Welcome to European Tribune. It's gone a bit quiet around here these days, but it's still going.
Display:
For the terminal station, the question is the number of services that might start or end at the station.

Let's look at CHSRA's figures in this projection (page 5/pdf page 9). It's only for Phase 1, but at full capacity, and foresees a more complex service pattern than you (one more comparable to SNCF's and JR East's service pattern than say RENFE's). The table for peak hours is formulated a bit confusing about frequency, so I re-calculated it for every two hours and every direction:

  1. (express with multiple stops only in the megapolises): 1 train
  2. (express with stops only in 4 main stations): 2 trains
  3. (semi express, with more stops in the Central Valley): 1 train
  4. (semi express only to LAUS, with evenly distributed stops that include Sylmar): 4 trains
  5. (semi express only to LAUS, with evenly distributed stops that include Palmdale): 4 trains
  6. (semi express, with more stops nearer to LA): 1 train
  7. (all-stations SF-Merced): 3 trains
  8. (all-stations LA-Merced): 3 trains

That's 16 trains from TBT every two peak hours, or 8 every hour.

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Mon Mar 23rd, 2009 at 11:07:12 AM EST
Thanks for that. I think there is likely to be more ridership than the CAHSR schedule contemplates in the origin/destination pairs in the "tail" of the ridership distribution, especially in the CV, so put more emphasis than they did on building up the all-stations frequency ... but yes, going from a two-tier to a three tier Express, Limited, and all-stops would also add tph.

All of which says that an upper ceiling of 6tph is awfully low ... and the incremental cost of expanding the box, that is, the addition to the cost from doing it in two stages, can easily be greater than the original cost of building the three-island box in the first place.


I've been accused of being a Marxist, yet while Harpo's my favourite, it's Groucho I'm always quoting. Odd, that.

by BruceMcF (agila61 at netscape dot net) on Mon Mar 23rd, 2009 at 11:31:54 AM EST
[ Parent ]
After reading the rest of your diary, I am very much puzzled.

  • What is the (expected future?) train frequency on Caltrain?

  • If only two platforms are for departing trains, that means that just to support 8 tph, maximum platform waiting time is 15, not 30 minutes. (For 12 tph, just 10 minutes.)

  • (Sidenote: a 15 minute period at each departure platform allows for a 40-45-minute train turnaround with uniform 15-minute cycles at each of two arrival, tail, and departure tracks. A 10 minute period at departure platforms would require four tail tracks operating in a - perhaps more realistic - 20-minute cycle.)

  • I must ask, what is the planned curve radius? For, curved platforms even where you have cars is not at all unheard of, so I am not sure why they require straight platforms. (Below: Cologne main station, and my own photo at Le Mans)


* Those tail tracks must be pretty long already to accomodate 400m trains. Has no genius considered a loop-back through station, at least for the Caltrain part?

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Mon Mar 23rd, 2009 at 11:47:47 AM EST
[ Parent ]
In reverse:

* Yes, a loop-back through station was originally a design option, it seems to have died somewhere early in the process, pre-2003. The tail tracks are shallow, cut-and-cover tunneling, I believe they will probably have had to bore to get a loop-back ... the loop-back would be extending well beyond the property being re-developed for the TBT project, so that would be much more expensive eminent domain for cut-and-cover on the turn.

And, yes, the original design spec was to stable four Caltrain trains, two per track ... they tend to run fairly short, but if they get to ten cars EMU, there's the capacity to hold a single full-length HSR set. AFAIU, they run the tail track through to a foundation of a freeway.

  • I am unclear about the CHRSA demand for 400m of straight platform ... if I was to guess, I would guess that they can not provide detailed specification of what would be an acceptable platform curve radius for the extremes of the platforms (the middle portions are straight in any event because of structural support columns extending through the box to foundations beneath) and so are laying out an ambit claim that is on the safe side.

  • On turn-arounds (that is, the Sidenote), once I hit arrival, tail, departure station and platform dwells that seemed comfortable to me as maximum allowables, I didn't press it any more. There's nothing I have heard about maximizing tunnel track throughput, and with 3 minute headways, that seemed to me on the access track to be 4 service slots, alternating between Caltrain and HSR, an open slot, repeated four times.

The alternation is because of the trainbox throat track design ... with independent egress track there is no need for Caltrain access to crossover HSR egress, and visa versa, so alternating HSR and Caltrain access means there is an egress slot for each service immediately preceeding each access.

