Severn Tunnel Collision 1991

The month of December brings the anniversary of a number of major railway accidents, notably Lewisham (4.12.57), Clapham (12.12.88) and Castlecary (10.12.37) plus equally tragic events at Elliot Junction, Hawes Junction and Ealing.

The official reports into these events are highly detailed and, in the main, the Inspecting Officer was able to identify the cause with little room for doubt.

The report on a further accident, that in the Severn Tunnel on 7th December 1991, is different. It concludes that the cause was an ‘unaccountable’ error either on the part of the driver; or by the S&T technicians working in the relay room to fix a signalling fault.

These alternative explanations are radically different. Having read the report several times, I find it difficult to form a view as to which explanation is the more likely.

The report concentrates on the errors which came to light in the circumstances leading to the accident (a failure of remote control equipment) and on the emergency response. The actual cause of the rear end collision, however, remains undetermined.

The report can be seen at http://www.railwaysarchive.co.uk/documents/HSE_Severn Tunnel1991.pdf. It is lengthy and very detailed, running to 39 pages and 308 paragraphs, plus several appendixes.

It does seem that several things may have happened to combine to give the driver of the train a brief green signal when it should have been red.

But the HM Chief Inspector could not nail down the exact reason, and quite correctly did not therefore pick on any one item as the definitive cause of this accident.

I think Adrian Vaughan's book on accidents which has a chapter on this gets close to what happened. The official report seems to have given up when the Sprinter driver's lawyer said the driver was not going to give any evidence.

Apart from the actual collision, the operation of the line both before and particularly after the collision was a shambles. It wasn't the finest of times for the Inspectorate either.

Taunton said:

I think Adrian Vaughan's book on accidents which has a chapter on this gets close to what happened. The official report seems to have given up when the Sprinter driver's lawyer said the driver was not going to give any evidence.

Apart from the actual collision, the operation of the line both before and particularly after the collision was a shambles. It wasn't the finest of times for the Inspectorate either.

Click to expand...

For those of us without the book would it be possible (without breaching copyright) to summarise the conclusion he reached?

PG said:

For those of us without the book would it be possible (without breaching copyright) to summarise the conclusion he reached?

Click to expand...

That the technicians working on the failed axle counter controls at the moment the Sprinter approached the tunnel likely caused a false green.

Taunton said:

That the technicians working on the failed axle counter controls at the moment the Sprinter approached the tunnel likely caused a false green.

Click to expand...

Thank you, appreciated.

With respect to ‘Taunton’, that is not what Adrian Vaughan says (in his book, ‘Tracks to Disaster’).
His concluding paragraphs include the following:

‘Could experienced technicians have forgotten to remove the vital fuse from the WDH circuit before they started work? It seems very unlikely, but it is not impossible.’

‘Could a driver slow down to almost a stand at a red light and then drive past it? That seems impossible, and yet we know that at Newton, Hyde and Belgrave a driver . . . facing a red light had moved away . . .to pass the signal at ‘Danger’.’

and

‘We shall never know for certain what caused the first ever collision in the Severn Tunnel.”

Adrian Vaughan and the Official Report describe the shambles of the rescue operation but the fact remains, as John Webb has pointed out, the Inspector could not identify any one item as the definitive cause of the collision.

I think Vaughan is being generous referring to "experienced technicians" when his preceding pages describe how there was a substantial lack of knowledge about the (new) axle counters, along with an extraordinarily excessive number of failures the system had achieved. This accounted for a whole crew (eight, says the report) working on the control unit at the moment the train was approaching the signal.

The (old) GWR knew just how to run the tunnel safely, there was a signalbox at each end with independent box-to-box circuits and all sorts of procedures to manage any failures. The tunnel rescue train was always provided, it was located at Severn Tunnel Junction loco shed and could be dispatched in minutes, not that I think it was ever required. Putting the rescue train down at the end of the obscure long siding to Sudbrook tunnel pumping station from where (as the book describes) it took for ever to get out and reverse into the tunnel, was just a nonsense, as were the arrangements for hand-signalling, which were just given up shortly before the accident.

Taunton said:

I think Vaughan is being generous referring to "experienced technicians" when his preceding pages describe how there was a substantial lack of knowledge about the (new) axle counters, along with an extraordinarily excessive number of failures the system had achieved. This accounted for a whole crew (eight, says the report) working on the control unit at the moment the train was approaching the signal.

