Page 2 of 2
| When is a train "on-time" 08/03/2013 at 22:03 #42188 | |
benstafford
88 posts |
At Redhill they report departure times to the second through radios. Where does the info go on this. I travel on a train each day that was always between 1 and 5 minutes late. They retimed it in December and it's now to the minute 9 times out of 10.
Log in to reply |
| When is a train "on-time" 08/03/2013 at 22:58 #42193 | |
|
rodney30
165 posts |
on the southern railway network (ie redhill) the put the train departure times into a program called (sooti) it stands for station occerance (wrong spelling i think) over time something. it is all to do with delay attribution. u have to put the reason for any delay e.g. waiting signal, waiting train crew, assisting passengers etc.
Log in to reply |
| When is a train "on-time" 09/03/2013 at 00:34 #42198 | |
|
AndyG
1842 posts 
|
What about that train in Kent last night? Delayed due to an early arrival........ mother and baby doing well. I can only help one person a day. Today's not your day. Tomorrow doesn't look too good either. Log in to reply |
| When is a train "on-time" 09/03/2013 at 13:33 #42205 | |
|
metcontrol
227 posts |
" said:" said:I can remember Royston-Ashwell being 4, Ashwell-Baldock being 4, Baldock-Letchworth 2, Letchworth-Hitchin being 3 - a total of 13 minutes compared to todays average 17-18 minutes. There is a long-term fault with the train reporting between Royston and Baldock, which also means for large chunks of time the platform indicator at Baldock simply shows "Fast Train Approaching." Connected with this fault, it is not uncommon for trains to "lose" 2 minutes between Ashwell and Baldock (I am basing this on where the train "should" be and the actual arrival of the service). I have noticed the same data is used on the NR / FCC websites to predict train running information. As an aside, but connected to the thread as a whole, I have noticed that certainly as far as platform indicators are concerned, there are no "delays" to any service until a train is 2 minutes late. From then on, until a train is 4 minutes late, it continues to show as 2 minutes late (or On Time sometimes). And so on. Noticed this yesterday at Stevenage. My train arrived 1½ minutes late, and we were shown on time. After a slightly lengthy dwell, the indicator updated to "2 minutes late." I have also seen trains be shown as left Hitchin 5 late, with an expected arrival time as "on time" at Letchworth - with only a 4 minute run time given. Log in to reply |
| When is a train "on-time" 09/03/2013 at 15:07 #42212 | |
guyh
54 posts |
I think the combination of responses has pretty much covered it all. TRUST is the official source of information. It does indeed record most of the network's movements based on TD movements. For all timing locations where trains are to be reported, the various related / nearest TD movements are used, with a "berth offset" in seconds - used to be whole minutes until sometime around 10 years ago. So, having a quick glance at the Exeter sim: the berth movement 0060 to 0062 will be set up for Exeter St Davids departure with a negative offset of not very many seconds; but also set up for reporting arrival at Exeter St Thomas with a positive (and much higher) number of seconds. The same will be set up for 0360 to 0062, 0260 to 0062, 0160 to 0062, and of course each might have very slightly different values. At the other end, 0137 to 0041 etc etc will be set up but only for EXD departures. TRUST can only report the times against a train schedule, so if a train doesn't stop at Exeter St Thomas, and also doesn't have a passing time in its schedule, then TRUST will discard that arrival time. If it passes and has a passing time in its schedule, then TRUST would show "At EXETSTTOM hh:mm" until the berth movement 0062 to 0196 reports a departure time, and then the departure report will be used as the passing time: "pass EXETSTTOM hh:mm". Exeter St Thomas happens not to be a compulsory timing point, so it would be rare to see a timing in the schedule and therefore rare to see the pass report in TRUST - this is just an example to explain how it works! I happen to have recently been investigating an issue Seer Green and was sent the berth offset values: 0126 – 0124 is +60 (ie when the train passes ME126, it takes on average 60 seconds to come to a stand in the platform at SRG). 