Return to Archive -by date - by topic - 2001 Archive.
Diffuse and prolix it may be, but Lord Cullen's report into the Ladbroke Grove accident paints a disturbing picture of a railway left leaderless.
Given that Moses came down the mountain with just 10 commandments to govern the conduct of an entire people, it seems somewhat self indulgent for Lord Cullen to come up with 89 Recommendations in his report into the inquiry into the Ladbroke Grove accident. True, that is four fewer than the Hidden Report into the Clapham Junction accident, but one would have hoped that by now the lesson had been learned that less is more.
Hidden also digressed into areas that were unrelated to the cause of the accident and its consequences – notably his recommendation on fitting Automatic Train Protection – and inevitably the flood or recommendations meant that the important issues lost focus.
Topic Support of the bereaved & injured Track & signalling changes Implementation of formal inquiry Recommendations Signalling in the Paddington area ) Driver management & training Safety auditing Signal sighting SPAD investigation Signaller's instructions Signallers training and briefing Signallers working conditions IECC equipment Automatic controls Radio communication Preservation of data Crashworthiness Fire mitigation Passenger protection, evacuation & Escape Implementation of these Recommendations |
Number 4 1 2(1) 3(2) 9 2 11 3 3(3) 4 4 2 2 1 1 8(4) 1(5) 27(6) 1 |
1) Including nine sub clauses
2) Including seven sub clauses
3) Including two sub clauses
4) Including six sub clauses
5) Including eight sub clauses
6) Including three sub clauses
For example, Recommendation 51 says ‘There should be a national system of direct radio communication between trains and signallers'. This has been blindingly obvious to all except the rail signal and telecommunications profession for years. Something must be done
Going from the profound to the footling, Recommendation 75 says ‘The provision of ‘snap wands' should be considered as a supplementary means of providing lighting in an emergency'. True, this is not one of the ‘key recommendations' which are italicised in the report, but it is graded ‘1' in the timescale for implementation. That is within ‘up to six months' and Lord Cullen emphasises ‘up to'.
Running through the groups of recommendations, ‘Support for the bereaved and injured' is what the real world calls casualty lists. If a jet approaching Heathrow runs out of fuel and crash lands in Hyde Park , you know who is on board. Trains don't have passenger lists and a big accident generates such a mass of data that it overwhelms paper systems. Cullen calls for computerisation.
‘Track and signalling changes' (Recommendation 5) is intended to make infrastructure engineers think twice before designing layouts that are incredibly versatile, but are horribly complex and place unreasonable demands on drivers. It calls for safety assessments on an holistic basis throughout the specification and design process.
There were suggestions during the Inquiry that the Paddington bi-directional layout went ahead only because the designers assumed trains would be protected by Automatic Train Protection (ATP): I don't buy this. There was no flank protection for SN109, the signal passed at danger in the accident, because no one believed that a driver, having SPADed, would not realise his error.
Nor will the presence of ATP or the Train Protection & Warning System, diminish the importance of driver friendly layouts and signalling. One driver told the Inquiry that he drove out of Paddington from signal to signal at 15mile/h. Thus, the complexity of the layout had negated the capacity benefits in his case.
With automatic protection, drivers would still be running from signal to signal. Driving on the assumption that ATP/TPWS will stop you if you make a mistake is not good, or even acceptable, practice and incompatible with defensive driving.
'Implementation of formal inquiry recommendations' (Recommendation 6) highlights the fact that in Great Western Zone recommendations repeatedly went missing, to the extent that people tasked with action didn't know about it months after the recommendation was made. This single recommendation, with its nine sub-clauses, defines the management of formal inquiry recommendations and has a 12 month timescale.
'Signalling in the Paddington area' essentially calls for a safety examination and independent risk assessment of the route out to two miles six chains from Paddington, that is, just west of the accident site. Any changes can be made only after Her Majesty's Railway Inspectorate is satisfied that they are safe and any associated mitigating measures implemented and shown to be effective.
Measures suggested by Cullen include the conversion of four aspect signals to three aspect, flank protection for SN109 and possibly other signals, the installation of ‘standard, simple, non-distracting and consistent means of line identification', changes to signal head installation and the erection of a new gantry on the Paddington side of Portobello Bridge carrying the down signals, including SN109, currently mounted on gantry 8 to the west of the bridge. It is clear that the bridge obscured the signals on gantry 8 and SN109 in particular.
