|
|
| Re: SAS termini questions (127346) | |||
 |
|||
| Home > SubChat | |||
|
[ Read Responses | Post a New Response | Return to the Index ] |
|
||
|
Re: SAS termini questions |
|
|
Posted by stephenk on Mon Aug 15 17:47:09 2005, in response to Re: SAS termini questions, posted by Stephen Bauman on Mon Aug 15 15:04:56 2005. "There is a window by which trains will not encounter any delays. I also showed how to calculate this window, based on the time it takes to traverse the crossover and the headway. It makes most sense to make sure that all trains operate within this window."But realistically, how many metro trains run within seconds of their schedule during the peaks? "There's a safety concern here. The signal system must be able to stop a train before it hits the crossover. This distance is essentially the same for a block system and systems with speed controls or ATO. The signal system was also designed for degraded operation - the possibility that the brakes might be disabled from the lead car. The IND designers never thought that their successors would be so stupid as to introduce operating equipment were a single brake system would control more than one car. Two passengers were killed at Roosevelt Ave because of this." With many modern ATO systems the trains can move up to about 50m from a crossover which is set against that train. However for safety reasons the train will move very slowly up to that mark so as to not compromise the overlap. One of the problems with Brixton crossover on the Victoria Line in London is that the ancient ATO system can only run trains at approx 0,25, and 50mph. Thus the overlap on approach to the crossover is well over 100m (for a 25mph overrun). Because of this a train waiting to get through in through the crossover (which is a pretty much guaranteed occurance during the peaks!) takes over 15secs to even reach the crossover once it route has been selected. This is one of the reasons why it takes 95secs+ to get a train out and in to a platform at Brixton despite a high speed (50mph) crossover. At Elephant&Castle on the Bakerloo Line, there is a 20mph speed control signal ahead of the signal protecting the crossover, thus the overlap is quite short (at a guess around 60m). The 20mph crossover is also very compact compared to the crossover at Brixton as the curve radii are tighter. Both of these factors allow for just 80-85secs needed to get a train out out and into a platform there, despite having a much lower speed crossover than Brixton. "Try to get some perspective on how much a factor the switch is. Let's assume it is 200 feet long (including signal protction) and that we are dealing with 600 foot long trains. This means that a train has to traverse 800 feet to enter and leave the switch. If this train takes 60 seconds that means an average speed of 9 mph. Let's assume that the average speed is to be increased to 15 mph - hardly a "high speed" crossover. This reduces the 60 second traversal time to 36 seconds." Not quite, as the more shallow curve radii of the higher speed crossover would lengthen the crossover, and thus the train would have to travel further to clear it. "How much margin did those Flushing trains have? They operated at 100 second intervals and spent 72 seconds in the crossover. That gave them a margin of 28 seconds. So each train was expected to be within 14 seconds of its schedule. That is reasonable. That is what Moscow does." However Moscow uses a different method of reversing trains, which is realistically better at running high tph's. Some of this may be because reversing in platforms always suffers from time being wasted from the departing trains getting the green light, to actually departing. By the way, is the reason why the Flushing Line now only runs 27tph, due to less demand, or TA cost cutting? |
|
|
|