Derailment at East Somerset Jct (20/03) [RAIB report now released]

[Sunset route]

If the signaller set another route as the train was passing over them it would not move the points because of the track circuits and interlocks would prevent it!

I can vouch for that on many many occasions, "it's called testing the interlocking" lol

 

[mr_moo]

Some basic info from what I know.

The isn't the same cause as Lewisham. I know the 'immediate cause' (to use RAIB language) for Lewisham (although why that happened is still up for debate) and it's different to here.

I also know a bit about this one. The train was exiting the yard, not entering, so the people saying they can see the points moved under the train need to be clear about which set of points they are referring to. The main set of points you can see in the photos are the ones where the exit to the sidings join the main line. The derailment occurred earlier than that - the set was already running de-railed before it reached those points. The points that are more in question are the traps out of the yard, and I've heard (but am not sure of this) that they were plain-lined, so these did not move under the train either! Track condition is a primary focus.

Sorry to be a bit vague and not quote sources etc but I'm trying to give out what info I can whilst also being well aware that investigations into both incidents are still very much ongoing, so I need to be careful how much info I release.

 

[GB]

Can you clarify what you mean by exiting the yard? From what I understand the train was exiting the quarry branch line.

 

[mr_moo]

sorry, that is what I meant! the branch, not the yard. I'm tired...

 

[GB]

Ha ha

 

[MarkyT]

No not really!
"The points moved" means they moved of their own accord
"The points were moved" means something (like a derailed axle) forced the points to move

Maybe a chain or similar dragging under a wagon that got caught in the mechanism and forced the switch open.

A derailed axle would be diverted from the signalled route when encountering a correctly secured facing point anyway as one wheel will be outside the 4 foot and could follow the outside of the diverging rail pulling following vehicles with it. A facing point leading to a goods loop is what magnified the consequences of the high speed Ufton Nervet derailment.

If the signaller set another route as the train was passing over them it would not move the points because of the track circuits and interlocks would prevent it!

Completely agree opening of the trap switch by motoring of the point machine is extremely unlikely as there would had to have been a conscious attempt by the signaller to initiate such a movement, as well as a complete failure of the interlocking to prevent it. Track circuits can sometimes go falsely clear under short lightweight trains like pacers, resulting in early release of route locking, but this is extremely unlikely here as the train is long and heavy. Even if that happened, the signaller would have had to be simultaneously trying to set a conflicting route or attempting to move the points using the individual point switch.

 

[Signal Head]

Some basic info from what I know.

The train was exiting the yard, not entering, so the people saying they can see the points moved under the train need to be clear about which set of points they are referring to. The main set of points you can see in the photos are the ones where the exit to the sidings join the main line. The derailment occurred earlier than that - the set was already running de-railed before it reached those points. The points that are more in question are the traps out of the yard, and I've heard (but am not sure of this) that they were plain-lined, so these did not move under the train either!

I don't see how the traps could be plain-lined, not without losing use of the connection anyway, you can't just effectively remove a set of traps as they're a safety requirement.

However, in one of the official NR photographs, it does look like the set 'behind' the traps (in direction of travel) has had its switches plain lined. This is a connection into two sidings, and was trailing for this train.

 

[Signal Head]

I'm not sure how well this will come out given the size restrictions for attachments, but here's the relevant portion of the local control panel in Witham RR. The points which appear to be plain lined are 945.

 

[MarkyT]

I'm not sure how well this will come out given the size restrictions for attachments, but here's the relevant portion of the local control panel in Witham RR. The points which appear to be plain lined are 945.

The siding points then, trailing with respect to the derailed train. I don't recall ever seeing anything in those sidings.

 

[edwin_m]

It is debatable whether traps are needed for the convergence of a freight-only line with a passenger line but they are definitely needed to protect a passenger line from rogue movements originating in sidings. So if the sidings are being recovered it may be that the traps can be (and have been) plain-lined too.

 

[GW43125]

It is debatable whether traps are needed for the convergence of a freight-only line with a passenger line but they are definitely needed to protect a passenger line from rogue movements originating in sidings. So if the sidings are being recovered it may be that the traps can be (and have been) plain-lined too.

