Two St Pancras Platforms

[thelem]

However in years to come when the platforms are crowded...

Have you seen the platforms at St Pancras? They are huge. While there may be an argument to have more platforms, the ones that have been build are certainly wide enough.

 

[A60K]

Please can someone quantify this for me, because I really don't see the logic.

Where you have converging lines you will inevitably get a bottleneck where traffic density is high. You have to carefully timetable and control your train running to ensure that services don't delay each other. In the event that two trains arrive at the junction at the same time, one will always have to wait. All an extra platform does is move the junction from before the station to after it.

Say you have a train scheduled to arrive every 2.5 minutes, with the dwell time at the station being 1 minute. The minimum time between a train starting and the next one arriving at the platform is 1.5 minutes.

You have a 0900 arrival 0901 departure, an 0902.5 arrival 0903.5 departure, an 0905 arrival 0906 departure and so on. Ideally there are coming alternately from the Midland and from the GN.

With the junction before a single platform: If the 0900 arrival 0901 departure is running 1 minute late, then the 0903.5 departure has to follow it into the same platform and then allow passengers to board and alight, making it automatically 1 minute late. Meanwhile the following train is also being delayed and it's a slow process to catch up the delays.

With the junction after a two face island platform: the 0903.5 departure is able to arrive on time, and passengers have boarded and alighted to allow it to leave on time while the late-running 0901 is still clearing the section towards Farringdon. You will still have the trains delayed further down the two-track railway, but not by as much as if you have a single platform after the junction.


If you run the same tests with a 2 minute delay to the first train the benefit of the two-platform layout is even greater.

Basically the way it's working is to get the dwell time over and done with rather than waiting for the train in front to get out of the way first, therefore clawing back the odd minute or so of delay without so much of a knock on effect back down the line.

Trust me that this layout design is the ideal on a rapid transit system wherever it is physically possible - as well as the London examples above, you'll find the same in Paris on both the Metro and RER, in Berlin, and I'm sure on other systems.

I state again that I think there was a structural reason why it wasn't possible to have a junction south of the platform at St P - there would however have been plenty of space to make the platform itself wider, as it lies to the north of the old trainshed under the area that was completely redeveloped for the main station extension.

 

[O L Leigh]

The maths is compelling, but you make a number of assumptions.

The planned frequency is indeed an average on 1 train every 2.5 minutes (24 trains per hour in total). However, what your maths ignores is the length of headways permitted by ATO. You are assuming that the next train cannot access the station until 1.5 minutes after the previous one has left.

So lets change the parameters a bit and assume a 1 minute headway. The theoretical 1 minute delay accrued by your 0900 arrival can be recovered very quickly to the extent that the 0905 arrival comes in on time. If it is possible to keep the dwell times down to less than a minute, service recovery can be made even faster.

To be entirely honest, I don't know what the headways and dwell times are likely to be and can find no information quickly. However, I suspect that with a very high frequency dwell times will be modest. Then again, the frequency of services outlined in the Thameslink programme seems to be frighteningly high, so perhaps the number of trains per hour won't be quite so great when the service eventually opens.

O L Leigh

 

[A60K]

I agree that ATO will help, and moving block signalling would help even more, but they don't completely compensate for the flexibility of two platforms - we'll agree to disagree!

I think you've hit the nail on the head though - I don't think the proposed service pattern and frequency is going to be practical, and I think the 2015 actual service pattern will be simpler.

The RER in Paris achieves very high frequency, but with dual platforms at key stations, and all services being about the same length of journey - basically just to the edge of Greater Paris.

 

[O L Leigh]

I think we actually agree on a number of factors, but I just don't find the case for additional platforms sufficiently compelling.

We shall see.

O L Leigh

 

[tbtc]

The big difference between this and a "tube" is that most people on a tube are going to jump on the first train (so the platform is almost empty immediately after a train has departed), whereas a lot of Thameslink passengers (going further than London Bridge) aren't going to want the "first" train, meaning a lot of queuing passengers.

Still, the proposed eye watering frequencies on this long Thameslink tunnel is a contrast to some routes elsewhere in the country that don't have capacity for a fourth train an hour etc (different circumstances, I know, I know)...
--- old post above --- --- new post below ---
Also, has there been any guide given to where these 24 trains an hour are to go?

Thinking about it northbound, there's eight an hour on the existing Thameslink route (St Albans/ Luton/ Bedford), plus I understood the four "semi fast" Great Northern services an hour (two to Peterborough and two to Cambridge) would go into the tunnel too. But are the additional paths going to be peak only or daytime? Or going elsewhere?

