Return to Archive -by date - by topic - Archive 2002.
While the HSE's latest progress report told us less than we knew already about 2182A it gave the media another railway safety field-day
Much as I would like to, there is no point in flacking the Health & Safety Executive over the progress report published on 4 July by its Investigation Board for the Potters Bar derailment. No, I must try my hand at cool, clinical, analysis for once.
On the condition of points 2182A before they failed, causing the derailment, there is less in the Report than you read in this column last month. When I tried to elucidate technical detail at the press conference for the report, the HSE team declined to answer on the grounds that such information was evidence which, if released, could prejudice the British Transport Police inquiry.
According to the HSE, its investigation is being undertaken jointly with the BTP ‘who are in the lead while criminal charges, other than those under the Health & Safety at work Act 1974, are a possibility. The contrast with the work of the Department for Transport's Accident Investigation Branches, where the primary task is to determine the cause and then propose remedies, should be noted.
As this was being written an announcement about the formation of the Rail Accident Investigation Board was imminent. The Potters Bar investigation shows why an independent professional investigation body for the railways is essential.
‘Lost Motion' differences
As designed, UIC-54B installations should be fitted with two lost motion devices, or escapements; one on the connection between the points machine drive and the first (front) adjustable stretcher bar and the other on the connection between the back drive and the second (rear) adjustable stretcher bar. These devices allow a precise amount of slack to be introduced which absorbs the excess motion of the points machine, which will always deliver a standard movement. This allows the movement of the switch rails to be set up irrespective (within range) of the movement provided by the points machine. The set up of points 2182A was found to be not as designed in that there was no lost motion adjustment on the second (rear) adjustable stretcher bar. The lost motion on the second (rear) adjustable stretcher bar allows for minor movements to occur at the rear drive position of the switch rails without transferring these movements to the front drive position. The implication of the lost motion being at the wrong connection point to the adjustable stretcher bars is that any movements of the second (rear) adjustable stretcher bar are transferred to the first (front) adjustable stretcher bar. Should this movement be induced by lateral vehicle wheel forces, i.e. by a train travelling over the points, this could result in a higher rate of fatigue loading or vibration to components. Given the concerns about these differences, and that it may contribute to nuts loosening, this matter has been brought to the attention of HMRI who have discussed the matter with Railtrack. Appendix E Progress Report on the Potters Bar derailment
|
There are several anomalies in this Appendix.
irst, it says that UIC-54B points should be fitted with two lost motion escapements. To provide an escapement two locked nuts on a drive rod are not clamped to their bracket: instead, a gap is left between the nuts to allow play.
Note, the Appendix says two escapements, where the HSE's diagram shows three, which I have marked as A, B and C.
econd, it says that there was no lost motion adjustment in the rear stretcher, what I call tie bar 2 for clarity. This is not strictly true.
hile there was no escapement at C, one was provided at B in the drive rod from Tie bar 1 to the back drive bell crank. The implication that the lack of an escapement at C would have allowed movement of Tie bar 2 to be transmitted to Tie bar 1 ‘resulting in a higher rate of fatigue loading or vibration to components', is thus mistaken.
n my view, an escapement is required at either B or C, but not both. As reported last month, the escapement in the drive rod to the back drive, lacked the sleeve which protects the exposed threads on the rod and also appeared to be wider than necessary.
Other than the ‘missing' escapement, the Investigation Board would not discuss other variations from standard in the condition of 2182A. In particular, the asymmetric position of Tie Bar 1, which reduced the length of thread available at the end where the nuts were missing.
However, thanks to an Omnicom survey in July 2001, we know that the 2182A was in the correct configuration at that time. According to Informed Sources, Tie bar 1 was symmetrical and the linkage connecting the back drive bell cranks was correctly aligned.
Clearly changes had been made to the points in the 10 months before the derailment. A key issue is whether the Operation Galaxy inspection by a Railtrack/Jarvis team of engineers would have spotted the asymmetric Tie bar 1 dur9ng their inspection on 20 March.
On the possibility of sabotage, the Progress report is ambiguous ( see box) . At the press conference the point was made that Jarvis, the contractor responsible for the points, had shown the media photographs which suggested that the inboard nuts had indeed been forced outward – the ‘shiny thread' theory. However the Board would not comment.