0124 – 0120 is -80 (ie when the train passes ME124, on average it will have started moving from SRG 80 seconds ago). 0123 – 0129 is +47. 0129 – 0131 is -99. Each location has a varying importance to getting the timing spot on, and will be audited by NR's data quality people at a frequency appropriate to its importance (listed below higher to lower priority): - Monitoring points: these are locations where the lateness of the train is monitored in at least one TOC's schedule 8 (NR / TOC performance contract). Contrary to what I said above - if a TRUST report is generated for one of these points, it gets noted in the TRUST schedule even if that location doesn't appear in the schedule (gets shown as an off-route report at the bottom of the schedule). - Delay recording points: these are locations where the lateness of the train is compared to the lateness at the previous delay recording point, and a TRUST delay is created if the difference in lateness is 2 mins or greater (class 1's) or 3 mins or greater (all other classes). - Information: Any other location which is set up to report in TRUST, but where the data isn't generally used for anything contractual / delay reporting etc. In reference to Peter's case, it is the offsets which often cause bang-on-time trains to report 1 early. If on average a train takes 25 seconds to make the TD berth movement, that is probably an average of values ranging around 23 seconds to 27 seconds. If the train leaves at 07:00:00 and the berth movement happens at 07:00:24, then TRUST will report as 06:59 (truncated from 06:59:59). Taking a station like King's Cross, a 12-car train that starts near the signal is going to get to the signal considerably earlier than a 4-car that has to come all the way down the platform. This is a particular flaw of using the exact numerical average that I'd recognised but never considered debating with NR; I'm not sure I want to ever have the debate! In terms of when is a train on-time? It depends on the context. Punctuality data is often reported in PPM terms, which has been correctly described already. Of course, this is an average of all trains at destination, so doesn't necessarily bear any relationship to the individual's journey which often won't be to the terminating point of the train, and will usually only be on one or two trains a day, not several hundred! At Chiltern, we have many trains which have an individual PPM of 100%, but we regularly monitor our worst performing trains - these do not have a record that matches our average PPM (around 94%) - they are substantially away from figures beginning with a 9!! We also monitor and report (internally - there is growing pressure for the industry to report publicly as well) the right time statistics - ie the percentage of trains that arrive on time or early (uses TRUST reporting, so the truncating effect still allows you to be 59 seconds late). There is a very interesting article by Chris Gibb (I think it's in this month's Modern Railways) about the subject of what we should be trying to achieve - and what it is sensible to wish for! Data only stays in TRUST for 10 days from the date of operation, but TOCs and NR can download data via the PHIS (Performance Historical Information System) - those with TRUST access can just send the command "PHIS" to see this, although only very few #-sign on codes give access to the menus after that! The downloaded running and delay data is then read into various databases (Chiltern use Compass Delays, a lot of TOCs use Bugle, I suspect there are others), from which we can run reports to investigate trends. Whilst most think that TRUST and delay attribution is largely a financial game, there's actually (certainly in my TOC anyway) more people analysing data and dealing with performance improvement than there are involved in doing the financial bits! If anyone wants a basic description of how schedule 