Surely you'd still need them off a freight line as a 1500 tonne stone train sliding at 10mph is going to do a LOT of damage to a 150.

 

[GB]

Surely you'd still need them off a freight line as a 1500 tonne stone train sliding at 10mph is going to do a LOT of damage to a 150.

I suspect that as there is at least a mile from the protecting signal (275) to the point of conflict on the up road trap points are not deemed a requirement for that purpose.

 

[furnessvale]

Surely you'd still need them off a freight line as a 1500 tonne stone train sliding at 10mph is going to do a LOT of damage to a 150.

If the freight line is properly signalled, it is no different to that same 1500 tonnes train being out on a passenger line which converges on another at a typical junction.

 

[MarkyT]

If the freight line is properly signalled, it is no different to that same 1500 tonnes train being out on a passenger line which converges on another at a typical junction.

Having been out of the loop professionally for a while I no longer know where to look for such familiar principles as the old 'blue book' requirements for trapping in layout design, yet I'm also seeing modern signalling schemes such as Banbury that still look very familiar and appear to follow traditional layout and signalling rules, including retaining conventional UK style trapping. So I believe NR still requires traps as a company rule for where freight lines and sidings join passenger lines to adequately control safety at such interfaces. At East Somerset Jn, the single trap accomplishes this nicely for both the local sidings and the branch to both Merehead and Cranmore. Without the trap there, no means of protecting the passenger main line from any potentially uncontrolled movement coming out of Merehead quarry sidings would exist, as I can't see any controlled traps there visible on Google Earth, nor are any shown on the Simsig numbering plan:

https://www.simsig.co.uk/Media/Wiki//usertrack/mans/westbury-signal-number-plan.pdf

I could understand no trapping at East Somerset Jn if the sidings were removed there and the Cranmore branch was reclassified as a 'passenger' line, but the siding connections onto the single line at Merehead would then require safely controlled trapping protection, actually very easy to provide as a second 'end' to the existing siding turnout. The single line looks like it is otherwise fully TCB signalled from Westbury with overlaps, so could probably be modified and reclassified fairly easily as 'passenger/general use'.

 

[alxndr]

I suspect that as there is at least a mile from the protecting signal (275) to the point of conflict on the up road trap points are not deemed a requirement for that purpose.

Is it? From what I can ascertain, W275 is only just over a third of a mile from the point that the branch meets the Up Westbury.

 

[Signal Head]

I think the Blue Book requirements for trapping off freight-only lines have a long-standing relaxation that they can be dispensed with if 'other means' of protection are provided, which I take as a reference to overlaps. I can think of several freight-only (though probably 'passed' for passenger traffic) lines which join bona-fide passenger lines without traps. Equally, there are examples where traps are provided, so provision is inconsistent.

I think it unlikely that NR would have carried out work to recover traps in this instance just because the sidings are out of use, there's no point spending money to make the railway (theoretically) less safe.

 

[AndyW33]

According to National Rail Enquiries and Network Rail, the main line has now reopened, initially with a speed restriction.

 

[GB]

Is it? From what I can ascertain, W275 is only just over a third of a mile from the point that the branch meets the Up Westbury.

Edit: sorry I think your right. The resoultion of my map makes the double track section look a lot longer than it actually is!

 

[70014IronDuke]

According to National Rail Enquiries and Network Rail, the main line has now reopened, initially with a speed restriction.

24 hours ahead of the original estimate? Well done that team!

The overall result, however, was rather spoiled on the main line early in the day when it seems a person was hit at or near Newton Abbot.

 

[GB]

24 hours ahead of the original estimate? Well done that team!

The overall result, however, was rather spoiled on the main line early in the day when it seems a person was hit at or near Newton Abbot.

2200 Friday 24th was always the deadline they were working to, though I don't know what deadline they had published publicly.

Either way the site was a mess and they did a good job.

 

[MarkyT]

I think the Blue Book requirements for trapping off freight-only lines have a long-standing relaxation that they can be dispensed with if 'other means' of protection are provided, which I take as a reference to overlaps. I can think of several freight-only (though probably 'passed' for passenger traffic) lines which join bona-fide passenger lines without traps. Equally, there are examples where traps are provided, so provision is inconsistent.