The (old) GWR knew just how to run the tunnel safely, there was a signalbox at each end with independent box-to-box circuits and all sorts of procedures to manage any failures. The tunnel rescue train was always provided, it was located at Severn Tunnel Junction loco shed and could be dispatched in minutes, not that I think it was ever required. Putting the rescue train down at the end of the obscure long siding to Sudbrook tunnel pumping station from where (as the book describes) it took for ever to get out and reverse into the tunnel, was just a nonsense, as were the arrangements for hand-signalling, which were just given up shortly before the accident.

Click to expand...

At the end of the day it’s one of those accidents where we will never know the cause.

For me either scenario is possible.

It’s quite possible for a driver to approach a red signal and then drive through it - it happened at Ladbroke Grove, and a signal often displaying an adverse aspect (as the signal protecting the Severn Tunnel was known to be due to the long section) is ripe for humans to let their guard down by normalising the situation. It’s a bit of a coincidence that work was taking place in the relay room at the time, but coincidences do happen. I agree with the general feeling that it’s unlikely, but unlikely doesn’t equal impossible.

There again the fact that technicians were working on the axle counter is naturally suspicious, especially the fact that the reset was done more-or-less exactly at the time in question - although of course things depend on how quickly that 8 count came in, if immediate really was immediate then it couldn’t have been the technicians but if immediate was a minute or two then it perhaps swings the other way.

I find it very hard to make a judgement call either way, the whole thing remains quite a matter of intrigue. Nowadays data recorders on both train and signalling would have given the answer, and there would of course have been the safeguard of TPWS against a SPAD. The 1990s had a string of accidents for which slapdash safety management was the root cause, and that’s perhaps the biggest thing to take away from the report.

bramling said:

There again the fact that technicians were working on the axle counter is naturally suspicious, especially the fact that the reset was done more-or-less exactly at the time in question - although of course things depend on how quickly that 8 count came in, if immediate really was immediate then it couldn’t have been the technicians but if immediate was a minute or two then it perhaps swings the other way.

Click to expand...

I think it could have been well under a minute if the train was crawling up to the signal and the driver applied power straight away on seeing it clear to green. That would probably have qualified for the description "immediate" considering that the person resetting had to check the lights on the front panel and have a brief conversation with someone else before noting that a count was coming in. Especially as the shock of what came after may have distorted the memory of what happened in the minutes before.

If I am not mistaken the old East and West tunnel signalboxes also had detonator placers as well, linked to the starting signal into the tunnel. The old GW really did put all the precautions in. It was probably before there were risk assessments that someone just decided to do away with those.

To my embarrassment I'd never heard of this incident before or, if I had, had forgotten.

The report would make good reading for anyone who might be misty eyed for a return to the glory days of BR. It's the sort of story you would expect to read relating to the worst days of Railtrack six or seven years later.

Aside from the discussion around the direct causes of the collision, the main question in my mind is by what set of factors did standards of competence (of technical maintenance and repairs) and preparedness (for rescue) deteriorate to the point they had in the Severn Tunnel in 1991?

Was it pre privatisation cost cutting, or confusion created by recent organisational change within BR, or both? Or was it just local complacency? Would the same degree of dysfunction have been found in the Western Region 10, 20 or 30 years beforehand?

They're relevant questions for whenever anyone considers cost cutting or making changes in the future.

47271 said:

To my embarrassment I'd never heard of this incident before or, if I had, had forgotten.

The report would make good reading for anyone who might be misty eyed for a return to the glory days of BR. It's the sort of story you would expect to read relating to the worst days of Railtrack six or seven years later.

Aside from the discussion around the direct causes of the collision, the main question in my mind is by what set of factors did standards of competence (of technical maintenance and repairs) and preparedness (for rescue) deteriorate to the point they had in the Severn Tunnel in 1991?

Was it pre privatisation cost cutting, or confusion created by recent organisational change within BR, or both? Or was it just local complacency? Would the same degree of dysfunction have been found in the Western Region 10, 20 or 30 years beforehand?

They're relevant questions for whenever anyone considers cost cutting or making changes in the future.

Click to expand...

Not the only accident around that time where dubious safety management was a factor. A read of the background into the Cowden and Newton accidents is similarly depressing reading, both accidents having a depressing inevitability about them. Things were to get even worse ultimately leading to Southall and Ladbroke Grove, and the Watford accident had quite a story behind it too.