* On Caltrain, I think its about 6 tph, but on the other hand, there are other standard gauge regional rail services being pushed for, and they would of course want to connect to Caltrain as well as to the intermodel (Bus at the TBT, BART a block away) transfers there.

The present Caltrain terminus at 4th and (mumble) would continue to be a Caltrain station, so if Caltrain hits capacity they can always start alternating between services terminating at the current terminus and services terminating at the TBT.


I've been accused of being a Marxist, yet while Harpo's my favourite, it's Groucho I'm always quoting. Odd, that.

by BruceMcF (agila61 at netscape dot net) on Mon Mar 23rd, 2009 at 01:33:37 PM EST
[ Parent ]
Will it be possible to extend the line northwards after this ?

Un roi sans divertissement est un homme plein de misères
by linca (antonin POINT lucas AROBASE gmail.com) on Mon Mar 23rd, 2009 at 08:16:27 PM EST
[ Parent ]
There's not really where northward to go, this is downtown SF ... three short blocks and your at the Embarcadero and the piers.

There is the notional possibility to head east in a tunnel under the bay to Oakland, if they dug a new transbay tunnel. Locals from SF on the CA HSR blog seem to think this is not a politic issue to raise at this point in time.

The possibility exists of a people mover to connect underground to the BART stations from the TBT (1/4 mile, a block away) ... its penciled into the designs, but it seems as if without funding.


I've been accused of being a Marxist, yet while Harpo's my favourite, it's Groucho I'm always quoting. Odd, that.

by BruceMcF (agila61 at netscape dot net) on Mon Mar 23rd, 2009 at 11:38:47 PM EST
[ Parent ]
Oh, and on this point

"If only two platforms are for departing trains, that means that just to support 8 tph, maximum platform waiting time is 15, not 30 minutes."

... yes, with three minute headways in the tunnel, 12 minutes to debark, then to the tail track, then 15 minutes to embark, minus time to get from the tail to the platform.

OTOH, there is a dedicated mezzanine the full width of the train box, so with one island is a dedicated arrival island and one island is a dedicated departure island, and boarding an empty train after most passengers have been marshalled into the correct bunch in the mezzanine level seems like it could be made into a fairly quick proposition.

Still, there seems to be "mission creep" with some commentators at the California HSR blog in terms of shorter and shorter station dwells being "reasonable", including biased spin being tossed in by the TBT engineer, so maximum allowed platform dwells for the French, Spanish and German HSR services would be useful real world information.

(And, yes, Japanese terminal platform dwells seem likely to be Sugoi! ... amazing ... but if the system has to run to Japanese human tolerances to avoid delays, I don't think that will happen in California.)


I've been accused of being a Marxist, yet while Harpo's my favourite, it's Groucho I'm always quoting. Odd, that.

by BruceMcF (agila61 at netscape dot net) on Mon Mar 23rd, 2009 at 01:44:39 PM EST
[ Parent ]
maximum allowed platform dwells for the French, Spanish and German HSR services would be useful real world information.

Unfortunately, I don't have those data (if they are fixed at all -- I don't think they are in Germany at least), but I will make some notes.

First, it may be that East Asian high-speed is really a better comparison than Europe. CHSR traffic will be shuttle-like: keeping to dedicated lines with relatively short total travel times. However, In Europe, longer high-speed train runs continuing on conventional track are more typical, especially in Germany. (The exception is Spain, but there, station capacity is adjusted to an expected massive future growth, so currently they aren't pressed.)

So, in Europe, turnaround times may be lengthened by a buffer for delays, and shortened by a service concept foreseeing most of the work (like filling up tanks) during the night, when there is time anyway. As a rule of the thumb, 1-1.5 hours would be normal for stops including cleaning/checking. Some numbers:

  • Paris Gare de Lyon:

    The surface station has 22 tracks. TGVs depart at a maximum 9 tph, an additional 6 non-TGV trains also depart in the peak hour. I couldn't track down French circulation plans, but extrapolating for a maximum capacity from the minimum 4-minute headways, I get 60-minute turnaround times.

  • AVE:

    Barcelona Sants, Sevilla Estación de Santa Justa both have 6 normal-gauge tracks for high-speed trains, but only a maximum 3 tph; Málaga's Estación María Zambrano even 5 tracks for a max. 3 tph -- though, not evenly distributed. There are 10 minute headways, so one can assume a capacity of 6 tph with, again, 60 min turnaround times.

