VIA Rail

[44 North]

Single track with passing loops works just fine. Germany, Japan, etc, etc even run close to high speed trains over single track sections, with fifteen minute intervals...*both ways*. It means state of the art signalling and control. Hourly schedules would be a breeze for daytime running. Freight could be run temporally at night.

Surely, Canada could learn how to do that? Perhaps not...

Ok, I guess I just assumed we'd have to add 1-2 tracks onto the CP mainline between Malvern and Millwood to not obstruct CP. But you're probably right that sharing the existing two is more than sufficient considering Via's proposed frequencies.

 

[steveintoronto]

I can confirm that many of the Shinkansen stations in Japan have passing tracks in the center. Passing trains just zoom by while the stopping train gets rocked by the pseudo sonic boom. It can be done, they have 5 min headways at times as well

Many thanks! Google tags are a "Zen" and I'm struggling with them today. Have found a map of Germany's ICE though:

https://en.wikipedia.org/wiki/Intercity-Express

I'm still searching, but found this paper on High Speed:

[...]
This gives us the total time required for a train to clear a station platform and allow the
following train to approach without being offered a restrictive speed command. It adds
up to 341 seconds or 5 minutes 41 seconds. In capacity terms, this equates to 10.5
trains per hour.
And don’t forget, we must add our operating margin to this. Now our capacity drops to
8.5 trains per hour. This doesn’t seem a lot but we can improve on it by using passing
loops at stations.

Rule 10: Build 4-‐track stations

The obvious solution to the problem of the delays caused by station stops is to divert
stopping trains off the main line into a loop track built specially for the platform. One of
these loops on each side of the line will give a 4-‐track station layout. It’s a common
solution on high speed lines and it’s used on conventional routes too. Let’s see what
happens to the throughput if we try it.
The best approach would be the one where the stopping train diverges from the through
line at full speed but this is unlikely to be viable. No one has managed to engineer a
turnout to take a speed higher than 200km/h and even that was very expensive. As we
mentioned in our discussion on junctions (Rule 6), the operational maximum is
generally reckoned to be 160km/h, just over half our line speed of 300km/h.
If we assume a possible headway of 180s (from Rule 5), a train slowing for a diverging
route will increase our time by 16s to 196s. Effectively, this is the additional run-‐in time
for the station. If a non stop train is following, the closest it could get is 196s.
The run out time will also be limited by the converging turnout. If we assume an
average acceleration of 0.3m/s2, a speed of 160km/h is reached 3300m beyond the
station. This leaves 8200m to get to full speed and it will take 130s to do it. This is 32s
more than it would take to cover this distance at full speed. It’s double the time
required for the run in, so we can discount the run in as far as the overall headway is
concerned but we have to add the run out time to the 180s full speed headway, and this
will give us just over 3½ minutes. Add the operating margin and we get 4½ minutes.
Our capacity is now just over 13 trains per hour.
Just for the record, to get the maximum benefit of a 4-‐track station, the section between
the entry turnout and the exit turnout must be at least 2000m long on the run in side
and 3300m on the run out side.
[...]

http://webcache.googleusercontent.com/search?q=cache3IEynrRT3MJ:www.railway-technical.com/Infopaper%203%20High%20Speed%20Line%20Capacity%20v3.pdf+&cd=4&hl=en&ct=clnk&gl=ca&client=ubuntu

It doesn't directly answer the point of opposing headways using single track, but the methodology would be the same, just compounded.

Still searching for the ICE case(s).

Edit to Add:

Design of single-track rail line for high-speed trains
. Author links open the author workspace.E.R.Petersen†. Numbers and letters correspond to the affiliation list. Click to expose these in author workspace. Author links open the author workspace.A.J.Taylor†. Numbers and letters correspond to the affiliation list. Click to expose these in author workspace
School of Business, Queen's University, Kingston, Ontario, Canada
https://doi.org/10.1016/0191-2607(87)90023-9Get rights and content
Abstract
This paper presents a method for designing a single-track rail line for a reliable high-speed passenger train service. We first consider deterministic train performance and describe a general method for finding the best location and length of the passing tracks. The design is then modified to include slack, which is necessary to ensure on-time performance for trains encountering unexpected delay. An analysis of the robustness of the system to small delays is presented. Robustness to large train delays and slower traffic is incorporated through the provision of additional sidings. Finally, simulation results are presented which compare the performance of a single-track line to a fully double-tracked equivalent.

