How can the southbound approach to Peterborough be made safer to avoid another 'Lumo' incident?

[MotCO]

This thread https://www.railforums.co.uk/thread...orough-17-04-2022.230537/page-17#post-6313516 started with the news that a Lumo train passed a junction at too high a speed for the turn-off. The thread discussed the incident and the RAIB report, but also moved on to what could be done to avoid a similar incident in the future, which feels like 'speculative discussion' (and I'm as guilty as anyone!). This thread therefore deals with suggestions as to how the approach to Peterborogh could be improved to avoid a similr incident.

To kick off, as a layman I have two suggestions. Firstly, although there are some trains which do not stop at Peterborough, could a speed limit be imposed to reduce the speed on approach until clear of the junction to ensure that any diverging route could be taken safely?

Secondly, could a 'speed camera' be installed which flashes up the speed a train is travelling if it is too fast for the divering route? Much the same as the road signs which flash up your speed if you are exceeding the speed limt or show a smiley face if you're being 'good'.

 

[bluenoxid]

I think (not at all my area of expertise) the technology is already in place, which would be two TPWS grids set to activate the braking system when the route is set and the train travels above a certain speed at that point

 

[Ken H]

Replace the points with a higher speed turnout. Dont know if thats possible. But too many turnouts could do with replacing with higher speed ones. Probably a 30mph turnout is cheaper than a higher speed one.

 

[edwin_m]

I think (not at all my area of expertise) the technology is already in place, which would be two TPWS grids set to activate the braking system when the route is set and the train travels above a certain speed at that point

Technically possible but may be tricky to pick the loop speed and position. If the train is accelerating because the driver thinks they have the unrestricted route, then it could be below the trap speed when passing the loop but still too fast at the turnout.

 

[MarkyT]

Technically possible but may be tricky to pick the loop speed and position. If the train is accelerating because the driver thinks they have the unrestricted route, then it could be below the trap speed when passing the loop but still too fast at the turnout.

Perhaps a number of speed traps between signal and turnout might help with that. I'd expect the first OSS encountered after the signal would be placed and have its set speed calculated in accordance with standard speed restriction installations, with extra OSSs between that and the turnout for the fairly likely scenario you reference. They'd all be switched by point position.

 

[Snow1964]

The alternative is to relay the turnout with a longer higher speed version. There is something like 700m between signal and current low speed turnout, and is far below the normal braking curve of train stopping at the platform.

There is not lot of common sense is allowing 125mph for miles, then expect crawling along at about 20% of that for a mile to a platform taking 2-3 minutes longer.

 

[furnessvale]

Hasn't action already been taken, in that a train taking the diverging route the Lumo took will no longer receive a green signal, but will be given a yellow.

 

[edwin_m]

The alternative is to relay the turnout with a longer higher speed version. There is something like 700m between signal and current low speed turnout, and is far below the normal braking curve of train stopping at the platform.

There is not lot of common sense is allowing 125mph for miles, then expect crawling along at about 20% of that for a mile to a platform taking 2-3 minutes longer.

I believe the report mentioned a plan to replace it with a 40mph turnout when it is next renewed.

Hasn't action already been taken, in that a train taking the diverging route the Lumo took will no longer receive a green signal, but will be given a yellow.

That's unlikely to be more than a stopgap, and possibly creates a hazard of anticipation - if they always get a single yellow there is a risk they will expect a proceed aspect at the south end of platform 1, only to find on one occasion that it remains at red. However, that hazard here is probably less of a risk than in some other places, because more often than not a train taking this route will be heading for a station stop in platform 1.

 

[HSTEd]

Unfortunately, sophisticated bespoke solutions may struggle to actually get into service before ETCS renders the whole problem moot.

The most important takeaway from this incident is that ETCS fitment must not be delayed further.
Only ATP can truly prevent this from happening again.

 

[D365]

The most important takeaway from this incident is that ETCS fitment must not be delayed further.
Only ATP can truly prevent this from happening again.

Alas, whether the rolling stock will be fully fitted in time…

 

[eldomtom2]

I don't know how simple it would be, but introducing a "clear over diverging route" aspect as many other countries have would solve the issue.

 

[30907]

I don't know how simple it would be, but introducing a "clear over diverging route" aspect as many other countries have would solve the issue.

How does that differ from green with a feather, though?

 

[yorksrob]

The alternative is to relay the turnout with a longer higher speed version. There is something like 700m between signal and current low speed turnout, and is far below the normal braking curve of train stopping at the platform.

There is not lot of common sense is allowing 125mph for miles, then expect crawling along at about 20% of that for a mile to a platform taking 2-3 minutes longer.

Indeed. It's not as though NR don't already own the land required.