8 works, please ask - it'll probably take me an hour or two to write up the basic description! There is also a separate national database of train running called DARWIN, run by the National Rail Enquiries team at ATOC. It also takes the TD stepping data, and backs this up with TRUST reports (mainly to fill in for those few areas where signallers still manually report the train movements in TRUST (those with TRUST access: bring up a TSID report for a location and type "R" next to a train and you'll end up on the movement reporting screen - although unless it's part of your role, your input will be ignored). DARWIN also makes a pretty good stab at predicting when a train will get to every location further down the route. Any online system you use to look at train running (apart from TRUST) will be using DARWIN data, whichever TOC's website you use. I say that, but Ben (sacro) might tell me there are other sources!! The predictions are limited to normal running times of a train - DARWIN can't spot that a train is following another service, so if you're on a non-stop behind a stopper, DARWIN will predict you'll leapfrog it, although as you make each berth movement, the prediction will update to show you getting a bit later. DARWIN is heuristic - it records the time it actually takes trains on average to go through signal sections and stop at each station, and learns as it gets each new report. So, if a damaging speed restriction is applied somewhere, DARWIN will learn that it is taking longer to get from A to B and predictions for future trains will gradually build up to the longer value. It's quite interesting to watch it when there are diversions - for example Cross Country having to run via Whitacre instead of the direct lines between Kingsbury and Water Orton, after a couple of trains it will "correctly" show the expected delay on following trains. The staff versions (such as the ones vontrapp showed) show the full schedules including passing times, and consideration is being given to building freight and empty stock moves in as well, so that font line staff can see an equivalent picture to what can be viewed via TRUST. The public versions (eg live departure boards on NRES) only show the stopping points of the train, and therefore it's slightly less obvious what is going on. As it knows how long each train should be in each signal section, it can also recognise when a train is being delayed. As soon as the train has not reported 5 mins after DARWIN would expect to have seen a report / TD movement, it will revert the train to "Delayed" - then when it gets the next report / TD movement it will recalculate when the train will get everywhere. Virgin's and Chiltern's Customer Info Systems (platform screens) are linked to DARWIN - whatever DARWIN says is what will be on the platform screen (for delay values and any changes to stopping patterns). Most other systems have their own prediction mechanisms - the odd few TD movements might be used close to a station, and most systems can grab the latest reports from TRUST, however they're generally not very good at spotting that a train is *being* delayed, and only update when the next recognised TD movement or TRUST report takes place, which might be the arrival of the train at the station you're waiting. Hope that's useful. Happy to answer any questions! Guy. Log in to reply The following users said thank you: postal, GeoffM, Peter Bennet, Tempest Malice, AndyG, guidomcc, JamesN, GoochyB, sorabain |
| When is a train "on-time" 09/03/2013 at 20:29 #42223 | |
jc92
3685 posts |
" said:I think the combination of responses has pretty much covered it all.how does this work in areas without TD berths? for instance on the hope valley, does a train report its status at totley tunnel easts section signal (end of TCB) then not update until chinley norths home signal (recommencment of TCB) or is there an alternative arrangement? similarly with lines like Pbro - Leicester and the harrogate loop. Joe "We don't stop camborne wednesdays" Log in to reply |
| When is a train "on-time" 09/03/2013 at 20:57 #42225 | |
|
Late Turn
699 posts |
