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

[Deepgreen]

Not always. My driving style is such that I use the same technique all year round.

Just out of interest, is that technique based on leaf fall conditions or not? If the latter, how do you deal with leaf fall?

 

[HSTEd]

This is going to lead to a real mess, tightly tuned timetables interacting with ever-increasing maintenance costs for vegetation control interacting with a corporate culture that strongly favours reduced risks.

I suppose the technical solution to this problem is to fit trains with either magnetic track brakes or outright eddy current brakes for emergency use.
But any solution is going to be very expensive, and its highly unlikely the railway industry would be able to carry through vegetation removal to the required extent even if it was inclined to try.

 

[Ken H]

When BR started to introduce disc brakes on short trains, there was concern about adhesion, both on starting and stopping. The scrubbing effect of tread brakes on the wheel tread was supposed to be beneficial.
I am not sure as to the why and the wherefore, but when the 158's were first introduced, they were split to male 156/158 2 car sets, so the 156 tread brakes could keep the railhead a but cleaner. And keep the track circuits working.

 

[TheLastMinute]

I'm surprised there was no comment on the misinformation that "a Network Rail spokesperson said: "At around 19:00 GMT this evening, the rear carriage of the 17:08 Great Western Railway service from Portsmouth Harbour to Bristol Temple Meads derailed after striking an object on its approach to Salisbury station. "The derailment knocked out all of the signalling in the area". (post 256)

Since this is clearly not what had happened, and caused unneccessary concern about the robustness and safety of the signalling system (which actually performed to specification, the faults were elsewhere), was there ever an investigation as to who gave out this information, and why?

I think it's a little disengious to characterise the NR statement in that way. I suspect it was issued in the first few minutes after the incident based on information in the inital REC call from the GWR driver who stated their train had derailed. It took about 17 minutes for the WICC (Wessex Intergrated Control Centre, the joint NR/SWR control)to become fully aware it was actually a collision involving the SWR train.

 

[Grumpy Git]

.............. its highly unlikely the railway industry would be able to carry through vegetation removal to the required extent even if it was inclined to try.

But the line side vegetation has got into it's current state by budget constraints in the first place. As per normal in the UK, its penny wise, pound foolish.

On the CLC, the passing trains are the biggest controller of line side vegetation. It's an absolutely ridiculous state of affairs, when the consequences aren't just damaged train bodywork, but accidents with serious injuries (or worse).

 

[GC class B1]

That's what running brake tests are for.

I understood that running brake tests were to check that the train brake system performance was satisfactory and not primarily to assess available adhesion. This is stated in clause 54 of the report. The available adhesion at the location of the running brake test is unlikely to be the same at all locations where braking will be required. Sorry I can’t copy the text from the report.

 

[Watershed]

Aside from the clear evidence of NR's poor internal communications and lack of adherence to policies, one aspect I'm surprised by is how the driver decided to brake later based on his knowledge of the fallen tree.

The driver's actions seem far from compliant with the defensive driving technique that has been espoused for about the last 20 years - particularly given the weather forecast and the fact that another driver specifically took the effort to warn the accident driver of poor railhead conditions.

One imagines his exceptionally long experience of the line (nearly 60 years!) led him to believe there wouldn't be any problems stopping, but complacency can catch out anyone, as this incident demonstrates.

One of the concerning aspects here remains how the incident could so easily happen again - DVRS equipment is incredibly effective in reducing stopping distances during poor adhesion, making them nearly as good as in clean, dry rail conditions (as noted in the report). Yet it is currently only slated for installation on two fleets.

Given its limited budget and resources, NR can't possibly hope to avoid all instances of poor adhesion - therefore I would have thought there is far more "mileage" in fitting all fleets with DVRS. Whilst that has a cost, it is surely much less than an extensive programme of works to remove lineside vegetation and increase the frequency of RHTTs/MPVs.

 

[Archvile]

Shot in the dark here, but is it possible the driver decided to brake later because of local knowledge? I.e. he thought he knew where the poor adhesion would be and wanted to wait until passing it before braking harder on what he perceived to be the section of good adhesion, to avoid sliding from the poor section into the good section?

 

[HSTEd]

But the line side vegetation has got into it's current state by budget constraints in the first place. As per normal in the UK, its penny wise, pound foolish.

On the CLC, the passing trains are the biggest controller of line side vegetation. It's an absolutely ridiculous state of affairs, when the consequences aren't just damaged train bodywork, but accidents with serious injuries (or worse).

