Page 1 of 1
| Draft graph for TT 4.1.4 19/09/2024 at 15:29 #158584 | |
Anothersignalman
89 posts |
Graph attached, prepared in Libreoffice Calc. Uses a repeating sequence of thirteen colours to represent the thirty-nine train sets operating on the mainline; does not show moves on the Picadilly line. I may be able to adjust the specific colours of each trip later to avoid following trips sharing colours, but it has to be done manually. The horizontal scale varies as required to show detail. I'm calling the graph v4.1.4 because it's a tweaked version of 4.1.3, though there may be other changes necessary before a new timetable file can be issued. The changes against v4.1.3 as it presently exists are: • V226-27 should depart Walthamstow 00:23:00 • V226-19 should depart Seven Sisters 18:55:30 • V236-19 should depart Walthamstow 18:50:30 • V211-31 should depart Walthamstow 00:13:00 The graph also demonstrates that V240-18E needs to be rescheduled, because it departs Seven Sisters five minutes before the previous trip (V240-18) arrives. There are fifteen further trips which look suspiciously close, and may need their times revised: • 08:57;30 ex Walthamstow to Brixton or 09:00;30 ex Northumberland Park Depot to Brixton • 10:20;00 ex Walthamstow to Brixton or 10:14;00 ex Northumberland Park Depot to Brixton • 12:22;30 ex Walthamstow to Brixton or 12:17;00 ex Northumberland Park Depot to Brixton • 12:36;30 ex Walthamstow to Brixton or 12:32;00 ex Northumberland Park Depot to Brixton • 13:22;30 ex Walthamstow to Brixton or 13:17;00 ex Northumberland Park Depot to Brixton • 14:36;30 ex Walthamstow to Brixton or 14:32;00 ex Northumberland Park Depot to Brixton • 15:05;00 ex Walthamstow to Brixton or 15:00;00 ex Northumberland Park Depot to Brixton • 15:57;00 ex Walthamstow to Brixton or 15:55;00 ex Northumberland Park Depot to Brixton • 16:42;30 ex Walthamstow to Brixton or 16:49;00 ex Northumberland Park Depot to Brixton • 17:15;30 ex Walthamstow to Brixton or 17:22;00 ex Northumberland Park Depot to Brixton • 17:35;00 ex Walthamstow to Brixton or 17:41;30 ex Northumberland Park Depot to Brixton • 19:22;30 ex Walthamstow to Brixton or 19:17;00 ex Northumberland Park Depot to Brixton • 21:22;30 ex Walthamstow to Brixton or 21:17;00 ex Northumberland Park Depot to Brixton • 21:38;00 ex Walthamstow to Brixton or 21:32;00 ex Northumberland Park Depot to Brixton • 22:57;00 ex Brixton to Walthamstow or 23:05;00 ex Victoria to Northumberland Park Depot (These times are from the graph, so may be +/-30sec from actuals.) @Metcontrol, are you able to check the departure times for the 31 trips listed above, and provide corrections where required? There are also plenty of trips which must arrive 1-2min late at Brixton and Walthamstow, particularly in peak hours, because their arrival platforms are occupied. 
