Derailment near Watford Junction (16 September 2016)

[D365]

Could tilt sensors be added to trains so that if a tilt of more then a certain amount is detected, an emergency GSM-R call gets made? Probably won't catch every derailment but would hopefully catch the worst ones

I thought that you had recently insinuated that Siemens engineers (on the Class 700 at least) are useless.

 

[Nippy]

We don't want GSM-R putting signals back as well. We get enough grief when an emergency call in the Willesden area comes through to the Western side as well.

Also recently, we had an emergency call from a driver to alert us to the fact he didn't think he'd stop at a signal due to wheel slide. The problem with that was that the train in front leaving the station stopped on receipt of the emergency broadcast. As it happens, the driver did stop before the signal.

 

[ainsworth74]

Also recently, we had an emergency call from a driver to alert us to the fact he didn't think he'd stop at a signal due to wheel slide. The problem with that was that the train in front leaving the station stopped on receipt of the emergency broadcast. As it happens, the driver did stop before the signal.

Ah whereas I asume in that scenario you'd actually want the train in front to keep going not stop!

 

[causton]

Ah whereas I asume in that scenario you'd actually want the train in front to keep going not stop!

Yes, I have been told a new signalling control system automatically replaces all signals to red in the event of a SPAD being detected. Yes, once again, surely you want the trains in front to keep going, Northern line style!

 

[carriageline]

Yes, I have been told a new signalling control system automatically replaces all signals to red in the event of a SPAD being detected. Yes, once again, surely you want the trains in front to keep going, Northern line style!

The workstations I work have SPAD sub areas, which will put CONFLICTING signals to danger. They generally won't put any signals at front in danger. It depends on where trains are and what routes are sets


Sent from my iPhone using Tapatalk

 

[edwin_m]

It's pretty much a standard feature of interlockings to replace conflicting signals to danger if a track circuit goes occupied that indicates a SPAD.

 

[carriageline]

It's pretty much a standard feature of interlockings to replace conflicting signals to danger if a track circuit goes occupied that indicates a SPAD.

One of ours even replaces the group replacement in the station throat


Sent from my iPhone using Tapatalk

 

[oversteer]

RAIB report released
https://www.gov.uk/government/news/report-112017-derailment-and-subsequent-collision-at-watford

 

[SpacePhoenix]

Any new news on the units involved?

 

[47421]

The photo on p14 of the report is terrifying - camera on the southbound unit takes photo of landslide with headlight of northbound unit clearly visible in tunnel, going at 79mph at the time. Almost unbelievable that this was not much more serious

 

[YorkshireBear]

The photo on p14 of the report is terrifying - camera on the southbound unit takes photo of landslide with headlight of northbound unit clearly visible in tunnel, going at 79mph at the time. Almost unbelievable that this was not much more serious

As the RAIB said. Luck rather than judgement.

 

[Bletchleyite]

The photo on p14 of the report is terrifying - camera on the southbound unit takes photo of landslide with headlight of northbound unit clearly visible in tunnel, going at 79mph at the time. Almost unbelievable that this was not much more serious

Surprised by the poor quality of the camera, but then I guess these are now over 10 years old.

Interesting that the procedure on seeing a hazard light / getting a GSMR stop instruction are so vague. While the investigation says the use of full service vs emergency probably made no difference as the wheels were slipping anyway, this surely needs defining or even fully automating to use emergency.

 

[GW43125]

Interesting that the procedure on seeing a hazard light / getting a GSMR stop instruction are so vague. While the investigation says the use of full service vs emergency probably made no difference as the wheels were slipping anyway, this surely needs defining or even fully automating to use emergency.

It did say though, try to avoid stopping in a tunnel, thus the driver may have kept it in full service so as to be able to release the brakes later to roll out of the tunnel.

 

[Bletchleyite]

It did say though, try to avoid stopping in a tunnel, thus the driver may have kept it in full service so as to be able to release the brakes later to roll out of the tunnel.

Possibly so, though I wonder if that should actually be the case.

It's best if an emergency procedure is absolute and doesn't require any human discretion. To me, GSMR emergency = brake to emergency would be simplest - or even automate that.