  • Germany:

    The shortest turnaround I found with a brief search was 53 min, for turnarounds including a visit to the maintenance yard: 109 min(!). More relevant for CHSR is the example of the onetime Metropolitan, which ran a short tight schedule between Cologne and Hamburg: turnaround times were down to 34 minutes.

As for East Asia: in Tokyo, both JR Central and JR East operate up to 11 tph, but the former from 4 tracks while the latter from 6. Taipei Main Station has 4 tracks for THSR trains, with a maximum 5tph right now.

Now, it is one thing whether Japanese precision can be copied in SoCal, another is passenger discipline, yet another the size and efficiency of maintenance crews. So I don't think 12tph is possible in the TBT on four platforms (e.g. 20 minute turnarounds).

By the way, is the operation of the 3 access tracks in the tunnel with one track strictly for the total incoming traffic, and two for departures, an explicit design feature, or your interpretation? Because I don't like that. Three tracks with a proper train signalling in bi-diretional operation might be better, but unless Caltrain is electrified and gets modern EMUs, I would prefer to not run 'em on the same track at all.

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Tue Mar 24th, 2009 at 07:01:25 PM EST
[ Parent ]
First, yes, this is part of Caltrain getting modernized and getting all modern EMU's.

As far as bi-directional track, its a mile plus of tunnel with two very tight turns, and transit time to clear the track is likely to be fairly long. With each tunnel track directly connecting to one island (pair of platform tracks), clearing a pair of platform tracks then filling them would be an even longer period the platform is idle.

With the system pressed for platform dwell time, the freedom of completely parallel movements for each island comes at a substantial reduction in total train capacity.

Now, without a fracking through track looping back to the present terminal (clenches fist and shakes it in a vaguely western direction at the TBT authority), dedicated two-way track has substantial higher throughput capacity than bi-directional track.

The penalty, of course, is that there is no longer completely parallel movements per island.

However, with a central access track and an egress track on each side, then it can be set up to allow parallel HSR access and Caltrain egress, and parallel Caltrain access and HSR egress.

That means that if Caltrain and HSR services enter in alternation, there is an open slot for HSR egress before each HSR access, and an open slot for Caltrain egress before each HSR access.

It could equally well be two access and a central egress, but since the total station dwell capacity limit is at the TBT trainbox side of the tunnel, if the access and egress capacities are to be unbalanced, the larger capacity should be allocated to egress.

I've been accused of being a Marxist, yet while Harpo's my favourite, it's Groucho I'm always quoting. Odd, that.

by BruceMcF (agila61 at netscape dot net) on Tue Mar 24th, 2009 at 08:51:56 PM EST
[ Parent ]
two very tight turns

How tight?

With each tunnel track directly connecting to one island (pair of platform tracks), clearing a pair of platform tracks then filling them would be an even longer period the platform is idle.

From the arrangement described in your diary, at least the Caltrain platforms can be reached from two of the tunnel tracks. Making just one of the tunnel tracks bi-directional, e.g. the Caltrain departure track, would already be an improvement. It would add the possibility of a Caltrain and a CHSR train arriving in parallel, which, due to eventual delays, is more important than the possibility of two departures in parallel.

E.g., in addition to these possibilities:


...now you would gain this one:

Even better would be another connection between the two outer tracks, so that a Caltrain departure opposite a parallel Caltrain+CHSR arrival is possible.

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Wed Mar 25th, 2009 at 04:40:44 AM EST
[ Parent ]
How tight?

I don't know ... I have seen claims they are too tight to be workable. Other than vague happy talk, all I can find is the fuzzy diagram from a 2003 document that was part of the Environmental Impact Report process.

When I say in the post that I don't know if its physically possible ... that's part of it. Swinging the central platform tracks so they come in #2 and #3 from the outside, and the tail track so they go out #1 and #2 from the outside, buys some straightening. Probably not enough, though.

When I was sketching out an operations table, I included one empty slot for four full slots on the access track for delays.

On the CAHSR blog, Brandon in San Diego pointed out that Prop 1A requires 5 minute headways on the HSR network, so on the operation model above, with alternation between HSR and Caltrain built in, an argument can be made that 2 1/2 minute headways would be required by the funding legislation. Tighter headways does not increase platform capacity, so if they can arrive at 2 1/2 minute headways, that could be one open slot for each pair of inbound services.

