Runaway freight train incident at Tebay - 17 August 2010

[whole diary]

[JayMac]

From Railnews (23/08/2010):

THE Rail Accident Investigation Branch has started an inquiry into an incident on the West Coast Main Line, in which it is said a DB^» (Deutsche Bahn - German State Railway - about) Schenker freight train rolled backwards for more than 1.5km towards Tebay from Shap Summit last Tuesday.

Four track workers were killed at Tebay when a trailer being used for permanent way maintenance ran away in 2004.

The RMT^▸ (National Union of Rail, Maritime & Transport Workers) claimed only quick action by signallers at Carlisle had prevented another accident on Tuesday.

The union's general secretary Bob Crow said: "This is just the latest runaway rail vehicle, following incidents on London Underground and in Scotland, yet Network Rail is playing fast and loose with safety in the name of cutting costs. We have been calling for secondary protection for track workers during track possessions since Tebay."

Network Rail promised that safety would not be compromised. A spokesman said: "The Rail Accident Investigation Branch is looking into an incident in the early hours of Tuesday 17 August involving a freight train operated by DB Schenker. There were no people working on that section of railway and there were no passenger trains in the vicinity either."

He also said lookouts were still being used, and that workers were protected elsewhere in any case.

"Nonetheless, we are happy to talk to the RMT about any legitimate concerns it may have," he added.

Edit note: This post has now been split off from the topic about the London Underground incident, to form a new topic here. CfN.

[Chris from Nailsea]

From the RAIB^▸ (Rail Accident Investigation Branch):

Investigation into an incident involving a freight train near Shap, Cumbria, on 17 August 2010

The RAIB is carrying out an investigation into a serious incident involving a freight train that occurred on the West Coast Main Line in Cumbria.

At around 01:51 hrs on 17 August 2010 and around two miles from Shap summit, freight train 4S25, the 21:32 hrs northbound container service from Hams Hall to Mossend Up Yard, began rolling backwards. The train continued to roll back down a gradient of around 1 in 75 for nearly 4 minutes, covering 2.1 miles (3.4 km) and reaching a speed of around 50 mph (80 km/h).

The signaller became aware that the train was rolling backwards from indications provided by the signalling system.  He responded by arranging for the transmission of an emergency radio message to the train^s driver and setting a route into sidings at Tebay to divert the train clear of the running line. There were no other trains closely following the freight train at this time.

The driver realised that his train was moving backwards, applied the brake and brought the train safely to a stand more than a mile north of Tebay and before the rear of the train had reached the sidings. The train comprised a class 92 locomotive and 13 twin-container wagons with a total length of 498 metres and a trailing load of 663 tonnes.

The investigation will identify the sequence of events before, during and after the incident and the reasons for the driver being unaware that his train had lost forward motion and was rolling backwards.

The RAIB^s investigation is independent of any investigations by the safety authority.

The RAIB will publish a report, including any recommendations to improve safety, at the conclusion of its investigation. This report will be available on the RAIB website.

[eightf48544]

Apparently there was a similar incident 30 years ago, only the runaway collided with the following train.

[Chris from Nailsea]

An update, from the RAIB^▸ (Rail Accident Investigation Branch):

Investigation into an incident involving a freight train near Shap, Cumbria, on 17 August 2010

Updated 17 February 2011

The RAIB is carrying out an investigation into a serious incident involving a freight train that occurred on the West Coast Main Line in Cumbria.

At around 02:05 hrs on 17 August 2010 and around two miles from Shap summit, freight train 4S25, the 21:32 hrs northbound container service from Hams Hall to Mossend Up Yard, began rolling backwards. The train continued to roll back down a gradient of around 1 in 75 for nearly 4 minutes, covering 2.1 miles (3.4 km) and reaching a speed of around 50 mph (80 km/h).

The signaller became aware that the train was rolling backwards from indications provided by the signalling system.  He responded by arranging for the transmission of an emergency radio message to the train^s driver and setting a route into sidings at Tebay to divert the train clear of the running line. There were no other trains closely following the freight train at this time.