 

[CosherB]

Any new news on the units involved?

Both units undergoing repairs at Brodies Kilmarnock. No idea if they have been completed.

 

[Blindtraveler]

If their at Kilmarnock we'lle expect them back just in time for the 350 farewell tour in about a decade or so!

 

[ainsworth74]

or even fully automating to use emergency.

I'm not sure that's wise. There's a reason why pulling the cord on modern trains doesn't just dump the brake for example. It allows drivers to use their judgement to bring trains to a stand in the best possible location.

 

[Bletchleyite]

I'm not sure that's wise. There's a reason why pulling the cord on modern trains doesn't just dump the brake for example. It allows drivers to use their judgement to bring trains to a stand in the best possible location.

But pulling the cord doesn't mean "there's something you might hit", whereas GSMR Emergency pretty much does. The cord is much more likely to mean an ill or errant passenger or perhaps a small (but controllable due to modern materials) fire.

Though I prefer the German model of having a separate passcom for such situations, and an emergency brake for "the train must stop now", e.g. someone being dragged in the doors or a derailed bogie.

 

[ainsworth74]

But pulling the cord doesn't mean "there's something you might hit", whereas GSMR Emergency pretty much does. The cord is much more likely to mean an ill or errant passenger or perhaps a small (but controllable due to modern materials) fire.

True but it still seems far more sensible to me to leave that sort of control to the driver of the train. By all means alter the rules so that it should be emergency brake used but I'd be hesitant to take that control away from the driver.

 

[theageofthetra]

Having read the report credit must go to the train crew for preventing what could have been a much more serious incident. The pure luck that kept the derailed unit mostly in line is more disturbing.

 

[Dieseldriver]

Possibly so, though I wonder if that should actually be the case.

It's best if an emergency procedure is absolute and doesn't require any human discretion. To me, GSMR emergency = brake to emergency would be simplest - or even automate that.

If you use emergency, you are unable to release the brake until you've come to a stand (I'm actually pretty sure that on Desiros there is actually a time delay in addition). If I were in the position of the Driver on the train that received the REC, I would have been pretty reluctant to select emergency aswell. The call could have been initiated due to a fire on the oncoming train or even due to a serious passenger incident that would pose a significant risk to the passengers on board my train if we stopped within the tunnel alongside. I have received RECs a few times, on all of those occasions I have used the emergency position, in the Watford tunnel incident, I would have used full service.

 

[Eccles1983]

If you use emergency, you are unable to release the brake until you've come to a stand (I'm actually pretty sure that on Desiros there is actually a time delay in addition). If I were in the position of the Driver on the train that received the REC, I would have been pretty reluctant to select emergency aswell. The call could have been initiated due to a fire on the oncoming train or even due to a serious passenger incident that would pose a significant risk to the passengers on board my train if we stopped within the tunnel alongside. I have received RECs a few times, on all of those occasions I have used the emergency position, in the Watford tunnel incident, I would have used full service.

Exactly.

There lies the difference between "bashers" and drivers.

On paper his idea is sound. In the real world it doesnt fit.

 

[bramling]

Possibly so, though I wonder if that should actually be the case.

It's best if an emergency procedure is absolute and doesn't require any human discretion. To me, GSMR emergency = brake to emergency would be simplest - or even automate that.

I'm not really keen on the idea of a train being forced to come to a stand. Sure, in this specific incident, an emergency stop might have been preferable (*), but there are other situations where it's definitely not preferable for a train to come to a stand in a random location. Next to a fire is a good example - and there have been such incidents in the past where an enforced stop has contributed to a much worse outcome for a particular incident. The fire near Goodge Street in the 1980s is one such example.

(* although in reality the difference between emergency and full service was probably minimal in any case)

The report as a whole does make sobering reading for those keen on automation taking the driver away from the front of the train, and it does show the value of guards from a practical if not financial point of view. In this case having the human at the front of the up train quite possibly saved many lives.

 

[Bletchleyite]

(* although in reality the difference between emergency and full service was probably minimal in any case)

The report does confirm this.

