Northern line service levels

[bramling]

Ah I see, is that a new methodology that Seltrac has introduced?

Not really, all the lines run like that, just that with increasing input from computers it’s possible to manage things that bit more closely.

For example under the old signalling, a Northern Line train would only be held to time at a few places, and then only if the signalmen had all the machines in the right mode to do that (which generally would the the case by default, but soon out of that mode during significant disruption).

By contrast, the Seltrac system holds a train to time at *every* station, and even imposes speed ceilings if it calculates that a train is going to arrive at a station early.

In terms of the end user, it may be the case in theory that a train is 20 minutes late but providing the headways are maintained then they might not be aware that their train is actually late. But in a lot of cases this doesn’t end up being true in practice, as when the service is running late then it’s quite likely conflicts will occur at junctions which wouldn’t otherwise have happened, also there’s a likelihood that a crew relief issue will occur at some point, and of course the control staff may need to turn the train short in order to get it back on time, or “reform” it (which if you’re lucky may still mean it goes to the original destination, but in some cases it may not). So, in essence, even as an uninterested end-user you still want your train to be on time if possible.

Also worth bearing in mind that not all Underground services run to high frequencies. For example Mill Hill East only has 4tph, and these are divided equally between Bank and Charing X trains. So if I’m travelling from Warren Street to Mill Hill East I kind of want the 2100 train to be on time and to run… And people at these sorts of stations do have some idea what time the trains are timetabled to run.

 

[Basil Jet]

This isn’t really true. For the Northern Line the system attempts to run everything *exactly* as per the WTT, essentially to the nearest second.

The WTT on the Northern Line only specifies times to the nearest quarter minute. Even the Victoria Line only specifies times to the nearest 5 seconds, and I think I heard somewhere that that is the most, er, granular railway timetable in the world.

 

[bramling]

The WTT on the Northern Line only specifies times to the nearest quarter minute. Even the Victoria Line only specifies times to the nearest 5 seconds, and I think I heard somewhere that that is the most, er, granular railway timetable in the world.

What I mean is that the system gives a countdown to the driver, with the intention being that the train starts moving at zero seconds.

Service control can manually adjust the timings to the nearest second IIRC.

 

[Daniel]

Service control can manually adjust the timings to the nearest second IIRC.

Correct

 

[D7666]

The WTT on the Northern Line only specifies times to the nearest quarter minute. Even the Victoria Line only specifies times to the nearest 5 seconds, and I think I heard somewhere that that is the most, er, granular railway timetable in the world.

The paper WTT as made available to the public, yes, are that granularity, but as well as the ability for service control to manually adjust to nearest second, the internal computer timetables are to the second on all lines, even on the 'clockwork' non ATO lines (leaving the remaining cabins out of this). The "paper WTT" are those available in PDF form in the public domain; the electronic timetables loaded on to the assorted different computer types on different lines are generated in a bespoke LU format and are not in the public domain and are not for publication; these are then loaded locally and compiled at site for their specific computer type. And they were one of the banes of my life for several years on one certain lines timetable servers, and I've worked at one point or another on every line's timetable computers except Bakerloo and Piccadilly. Edit not Victofiea either.

The "paper WTT" are not used on the timetable servers. Think about it. Take part of the Central Line core, WTT timing points are Queensway / Marble Arch / Holborn, but nothing in between. If every intermediate station were resolved to the nearest 1/4 min, the timetable would be too klunky and not permit the TPH demand. So in reality, the intermediate stops will all be arrive and depart to the nearest second, and, IIRC, the arrival time at those named locations is also to the nearest second.

Working in LU control rooms as I have now, as a engineer, for almost the past 20 years now, I am still impressed by the whole gamut of timetable and service and incident recovery, from one train out of turn through to major disruption - and especially the speed with which it is carried out. I can't do it. And I have to say I'd not want to do it.

One of the pieces of nonsense that I unfortunately got involved with trialling was one line was co-erced into trying under PPP a sort of indirect regulation but information only system that had inputs from the timetable server and from the signals control system trains positions; this was supposed to remove workload from service controllers; it fell down because one of the master keys to all keys are the train operators, without which you can do nothing, and a simple radio call to do something like a "stock and crew" or a "quick spin" similar can be done in anticipation ahead of an issue (i.e. to prevent it actually becoming an issue) proved (TBH it did not proving to anyone but software designers and systems analysts) far less clumsy and far more efficient than trying to get computer algorithms right for every possible "stock and crew" permutation. OR that system had to wait until the predicted issue had become an actual issue then it lit up like Blackpool hallucinations - and too late - and NBG to anybody. I took great delight one day on being officially authorised to shut it down, isolate it, and remove the data links.