The cost of maintaining steam-era trackside conditions would far exceed the savings from avoiding these problems.

This is not the 1920s, you can't keep an enormous gang of Irish peasants around (paid almost nothing) and not really worry about casualties from inevitable accidents caused by work next to open lines.

Trackside workers are now extremely expensive and safety concerns make large-scale vegetation work on tracksides enormously disruptive.

That's just the world we live in now. We just have to build a railway that can function better in the presence of degraded adhesion conditions.

 

[Western Gaffer]

Shot in the dark here, but is it possible the driver decided to brake later because of local knowledge? I.e. he thought he knew where the poor adhesion would be and wanted to wait until passing it before braking harder on what he perceived to be the section of good adhesion, to avoid sliding from the poor section into the good section?

It's explained in the report that he believed that adhesion would be poor where the tree had fallen earlier, so he made a conscious decision to wait until passing that, for exactly the reason you refer to, but unfortunately left it until 750m past the site of the tree.

 

[TurboMan]

I understood that running brake tests were to check that the train brake system performance was satisfactory and not primarily to assess available adhesion. This is stated in clause 54 of the report. The available adhesion at the location of the running brake test is unlikely to be the same at all locations where braking will be required. Sorry I can’t copy the text from the report.

Railhead adhesion is an intrinsic factor of braking performance, so one of the purposes of running brake tests is to gauge braking performance and how this is affected by the level of adhesion at the time. Clause 54 discusses a running brake test soon after the driver left Waterloo, but there should be nothing stopping a driver carrying out additional running brake tests at any time should they perceive that adhesion may have worsened.

 

[Deepgreen]

I suspect the Salisbury Tunnel Junction vicinity may be one of the worst examples of trees clustering over the railway at a junction, so perhaps an exercise to keep such junction areas clear(er) of vegetation may be feasible. There may be reasons why the trees are as they are there (TPOs, ewtc.), but as a principle, keeping junction areas clear would be hugely cheaper than whole routes. I also wonder about the practicality of installing retarders at such junctions, similar to those that used to be installed at marshalling yards to brake loose wagons being shunted. These would fufil a similar role to track brakes on trains, but might be a lot cheaper than fitting all vehicles with magnetic brakes.

Shot in the dark here, but is it possible the driver decided to brake later because of local knowledge? I.e. he thought he knew where the poor adhesion would be and wanted to wait until passing it before braking harder on what he perceived to be the section of good adhesion, to avoid sliding from the poor section into the good section?

It's very possible, but I don't think defensive driving techniques include the choice to delay braking and then brake hard, even if instinct and/or reasoning prompt it - it's more or less the opposite of what is required.

 

[david1212]

I find it interesting that RAIB have included this image of the change in scene over the last 60 years.

https://www.railforums.co.uk/attachments/screenshot-2023-10-24-112209-png.145323/

... Lets not lose sight of the fact that this area has a lot more vegetation around it compared to the past. It is lack of maintenance, we can all see vegetation growing out of the brickwork in viaducts.

Blatantly obvious the lack of vegetation management is the primary cause, not mitigated by lack of rail treatment. The rest is all consequences after these.

Has or had a lot more vegetation around it compared to the past ? i.e. During almost two years has the vegetation been significantly cut back not just above the tunnel portal but along the approaches ?

 

[guard1]

Fair enough although I think if we used the same technique all year round then the trains would constantly be late!

I use the same technique all year round and never get delay slips. It's the drivers that tear around half the year using too much brake in good conditions that struggle in leaf fall.

 

[GC class B1]

Shot in the dark here, but is it possible the driver decided to brake later because of local knowledge? I.e. he thought he knew where the poor adhesion would be and wanted to wait until passing it before braking harder on what he perceived to be the section of good adhesion, to avoid sliding from the poor section into the good section?

I am not a driver but I have an interest in understanding the mechanics of wheel/rail adhesion. I can see the logic in this driver waiting until he had passed the area where he understood the poor adhesion to be. It requires a higher level of available adhesion to restart a stationary wheelset than to prevent a rotating wheelset from sliding. Therefore if the driver had commenced braking in step 2 on the area of better adhesion he may have achieved (or at least believed he would achieve) a shorter stopping distance than relying on the older design of WSP (which may not be as effective as more modern systems) with sanding to control the deceleration. This is a judgement that cannot be quantified.
I don’t understand though why knowing he was approaching a conflict point he didn’t brake earlier and approach the poor adhesion site at a slower speed so later defensive braking was still possible after passing the poor adhesion site.