[...]

The authors gratefully acknowledge the support of the Natural Science and Engineering Research Council of Canada and VIA Rail Canada Incorporated for their support of this study.
Copyright © 1987 Published by Elsevier Ltd.
http://www.sciencedirect.com/science/article/pii/0191260787900239

Attempting to gain entire paper.

Edit: Author's publisher has approved I get full access to the above paper. Have confirmation, will read and quote sections later.

Meantime, finding a lot more on-line on this:

An Alternate Double–Single Track Proposal for High-Speed Peripheral Railway Lines
First published: 20 May 2014Full publication history

• Abstract
  This article presents alternate double–single track (ADST) lines as an alternative to double-track lines. The idea consists of using single track where the infrastructure is very expensive (tunnels and viaducts) and double track where it is cheaper (smooth orography) combined with small changes in departure times so that trains may cross in the double-track segments with no reduction in travel times. The solution is shown to be very efficient for traffic demands between 30 and 40 trains per day and the costs are reduced substantially (close to 40%). A linear programming program is given that (1) decides the optimal sequence of single and double tracks and (2) optimizes the timetables for the optimal or other alternative sequences. The Palencia–Santander line is used to illustrate the proposal and some suggestions are given to use the ADST lines in several countries. [...]

• http://onlinelibrary.wiley.com/doi/10.1111/mice.12083/abstract

 

[crs1026]

Ok, I guess I just assumed we'd have to add 1-2 tracks onto the CP mainline between Malvern and Millwood to not obstruct CP. But you're probably right that sharing the existing two is more than sufficient considering Via's proposed frequencies.

It's far from certain that CP would entertain any encroachment on its double track main line. The situation is really no different than CN in Halton (except that ML has been talking about using that line for so long, you'd think they'd have hit the impasse years ago. Still not clear why ML was caught by surprise so recently on that). If VIA plans to electrify, the catenary will conflict with the vertical clearances for double stack carriers and potentially auto carriers also. I would bet a box of donuts that CP's response would be "no bloody way".

You want the first few miles out of the terminal to be double track, so that a late incoming train does not propagate the delay onto outgoing trains. Perhaps meeting points at Kennedy and Leaside would be sufficient, leaving the Don bridge as single track, and requiring only one new track between Kennedy and Leaside. A quick peek at Google Earth says that many of the grade separations along that stretch are only as wide as the existing double track, so new overpasses at these would be needed.

Single track in the intermediate territory is a much easier proposition.

- Paul

 

[steveintoronto]

The ostensible plan discussed by some is to tunnel or bridge over the CP yard. (Different yard, but same approach: http://forum.skyscraperpage.com/showthread.php?p=7507728 ) Not cheap, but far cheaper than any other route option proposed so far. Needless to say, with CP given access to a higher speed route for premium express freight, and some overhead traffic, and the present line to Peterborough vastly improved, they'd be amenable to talking an arrangement.

I found the note my friend left me on the ICE:
[...]
ICE, Braunschweig-Wolfsburg (One ICE and one RE (Regional Express) train per hour and direction):

View attachment 114868

This is the timetable with 6 trains per hour and direction (the orange trains overtake the preceding train at Uji) on what is a single-tracked line between Momoyama and Uji (all stations equipped with passing loops, but single-tracked sections are up to 2.2 km long):
View attachment 114869

[...]
He had drawn a map which I can't find at this time, but the former attachment above displays it circled on the official map. I'm still looking for more examples.

 

[crs1026]

Steve, thanks for the links. Interesting stuff, and actually understandable.

I wonder what kind of "buffer zone" TC would want for two high speed trains approaching a short siding for a meet. I bet they would want fairly conservative braking distance parameters. The longer the braking distance assumption, the longer trains are delayed at a meet or the longer the sidings must be.

Hourly service at 110 mph on a rigid schedule only needs one meeting point every 55 miles. Put a further meeting point 10 miles on each side and you have contingency for a train that is 6 minutes late. Put another meeting point 20 miles on either side and there is contingency for trains 12 minutes late, or a following train running 12 minutes behind the first.

The challenges in the Canadian setting (besides getting timekeeping that much more reliable) would be keeping switches unfrozen and ice-free in winter, and possibly implementing higher speed turnouts than what we have now.