Possibility of interference with points 2182A• It has been suggested that the nuts on the left-hand end of the front stretcher bar and on the right-hand end of the rear stretcher bar may have been removed deliberately. In addition, it has been further postulated that the inboard nuts on the right-hand end of the rear stretcher had been moved deliberately to the right in order to narrow the flange-way gap on the right-hand side. No evidence has yet been established to support speculation or theories that vandalism or deliberate unauthorised interference caused the derailment. Excerpt from Progress report
|
However, the claim that ‘No evidence has yet been established to support speculation or theories that vandalism or deliberate unauthorised interference caused the derailment', was widely interpreted as dismissing the possibility of sabotage altogether. The corollary is that the changes from the condition recorded in July 2001 were the result of ‘authorized interference', in other words adjustments made during maintenance.
Even so, if the estimated time for corrosion to build up on the shiny thread – Jarvis' metallurgists suggest under 24 hours, is discounted, it is still likely that the nuts had not been off the tie bars for long.
As part of its response to Potters Bar, Railtrack examined the effect of removing the nuts on the ends of the tie bars on a set of UIC-84B points at Ashford. It was immediately apparent that the Lock Stretcher, was ‘under considerable distress' as it held the curve of the switch blade on its own.
As noted last month, the two pairs of nuts from the tie bar ends were found lying on the ballast, rather than inside the brackets where they would have fallen had they vibrated off. Once again, the HSE would not discuss this, but the theory appears to be that the nuts vibrated loose, perhaps because of the ‘missing' escapement or some other phenomenon inherent in the design.
To support this theory, the HSE adduces tests carried out on an unspecified number of lock nuts on Tie Bars on other points in the Potters Bar, Marshmoor and Hatfield station areas. According to the Report, ‘ Around 20% of the locknuts tested were considered to be not fully tight, though the root cause for this has not been determined'.
Here too, the wording of the report gave the media more ammunition. The survey was undertaken ‘using tools supplied by the maintenance contractors'. The Report added, ‘Some fastenings could not be tested with the tools available. This was a consequence of the difficulty, given the design gaps, in getting the adjustable spanners supplied by the contractor onto both the main nut and lock nut'.
This led the media to infer that Jarvis's maintenance teams didn't have the proper spanners for the job and used adjustable spanners for all tasks. Asked about this Frank Hyland of the HMRI said ‘Whether in maintenance other tools are available, I don't know'. But Jarvis tell me that the spanners were borrowed from their emergency response team, not a maintenance gang.
In any event, the HSE concedes that the tests were only qualitative, since torque spanners, which measure the force applied, were not used. An HMRI inspector put one spanner on the main nut, another on the lock nut and saw how much force was needed to undo the lock nut.
Around 20% of the locknuts tested were considered ‘not fully tight'. However, none were finger tight, so they were not ‘loose' as was widely reported. And, of course, there is a difference.
The condition of points in the Potters Bar area, especially the points which caused the accident, was appalling' Transport Secretary Alistair Darling Newsnight 4 July
|
As an avid reader of Air Accident Investigation Branch (AAIB) reports I find these tests amateurish and their publication irresponsible. As I prepared for a TV interview after the press conference, a reporter beside me live to camera was telling his viewers that 20% of points had loose bolts. You can imagine what hay bereaved and victims lawyer Louise Christian made of this?
An AAIB investigation, I believe, would have measured the torque needed to free each lock nut and noted the figure. It would then have established the minimum torque for acceptable tightness for locknuts of this size.
In the subsequent report the data would have been presented in the form ‘The minimum desirable release torque is 50 ft lb, of the 83 nuts tested, 66 were within the range 50-60 ft lb, 8 were in the range 40-50 ft lb, 6 between 30-40ft lb' and so on, with a table of the results.
Given that nearly two months had elapsed since the accident, one would also expected a quantification of what constituted an unacceptable degree of tightness. Of course, it is also a criticism of the railway that, as remarked last month, no tightening torque is specified for the locknuts.
Media outrage was further fuelled by the revelation in the report that while in the subsequent programme of checks on points across the rail network ‘no other single set of points was judged to be in the condition that pertained at Potters Bar', in one location ‘such was the condition of points that HMRI took formal enforcement action to ensure that improvements were made'.