 

[Brush 4]

Technical solutions are not needed, just cut the trees down and then, keep them from re-growing. This is a 50 year long literally growing problem, caused by first BR giving up lineside maintenance as soon as steam finished and then Railtrack and NR doing even more nothing. Paul Clifton on BBC South tonight, quite rightly highlighted the tree growth problem and used the Warship picture on screen to compare with a scene from today. He pointedly said that the trees are still there 2 years after the crash and that NR when contacted, refused to comment.

I would think, if they still won't cut them down, this may be criminal negligence, certainly mismanagement. If DfT won't fund vegetation management, then they should be prosecuted. Putting cost cutting over safety is never right, whether it is rail, road, air or health.

 

[Bald Rick]

I would think, if they still won't cut them down, this may be criminal negligence, certainly mismanagement. If DfT won't fund vegetation management, then they should be prosecuted. Putting cost cutting over safety is never right, whether it is rail, road, air or health.

There are a whole series of controls that mean it is patently not criminal negligence.

 

[Brush 4]

Depends if it can be proved that they choose not to take preventive action when the solution is relatively simple.

 

[bramling]

I use the same technique all year round and never get delay slips. It's the drivers that tear around half the year using too much brake in good conditions that struggle in leaf fall.

Have to say I’ve always observed the opposite. Those who tear around in good conditions tend, IME, to be good at anticipating issues and knowing when to tone down their driving.

On the more general subject of Salisbury, from first seeing the transcript of how the train was driven I was surprised to see the train being driven like it was, even taking into account that it was an “old school” driver. I’d have expected the train to have been being brought down extra early, indeed the download seemed to show no allowance for the potential conditions at all, if anything the opposite - as the final report now seems to have picked up. I don’t write this as a criticism of the driver, but I still find the way the train was driven surprising.

 

[dctraindriver]

Or just listen to the colleagues who told him on the day. It’s in the report.

I know the view is - particularly in the driver grade - that drivers are more important than everyone else. But they are not above criticism.

Your first sentence is spot on, your second paragraph diminishes your opinion somewhat. Drivers are a small cog in the bigger wheel and the majority of drivers are of the same opinion.

Fair enough although I think if we used the same technique all year round then the trains would constantly be late!

I drive cautiously throughout the year and there’s plenty of padding within the timetable to be on time.

 

[HSTEd]

Technical solutions are not needed, just cut the trees down and then, keep them from re-growing. This is a 50 year long literally growing problem, caused by first BR giving up lineside maintenance as soon as steam finished and then Railtrack and NR doing even more nothing. Paul Clifton on BBC South tonight, quite rightly highlighted the tree growth problem and used the Warship picture on screen to compare with a scene from today. He pointedly said that the trees are still there 2 years after the crash and that NR when contacted, refused to comment.

Who will pay for the many thousands of new staff that will be required?
Who will pay for the massive disruption from the line closures that will be required, for the rest of time, to allow lineside vegetation to be taken back to the stema era?

This is not the 1920s, we do not have a vast underclass of cheap Irish labourers to throw at the problem. And modern safety culture will not tolerate the inevitable increase in workforce fatalities from working next to an open railway - so the railway will be forced to adopt a lower availability maintenance scheme with the huge attendant costs.

Putting cost cutting over safety is never right, whether it is rail, road, air or health.

The reality is that benefit cost ratio analysis exists precisely because safety is not infinitely valuable.
If it was, industrial society simply could not exist.

The real answer here is better sanding equipment, magnetic track brakes and eddy current brakes. Not trying to spend billions a year on a huge increase in staffing and reduction in sustainable railway traffic.

 

[Purple Train]

I think that the BBC report is rather skewed against the driver and away from the root (sorry) cause (leaves on the line). Cue, of course, the mainstream media complaining that "leaves on the line" is a cover-up for disruption, has no effect on braking, etc., etc.

I can see the sense in the driver's logic given his very long experience, but it's not a technique I'd want to use in the dark were I a driver (based on my understanding of physics). Ultimately, though the severity of the incident was exacerbated by the driver's decision, it shouldn't detract from the real issue - leaves. The statement issued by SWR post-accident (when I assume they would have been aware of the braking pattern) also suggests that the low adhesion was the main cause, exacerbated by the driver's error (though, to at least this layman, it illustrates how much rests on an error in leaf-fall season compared to normal).