Its parallel arrival HSR / departure Caltrain and parallel arrival Caltrain / departure HSR that is the point of having an egress track on either side of the access track ... eliminating that crossover between access and egress tracks at grade (or at least, putting it off to the mouth of the tunnel, when its a "half-dive" ... cheaper because the tracks have to dive into the tunnel in any event) ... is what that accomplishes.

As for bidirectional operations on the Caltrain egress track, if the Caltrain egress track is taken down first, to dive under the common access track to the tunnel, it can be a central branch, switched to both sides of the two way track out of the current terminal at 4th, which is about 2km away.


I've been accused of being a Marxist, yet while Harpo's my favourite, it's Groucho I'm always quoting. Odd, that.

by BruceMcF (agila61 at netscape dot net) on Wed Mar 25th, 2009 at 11:42:03 AM EST
[ Parent ]
Even better would be another connection between the two outer tracks, so that a Caltrain departure opposite a parallel Caltrain+CHSR arrival is possible.

From the diary:

The only player that strikes me as having the opportunity to say, "wait a minute, here's a fix that won't cost all that much to implement" is Caltrain. But ... under the solution above, they are giving up a 900 ft. platform, connected to the tail tracks, for what could end up being a 800 ft. platform, with only one connection to the tail track, and that connection only available when the closest HSR platform track is empty.

So, yes, the above, which would make it operationally feasible to squeeze 8tph into four terminal platforms, and which may or may not be physically feasible, does knock on to make things even worse for Caltrain. Efforts to squeeze extra capacity out of the same infrastructure tend to generate bad trade-offs like that.

If there is going to be three tracks, the ideal would be an access track on the one side, an egress track on the other, a bi-directional track in the middle with switches at several points along the tunnel, and an adequate 3:6 fan out including as many dives as necessary to avoid blocking movements.

But the TBT is a property redevelopment project using a transit project as its hook, with the location itself the wrong place for a central rail terminus, and while cut and cover tunneling for the station throat can be useful as an excuse to take over and redevelop certain sites, extra money for an adequate platform access or extra tunneling for dives (and its basically tunneling in mud, so it'll be fairly expensive) ... they are kind of "only if someone else pays for that part".


I've been accused of being a Marxist, yet while Harpo's my favourite, it's Groucho I'm always quoting. Odd, that.

by BruceMcF (agila61 at netscape dot net) on Wed Mar 25th, 2009 at 11:56:43 AM EST
[ Parent ]
Málaga's Estación María Zambrano even 5 tracks for a max. 3 tph

Typo. Should be: 2 tph.

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Wed Mar 25th, 2009 at 04:07:15 AM EST
[ Parent ]
Richard Mlyrnik says on the Caltrain HSR Compatibility Blog:
Start with "pretty much any ICE or EC stopping at Leipzig Hauptbahnhof."
There are thousands of other examples.
eg ICE 692: ... -- Stuttgart 08:47 08:51 -- Frankfurt(Main) 10:08 10:13 -- ...

Is there any way to find the turn-around times for HSR trains at Leipzig Hauptbahnof?

I've been accused of being a Marxist, yet while Harpo's my favourite, it's Groucho I'm always quoting. Odd, that.

by BruceMcF (agila61 at netscape dot net) on Mon Mar 30th, 2009 at 11:20:45 PM EST
[ Parent ]
Does "turn-around" refer to the train changing direction, or to emptying the train, cleaning it, and starting a new route. If the former, HSR can be turned very quickly indeed. I suspect the 30+ minutes refer to the latter, and you'll have a hard time finding these figures by simply studying the schedules.
by gk (gk (gk quattro due due sette @gmail.com)) on Thu Apr 16th, 2009 at 04:44:30 AM EST
[ Parent ]
... CHSRA is saying 40 minutes, they are talking about route terminus operations.

The most important route terminus operations for building ridership, of course, are the infrequent ones ... the operations to prevent a train swap from creating a service delay, and to prevent one service delay from cascading to multiple service delays.

But if they are through services through a terminal, with the route terminus located somewhere else, then that time buffer is also located somewhere else.