In (most?) AB areas, and others without TDs, TRUST reporting is done manually by Signalmen - see the reference to TSID + "R" in guyh's excellent and informative post (thanks for taking the time!), although it's more readily done from a TRJA or TRJS lineup (or TRJB for an off-route report). On our line, eight out of the nine boxes report trains in this way. Tom Log in to reply |
| When is a train "on-time" 09/03/2013 at 23:09 #42231 | |
|
guyh
54 posts |
As Tom says, the signallers report manually. Below is a report from a Manchester to Sheffield stopper via the Hope Valley. The timings on the right are the actuals - each time that has "A" next to it has been reported automatically by the signalling system (ie TD berth movements), and those with "M" have been manually reported, probably by signallers in the relevant boxes. Location Booked Arr Dep 32000 MANCR PIC 08:45 08:50A 5 LATE X LINE 32009 ARDWICKJN PASS 08:47 08:52A 5 LATE E LINE 32050 ASHBURYS PASS 08:48 08:55A 7 LATE B LINE 32063 REDDISH N 08:52 08:52 N/R 08:59A 7 LATE 32064 BRININGTN 08:55 08:56 09:02A 09:02A 6 LATE 32065 BREDBURY 08:58 08:59 09:05A 09:06A 7 LATE 32552 ROMILEY 09:02 09:02 09:08A 09:08A 6 LATE 32551 MARPLWHFJ PASS 09:04 09:10A 6 LATE R LINE 32553 MARPLE 09:05 09:06 09:11A 09:12A 6 LATE 32554 STRINES 09:09 09:10 N/R N/R 32555 NEWMLSCEN 09:13 09:13 09:19M 09:20M 7 LATE 32557 NEWMLSSJN PASS 09:16 N/R 34001 CHINLEY 09:20 09:21 N/R N/R 34002 CHINLYNJN PASS 09:23 09:29M 6 LATE 33998 CHINLYEJN PASS 09:23 N/R 34005 EDALE 09:29 09:30 N/R N/R 34009 EARLES SB PASS 09:34 09:41M 7 LATE 34010 HOPE DY 09:35 09:36 N/R N/R 34011 BAMFORD 09:38 09:39 N/R N/R 34015 HATHERSGE 09:42 09:42 N/R N/R 34022 GRINDLEFD 09:45 09:46 09:53M 09:54M 8 LATE 25710 TOTLEYTUN PASS 09:53 09:59A 6 LATE S LINE 25708 DOREWESTJ PASS 09:53 N/R 25704 DORE 09:54 09:57 N/R N/R 25705 DORESTNJN PASS 09:58 10:02A 4 LATE 25701 SHEFFIELD 10:04 10:07A 3 LATE [Edit - if anyone knows how to specify a font, please let me know - it looks rubbish in the displayed font!] Last edited: 10/03/2013 at 09:19 by guyh Log in to reply |
| When is a train "on-time" 10/03/2013 at 02:53 #42233 | |
|
Danny252
1461 posts |
Try the code tags:(Presumably the forum has also stripped out multiple spaces too  )
Log in to reply |
| When is a train "on-time" 10/03/2013 at 11:48 #42238 | |
headshot119
4869 posts |
" said:As Tom says, the signallers report manually. Below is a report from a Manchester to Sheffield stopper via the Hope Valley. The timings on the right are the actuals - each time that has "A" next to it has been reported automatically by the signalling system (ie TD berth movements), and those with "M" have been manually reported, probably by signallers in the relevant boxes.it might be worth editing your post, and wrapping it in "code" tags, that should preserve the multiple spaces as well. [CODE] Location Booked Arr Dep 32000 MANCR PIC 08:45 08:50A 5 LATE X LINE 32009 ARDWICKJN PASS 08:47 08:52A 5 LATE E LINE 32050 ASHBURYS PASS 08:48 08:55A 7 LATE B LINE 32063 REDDISH N 08:52 08:52 N/R 08:59A 7 LATE 32064 BRININGTN 08:55 08:56 09:02A 09:02A 6 LATE 32065 BREDBURY 08:58 08:59 09:05A 09:06A 7 LATE 32552 ROMILEY 09:02 09:02 09:08A 09:08A 6 LATE 32551 MARPLWHFJ PASS 09:04 09:10A 6 LATE R LINE 32553 MARPLE 09:05 09:06 09:11A 09:12A 6 LATE 32554 STRINES 09:09 09:10 N/R N/R 32555 NEWMLSCEN 09:13 09:13 09:19M 09:20M 7 LATE 32557 NEWMLSSJN PASS 09:16 N/R 34001 CHINLEY 09:20 09:21 N/R N/R 34002 CHINLYNJN PASS 09:23 09:29M 6 LATE 33998 CHINLYEJN PASS 09:23 N/R 34005 EDALE 09:29 09:30 N/R N/R 34009 EARLES SB PASS 09:34 09:41M 7 LATE 34010 HOPE DY 09:35 09:36 N/R N/R 34011 BAMFORD 09:38 09:39 N/R N/R 34015 HATHERSGE 09:42 09:42 N/R N/R 34022 GRINDLEFD 09:45 09:46 09:53M 09:54M 8 LATE 25710 TOTLEYTUN PASS 09:53 09:59A 6 LATE S LINE 25708 DOREWESTJ PASS 09:53 N/R 25704 DORE 09:54 09:57 N/R N/R 25705 DORESTNJN PASS 09:58 10:02A 4 LATE 25701 SHEFFIELD 10:04 10:07A 3 LATE [/CODE] "Passengers for New Lane, should be seated in the rear coach of the train " - Opinions are my own and not those of my employer Last edited: 10/03/2013 at 14:12 by headshot119 Log in to reply The following user said thank you: guyh |