 

[philjo]

When the GN route is integrated, the northbound platform in particular will have people waiting for a number of different services - i.e. fast to St albans or all stations TL route etc or fast to Stevenage & onwards to cambridge or peterborough or the GN semi-fast services - I assume that all the GN route services will call at Finsbury park.
Therefore, people will be waiting for their required service on the platform - though as the train is now coming through london they all needn't get on at st Pancras as they have to do at Kings cross at the moment

 

[Fincra5]

More platforms could have been built but at what cost? I think thats what it largely came down to.

 

[Edvid]

Has there been any guide given to where these 24 trains an hour are to go?

Thinking about it northbound, there's eight an hour on the existing Thameslink route (St Albans/ Luton/ Bedford), plus I understood the four "semi fast" Great Northern services an hour (two to Peterborough and two to Cambridge) would go into the tunnel too. But are the additional paths going to be peak only or daytime? Or going elsewhere?

Here's a link to an indicative diagram for morning peak hour 'Thameslink South' journeys.

As for their 'Thameslink North' counterparts, there are no equivalent diagrams, although it's most likely AFAIK the diverted morning peak hour GN services will be 8 outer-suburbans (Cambridge/Peterborough semi-fasts). Off-peak I guess would be 4 of the same.

As for the MML services, morning peak hour/off-peak tables of indicative services were recently published by Network Rail and I've attached them.

Therefore the off-peak TL core frequency may be 16tph, whereas 24tph is definitely limited to the peak hours.

 

[PhilipW]

For southbound trains, if (and it's a big 'if') departures were co-ordinated at Kentish Town and Finsbury Park then there would never be any train held up in the tunnel just north of St Pancras as each train would have a clear run. With common rolling stock the timings from those two stations to St Pancras should be known to the second.

I'm sure passengers would prefer to wait at those stations rather than in the tunnels.

 

[jon0844]

Nice as it will be to have GN trains going on south to places like Brighton, I actually do see an issue with the fact that people at King's Cross currently have the luxury of going to a platform, boarding an empty train and waiting for its departure northbound. Running via STP, this would now mean lots of people waiting on a single platform - and I guess, boarding a train that could actually be quite crowded at certain times (they can't start a train at St Pancras can they - where would it come from?).

Also, the tube-running frequency is fine for services that follow the same stopping pattern (e.g. when FCC begins to run a 15-minute service from WGC to Moorgate, stopping all stations) but not when the trains all have different destinations or stopping patterns. As said above, people can't just jump on the next available train. I guess at the moment you can if you're going to some destinations, but not when it will split and the train goes along a totally different line.

Although all northbound trains will run in one order - there's only the one track, there's the question of what happens southbound - do trains overtake others by having one held in one tunnel, or will they still have to run in a set order wherever possible?

I hope the system will work, but the next generation, or the one after that, will no doubt be forced to look into expansion - and St Pancras doesn't seem to offer much on any (literal) level!

 

[David Sinnett]

When the GN route is integrated, the northbound platform in particular will have people waiting for a number of different services - i.e. fast to St albans or all stations TL route etc or fast to Stevenage & onwards to cambridge or peterborough or the GN semi-fast services - I assume that all the GN route services will call at Finsbury park.
Therefore, people will be waiting for their required service on the platform - though as the train is now coming through london they all needn't get on at st Pancras as they have to do at Kings cross at the moment

Someone has already mentioned the wide platforms at ST P which are good for avoiding people waiting for later trains. In my opinion though, more people change from rail to tube at Farringdon than at St P and if they use Farringdon also to change from GN to TL then their platforms are going to become dangerously crowded.

 

[mackenzie_blu]

Right lets throw in some of the stuff Ive gathered but hasn't been discussed. The NXEMU are designed to cope with 1,000 people getting on or off in 37 seconds. This means roughly a train should be in the platform (or atleast part of it) for 90 seconds.

This doesn't sound alot but actually it isn't bad. Now before someone disagrees I work in the core and have been timing passenger broading times in real time in the peaks during various days to give me a decent sample of times. Most trains currently empty out around the 40-50 second mark depending on stations. IE City has a large number of customers who refuse to leave the train expect one door only leading to longer dwell times when the other door in the same carriage is empty of people using it for 20-30 seconds because they want to be the closest to the way out. So actually having trains designed to cope with this should allow for the needed dwell time to lower. Interestingly the 377/5 auto shut on the doors is 45 seconds which normally is fine as we dispatch the train as they close automantically.

