelled. I looked at the Saturday and saw the 0732 Eastleigh to Eastleigh (which didn’t run) that has all the hallmarks of being a RHTT somehow missed the one you are referring to. How I missed it I don’t know, it wasn’t a casual look, I went over it a number of times. The one you refer to “cleaned” the area at around 1105. The Sunday one, shown in RTT as being cancelled by the operator, would have cleaned around 1700 so not too long before the Honiton. So that’s circa 31 hours the down line wasn’t treated, still quite some time given the season and weather. Of course, it will be one aspect RAIB will no doubt investigate, I am not saying or inferring it is the cause.

I apologise to all on the forum for my inaccurate original post, I am usually very diligent in making sure I only post facts (unless it is a general discussion of opinions etc).

 

[Llama]

The standard TCB overlap length is 180 metres (197 yards) and has been for more than twenty years.

Thanks for that, as ever.

I was rounding up 200m to ~200yds, should've been 200yds to ~200m.

 

[rebmcr]

For any journalist snooping on these forums I suggest they read the Rail Delivery Group's 'Adhesion Working Group' report: Managing Low Adhesion Sixth Edition, January 2018

All 351 pages.

Superb thinking on your part. The BBC have reproduced this link verbatim!

 

[Crossover]

Superb thinking on your part. The BBC have reproduced this link verbatim!

Salisbury train crash: Driver acted impeccably, says operator

South Western Railway says a driver involved in the Salisbury crash has an "excellent track record".

www.bbc.co.uk

<snip>
The crash has caused major disruption, with lines through the city expected to remain closed until at least the end of Monday.
Of the 92 passengers on board the two trains, 14 required hospital treatment, with most suffering minor injuries.
The Rail Delivery Group said low adhesion can be caused by "moisture on the rail mixing with the film produced by 'leaves on the line' or other contaminants, such as rust or grease."
They add the problem can be worse in autumn and that "it can also cause safety risks, such as signals passed at danger and station overruns".

 

[dctraindriver]

Interesting. I've seen plenty of reports whereby one driver has reportable rail conditions, then 15 minutes later another driver goes over the same section and reports that it is nothing that wouldn't be expected for the time of the year - or indeed there's any issue at all. Obviously drivers are human and have different interpretations of the condition of the track and in different circumstances (i.e the second driver in my example knows he might be approaching an issue and prepares accordingly or is on different aspects). Perhaps the reverse happened here, but that is just speculation. I'm sure the RAIB will look very closely at the prevailing weather conditions in the hours preceding the incident.

As an aside it is interesting that the two most recent serious incidents on the railway regarding passenger services (this and Carmont; Enfield Town we shall see if the newspaper reports are true) have both seemingly been directly due to extreme or strange changes in weather / weather conditions as opposed to any specific signal / track / crew failings. It's already known, but it's going to become a serious issue over the next 10-15 years if the climate continues to change as expected with wetter seasons and higher chances of downpours & flash flooding.

Your first paragraph makes the point that maybe the interpretation of each driver to reportable rail conditions is different; just remember that many times the next train through that same section will be running on a railhead that has been sanded by the train experiencing reportable conditions. That sanding between the wheels and railhead will alter the retardation characteristics for the next train.

I’ve been asked to do controlled brake stops due to the previous train experiencing low adhesion, I’ve braked as specified in the rule book and each time I’ve stoped ok which I’ve reported back. I didn’t brake more cautiously than I would have for normally for that station stop and the time of year. Slippery rail can clear up as quick as it starts. Ask any driver whose immediately following the RHTT how terrifying that can be trying to stop normally.