The report as a whole does make sobering reading for those keen on automation taking the driver away from the front of the train, and it does show the value of guards from a practical if not financial point of view. In this case having the human at the front of the up train quite possibly saved many lives.

Indeed so. But you can offset that against the times when the driver has been the cause of a serious accident which would not have occurred on a fully-automated service. (Similarly, for each time a pilot has saved a severely damaged aircraft through skilled manual flying there are probably at least two cases where a pilot has killed everyone on board due to human error).

I suppose the "best of both worlds" is highly sophisticated ATP of some kind (for planes, the analogy is perhaps the Airbus control philosophy).

 

[najaB]

I suppose the "best of both worlds" is highly sophisticated ATP of some kind (for planes, the analogy is perhaps the Airbus control philosophy).

All that's needed is a short delay during which the driver can override the automated stop - so the brakes immediately go into full service and after (pulling a number out of thin air) five seconds they go into emergency.

 

[eman_resu]

Possibly already discussed, and apologies if so.

Looking at the comments by Simon French, from the report summary ;

everyone in the railway industry continues to work so hard to manage risk - the collision of two trains in a tunnel is a scenario we all hoped never to witness.

Is there any recourse to look at the diagrams to prevent trains passing each other in single bore tunnels...

Clearly this all works great until there is running out of schedule.....

And regarding the drivers, once again route knowledge, experience and situational awareness helps to save the day.

 

[bramling]

Possibly already discussed, and apologies if so.

Looking at the comments by Simon French, from the report summary ;



Is there any recourse to look at the diagrams to prevent trains passing each other in single bore tunnels...

Clearly this all works great until there is running out of schedule.....

And regarding the drivers, once again route knowledge, experience and situational awareness helps to save the day.

Yes one could certainly arrange the timetable so trains don't pass in tunnels, the signalling could quite easily be designed to enforce this. However this would seriously reduce the capacity of the line such that many routes would have to see a reduction in the number of services, quite significantly so on busy routes. Just think of the effect on something like the Welwyn tunnels on the east coast main line as an example. Meanwhile there is still just as much chance of two trains colliding elsewhere for whatever reason. For these reasons I can't see a case for doing it; whilst a head-on collision in tunnel scenario is something that doesn't bear thinking about, the chances of it happening are extremely low. It's hard to think of many incidents that have occurred in tunnels at all, although I'm sure statistically such an incident will surely occur one day.

 

[swt_passenger]

It was interesting to read about stuff on the bogies having the side effect of often helping to keep derailed trains "near" the rails, presumably by snagging on the inside of the rails. They mention that such an effect is not designed in or documented, so it could just as easily be accidentally removed by redesign.

From the press release/summary:

The second mitigating factor was the slotting of one rail of the track in the gap between a gearbox and a traction motor on three of the axles, so preventing the derailed train deviating any further into the path of the approaching train. This unintended consequence of the train’s design probably made the difference between a glancing blow and something closer to a head-on collision.

Perhaps within tunnels there should also be guard rails in the four foot, like the ones on some bridges, I assume they might only need to fit one rail as the main hazard is only in one direction?

 

[DarloRich]

The central point of that report is that we were very lucky. Good work by the train crews, prompt communications and a unknown design feature combined to prevent a much worse accident.

The next one might not be so lucky.

 

[ainsworth74]

Good work by the train crews, prompt communications and a unknown design feature combined to prevent a much worse accident.

Those are the big takeaways for me I have to say.

The reaction of the train crews and in particular the way that not only the drivers and guards works together to support each other and take control of the situation as a team was very impressive (particularly on the derailed train where the driver was trapped in the cab).

It was also good to see GSMR, for once, working as advertised in an emergency situation and enabling rapid and effective communication from the emergency stop transmission to coordination the response. To me it reinforces the key importance of making sure that the GSMR is absolutely bomb proof. It was also interesting that the GSMR presumably had no problems in the tunnel reaching the outside world but the radios brought by emergency services did.

And then just dumb luck to cover everything else and prevent a much worse accident. It's interesting that the RAIB recommend that the interaction of the bogies and track is looked into further with a view to looking at the standards.