In the New Lines Programme from Network Rail, they looked at the case for HS to relive the MML and gave KO2 tph north of the Core. Added with the South London RUS you get;

Northbound

16tph MML
8tph to Bedford
4tph to Luton
4tph to St Albans

8tph on ECML
4tph to Peterbrough
2tph to Cambridge
2tph to Letchworth

Soundbound

8tph '8 car' trains to Blackfriars bays through/via Elephant & Castle
4tph from Sutton
2tph from Kent House
2tph from Rochester

6tph '8 car' trains to TL Core through/via Elephant & Castle
2tph from Orpington
2tph from Sevenoaks
2tph from Maidstone East

18tph '12 car' to TL Core via London Bridge
4tph from Brighton
2tph from Horsham
2tph from Three Bridges
2tph from East Grinstead
2tph from East Croydon
2tph from Norwood Junction
2tph from Tunbridge Wells
2tph from Paddock Wood

Also note that additional trains are not going to Brighton as someone pointed out. Thats being kept as Bedford-Brighton as now. New destinations will be linked ie maybe Cambridge to Three Bridges or Maidstone to Letchworth.

 

[tbtc]

Here's a link to an indicative diagram for morning peak hour 'Thameslink South' journeys

Thanks for the link - very interesting

 

[Lee_Again]

There are a couple of points about St Pancras. Firstly, four platforms would have been a huge advantage for when things go wrong. Regular users of Thameslink will know that trains all to frequently break down in this section; causing chaos. Stabling one on a loop would help a lot. Further, I would extend the 3rd rail. Using St Pancras as the change over point would allow one train to arrive while another is changing traction type.

The main point is that the platforms (as large as they are) will be mobbed. 24tph. No chance; "you read it here first". Well, maybe not first, but all to predictable all the same.

Also, slightly off topic but nonetheless related, Farringdon, CityTL and Blackfriars. Three stations all within 1 mile. You can walk from Farr > Blackfriars in <10 mins. Maybe time to consider loosing CityTL. Remove the platforms and replace with loops, further improving the operational side.

 

[jon0844]

The distance between City Thameslink and Blackfriars is rather stupid, although I guess the northern entrance to City Thameslink does mean quite a long walk - although you can just as easily walk north to Farringdon.

I do think there will be many issues, and unless the brand new trains are somewhat more reliable than the 377s, it will be chaos. Can you imagine the amount of cancelled/terminated short trains when one of them breaks down and takes out about 10 services in quick succession?

And as I only really realised when I mentioned it earlier, what is going to happen to all the people who usually spread apart to board their various trains at King's Cross in the rush hour, now they all have to wait in one place for each train to come in?

That isn't even considering the fact you used to board a nice empty train, but now you've got a minute or so to board a potentially packed train that has picked up loads of people before - and it's not as if they can make the train skip loads of stops before, as it still has to queue up through the central core. Instead of getting my train from St Pancras, I'd be better off going back to Blackfriars or London Bridge and then going north!

 

[Aictos]

There are a couple of points about St Pancras. Firstly, four platforms would have been a huge advantage for when things go wrong. Regular users of Thameslink will know that trains all to frequently break down in this section; causing chaos. Stabling one on a loop would help a lot. Further, I would extend the 3rd rail. Using St Pancras as the change over point would allow one train to arrive while another is changing traction type.

The main point is that the platforms (as large as they are) will be mobbed. 24tph. No chance; "you read it here first". Well, maybe not first, but all to predictable all the same.

Also, slightly off topic but nonetheless related, Farringdon, CityTL and Blackfriars. Three stations all within 1 mile. You can walk from Farr > Blackfriars in <10 mins. Maybe time to consider loosing CityTL. Remove the platforms and replace with loops, further improving the operational side.

But City is one of the flagship stations for FCC much like St Pancras is,

As to loops, well it would be nice to have one at City but it won't happen full stop!

What they should do if possible is remove the platforms at Kings Cross Thameslink and put track where the old platforms used to be as loops which won't help the issue in the core section but would help to a certain extend.

 

[dan_atki]

What they should do if possible is remove the platforms at Kings Cross Thameslink and put track where the old platforms used to be

As far as I'm aware the platforms at KCM were retained for use in an emergency.

 

[A60K]

It wouldn't be possible to put in loops at KXT - space available is not long enough, and possibly not wide enough either.

 

[Tomnick]

The idea of putting loops in at the former KX Thameslink seems rather bizarre to me - what's it going to achieve? It'll do absolutely nothing to increase capacity, and it won't be much use in the case of a failure unless (which is rather unlikely) the failed unit happens to come to a stand close to the loop and is able to limp inside (surely it'd either be a complete failure, in which case you might as well assist it right out of the tunnels, or it'll be able to work forward to Kentish Town or Blackfriars on its own).

I could see the logic behind providing loops at St. Pancras if dwell times there were likely to be considerably longer than those further south through the centre - as that would allow a following train to be signalled into the station while punters were still sorting themselves out on the first. If other stations had similar dwell times though, the loops would be pointless (no pun intended), as your second train is just going to end up sitting behind the first when it gets to the next station - so that'll become the limiting factor for capacity instead. However, as has been pointed out, they'd be rather expensive loops!

