Return to Archive -by date - by topic - 2001 Archive.

INFORMED SOURCES March 2001

Hatfield Inquiry – HSE inexperience shows

There was a worrying lack of railway experience when the Health & Safety Executive unveiled it's second interim report into the Hatfield derailment

For some time now, there has been a debate going on between what might be called the traditionalist and modernist view of Her Majesty's Railway Inspectorate. The traditionalists hark back to the days before the transfer of HMRI to the Health & Safety Executive. Railway inspectors then were often retired army officers with railway experience who had the authority and respect to make things happen without the force of legislation.

Modernists claim that a)the source of old school Inspecting Officers has dried up, b)that Southall and Ladbroke Grove showed that the HMRI was unable to deal with the privatised railway and c)that the railway today comes under the Health & Safety at Work Act, with prosecutions and fines, which requires different specialist skills.

There was also the sub-text that the inspectors were a bit, how shall we say, ‘collegiate' and not really the sort of person you wanted as the shock troops of the new model safety regime.

So, at the press conference to announce the publication of the second interim report into the Hatfield derailment the top table was solid HSE, except for the Chief Inspecting Officer Vic Coleman who, I later learned, is not involved in the Hatfield Inquiry.

And did it show.

No gricers

Take Table 1 (reproduced here), which, at the trailing end seemed unable to differentiate between a Class 91 locomotive and a Driving Van Trailer. Worse, this basic error was repeated in the separate table which gave the condition of each vehicle in the train – ‘vehicle' also applying to locomotives.

Passing over the spurious Class 91 number, the painted numbers on the Mk 4 coaches are in the correct series for the standard IC225 formation of six Standard Class, a service vehicle and two First Class. However, the columns of ‘Vehicles Types' and ‘Seats' contain pure gibberish.

When I queried this subsequently, the HSE admitted to the ‘trailing Class 91' error but said that Table 1 ‘showed the vehicles and their designated numbers as stated on the identification labels attached to the vehicles'.

Now even if owner HSBC Rail and operator GNER had incorrectly labelled the vehicles in this IC225 rake, there is still no excuse for repeating the errors parrot fashion. The obvious response would have been to run two further columns with the correct information and give whoever was responsible for the mis-labelling a slap on the wrist.

Does it really matter? From a safety viewpoint, certainly, if, for example, you are the emergency services attending a future accident. The knowing the correct number of seats inside a vehicle on its side in darkness and pouring rain could be vital in prioritising rescue activities.

Dates unclear

Then, we all wanted to know when the track had been inspected prior to the accident. It had been ‘walked' the week before (visual inspection) came the answer. ‘What about ultrasonic inspection' I asked? ‘We haven't got the dates of inspections with us' replied Sandra Caldwell, the HSE director who is Chairing the Board overseeing the investigation.

Ms Caldwell is the Director of the HSE's Health Directorate and a former member of the Channel Tunnel Safety Authority. The old school HMRI view is that she was ‘imposed' on the investigation.

So I tried Frank Hyland, the HSE's ‘Team Investigation Manager' for Hatfield. He didn't know either. ‘I don't keep it in my head'.

Afterwards I tried Chris Willby another investigation Board Member. First of all he said that the trouble with ultrasonic tests was that you could only carry out the test where the surface of the rail was clear. No that wasn't the question, because ‘unreadable rail' should generate a report. Did the man use his ‘hockey stick' on the rail and fail to get a reading? ‘I don't have the dates, I don't carry the dates in my head' Willby replied.

At which point one can only shout ‘oh come on'!

When I write an article for MR I become the expert's expert on that subject for the two or three days of research and writing. The HSE team has been living and breathing Hatfield since mid October. The most vital detail is when the section of track that failed was inspected and, above all, ultrasonically inspected. And they haven't got the dates at their finger tips?

Well, probably, but the HSE view is that releasing this information might influence the statements of those giving evidence; or to put it another way, memory might be adjusted to suit the known dates. So why the mass amnesia when ‘sorry, can't reveal that' would have sufficed.

Anyway, former Railtrack Chief Executive Gerald Corbett gave the dates in evidence to the Parliamentary Transport sub-committee after the accident. He said that the track was ultrasonically tested in June or July 2000 – he said June first then changed to July. What the HSE did confirm that Hatfield was on three monthly ultrasonic inspection because of condition due to gauge corner cracking. This means that it would have been tested previously in March or April.

Track circuit confusion

After the press conference, with Richard Hope, the doyen of the railway writing community, we quizzed Chris Willby, on a detail of some importance. We wanted to know was whether the rail that failed was the track circuited rail.

Now the beauty of track circuits is that if the rail fractures, the electricity can't get through, the relay drops, the interlocking thinks the broken rail is a train in section and turns the signal in rear to red. But on electrified lines, one rail has to carry the traction current to earth, which means it has to be electrically conductive along its length and, perhaps, even cross bonded to other track.

As a result, only one rail is track circuited and the other rail is termed the traction rail. We understood that the high rail on the curve at Hatfield – and, oddly enough, the second interim report doesn't actually say which rail failed, was the traction rail.

Equally, there are sources, I can't comment on their informedness, which suggest that the rail broke before the derailment and several trains passed over it before the second fracture which caused the derailment. And said sources suggest that the break may have shown up on the maintenance data tapes in the signal box while not shorting out the track circuit.

Mr Willby confirmed that the HSE was studying evidence that the track may have had a break in it before the train ‘capsized'. No conclusion had been reached at the time of the Report (23 January) because signalling data tapes were still being analysed.

Marks on the leading wheels suggest that the locomotive hit the broken rail. But, complicating the investigation is the fact that the signalling indications are ‘not necessarily at the point where the break occurred', according to Willby.

So far so good, But then we got onto whether the track circuit would have detected the broken rail. ‘That would be expected' said Willby, adding ‘but if it had broken and the ends were still giving intermittent contact then you would come up with a situation that is much more difficult to analyse'.

‘But was the high rail track circuited?' I asked. ‘That is my understanding'. ‘So', said Richard, ‘had the rail parted the signal would have gone to red'.

‘We would have expected (that), had there been a clear indication from the parting'. ‘But', Richard pointed out, ‘the rail would have parted because it was in tension'.

‘I don't see how there can be any ambiguity, because what I'm saying is that we don't have an indication of a clear break so that I don't have the indication at the moment that there was a clear break which would have been strong enough to send a signal', Willby replied

And there it petered out in a non-meeting of minds. Anyway, I asked the unfailingly helpful HSE press office whether the high rail was the track circuited rail or the traction rail.

Back came the prompt reply, it was the traction rail. And my signalling chums tell me that more often than not in the event of a broken traction rail the track circuit current will find its way round the break, depending on the earth bonding. So the question now is what, if anything, the data tapes picked up.

Continues.........