Driver management and training covers many of the issues raised in my article on ‘The forgotten driver' (Modern Railways January 2000). The first recommendation (No 11) raised a wry smile. It says that signallers and drivers should jointly attend away days and other training processes to develop their mutual understanding'. Back in the late 1980s Ivor Warburton and Peter Rayner (General Manager and Operations Manager of the London Midland Region of British Railways) did just this, as I reported in Modern Railways. Well, not quite, the signalmen and drivers went to a country pub after their shifts had ended.
However, the meat of this section is all about training, periodic revalidation and central licensing of drivers. And it seems to me that it misses the big picture.
Only two recommendations are italicised. One requires Thames Trains to increase face to face driver briefings to at least monthly and the other (Recommendation 14) says that ‘TOCs should review the effectiveness of the system in place to deliver the required level of driver competence at least once very thtee years and should retest the driver against the revised systems at the same frequency'.
Talk about circumlocution. With drivers changing employer frequently as they chase this month's best pay packet, not to mention the weaknesses in Thames training revealed in the report, training is too important to be left to 25 train operators.
What is needed is a common national syllabus which established basic competence, leaving the TOCs to provide specialised traction knowledge and route learning. This should be backed up by a central driver licensing scheme (Recommendation 15 says that the ATOC study of central licensing should be ‘progressed expeditiously') and two yearly refresher training in emergency procedures combined with license validation.
ATOC cannot be expected to take on this responsibility which since it is an association does what its members want and not vice versa. The current attempts to develop a Railway Group Standard on driver training are making heavy going.
As for Recommendation 19 calling for further research into human factors – good grief, are there any human factors consultants not studying driver already?
Skipping over the two recommendations on safety auditing which are redundant, we come to signal sighting where Railway Safety Ltd is tasked with six recommendations requiring implementation within six months.
Here Lord Cullen has cast new light, for me at least, on the old canard that all the driver has to do is obey the signals. He quotes from a W.S. Atkins report: ‘A number of recent studies have emphasised the need for the driver not only to have a clear view of all the relevant signals but also to have the information from those signals presented in a way that enables him/her to gain a clear understanding of the message being conveyed'. In other words ‘signals require to be not only visible but also readable'.
‘Readable' defies an easy definition, but Gantry 8 provided a classic example of ‘illegibility'. At the most basic level, as a train wound its way through the six bi-directional tracks out of Paddington, including curves, the driver would see the signals on a gantry, but it was not immediately apparent which one applied to the line the train was on.
Gantry 8 had five signal heads and to identify ‘their signal' drivers would count across from one side. With the signals being temporarily obscured by overhead line electrification equipment, a driver leaving Paddington had to be super alert – and on the day of the accident the low bright morning sun shining into the lenses could have made life even more difficult.
Listen to these Thames Trains drivers giving evidence. Driver Robinson ‘there is just so many you have to keep working out which one is yours all the time'. Driver Instructor Wells, ‘I take signal to signal, you have to take it really slowly and you need to know which line you are actually on and you need to count from left to right and right to left. You need to double check again and again and you have to be certain. If you are not certain, you need to stop and you won't proceed'.
Driver Trainer Brookes (Heathrow Express) said that the view of the red aspect of SN109 (which was beside the lower yellow in an L configuration) was obstructed by an insulator in front of it. If a red was expected, she always approached the signal with extreme caution, at a speed of about 15mile/h, in accordance with defensive driving'.
These professionals knew that the watchword is ‘never assume'. But Driver Hodder's 200 odd hours of training had clearly left him unprepared for a very difficult station layout.
Commenting on the lack of any formal route maps of the station approaches Driver Hodder's practical handling Driver Instructor Mr Adams told the inquiry, ‘I have grown up with the system. I have taken it as it has evolved, but someone like Mick (Hodder) is just thrown in at the deep end and said well, this is Paddington and sort of make the best of it really'. And as a traction instructor Mr Adams did not consider that it was his job to teach the routes. This has been criticised, but under BR, route learning and traction handling were separate subjects.