 

[Waddon]

It's just going to get worse when the Moorgate branch is finally disconnected. There was a Sutton service with a fault this afternoon at Farringdon, and they ran it up the Moorgate branch to get it out of the way. I've seen this manouvre carried out several times in the last few weeks. In a few month's time where will they be able to put that train?

 

[generalvegitable]

your second train is just going to end up sitting behind the first when it gets to the next station - so that'll become the limiting factor for capacity instead. However, as has been pointed out, they'd be rather expensive loops!

if the junction is to be at st-pancras low level, then the loops will be something like one side for mml another for ecml

 

[dan_atki]

It's just going to get worse when the Moorgate branch is finally disconnected. There was a Sutton service with a fault this afternoon at Farringdon, and they ran it up the Moorgate branch to get it out of the way. I've seen this manouvre carried out several times in the last few weeks. In a few month's time where will they be able to put that train?

The faults at Farringdon are frequent enough when there are problems going from AC to DC and as there are overhead lines on the Moorgate branch trains were able to be moved there to clear the route.

Saying that, they are going to put wires up to City Thameslink (and presumably City's sidings) presumably for this exact reason.

 

[Waddon]

The faults at Farringdon are frequent enough when there are problems going from AC to DC and as there are overhead lines on the Moorgate branch trains were able to be moved there to clear the route.

Saying that, they are going to put wires up to City Thameslink (and presumably City's sidings) presumably for this exact reason.

Thats all fair and good in the long term, but the Moorgate Line is going to be cut off in a few months time - what happens if a train has difficulties switching power supplies before the new wires go up? There doesn't appear to be any way of reversing trains at Farringdon at the moment if this happens (or is there?)

 

[dan_atki]

Thats all fair and good in the long term, but the Moorgate Line is going to be cut off in a few months time - what happens if a train has difficulties switching power supplies before the new wires go up? There doesn't appear to be any way of reversing trains at Farringdon at the moment if this happens (or is there?)

I'd certainly hope they're going up before closure of the branch otherwise if something does happen at Farringdon, chaos will ensue for several hours.

 

[Broken Viking]

Just my 0.0002p's worth...But what about reviving the old KX Thameslink platforms and using them as "seconds" to suppliment the ones at St. Pancras?

What I mean by that is using the old KXT platforms to allow extra capacity during peak hours. If two southbound services are due close together, the first could be signalled right into the KXT platforms, with the second pulling into St. Pancras behind it.

It wouldn't solve the "jam" problem that well, with neither train being able to pass by the other along that section...But it'd allow STP/KXT to load/unload two trains at the same time, and would allow the increasing numbers of passengers to spread themselves out across two platforms insted of one.
As an added bonus, the only expense needed might be an NSE-style brush-up of KXT and maybe a connecting passage run between there and STP... 8)

 

[Tomnick]

if the junction is to be at st-pancras low level, then the loops will be something like one side for mml another for ecml

Maybe so, but - unless the dwell time at St Pancras is more than elsewhere - the provision of loops there will still do nothing for the capacity of the line.

 

[jon0844]

Will ALL trains run through London, or could some terminate upstairs? Say 1 in 4 (or whatever, maybe 1 in 8 depending on the frequency) trains from Bedford will terminate and go back, which also means a service running to a more reliable timetable if things have gone wrong through the core? I guess for emergencies and the weekend engineering, this sort of happens already.

The problem here is obvious; the lack of platforms upstairs and any likely expansion of them in the future.

Obviously the problem with having some trains upstairs and downstairs is another issue.

St Pancras will probably find itself being too small in not that many years, especially when/if Eurostar begins to expand with services to/from more distant destinations.

Out of interest; what are the plans, if any, to expand the station in the next 20-30 years?

 

[Waddon]

Out of interest; what are the plans, if any, to expand the station in the next 20-30 years?

I would guess there are no plans, as the station has just undergone one of the most comprehensive rebuilds of any rail station in living memory. If there was anything that wanted doing in that sort of timescale, they would presumably have done it at the same time.

 

[jon0844]

The station was restored and upgraded - but the building is what it is. It's a magnificent structure, but it was never used as the main destination for trains coming from Europe - which will hopefully grow. I wouldn't want the building torn down (and it won't be - it's listed and beautiful) but it will have to be extended somewhere.

Forget the building itself too. As and when more trains arrive, someone really must sort out the crazy situation where people all block the station waiting outside the arrivals door - which seemed to be placed there as an 'oh s**t, we forgot to design somewhere for the people that leave Eurostar!'