I remember years ago there was a policy change to incorporate passenger line signalling standards like overlaps and facing point locks on all new and resignalled freight lines by default. Part of the rationale was that no special measures would then be required for passenger diversions over such freight lines. If you wanted to deviate from this policy in any particular case for some good practical reason you had to fully risk assess and justify the decision. Older mechanical and relay installations omitted some of the features legitimately according to the rules of the day, and no retrospective change was required for them while they survive. The current Merehead/Cranmore branch looks like a hybrid, perhaps indicative of the signalling's age, part of the mid 1980s West of England Resignalling scheme. While the line appears to incorporate overlaps according to the Simsig numbering plan, the siding connection at Merehead does not have a trapping function to protect the single line, something it would need if the branch was nominally 'passenger'.

I think it unlikely that NR would have carried out work to recover traps in this instance just because the sidings are out of use, there's no point spending money to make the railway (theoretically) less safe.

First NR would have to decide whether the sidings were ever likely to return to use as the traps would still be essential if they were reinstated. If not, a permanent interlocking and panel change might be planned to recover the siding connections that could also remove the trap. As part of that project, overrun risk assessments for all signals affected by the change and concerned in potential junction conflicts would need to be carried out. It is my opinion that a trap or other physical restraint also should be provided at Merehead sidings outlet in order that any unauthorised movement in the non ROGS controlled sidings could never get out onto the general main line network, newly defined insofar as trapping is concerned to include the Cranmore branch.

 

[70014IronDuke]

2200 Friday 24th was always the deadline they were working to, though I don't know what deadline they had published publicly.

Either way the site was a mess and they did a good job.

Oh? Apologies - I thought I'd read it was until Saturday night.

 

[PHILIPE]

Oh? Apologies - I thought I'd read it was until Saturday night.

It was originally, but obviously progressed quicker than first estimated

 

[PHILIPE]

Oh? Apologies - I thought I'd read it was until Saturday night.

That was quoted at first but obviously quicker than first estimated.

 

[GB]

Wednesday afternoon (22nd) was the first internal declaration of a 2200 Friday hand back that I am aware of. Like I say, I don't know what was communicated to the public or whether they added a day just to be safe...

 

[Johncleesefan]

Yep the job is up and running again ay east Somerset junc with linespeed on the down Westbury and a 50 tsr on the up at 121 mp (where the derailment happened)

 

[Signal Head]

OTT was showing the connection to the branch with "SHUT" in the berth earlier today. I'm wondering if they have simply reinstated the Up Main for now, as plain line. Even without this connection there is still access to the branch through the loop.

 

[Signal Head]

I remember years ago there was a policy change to incorporate passenger line signalling standards like overlaps and facing point locks on all new and resignalled freight lines by default. Part of the rationale was that no special measures would then be required for passenger diversions over such freight lines.

Yes, I think I recall that being implemented.

First NR would have to decide whether the sidings were ever likely to return to use as the traps would still be essential if they were reinstated. If not, a permanent interlocking and panel change might be planned to recover the siding connections.. .

I am reasonably familiar with the process.

 

[notverydeep]

The RAIB have announced their intention to conduct an investigation (rather than a safety digest) https://www.gov.uk/government/news/freight-train-derailment-east-somerset-junction. The announcement gives the speed of the train (20 mph) without comment, suggesting that this was within the line speed. They also give the point of derailment as close to the location of a previously removed trailing point. This would seem to indicate that they do not believe that points moved under the train.

Picked out for a mention are: "the condition of the track, its geometry and how it was maintained, how the wagons were loaded, the condition of the wagons and any relevant underlying management factors"...

 

[cossie4i]

The RAIB have announced their intention to conduct an investigation (rather than a safety digest) https://www.gov.uk/government/news/freight-train-derailment-east-somerset-junction. The announcement gives the speed of the train (20 mph) without comment, suggesting that this was within the line speed. They also give the point of derailment as close to the location of a previously removed trailing point. This would seem to indicate that they do not believe that points moved under the train.

Picked out for a mention are: "the condition of the track, its geometry and how it was maintained, how the wagons were loaded, the condition of the wagons and any relevant underlying management factors"...

The rear of the train was on a TSR of 20mph. The line speed over the junction is 25mph.