The driver realised that his train was moving backwards, applied the brake and brought the train safely to a stand more than a mile north of Tebay and before the rear of the train had reached the sidings. The train comprised a class 92 locomotive and 13 twin-container wagons with a total length of 498 metres and a trailing load of 663 tonnes.

The driver was working a first night shift following an early shift and a day off duty. It is most likely that this incident occurred because the driver was fatigued.

The rail industry has carried out a lot of research into the causes of fatigue and its consequences and there is a significant amount of information available from elsewhere. No further research is required to understand the influence of fatigue on this incident.

Rail and other industries use the Health and Safety Executive^s Fatigue and Risk Index calculator to calculate fatigue and risk scores for each shift in a shift pattern.

A report published on the Office of Rail Regulation^s website evaluates UK^▸ (United Kingdom) rail industry practice and concludes that night shifts with fatigue scores less than ~40-45 and risk scores of ~1.6 or less represent good practice. See report RSU/08/03 (http://www.rail-reg.gov.uk/upload/pdf/sres-EvalRailFRIT.pdf)

The freight train operator used the Fatigue and Risk Index calculator to calculate fatigue and risk scores for the driver on the shift in which this incident occurred. The scores were well below those identified as good practice in report RSU/08/03.

The RAIB is now working to understand how the Fatigue and Risk Index calculator^s scores compare with the findings of other fatigue assessment tools and actual accidents and incidents, particularly in regard to a first night shift after a day off duty.

[JayMac]

So, Bob Crow's quick to judge and blame comments are totally unfounded in this case. I wonder if he will apologise to Network Rail?

I'm not holding my breath.

[inspector_blakey]

I can't help but wonder if the RAIB^▸ (Rail Accident Investigation Branch) finds itself having to over-analyse incidents like this because it would not be regarded as acceptable simply to say "it was a simple accident, very rarely they do happen and there's little we can do to legislate against something like this happening again".

Death by Health and Safety has an excellent article here discussing the RAIB report on a track worker fatality near Leeds. As anyone who is familiar with the website will know, it's written by a qualified COSS and fights the corner of the track worker vehemently. But their take on the accident is that it was just that, and it's unlikely that RAIB's recommendations will make much of a difference, or even be particularly practical to implement.

There are so many variables in the Shap incident that it's difficult to know where to start. From what I can gather (although it's not stated directly) the driver of a stationary train fell asleep in his/her cab and as a result, presumably after the brake leaked off, the whole job rolled backwards with the driver still asleep at the controls. This is an incredibly unusual event, and obviously the rostering pattern is unlikely to be the only causal factor. Other factors include whether the driver had slept properly, whether there was anything preventing them from doing so, did they have a cold, what was going on at home, etc etc. It's naive in the extreme to think that revisiting the fatigue index (which is essentially a theoretical box-ticking exercise) is likely to prevent a recurrence here, because there are so many variables in any such situation that are completely outside the control of the RAIB or the rail companies involved.

[eightf48544]

One question, does the driver's vigilance device work when the loco/train is stationary but the loco "switched on"?

[Tim]

One question, does the driver's vigilance device work when the loco/train is stationary but the loco "switched on"?

Good question.  I tend to agree with Inspector Blakey that a box ticking risk assessment exerice is not an effective way to prevent a reoccurance.  But surely there is some technical solution to this kind of problem. 

(I am about to show my ignorance of how trains work my driving experience being limited to 5 minutes on a class 08 on the GWR^▸ (Great Western Railway), so please put me right if I make any wrong assumptions but...)  surely if you deliberately want a train to go backwards you need to do something akin to "putting it into reverse",  If that isn't done and the train simply coasts backwards then is it not possible for a loco to be designed to apply the brake automatically?  Is this not infact something that could be done via a software modification in a modern computer controlled loco?

Certainly a class 92 should not be doing 50mph with the driver's key in the cab at the wrong end.  Isn;t that something that the train itself ought to be able to detect and correct for?

As I say, I am sure I am displaying my ignorance,  but IMHO^▸ (in my humble opinion) the RIAB might be better proposing engineering solutions than yet more boxes to tick. 