There are various theories for Driver Hodder's confident acceleration before the red aspect of SN109 became visible. Given that he engaged Notch 1 255m before SN109 and then Notch 5 at 239m suggests that he believed he had a proceed aspect and, as Table 2 shows, the obscuration of the signals meant that readability was almost non-existent.
| 433m | |
| LY | very heavily obscured |
| 299 | |
| LY | heavily obscured |
| 282 | |
| LY | partially obscure |
| R | brief & heavily obscured glimpse |
| 258 | |
| LY | Clear |
| UY | heavily obscured |
| 219 | |
| R | heavily obscured |
| LY | Clear |
| UY | partially obscure |
| 188 | |
| R | partially obscure |
| LY | clear |
| UY | partially obscure |
| 168 | |
| R | clear |
| LY | clear |
| UY | clear |
Key:
R Red
UY Upper Yellow
LY Lower yellow
Lord Cullen recommends that the standard on signal sighting should require that ‘explicit consideration' should be given to the readability of a signal. He adds ‘it should be made clear that the fact that a signal complies with the minimum requirements is not of itself to be taken as meaning that it is adequate'.
Cullen also calls for this standard to deal ‘explicitly' with the additional time taken for the reading of certain signals such as those mounted on gantries. How you do this is not clear, since experience will be a factor. But the thought of those experienced drivers treading carefully from signal to signal suggests that readability could be aligned with reduced approach speeds.
In the current Group Standard, obscuration of a signal by overhead line equipment is permitted for a ‘very short durations'. Lord Cullen says that OHLE should be restricted to wires and ‘droppers' and the dispensation should not apply to structures and larger components such as registration arms and insulators.
He also want ‘very short duration' to be clarified and
Railtrack is required to identify areas where any ambiguity ‘may have caused or may still cause problems'. There should be a retrospective review of all such locations.
These recommendations are already affecting current signalling schemes, including the Leeds remodelling, where a gantry had to be re-sited, and the upgrading of the West Coast Main Line to 125mile/h (page xx this issue)
On SPAD investigations the main recommendations are that it should not be presumed that driver error is the sole or principal cause or that any part played by the infrastructure ‘is only a contributory factor'. Lord Cullen doesn't like the use of the word ‘Disregard' when formally categorising SPADs. Since it means ‘pay no attention to', which is what can happen, I assume this is just his Lordship succumbing to ‘more-PC than-thou'.
When it comes to the role of Slough IECC in the accident we get the unreconstructed bloke railway. According to Informed Sources, Slough did not have a good reputation in its dealings with drivers. Nor was it well run: during one visit a senior railwaymen was not impressed by staff chatting with their backs to the display screens.
Attitudes to SPADs echoed those of signalling management and the public – all drivers have to do is obey the signals. Here is signaller Siddell being cross examined:
A:‘We were hoping the train would come to a stand'
Q: You do not rely on hope do you?
A: Well, the driver is supposed to stop at a red signal
And again
Q: What about the near miss in February 1998? Was there talk between you and your colleagues about what had happened that day and how matters could have been improved?
A: Generalisations, but it was still basically down to the driver to stop at a red signal.
Q: Was that the overall feeling? If a SPAD happens, that is essentially a driver problem?
A: Well how can it not be? If a person at a traffic light goes through a red light is it not his fault?
Q: What about the problems that there had been in the Paddington area of drivers going through not just SN109, but other signals on a number of occasions when they were red against them?
A: Yes.
Q: Was it still felt that the problem was exclusively the driver's?
A: Well if a driver fails to stop at a red light, he is at fault.
But apart from the management and ethos of Slough IECC, described by Cullen as a ‘slack and complacent regime' a bigger issue was the obvious confusion over the actions signallers should take in the event of SPAD. This centres on two regulations, Signalling General Instructions (SGI) No 47 and the Track Circuit Block Regulations (TCBR) No 6
In the event of a SPAD SGI No 47 requires the signaller to ‘immediately arrange for the movement to be stopped by the most appropriate means and take any other emergency action'.
TCBR Regulation 6, which covers ‘train divided or a train or vehicle running away', says that if the signaller is aware or suspects that a train is running away or proceeding without authority the first task is to
place or maintain at danger the signals for any line which may be affected in order to stop any other train which may be endangered.
So that's confused to start with
At Slough IECC, the policy in the event of a SPAD was that SGI No 47 applied until the train ran beyond the 200 yard signal overlap when TCBR Regulation 6 took priority.
Meanwhile the Instructions to Signallers at Slough IECC contained a section for “Driver Only Operation of Passenger Trains ‘Cab Secure Radio' system”. This said that if a signaller required to stop a train in an emergency ‘you must immediately carry out the procedure for sending a STOP message'. This was in addition to ‘placing or maintaining signals at danger and carrying out any local instructions'.
Despite the introduction of Cab Secure Radio (CSR), none of the signallers at Slough had been trained in its use in an emergency and none had used it in the event of a SPAD, of which there had been 46 between 1993 and 1999, including the ‘dress rehearsal with the IC125 In February 1998.