Post has attachments. Log in to view them. Last edited: 19/09/2024 at 15:31 by Anothersignalman Reason: None given Log in to reply |
| Draft graph for TT 4.1.4 21/09/2024 at 12:53 #158604 | |
|
metcontrol
225 posts |
Interesting work there. Unfortunately any of the information I can provide may have to wait a couple more days as I am nearing the end of a week of nights controlling the real thing (not the Victoria but another - I wonder which ;-) )
Log in to reply |
| Draft graph for TT 4.1.4 22/09/2024 at 18:37 #158622 | |
|
metcontrol
225 posts |
Right I've had a brief opportunity this evening to look at the trips you are querying. V240-18 / V240-18E: At some point I think a tweaking of the trip from Brixton may have occurred, which has resulted in a later arrival at Seven Sisters. The following trip has not been tweaked as it is merely a trip to the depot and has not highlighted itself as an issue. It does the job of moving said train to the depot. Of the other 31 trips you mention, taking the following as just 1 example (I simply do not have time to go through each and every trip): 10:20;00 ex Walthamstow to Brixton or 10:14;00 ex Northumberland Park Depot to Brixton These are 2 separate trains. T225 departs Walthamstow 10:20:00 and runs to Brixton T211 departs Northumberland Park depot at 10:14:00, calls at the staff platform* then runs onwards (on effectively siding roads so at reduced speed) to Seven Sisters, in time to form a service to Brixton at 10:25:30 ex Seven Sisters. *The staff platform is a "stop as required" platform to transport staff to/from the depot at Northumberland Park. I am not sure if you know but the depot is located nowhere near any LUL station, and so some of the service/depot working etc. is tailored to provide a shuttle service to/from Seven Sisters and the depot. Dependant on whether a train needs to stop and pick up staff, the train may arrive bang on tiem or early for the onward trip towards Brixton. At Seven Sisters there are 3 minutes between 211 and 225 departing, though 211 may arrive before its booked departure. At the end of the day, what you have created is interesting in terms of viewing the timetable and any conflicts / service patterns throughout the day. That said, the timetable you are analysing was created to give the sim a pretty much spot-on timetable to use. It was created using various means including the actual Working Timetable. There are over 1900 trips within the timetable, and to form the first drafts, certain methods were used to insert everything without the need to enter each and every individual train. So some trains - certainly between Seven Sisters and the depot - used template trips which were adapted and appeared through testing to all work correctly. Whilst the timetable is pretty much spot on for the core line itself (i.e. passenger workings) you have indeed found the odd discrepancy in timings for some trips. The timetable was tested many times - trust me I tested it a lot! I would often run through most of it before Peter would add something else and I would then test again. However your testing is obviously of a much different nature which wasn't foreseen. I would like to use the Joker card which, if you run enough sims and timetables you will come across, which is whilst the planners think things will work, and whilst everything on paper appears fine, signallers are employed in real life to overcome those very occasional times where in real life things have to be manipulated. Log in to reply The following users said thank you: Anothersignalman, mldaureol |
| Draft graph for TT 4.1.4 23/09/2024 at 08:15 #158631 | |
|
TimTamToe
659 posts |
For those in the UK and interested, U's programme Secrets of the Underground featured Northumberland Park Depot in the last series and mentions about the depot trains that MetControl was referring to. The episode can be found below https://u.co.uk/shows/secrets-of-the-london-underground/series-4/episode-7/6359599808112 Last edited: 23/09/2024 at 08:27 by TimTamToe Reason: None given Log in to reply The following user said thank you: metcontrol |
| Draft graph for TT 4.1.4 23/09/2024 at 15:12 #158638 | |
|
Anothersignalman
89 posts |
metcontrol in post 158622 said:Right I've had a brief opportunity this evening to look at the trips you are querying. Thanks metcontrol. I think it's evident from the last decade or so that V4.1.3 has been active that nobody would have found these issues without a graph as a reference. Unfortunately I don't have the skill set necessary to create a generalised XML to Graph translator; this required a week of manual data entry. It would also be a lot easier to spot this sort of thing if train number suffixes were a consistent number of characters, e.g. V211-02 instead of V211-2. That's something I'll look at changing in V4.1.4. Re the time issue with V240, I'm surprised it wasn't picked up by the Timetable Validation thing; so I've opened a thread on that - https://www.SimSig.co.uk/Forum/ThreadView/55825?postId=158637 With that said, and having had a better look at the graph, I think the problem is related to the last trips of V237 and V240 possibly being swapped. Both occupy Seven Sisters platform 4 at the same time (again, not picked up by the Validator, but this would be harder to detect automatically): V237: Finsbury 18:46;00, Seven 18:50;00-18:55;00, SevenRev 18:56;30, NorthDep 19:04;30 V240: Finsbury 18:50;30, Seven 18:54;30-18:49;30, SevenRev 18:51;00, NorthDep 18:59;00 So the solution is to push V240-18E to 18:57;00 departure, and also shift V237-18E to 18:52;30; in both cases, would 2min30sec be sufficient to clear passengers at Seven Sisters PL04? Re the example of V211-11E form V211-11 (10:23;30 to 10:25;30) and V225-7 (10:26;30), it's three minutes between arrivals, one minute between departures. So I'm wondering if 211-11 should start from Seven Sisters a minute earlier, at 10:24;30 instead, for a 120-second headway? The line is marketed as having a 36-second headway today, but that's fairly recent. This article - https://www.londonreconnections.com/2017/ninety-second-railway-making-victoria-frequent-metro-world/ - says it historically worked on a 120-second headway in peaks, 126-second headway offpeak. As a general principle, I'll probably end up moving the Seven Sisters trips, rather than the Walthamstow trips, to establish a minimum 120 second headway; does that sound reasonable? Log in to reply |