 

[abn444]

I use the same technique all year round and never get delay slips. It's the drivers that tear around half the year using too much brake in good conditions that struggle in leaf fall.

Genuine question, how is it possible to drive the same all year round and still keep to time when they put extra padding in the timetable from around October time specifically for leaf fall?

 

[Annetts key]

Who will pay for the many thousands of new staff that will be required?
Who will pay for the massive disruption from the line closures that will be required, for the rest of time, to allow lineside vegetation to be taken back to the stema era?

This is not the 1920s, we do not have a vast underclass of cheap Irish labourers to throw at the problem. And modern safety culture will not tolerate the inevitable increase in workforce fatalities from working next to an open railway - so the railway will be forced to adopt a lower availability maintenance scheme with the huge attendant costs.


The reality is that benefit cost ratio analysis exists precisely because safety is not infinitely valuable.
If it was, industrial society simply could not exist.

The real answer here is better sanding equipment, magnetic track brakes and eddy current brakes. Not trying to spend billions a year on a huge increase in staffing and reduction in sustainable railway traffic.

Keep in mind that during the growing season, hundreds of signals get reported because drivers find that they become or are becoming obscured by vegetation. In some areas, the vegetation is intertwined around the signal structure…

Staff cannot walk a safe distance (in the cess) away from the track to carry out work that they otherwise could do safely (for example, S&T, telecoms, electricians or off-track staff trying to get to equipment, fences or drainage systems that are safely clear of the line).

And when Network Rail does find money to pay (sadly normally contractors) for vegetation to be cleared, they do clear three metres from the running rail towards the boundary (or less obviously if the boundary is closer). And yes, the bulk of this work is done while trains are running.

Plus, trees growing on cutting embankments are supposed to be controlled, because if they fall during heavy rain, windy or stormy weather conditions…

The RMT has been asking for Network Rail to provide more resources on vegetation management for over ten years.

 

[LUYMun]

The real answer here is better sanding equipment, magnetic track brakes and eddy current brakes. Not trying to spend billions a year on a huge increase in staffing and reduction in sustainable railway traffic.

And how long would it take for the rolling stock to be retrofitted with such technologies, let alone the time to research and develop them?

There are also the other issues that are posed by increased vegetation, such as lower visibility and the risk of more line closures should one fall onto the line. Improved adhesion control won’t mitigate those scenarios. So it’s about time the garden railway needs some sprucing.

 

[43066]

Genuine question, how is it possible to drive the same all year round and still keep to time when they put extra padding in the timetable from around October time specifically for leaf fall?

Braking pretty conservatively is the default drivers are trained to these days and should, with decent RHTT and modern brakes, allow a reasonably timely performance all year round, especially if an additional margin is added to the timetable during leaf fall*.

As drivers I’m sure we have all taken the odd risk and slid into stations too quickly during leaf fall in full service, hoping the brake will bite, and wondering “how will this look on the download if I slip through”. Once you’ve scared yourself once or twice, you become conservative…

Ultimately the timetable isn’t even a consideration when it comes to matters of safety, or when it comes to maintaining a marketable CV as a qualified train driver!

Taking silly risks also makes precious little difference to the timetable: I was on a noticeably conservatively driven and early braked SE Networker earlier this evening (I say that as someone who drove them myself for a few years), and noted from RTT that the time lost was minimal, in the scheme of the various built in allowances etc.

*As an aside, it’s notably late this year! Most leaves are still on the trees due to the unusually mild start to autumn.

 

[Peter Sarf]

Shot in the dark here, but is it possible the driver decided to brake later because of local knowledge? I.e. he thought he knew where the poor adhesion would be and wanted to wait until passing it before braking harder on what he perceived to be the section of good adhesion, to avoid sliding from the poor section into the good section?

That is my understanding. Once the wheels are locked up it takes an age to get them rotating again and then able to apply the brake - that's my understanding.

Railhead adhesion is an intrinsic factor of braking performance, so one of the purposes of running brake tests is to gauge braking performance and how this is affected by the level of adhesion at the time. Clause 54 discusses a running brake test soon after the driver left Waterloo, but there should be nothing stopping a driver carrying out additional running brake tests at any time should they perceive that adhesion may have worsened.