As Mr W.G. Body, now a consultant but at one time Area S&T Engineer at Doncaster, told the Inquiry ‘The traditional way of trying to deal with an emergency has been to try and set back signals at danger and that seems to have carried on through, even though other methods such as CSR have become available and may be more appropriate. I do not really think that it has sunk into the culture of signallers as to how they should deal with such situations'.
Quite. To use the National Radio Network, the signaller at Slough was expected to contact the Railtrack Customer Service Centre at Swindon by using a specified phone number, state that it was an emergency and ask for an ‘all trains stop' message to be broadcast in the area affected.
‘Thank you for calling the Swindon CSC, if you would like an all train stop please press 1. Thank you for requesting an all station stop message, for the Swindon Area press 1, the Reading Area press 2'. Sorry to be flippant but a hundred years after Marconi bridged the Atlantic railways have not advanced much.
Anyway, so many caveats are applied to the use and availability of CSR that its value is debatable. Lord Cullen agrees that a CSR stop message should not be the mandatory reaction to a SPAD for a range of reasons. It might not be the right decision – for example stopping the train in mid section, CSR might not have been properly set up or might have failed in service so the message can't be received (!) or it might be faster to set a signal at danger ahead of the train.
That said , Signallers' instructions recommends that signallers' actions in the event of a SPAD should be clarified and set out in a single set of instructions. Separate local instructions might be necessary.
Instructions should state ‘explicitly' that in the event of a SPAD the signaller should make an assessment and take action ‘immediately'. The instructions should provide a set of options, including the use of CSR, ‘which should be supported by full and regularly repeated briefings as to the type of circumstances in which each option is or may be appropriate'.
At the press conference I asked Lord Cullen what he thought a signaller should do in the event of a SPAD. It will remain ‘a matter of judgement' he replied. Err on the safe side then.
Two recommendations FOR <B>Automatic controls<B> seem to conflict with the options approach for signallers faced with a SPAD. No 49 calls for a study into the possibility of signals being placed at danger automatically in the event of a SPAD and No 50 says that, subject to a risk assessment ‘an arrangement should be made whereby, when a train which is fitted with the CSR passes a signal at danger an audible warning sounds in the cab'.
This sounds technically iffy to me. Come to think of it, why not have a system linked to the on-board AWS which applies the brakes automatically if a train approaches a red signal too fast or SPADs? Gosh, yes, we could call it the Train Warning & Protection System, TWPS, or Twerps, for short.
Which brings us to the second half of the report which is all about crashworthiness, fire mitigation and passenger protection, evacuation and escape.
Quite frankly Cullen goes overboard on <B>crashworthiness<B> in an accident where doubling the crashworthiness of the Turbo front end would have been like a fart in a thunderstorm. When I pointed this out to Lord C at the press conference he agreed enthusiastically. ‘You are absolutely right, but that this does not mean that more cannot be done for lower speed accidents' he said.
Yes but, that way madness lies, with victims believing that the aluminium Turbo was a cheapo structure which should have absorbed its share of 350MegaJoules impact without breaking up. And Lord Cullen's report pays tribute to the performance of the Mk 3 coaches under such extreme conditions.
But it helps enormously if the Mk 3 coaches have a 70 tonne power car on the front to take the brunt of the collision and Great Heck showed what happens to a Mk4 coach, in the same league as the Mk 3, when it hits a Class 66 with a front end like a brick out house. The Mk 4 Driving Van Trailer, like the lead Turbo car at Ladbroke Grove, disintegrated.
Lord Cullen's recommendation on crashworthiness are based on the views of the experts in crashworthiness, (notably at WS Atkins) who reported to the Inquiry. The trouble is that crashworthiness is an integral part of a structure and retrospective modifications are costly, difficult and of uncertain efficacy. Hands up anyone who believes that Mk 1 cup and cone makes sense.
Cullen recommends that two retrofitted modifications should be considered for IC125 power cars - modified cab ends to improve ‘driver protection, energy absorption and compatibility' with other vehicles and enhanced bogie retention measures.
He concedes that these changes should be considered ‘subject to an assessment of feasibility, costs and benefits', but common sense tells you that designing and manufacturing brand new cab ends for power cars designed in the 1970s with a nominal 10 year life is not a sensible way to spend money. The laws of physics suggest that effective bogie retention in high energy accidents will probable depend on the availability of low cost industrial grade kryptonite.