- Paul

 

[steveintoronto]

The challenges in the Canadian setting (besides getting timekeeping that much more reliable) would be keeping switches unfrozen and ice-free in winter, and possibly implementing higher speed turnouts than what we have now.

It would be a challenge, but Japan also gets very cold, as do the Nordic countries. I think it's also safe to assume that given the Havelock Corridor upgraded, and grade separated, at least in the greatest middle section, TC would allow a considerably higher speed than 110 mph. If electrified especially, this would not be hard to achieve. God help us, it would also behoove TC to allow European (or Asian) off-the-shelf *proven* trainsets to run this route.

As a side note, I'm now leaning towards the belief that HFR can be financed sooner, with far less hurdles and 'ifs and buts' and compromise by Private Enterprise, and with corps like CP eager to participate ...and at this rate, possibly/probably outside of the Investment Bank, which is forthcoming, I have no doubts, but *when?*

For my rationale on this, and the impetus to do this, read the Moose Rail string. LeMine is big, with apparently very deep pockets. And yet tiny compared to other Chinese venture capitalists looking for investment here. And they love rail.

More on this later.

I wonder what kind of "buffer zone" TC would want for two high speed trains approaching a short siding for a meet.

The more I think about this, the more a flyover would be the cheapest way to do it, judging by the Davenport Diamond experience. The question would be "one track or two?" Normally the grades to do this in a tight envelope would be a challenge, but being electrified eliminates a lot of that challenge allowing much steeper grades. For the freight the Havelock Line might carry, it can cross through the yard and join the line further north. Since there would be a spur from the yard joining the line, there's even the opportunity for premium express freight to be electric loco hauled.

 

[nfitz]

Of all the problems in the VIA proposal, the CP tracks in Toronto are the least of the issue. Presumably they'd just run two new tracks parallel, and cross under/over somewhere, the same way there is that tunnelled cross-over just west of Union station. I don't see any show-stopping issues - the width is 75 feet or greater in most places from Leaside to Agincourt, though it narrows to 70 feet in a couple of small places, but there's nothing on the adjacent lands.

5 straight mainline tracks would require 70 feet minimum. So no reason 2 tracks for VIA wouldn't fit.

 

[kps]

[…] If VIA plans to electrify, the catenary will conflict with the vertical clearances for double stack carriers and potentially auto carriers also.

So put the catenary higher. GO peak capacity would also benefit from filling out the available loading gauge with two full levels instead of one and a half.

 

[smallspy]

So put the catenary higher. GO peak capacity would also benefit from filling out the available loading gauge with two full levels instead of one and a half.

You can only go so high before you run into structures. That said, the height of the catenary in an of itself isn't really much of a restriction in the grand scheme of things - double-stacks and autoracks are regularly run under a number of portions of the North East Corridor, for instance.

For the record, GO's BiLevels are as high as you could safely run considering the clearance restrictions of electrifying the Union trainshed.

Dan
Toronto, Ont.

 

[steveintoronto]

CP wouldn't be running double stack on the Havelock line, just as premium express freight runs on the NE Corridor aren't. CP could avoid the high costs assessed to freight operators on the NEC by being part of the consortium that builds and runs the Havelock route, or making a deal up front for the purchase/acquisition of the extant RoW for rate consideration later over the entire run to their existing track near Smith Falls.

When running next to the extant CP line near Smith Falls, a parallel track would be built. There was a track there prior, so room exists last time I Google Satellite'd the line. At Yards, best to dive under or bridge over so double stack isn't an issue. This is still far less imposition than doing same along the the lake route to Montreal, and up to Ottawa from there.

 

[dowlingm]

when considering ROWs and structures where VIA does not own, bear in mind that wider spacing between track centres may be insisted upon as CSX did for Amtrak's plan for expansion west of Albany
https://www.google.ca/amp/www.times...ses-high-speed-rail-on-its-tracks-5571510.php

 

[steveintoronto]

when considering ROWs and structures where VIA does not own, bear in mind that wider spacing between track centres may be insisted upon as CSX did for Amtrak's plan for expansion west of Albany
https://www.google.ca/amp/www.times...ses-high-speed-rail-on-its-tracks-5571510.php

That's a very good insight into how CN and CP look at VIA wanting to share RoW, let alone track, save that even the *State of NY* exacts greater control over the railroads than TC/CTA does in Canada!