 

[MarkyT]

A concern with faster turnouts is they usually require more backdrives and detection elements to accurately move the long switch rails and ensure they're correctly positioned. Thus the greater complexity could make the points less reliable. I was surprised they weren't already 40mph though, which seems a reasonable compromise and would make the divergence properly compliant for a flashing yellow approach sequence. Reliability was a major reason why the original 90mph turnouts for Heathrow Express at the GWML airport junction were replaced at renewal with the current shorter 75mph examples. The originals also had swingnose crossings I think.

 

[D365]

I don't know how simple it would be, but introducing a "clear over diverging route" aspect as many other countries have would solve the issue.

From what I understand of ETCS it will, in effect, provide this facility.

 

[MarkyT]

From what I understand of ETCS it will, in effect, provide this facility.

And, for trains suitably equipped, it will enforce compliance with the safe speed profile the onboard computer calculates for the movement authority. However, not all trains approaching Peterborough from the north will be ETCS-fitted, which is why NR have decided to go for an overlay solution for the station and ECML approaches, retaining colour lights and TPWS/AWS for traffic on the east-west axis through the wider station area. All ECML trains south of Peterborough will be required to be fully ETCS equipped as there will be block marker boards only and no signals for unfitted legacy traction.

 

[Northumbriana]

By-pass Peterborough?

 

[eldomtom2]

How does that differ from green with a feather, though?

In theory as it would be part of the regular signal aspect instead of the "additional" route indicator it would be more noticeable. Just a theory, don't know if any research into it's been done.

 

[MarkyT]

By-pass Peterborough?

That's a sledgehammer to crack a nut. Speed for through trains on the fasts has been dramatically increased over a series of remodelling and signalling projects since the 1970s. It used to be a crawl through the platforms back in steam days.

 

[Ken X]

That's a sledgehammer to crack a nut. Speed for through trains on the fasts has been dramatically increased over a series of remodelling and signalling projects since the 1970s. It used to be a crawl through the platforms back in steam days.

It could be argued they had it right back then. No problems with overspeeding in those days. Whole station 25mph, job done. Change can be a funny old thing. Progress maybe. Wonder what the time saving is?

 

[MarkyT]

It could be argued they had it right back then. No problems with overspeeding in those days. Whole station 25mph, job done. Change can be a funny old thing. Progress maybe. Wonder what the time saving is?

Passing New England Nth Jn. to Fletton Jn. takes 1.5m for a non-stop on the up fast. With a stop at platform 3, the time taken is 6 or 6.5 mins between these timing points, with a stop at the platform of 3 minutes. With signals cleared through, stopping trains can enter a platform on a fast line at quite a lick, at least 40mph, and still be able to come to a stand accurately. The old layout at Peterborough restricted speed over a distance either side of the platforms so I'd suggest that the modern high speed saves five minutes for a non-stop. Timings taken from Real Time Trains data for today (15/07/23). I'd need to look at an old timetable to see what A4s or deltics could achieve on the old layout!

 

[D365]

And, for trains suitably equipped, it will enforce compliance with the safe speed profile the onboard computer calculates for the movement authority. However, not all trains approaching Peterborough from the north will be ETCS-fitted, which is why NR have decided to go for an overlay solution for the station and ECML approaches, retaining colour lights and TPWS/AWS for traffic on the east-west axis through the wider station area. All ECML trains south of Peterborough will be required to be fully ETCS equipped as there will be block marker boards only and no signals for unfitted legacy traction.

the net effect is that all ’fast’ passenger services through Peterborough will be ETCS equipped.

 

[Halish Railway]

I seem to remember that there was a proposal a 100mph turnout at Felton Jn to assist in implementing the now shelved plan for 8 LNER trains for hour out of King’s Cross.

The ECML does have a lot of slow (<40mph) turnouts that I’m surprised haven’t been replaced with faster ones.

 

[Re 4/4]

I don't know how simple it would be, but introducing a "clear over diverging route" aspect as many other countries have would solve the issue.

In principle, a 4 aspect signal could show Y over G as a new aspect for this. The top Y is already there after all.

It would tie back to the days of stacked distant semaphore arms where, as far as I know, you'd get upper arm horizontal, lower one off for the diverging route - presumably the lamps in that case were G over Y for the main and Y over G for the diverge?

I suspect the cost of redoing everything except the signal lamps themselves will keep this firmly in the speculative thread though.

 

[HSTEd]

the net effect is that all ’fast’ passenger services through Peterborough will be ETCS equipped.

By the time 2029 rolls around very little outside the freight sector will not have ETCS fitted.

 

[D365]

By the time 2029 rolls around very little outside the freight sector will not have ETCS fitted.

I wish I could say more, but that’s by the by..!