The restock, full clean, high frequency safety checks, etc., that are scheduled as the normal terminus operations are an efficient way to use the time, but for corridor trains that are not operating around the clock, many of them do not have to be done when they are, but are more spreading the work around and not letting the time buffers built into the system go to waste.

More on this in Part 3.


I've been accused of being a Marxist, yet while Harpo's my favourite, it's Groucho I'm always quoting. Odd, that.

by BruceMcF (agila61 at netscape dot net) on Thu Apr 16th, 2009 at 01:40:18 PM EST
[ Parent ]
As gk said. Frankfurt Hbf is another terminal station where almost all long-distance trains (be them ICE or loco-pulled IC/EC) change direction, but don't end their run, and thus the time is short. Also true for most trains in Stuttgart Hbf, andmany in Munich Hbf.

To complicate matters, in Germany, many long-distance trains run in circuitous routes -- e.g. a train may run a big half-circle from Berlin through Dortmund Frankfurt, Stuttgart to Munich, then go North through Nuremburg to Hamburg; or alternatively through Leipzig back to Berlin... and in such a case, the "terminal station" is a matter of choice.

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Fri Apr 17th, 2009 at 11:26:38 AM EST
[ Parent ]
Uhh, and sorry for a very belated reply, didn't catch it.

*Lunatic*, n.
One whose delusions are out of fashion.
by DoDo on Fri Apr 17th, 2009 at 11:27:15 AM EST
[ Parent ]
That was the discussion on the Caltrain HSR compatibility blog ... if the TBT is used as a through-service terminal rather than as a route terminus, at least when trains are turning around, its not the platforms that are a bottleneck, but rather the requirement to offer 5 minute headways to the HSR.


I've been accused of being a Marxist, yet while Harpo's my favourite, it's Groucho I'm always quoting. Odd, that.
by BruceMcF (agila61 at netscape dot net) on Fri Apr 17th, 2009 at 07:06:18 PM EST
[ Parent ]
But, NOT using it as a terminal station is pretty stupid, aint' it.

On one hand, from the passengers' point of view, it will be a terminal station -- meaning that all passengers of the arriving train will get off, and all passengers on the departing train will board there. You can't use the 6 minute turns in Frankfurt as example for that -- 10-15 minutes in Vienna Westbahnhof is more like it.

Then, you'd have to service trains in LA after a full turn (that would be up to 7 hours 20 minutes without servicing, with runs in 213-minute train pattern #3 both ways), or devise an operation plan in which runs from LA to SF are always followed by a run from SF to Merced, where there would be time for servicing.

All in all, I have a very bad feeling about this TBT project. The seeds of several future problems will be planted, even with the improvements you sketch in the second and third diaries.

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Sat Apr 18th, 2009 at 04:57:26 AM EST
[ Parent ]
... four platforms and 15 minute turns permits 12tph with a 5 minute time buffer for each turn.

The TBT project will, of course, be studied in future years as an vestige of the Kamikaze Century, when the location of bus ramps onto a road bridge could determine the location of what was supposed to be the main rail hub of a city like San Francisco ... because the rail hub part of it was tacked on as something that sounded good, with the plan of getting trains in and out left for later.


I've been accused of being a Marxist, yet while Harpo's my favourite, it's Groucho I'm always quoting. Odd, that.

by BruceMcF (agila61 at netscape dot net) on Sat Apr 18th, 2009 at 07:13:12 AM EST
[ Parent ]
... capacity is much less of an issue at the Southern end, because as the system builds out, it becomes a "Y", with an Anaheim terminus and a San Diego terminus (well, its an "X" including Sacramento, but with the bulk of northbound routes heading to the San Jose / San Francisco side). Anaheim platforms are at grade, and they are building the platforms to the CHSRA request, while if the San Diego terminus is at the Airport, that will be able to act as a through platform when needed to avoid getting bottlenecked.


I've been accused of being a Marxist, yet while Harpo's my favourite, it's Groucho I'm always quoting. Odd, that.
by BruceMcF (agila61 at netscape dot net) on Sat Apr 18th, 2009 at 01:57:35 PM EST
[ Parent ]

Display:

Top Diaries

Ulster says NO!

by Frank Schnittger - Oct 17
27 comments

Trump's Presidency Transfixed

by ARGeezer - Oct 17
7 comments

Spain is not a democracy

by IdiotSavant - Oct 14
10 comments

Does anyone care?

by Frank Schnittger - Oct 10
71 comments

Occasional Series