[eightf48544]

(I am about to show my ignorance of how trains work my driving experience being As I say, I am sure I am displaying my ignorance,  but IMHO^▸ (in my humble opinion) the RIAB might be better proposing engineering solutions than yet more boxes to tick. 

This is an very interesting point, if you contrast the current RAIB^▸ (Rail Accident Investigation Branch) reports with the old HMRI^▸ (Her Majesty's Railway Inspectorate) reports, they are a completly different.

The HMRI reports are concise they give the circumstances and thus deduce the cause and who was responsible (named) plus making recommendations where necessary. I have the report on the head on crash at Ponthydrefen on the R&SBR
which unfortunately resulted in several deaths. It's only about 8 foolscap pages.


Bssically it was a combination of circumstnces, the guard probably didn't pin down enough brakes leaving the colliery siding, the driver came down the hill too fast. The signalman at Ponthydrefen pulled off his home signal too quickly on hearing the train whistle, instead of ensuring it was nearly at a stand and slowly lowering the signal to indicate the starter into the single line was at danger. Thus deceiving the driver into thinking he had a clear road through the station. He was then going too fast to stop on a steep down grade (unfitted loaded coal train) and ran head on into the DMU^▸ (Diesel Multiple Unit) coming up the hill.

In contrast I've seen the RAIB report on a SPAD^▸ (Signal Passed At Danger) at Purley by a stone train, which must be at least 20 pages long to say that depending on where the driver was in front of the signal he would be unsighted due to the station canopy.

The old HMRI inspectors were mostly ex RE^▸ (Religious Education) officers who had run trains in war zones round the world, what they diidn't know about the dodges and wheezes the boys got up to the boys didn't know either.

[bobm]

In this age of desktop publishing the eight pages of foolscap probably cost more to produce than the 20 page glossy version of today by the time it had been bashed out on a typewriter and then typset by the printers.

[Tim]

In this age of desktop publishing the eight pages of foolscap probably cost more to produce than the 20 page glossy version of today by the time it had been bashed out on a typewriter and then typset by the printers.

the time taken to write the report isn;'t the issue.  Short concise clear reports are earier to read and understand.  That is the issue

[inspector_blakey]

surely if you deliberately want a train to go backwards you need to do something akin to "putting it into reverse",  If that isn't done and the train simply coasts backwards then is it not possible for a loco to be designed to apply the brake automatically?  Is this not infact something that could be done via a software modification in a modern computer controlled loco?

Trains will certainly roll backwards on an uphill gradient, often surprisingly fast because their rolling resistance is so low when compared to a car. I'm not familiar with modern stock but certainly when starting older trains on a hill the driver may need to take power with the brake partially applied then release it more or less immediately to stop the train rolling. So it's certainly not implausible that a heavy train rolling back over a long distance could pick up significant speed.

What puzzles me is that I think the driver's safety device (DSD^▸ (Driver's Safety Device), aka deadman's) should prevent a train rolling if the driver nods off at a signal, as they will cause a brake application unless they're held down. I'm ready to be corrected though, as I'm not familiar with more modern locos/units than older BR^▸ (British Rail(ways)) heritage stock.

[Electric train]

I am sure the vigilance system is only active when the reverser is selected in a direction of travel, I have not read the report I do wonder if the driver was holding the train on the loco (straight) air brake and not on the train brake

[Chris from Nailsea]

The RAIB^▸ (Rail Accident Investigation Branch) full report has yet to be published - the above item on their website was just an update on their investigation.

Interestingly, it seems to have raised more questions than answers. 

[Chris from Nailsea]

The HMRI^▸ (Her Majesty's Railway Inspectorate) reports are concise they give the circumstances and thus deduce the cause and who was responsible (named) plus making recommendations where necessary. I have the report on the head on crash at Ponthydrefen on the R&SBR which unfortunately resulted in several deaths. It's only about 8 foolscap pages.

Thanks for posting that interesting comparison, eightf48544. 

That accident report is available on the Railway Archive website - as you say, it's nine pages, one of which is an excellent map of the scene.