Return to Archive -by date - by topic.

INFORMED SOURCES June 2005

Potters Bar – Formal Inquiry equally baffled

Four years on and the root cause of the Potters Bar derailment remains a mystery

To judge by the national media coverage of the report of the Formal Inquiry into the Potters Bar derailment, published by the Rail Safety & Standards Board 12 April, the failure of points 2182A was caused by an omission of formal procedures for tightening nuts on stretcher bars when the design was introduced by British Rail in the 1980s. So that's alright then.

Well, no. As Chief Executive Len Porter, admitted in the RSSB press release, ‘Unfortunately the report does not provide us with conclusive evidence of any one cause'. He added, ‘What it does point to is a combination of mechanical and human issues plus a lack of proper maintenance procedure'.

So we still don't know how this one set of points came to be in what was a uniquely dangerous condition. And it still seems clear to me, from examination of the evidence and talking to Informed Sources, that the state of the points immediately before the accident cannot be fully explained by sloppy maintenance or inadequately tightened nuts spontaneously vibrating loose in normal service.

Condition

As the HSE Report of May 2003 made clear, tests (admittedly pretty unscientific) on nine other sets of points at Potters Bar and nearby immediately after the accident revealed ‘no instances of nuts missing or gaps between main nuts and lock nuts'. Of 23 stretcher bars tested 19 were considered ‘tight' and the remainder ‘could be rotated after some initial resistance. In three cases the bolts attaching the stretcher bar brackets to the switch rails were found to be loose.

So, not the best maintained group of points on the network, a state of affairs to which maintenance procedures based on custom and practice undoubtedly contributed. But, equally, not posing the same risk as 2182A

Certainly, since Network Rail brought track maintenance in house, examples of inadequate staff training and equipment have been exposed. In the case of the much maligned HPSS points, for example, Network Rail found that diagnostics lap top computers had not been issued to frontline staff nor had staff training followed the Network Rail syllabus. And as I write Mowlem has just been suspended from track renewal work.

But whatever the general standard of maintenance in the Potters Bar area, it does not explain the condition of 2182A. It is not that they were the worst maintained: they were simply wrong.

Front stretcher

Photo 1, taken the day after the accident, shows the left hand end of the front permanent way stretcher, together with the bracket to which it should be clamped. Two things are apparent.

First, by the time you have allowed for the thickness of the bracket, there is barely going to be enough thread protruding on the other side to take the main nut, let alone a locking nut. Photo 2 shows the left hand end of the rear p-way stretcher correctly assembled for comparison. Second, the contamination on the thread to the left of the nuts suggests that they have been in this position for some time.

But that is not all that was wrong with the front stretcher. To have insufficient thread at one end, the stretcher must have been transposed to the right. At the right hand end (Photo 3), the nuts were up against the shoulder where the threaded section ends and the square stretcher begins. The stretcher had been displaced as far as it would go.

According to the Formal Inquiry report none of the nuts remaining on the front stretcher was tight. The Panel concludes that ‘vibration and/or movement of the points in service could have caused the nuts to migrate to the left from a symmetric position and/or the bar itself to rotate to some degree'.

Does that sound likely? The report continues ‘however, the Panel considers it extremely unlikely that all six nuts would have moved relative to the bar by the same amount (about 80mm), such that they remained correctly positioned relative to each other and to their respective support bracket'.

From the evidence available, the panel was unable to reach a ‘clear conclusion' as to how the front stretcher bar came to be in its as-found state. ‘There are a number of aspects of the as-found state of the assembly for which, if they were the result of in-service deterioration, the panel either has no explanation, or one appears very unlikely'.

‘However, in considering whether human intervention could have left the bar in its as-found state, there are a number of issues that the panel has been unable to resolve, such as why an intervention might have taken place, when and by whom. The possibility of a malicious act cannot be ruled out, although the panel considers this to be highly unlikely'.

Exactly. In other words, we can't see how the stretcher moved to the right under in-service conditions, but we can't get our brain round why someone would have wanted to do it either.

Rear stretcher

There is a similar situation at the right hand end of the rear stretcher (photo 4) (pic 9/11). Once again, the inner nuts are close to end of the threaded section. Once again there is barely enough thread left at the outer end for a main nut, let alone a lock nut.

