Met problems, night of 18th December

Loss of signal power 18.55, signals remain at danger both main and local lines.
Suspected failure in Northwood substation at 400v circuit.
Much investigation work to find the cause.
Shuttle service implemented Northwood-Amersham/Chesham/Watford 19.35
Limited non-advertised trains running under failure conditions through area.
22.00 3 train shuttle service in operation (461 Amersham-Watford; 465 Rickmansworth-Chesham; 435 Amersham-Northwood)
Fuses changed 23.30 and power restored.
Service remained advertised suspended Northwood-Harrow until close of day.

One stalled MET on approach to Moor Park NB, arrived Moor Park 19.50.
One stalled Chiltern on approach to Harrow SB, arrived Harrow 19.48.
One stalled Chiltern on approach to Pinner NB.

There's a number of reasons for pushing a train through a multiple signal failure. Whilst not an exhaustive list, some that spring to mind are:

A) Could have been a Rickmansworth based driver so was facilitating getting that driver back to their home location.
B) It may have been used as a "test train" to see if a fix had worked.
C) Could have been used to drop off technical staff or run technical staff through the area to see if they can spot an obvious reason for the failure.

As for why not run more than one train through.......well it depends on the number of signals that have failed for starters. The signaller has to speak to each driver at each signal to authorise them forward - this causes a high workload which aside from creating a higher risk of a mistake being made, takes time to do for each train and then each train when it is moving is doing so at less than 10mph for at least two minutes at a time - often longer.

Then there's the nature of the failure - on this occasion the whole signal main was lost - the signaller can't even see where the trains are, so from a safety/risk management point of view, its not desirable to pump trains into an area you can't see or maintain full control of. And if you did push enough trains through the failure area, how long will it be before they return back to the part of the railway that's operating normally? It's about balancing the need of the whole line rather than allowing the failed part to overly affect the non failed part.



A, B or C above could equally apply to this train (obviously wouldn't be a Rickmansworth based driver but could still be facilitating getting them to wherever they are based). It may also have been the case that the train was too far south to be able to reverse anywhere so was simply pushed forward via signaller authorisation.


Harrow is a driver relief point - if trains are turning up off timetable or relief drivers are still on trains stuck in a failure area, it doesn't take much for the booked driver reliefs to start going awry. This will obviously impact the ability to keep trains moving.



A stalled train, in railway vernacular, is a train between stations that cannot be moved because of an external factor not caused by the train itself such as a train in the section ahead or rear preventing further movement, a failure of points not in favour of the route required, a signal failing to clear, traction current turned off, physical obstruction on the track, etc.

Where a train cannot be moved because it is the direct cause of the issue, such as a being gapped, having an onboard defect, person ill on the train, no relieving driver, etc.......we call that a sit down.

Once the railway falls into failure mode, the railway can't always remain 100% customer focussed. There has to be a balance with attempting to resolve the root cause of the failure and also preparing operationally for resumption after a failure. Having said that, getting that driver back to their home location could potentially mean you now have a member of staff in position ready to give an earlier resumption of service - ditto with any rolling stock that has been moved through a failure area - so it could be argued in a roundabout way that there was customer focussed decision making going on; albeit not immediately obvious.

As for the need for the signaller to authorise separately past each signal......I'm not particularly familiar with the part of the Metropolitan line covered by the incident that led to this thread so unsure of exact type of signals in that area, but to perhaps expand a bit on your suggested potential corner cutting.....

In the case of automatic signals on LU........drivers must stop at each one, and unless previously advised that the signal is known to have failed, must wait two minutes before attempting to seek information as to why the signal is remaining at danger. The information they receive may be to continue to remain at that signal, or it may be that they're authorised to pass the signal at danger. If they cannot garner any further information after waiting those two minutes they can then pass that signal at danger under their own authority. Drivers must then proceed at a speed where they can stop short of an obstruction (ie. walking pace) for two consecutive stop signals showing a proceed aspect. If they were unable to gain any information, they must further attempt to do so at the earliest opportunity. If they come across another auto signal at danger, the process starts again.

Failure to follow this basic procedure correctly has led to collisions in the past - perhaps the most famous being on the Central line at Leyton back in the 80's when a train went into the back of another.

Semi automatic signals on LU are controlled by the signaller and often involve points, etc so drivers absolutely must get authorisation from the signaller before proceeding past a semi auto signal on LU. If a signaller were to attempt to authorise me past two or three signals in one go I would politely refuse their authorisation and request it individually at each signal. The same proceed at a speed where they can stop short of an obstruction for two consecutive stop signals showing a proceed aspect procedure then applies.

On the TBTC/CBTC ATO areas (Jubilee, Northern and much of the SSR) drivers can only ever be authorised from one fixed point to the next fixed point and never ever through multiple fixed points in one go.

Passing signals at danger is a safety critical action and anyone who attempts to speed the process up by cutting corners is opening themselves up to a catastrophic error that could easily put a train on the deck. I've seen it happen more than once. Especially with drivers receiving instructions in a train cab where it's often not written down and done "on the fly". They forget what they were told or get the instructions mixed up in their mind. Slow and steady always wins the race!