This statement is intriguing:
[...]
It (CSX) favored an alternative that would see top speeds reach 125 mph, but only because it would require an entirely new rail line separate from CSX's existing corridor.
[...]

The Water Level Route is diesel with absolutely no sign of being electrified. So this article being three years back, what did CSX have in mind for diesel running that fast? Talgo? The Siemens Charger was just a twinkle in American eyes at that time, albeit the Hudson saw the Turboliners https://en.wikipedia.org/wiki/Turboliner
which had a top speed of 125 mph, but they disappeared years before this story. So was CSX just ranting? Curious...

 

[crs1026]

The issue with double stack isn't that it isn't feasible, it's that CP and CN have both said they aren't interested in wires over their lines. Check out the RER BCA documents. Metrolinx has encountered this already. Simply for operational reasons, I can't see CP allowing high frequency passenger on its freight line. Hourly two way service would pretty much tie up one main line completely, and CP won't go for that. VIA will need to build its own track(s).

There is plenty of right of way width for more trackage - but check out the overpasses at Lawrence Warden Ellesmere GO Midland Sheppard Brimley and McCowan. These may be limiting.

Nothing that can't be solved with money, but the envelope will only stretch so far.

- Paul

 

[steveintoronto]

There is plenty of right of way width for more trackage - but check out the overpasses at Lawrence Warden Ellesmere GO Midland Sheppard Brimley and McCowan. These may be limiting.

Nothing that can't be solved with money, but the envelope will only stretch so far.

And the alternative is? I can't speak for others, but it's understood from the get-go that CN and CP won't allow catenary over their track. How does that make the Lakeshore route, for instance, any easier a choice?

There's two choices here: Getting your boots wet, or swimming across the lake.

 

[reaperexpress]

Single track with passing loops works just fine. Germany, Japan, etc, etc even run close to high speed trains over single track sections, with fifteen minute intervals...*both ways*. It means state of the art signalling and control. Hourly schedules would be a breeze for daytime running. Freight could be run temporally at night.

Surely, Canada could learn how to do that? Perhaps not...

The challenge is that the most trains on the Toronto-Ottawa route will have hundreds of kilometres between stations, which means that trains will mostly need to meet between stations. And these sidings would need to be extremely long in order to achieve the ambitious average speed claims.

To operate a single tracked line reliably, you need to build in buffer time to allow one train to be slightly late without delaying the other. Roughly speaking, if your double-tracked segment takes 2 minutes to cross, trains need to be within +/- 1 minute of schedule, otherwise they will delay the train going the other way. In that example where trains need to be within a minute of the schedule, you would need the schedules to be very generous, to minimize the chance of a train being late. That means that you need to decrease the scheduled average speed. But VIA's claimed travel times are extremely ambitious, which implies they assume only minimal schedule padding (or more likely, "forgot" to pad the schedule at all in order to make the numbers more exciting). With minimal schedule padding, there is a high likelihood of any given train being late, which means that the sidings need to be very long to avoid cascading delays. So perhaps we need 10-minute sidings, to allow up to 5-minute delays without cascading effects. At 110 mph, 10 minutes is 30 km. And with hourly service, you need passing points every 30 minutes, or 90 km (i.e. 1/3 of the line needs to be double-tracked).

In most the places you mentioned, the passing points tend to be at stations, which works well since that's where trains are travelling slowest (i.e. the highest time spent per track distance). So the passing places can be as short as a few hundred metres, yet naturally provide a fair amount of time within the siding (then usually that time will be artificially increased by scheduling a longer-than-necessary stop in that station). It also means that there's no lost time due to switching tracks: the trains are travelling slowly anyway to stop at the station.

So the bottom line is that bidirectional single-track running is definitely doable, and indeed it is currently employed on the VIA services between Ottawa and Montreal, as well as the off-peak GO train services to Aurora and Unionville, the Ottawa O-Train and countless other lines in Canada. It just means that trains will need to be as punctual as required by the length of the sidings, which in turn means that extra time will need to be added to the schedules to accordingly limit the likelihood of a train running behind schedule. And the HFR travel time figures are so ambitious that I find it hard to believe that they include this extra time that will need to be added to schedules in order to actually operate a reliable service in the real world.