‘One set of points was so dangerous that the HMRI has had to issue an enforcement order.
Solicitor Louise Christian Newsnight 4 July
|
This clearly implies that there was this dangerous set of points which the HMRI discovered and promptly took action. Sensing a new scandal the media asked what had been found?
‘No, I can't give you details of what the individual problems were at this time,' was the answer.
Where are the points? At ‘the south end' of the WCML. Where exactly? ‘In the Wembley-Watford area'.
So much for openness. Travellers on the WCML will have noticed that all four single-lead crossovers at North Wembley had been out of use for most of May and June. Questioned by the railway press, HMRI's Head of Operations, Bob Smallwood admitted that when his inspectors found the points they had already been locked out of use by the contractor, Carillion Rail.
Carillion maintainers had already noted a marked deterioration of these crossovers, possibly as the result of problems dating back to their installation 10 years ago. An action plan to rectify the situation had been developed and this was accepted by HMRI as part of its enforcement order.
Given all this, why was a formal Enforcement Notice issued? ‘It was precautionary action that was taken to ensure that the points do not present a risk,' Mr Smallwood explained.
On the basis of the evidence Potters Bar still seems to have been a ‘one off'. An HMRI inspection of 240 sets of points across the network showed a ‘spectrum' of differing standards on the condition of points and on maintenance arrangements including record keeping. but ‘no single set of points was judged to be in the condition that pertained at Potters Bar'. And the only points judged to merit enforcement action were already out of use and being rectified.
Clearly, uniform procedures for maintenance, including torque settings for lock nuts, are overdue. But the Report then set yet another hare running by querying the safety of the UIC-54B point design. Specifically, it questioned the safety of nut locking arrangements on the tie bars of points.
According to the report, the tightness tests indicated that there may be mechanisms that cause nuts to lose tightness. This will be an area of further technical investigation, says the report. I suspect it may find that the problem is one of variations in strength of the person doing up the lock nut, but we shall see.
However, the report then says that this illustrates the need to re-examine ‘the design and safety assessment of these items in the light of the reasonably foreseeable operating conditions within present and planned future usage of the rail network'.
This got my railway civil engineering chums fizzing. They pointed out that lock nuts are widely used in other safety critical applications, not least the steering systems of road vehicles.
One senior informed source described the UIC 54B points as an ‘utterly successful design'. It has been in service on high speed lines since the 1980 with no record of problems with tie bar lock nuts.
Yet there was a precedent for Potters Bar, pointed out by an erudite reader. T he Kingham derailment in 1966 was caused by a permanent way gang removing the nuts and lock-nuts from tie bars on a set of points in readiness for their planned removal.
But before they could be removed, the switch rail moved under a train in a carbon copy of the Potters Bar derailment, even to the last coach derailing and mounting the platform. With some feeling a senior engineer remarked , ‘are they saying that the system has to be safe when it is dismantled'?
Printbite
In addition, the use of the present nut locking arrangements is questioned given the operating environment. Progress report |
Anyway, the HSE Investigation Board recommends that Railtrack should, within a month of publication, review the design and safety analysis of points with adjustable tie bars ‘to identify any immediate modifications (such as changing the locking nut arrangements) to better ensure that the component can fulfill its safety functional requirements.
This review should take account of the ‘present and reasonably foreseeable future operating environment, the achievement of safety by design, the practicability of testing, maintenance, setting and adjusting, and any other relevant factors'. Railtrack should then implement any improvements identified by such a review to ensure that good engineering standards for such safety critical items are achieved and can be maintained.
Longer term, Railtrack is recommended to undertake a more in-depth review of the design and safety analysis of points with adjustable tie bars, ‘including consideration of replacing them with a points system that is more inherently safe by design. Any changes should then be implemented ‘as required to further rail safety'.
Note, there is no mention of ALARP.
It seems to me that the issue of lock nut security is easily resolved. First maintenance teams are given torque spanners with a fixed setting sufficient to ensure tightness. Second, after the lock nut has been tightened a simple visual indicator can be applied. You could draw a thin line of paint across both nuts or make center punch marks. Any subsequent movement of the locknut could then be inspected visually.
As for what caused the accident at Potters Bar, the suspicion is growing that we may never find out how the nuts came off.