But is the running brake test more to test the trains brakes ?. Its not really to test the rail conditions which vary from place to place ?.

Has or had a lot more vegetation around it compared to the past ? i.e. During almost two years has the vegetation been significantly cut back not just above the tunnel portal but along the approaches ?

Good question - answered later as NO.

Braking pretty conservatively is the default drivers are trained to these days and should, with decent RHTT and modern brakes, allow a reasonably timely performance all year round, especially if an additional margin is added to the timetable during leaf fall*.

As drivers I’m sure we have all taken the odd risk and slid into stations too quickly during leaf fall in full service, hoping the brake will bite, and wondering “how will this look on the download if I slip through”. Once you’ve scared yourself once or twice, you become conservative…

Ultimately the timetable isn’t even a consideration when it comes to matters of safety, or when it comes to maintaining a marketable CV as a qualified train driver!

Taking silly risks also makes precious little difference to the timetable: I was on a noticeably conservatively driven and early braked SE Networker earlier this evening (I say that as someone who drove them myself for a few years), and noted from RTT that the time lost was minimal, in the scheme of the various built in allowances etc.

*As an aside, it’s notably late this year! Most leaves are still on the trees due to the unusually mild start to autumn.

I breath of realism. We are talking about humans. This was a very experienced driver (both traction and route).

As for vegetation clearance I agree that it costs. But cutting back 3 metres from the running lines does not need to happen every year. It is the unchecked steady built up of trees over decades that is going to cost money. This problem has been in the making for decades and it is coming home to roost.

 

[GRALISTAIR]

This problem has been in the making for decades and it is coming home to roost.

I totally agree.

 

[bramling]

Braking pretty conservatively is the default drivers are trained to these days and should, with decent RHTT and modern brakes, allow a reasonably timely performance all year round, especially if an additional margin is added to the timetable during leaf fall*.

As drivers I’m sure we have all taken the odd risk and slid into stations too quickly during leaf fall in full service, hoping the brake will bite, and wondering “how will this look on the download if I slip through”. Once you’ve scared yourself once or twice, you become conservative…

Ultimately the timetable isn’t even a consideration when it comes to matters of safety, or when it comes to maintaining a marketable CV as a qualified train driver!

Taking silly risks also makes precious little difference to the timetable: I was on a noticeably conservatively driven and early braked SE Networker earlier this evening (I say that as someone who drove them myself for a few years), and noted from RTT that the time lost was minimal, in the scheme of the various built in allowances etc.

*As an aside, it’s notably late this year! Most leaves are still on the trees due to the unusually mild start to autumn.

No matter how many times I read this report, I still can’t make much sense of what this driver was doing. I find it bordering on incredulous that a driver would be scouring the line (at nearly 90 mph) in the pitch dark for a fallen tree to use as a braking marker, when he had a perfectly good braking marker in the form of the double yellow aspect, even if this meant the red aspect was heavily overbraked. If worried about adhesion, why would anyone not simply brake earlier, even if just getting the train down a more manageable speed? If the driver wasn’t concerned about adhesion then why change braking points in the first place?

I suspect this is going to be added to the list of incidents where we we will probably never know the true cause. It would neatly fit in to the box of commonly seen caution aspect, no need to react immediately (especially in this case with the extra distance), lose concentration (in this case possibly wondering about the fallen tree), and then the brown stuff moment - except in this case the driver still applied light braking. To me the whole thing made little sense in the first version of the report, and it still doesn’t.

 

[Peter Sarf]

No matter how many times I read this report, I still can’t make much sense of what this driver was doing. I find it bordering on incredulous that a driver would be scouring the line (at nearly 90 mph) in the pitch dark for a fallen tree to use as a braking marker, when he had a perfectly good braking marker in the former of the double yellow aspect, even if this meant the red aspect was heavily overbraked.

I suspect this is going to be added to the list of incidents where we we will probably never know the true cause. It would neatly fit in to the box of commonly seen caution aspect, no need to react immediately (especially in this case with the extra distance), lose concentration (in this case possibly wondering about the fallen tree), and then the brown stuff moment - except in this case the driver still applied light braking. To me the whole thing made little sense in the first version of the report, and it still doesn’t.

Lest we forget. Whatever the cause - bad things happen AND it was probably the design of the coaches in the train that resulted in less casualties. So there is progress.