In parallel, Recommendation 54 calls for current standards for crashworthiness for new vehicles to be reviewed in the light of experience at Ladbroke Grove ‘in order to ensure that there are adequate measures for the protection of drivers and the retention of bogies'. This is meaningless since 'adequacy' is determined by the collision speed and energy and increased protection can mean that more energy is absorbed by the other vehicle in the collision.
At Great Heck, the Class 66 freight locomotive cab, designed for structural integrity rather than energy absorption, successfully protected both drivers in the initial impact, but at the cost of the complete destruction of the Driving Van Trailer of the other train.
Also verging on the unrealistic is Recommendation 55 which calls for major surgery to the ends of Class 165/166 Networker Turbo vehicles to be considered.
‘Poor quality' welding of the bodyshell end pillars ‘may have' reduced protection to passengers in the middle car of the Turbo. Enhancement of these welds should be considered, recommends Cullen, together with the ‘possible enhancement' of crashworthiness by weakening the end and strengthening the saloons of vehicles.
Oh come on! Attacking a bodyshell fabricated from continuous extrusions with angle grinder is not like gas cutting the underframe of a Mk 1 coach as part of cup and cone (Informed Sources May). I find this recommendation anomalous, given that the Inquiry's panel of crashworthiness experts did not consider, given the circumstances of the collision, that the death and injuries would have been reduced had the Turbo been built of steel.
Of course, you can't be complacent about crashworthiness and improvements are being sought, but the detail in this section is a distraction from the real issues of why the accident happened. Recommendation 58 calls for revision of the crashworthiness Group Standard to be ‘pursued' within 12 months, with particular reference to design requirements for more realistic scenarios, high speed accidents and dynamic verification testing. What's a ‘realistic scenario' and what do you call ‘high speed'? And under the interoperability directive I suspect we have to go with European practice on testing.
A number of Fire mitigation measures are to be ‘considered'. These include a review of Group Standards relating to crashworthiness of fuel tanks and consideration of internal flexible linings or cellular fillings.
Once again, we run into practicality. Certainly, new fuel tanks could be improved as suggested, although flexible linings would be difficult to retrofit since modern fuel tank design incorporates internal ‘egg crate' style baffles to stop surge.
Similarly, Lord Cullen recommends that moving fuel tanks to the rear of leading cars should be considered for added protection in frontal impacts. The trouble is, the optimum layout for driving power cars is to have the engine and transmission driving the trailing (or No 2) bogie. This provides better adhesion for the powered wheels since the leading bogie conditions the track. It also means that the fuel tank and compressed air reservoirs have to go behind the leading bogie. Unless you have electric transmission, of course.
Equally operators are unlikely to support proposals for a possible reduction in ‘fuel inventories' and the use of smaller fuel tanks, given that maximum range is an important factor in the utilisation of both passenger and freight power units.
Finally we come to Passenger protection, evacuation & escape which provides more than a quarter of Lord Cullen's recommendations. Many of the recommendations are concerned with detail, such as announcements and signage. Present practice, introduced after the Southall and Ladbroke Grove, is endorsed.
But hiding in this superfluity of recommendations are a number of really important features. Emergency lighting that fails under sudden deceleration is useless, yet the relevant recommendation is not italicised.
Something that should have been thought of years ago are one piece internal doors which slide in the same direction at both ends of a coach. If you are unlucky, with the vehicle on its side after an accident, both doors will have slid shut. By December 31 2002 one of three remedial options should have been implemented, opposite handed doors, bi-parting doors or a removable panel fitted in the door large enough for a passenger to crawl through.
This requirement has provision for an extension if Her Majesty's Railway Inspectorate is satisfied that the deadline cannot be met realistically.
Feasibility studies are required within 12 months into the use of windows as a means of escape, including removable windows. Saloon end doors into drivers' cabs, normally kept locked, should be fitted with an override device to allow their use by passengers in an emergency within six months.
Both these measures can be opposed on the grounds of the potential for vandalism. This is old railway thinking. If passengers can endanger the train, then it is up to the operator to control those passengers. If that means putting staff back on DOO trains, so be it
Finally, fitting ‘escape hatches' both to existing carriages and new passenger vehicles is also to be the subject of a feasibility and risk analysis. I think it is a waste of time – we shall see.
Supposing I could have made it a round 90 Recommendations with one of my own? My Recommendation 90 reads 'Railtrack should appoint an experienced Chief controller to be responsible for the operating policy and day to day management of all signalling control centres and the new Network Management Centre'. Any better ideas?