| Draft graph for TT 4.1.4 23/09/2024 at 21:32 #158649 | |
|
Anothersignalman
89 posts |
I've now ensured that all trains are at least 2 minutes apart at Brixton, in both directions. Most changes were +/-30sec, a few slightly more than that. I now have about 18 Walthamstow, 8 Brixton and 22 Seven Sisters potential instances of platforms being scheduled for two trains simultaneously. For Walthamstow 1, Walthamstow 2, Brixton 1 and Brixton 2, and Seven 4, how would I go about calculating the minimum time between departure and next arrival in the same road, taking into account signal spacing, point throw time etc all under ARS? Or would it be easiest to just signal a few trains at half-speed simulation, note the times and use those measured values? Log in to reply |
| Draft graph for TT 4.1.4 23/09/2024 at 22:34 #158650 | |
postal
5238 posts |
Anothersignalman in post 158649 said:I've now ensured that all trains are at least 2 minutes apart at Brixton, in both directions. Most changes were +/-30sec, a few slightly more than that.You could just do what happens in real life which often is to give the signaller an imperfect timetable and let them get on with it. “In life, there is always someone out there, who won’t like you, for whatever reason, don’t let the insecurities in their lives affect yours.” – Rashida Rowe Log in to reply |
| Draft graph for TT 4.1.4 24/09/2024 at 06:43 #158651 | |
Peter Bennet
5393 posts |
If there's a train every 90 second do you actually need a timetable? To what extent do the trains keep to the timetable in reality? Peter I identify as half man half biscuit - crumbs! Log in to reply The following users said thank you: sunocske, Anothersignalman |
| Draft graph for TT 4.1.4 24/09/2024 at 22:17 #158660 | |
|
Anothersignalman
89 posts |
Peter Bennet in post 158651 said:If there's a train every 90 second do you actually need a timetable? To what extent do the trains keep to the timetable in reality? It's an interesting question. If you mean within the simulation, it does have an impact when considering which sets run to/from the depot tracks in what order. The same principle applies in real life for fleet maintenance cycles, keeping all sets doing roughly the same route distance on a rotating basis to smooth out wear and tear. The timetable can also be relevant when sets have to attach/divide, or if there are multiple conflict points enroute that need to be balanced; an example that comes to mind is Kensington and Newmarket junctions in Melbourne (Victoria, Australia), where you'd aim to have diverging trains crossing at both, and non-diverging trains also crossing at both. Diagram here - hhttps://victorianrailways.net/signaling/completedia/keness1918.html - so you'd want signals 4 and 23, or 4 and 25, cleared at the same time; and offset by a few minutes, you also want 3 and 15, or 3 and 48, cleared at the same time. And you space the signals, design the turnouts and set line speed at the intermediate automatics to make that scenario most likely. If you mean from a passenger perspective, Ingvardson et al. (2018) p.300 is an interesting reference. The math is over my head, but I *think* it demonstrates that somewhere between 6min and 8min service frequencies, intending passengers change from a mostly-planned arrival at origin stations to random arrivals. This, in my view, refutes the idea that 10min (or 15min!) headways are sufficient to be marketed as "turn up and go", because people don't actually do that. 
Post has attachments. Log in to view them. Last edited: 24/09/2024 at 22:22 by Anothersignalman Reason: None given Log in to reply The following user said thank you: Peter Bennet |
| Draft graph for TT 4.1.4 Yesterday at 10:25 #158663 | |
|
metcontrol
225 posts |
Peter Bennet in post 158651 said:If there's a train every 90 second do you actually need a timetable? To what extent do the trains keep to the timetable in reality?For customers no - it is why apart from on the Metropolitan Line only first and last trains are published. The system is meant to be "turn up and go" though obviously there are many exceptions to this. The Circle Line runs at 10-minute frequencies, which is way outside of people's perceptions of a frequent service when compared to the Victoria Line. In reality yes most definitely. It is one of the Line Controller's main responsibilities to keep the line running on time to the timetable. You have to have a timetable or you would very quickly have disorder. It governs when a driver should depart a station. It relates to each driver's duty and where they should be and when. It might all seem like a random shovelling of trains down the pipe (and sometimes such lines as the Vic can certainly seem it even to those in control) but the timetable has all sorts of things relating to it that keep everything moving as it should. Log in to reply The following user said thank you: Peter Bennet |