But here we may have a smoking gun. Immediately inboard of the two nuts there is a section of what the HSE called the ‘distorted thread' and I think of as the ‘shiny thread', with a light coating of rust.

This could have been caused by the nuts being screwed outwards, from their previously correct position, without being unlocked. This would have pushed the switch rail towards the stock rail, closing up the flangeway gap.

Scaling off the photo, I estimate that the length of the ‘shiny thread' is around 30mm. According to the HSE progress report of May 2003, the minimum flangeway gap should be 50mm. When 2182A was reassembled the minimum clearance, just past the front stretcher, was 30mm.

But note also, that when HSE removed the outer nuts from the rear stretcher after 2182A had been reassembled with new components, the switch rail bowed outwards anyway as stored energy was released.

With the points set to normal trains were coming through at up to 100mile/h. If the wheel flanges had to force their way through the restricted flangeway, the stress on the lock stretcher would have varied and at high frequency – ideal fatigue conditions.

That would have accelerated fatigue failure and explains the two periods of crack growth observed by HSE. But the Health & Safety Laboratory reported that there were no evidence marks on the back flanges of the right hand wheels of the first two coaches of the derailed Class 365.

Attention

Analysis of the condition of the rear stretcher is complicated by the fact that it received attention on 1 May. According to the Formal Inquiry report, t he Panel is ‘not able to reach a definitive conclusion as to the state of the rear stretcher bar assembly immediately after the technician's intervention on 1 May 2002 '. The inference is that the technician's evidence of what he found and the condition in which the stretcher was left is unreliable or, at least ‘inconsistent with a number of as-found' features.

For a start there were deposits in the threads at the extreme right hand end of the stretcher (Photo 5) [James Figure 13 in HMRI second progress report] and also on the nuts found on the ballast. And the elastomer washer of the inner bush was missing.

Note that the length of clean thread at the end is about the thickness of a single nut. And that the compacted grease further inboard suggests that a nut has not been along that thread for some time.

Understandably, the Panel is unable to explain how, if the stretcher was left with both the outer nuts on, the deposits on the threads and faces of the nuts could have appeared so rapidly after the nuts vibrated off.

Intervention

All this seems to confirm that the failure was due to what the Formal Inquiry calls an ‘intervention' rather than in service deterioration due to incorrect maintenance. But, if 2182A was in such a state why was this not noticed?

Remember that according to the HSE report a joint permanent way/signalling maintenance visit in March ‘appears to have taken place without permanent way people being present'. Might a signalling technician have failed to spot points grotesquely misadjusted mechanically and with only a single nut at one end of the two p-way stretchers?

Such maintenance generally involves cranking the points across manually and checking the gap at the toe when the movement is detected. Surely the distortion at the rear stretcher would have been noticed?

Later that month Jarvis and Network Rail made a one-off inspection of the points as a local initiative. No signalling faults or defects were noted.

Then, on 1 May, after the planned maintenance visit, the technician noted only that ‘2182A points needs a new rubber bush on the back lock stretcher. Existing bush perished'.

And finally, the points were inspected by a patrolman the day before the derailment.

Box 1

Box 2

Which brings us back to how 2182A got, or were put, into their final condition. RSSB's, secrecy means that we don't have the Formal Inquiry's views on the validity of the tests by Jarvis and Network Rail, reported here in July 2003. These put the spare set of points intended to replace 2182A, into the as found state.

A video shows this taking around 25 minutes. To maintain detection, the front stretcher had to be adjusted as the nuts on the rear stretcher were wound out. HSE claimed the changes could be made in six minutes

And the as-found condition implies that two stretchers on 2182A were left without space for a lock-nut at one end, when the other stretchers in the area all had two.

So, still no further forward. And the final report of the HSE investigation will not be published until a decision has been taken on whether there can be a prosecution for manslaughter.

Currently, the papers relating to the BTP investigation into the derailment are with the Crown Prosecution Service. According to Informed Sources, the CPS is unlikely to make a decision until the outcome of the Hatfield derailment prosecutions becomes clearer.