Tuesday, March 24, 2009

BruceMcF on Transbay Terminal

NOTE: We've moved! Visit us at the California High Speed Rail Blog.

Note from Robert: Bruce McF has been generous enough to allow us to post this excellent discussion of the Transbay Terminal trainbox and track issues that he wrote. In addition to the link to his own site he gave below, he posted it over at the European Tribune as well, where as usual there is a good discussion on this.

From Burning the Midnight Oil for Living Energy Independence

I was able to get an interesting look into the proposed future of Intercity Travel in the Bay at the Transbay Terminal (TBT) in San Francisco.

Senate Info Hearing on High Speed Rail in California

Note that I am not trying to give "objective reporting" on this issue but rather to give vent to my reaction to watching the hearing online ... see The Troubling Discord Between Transbay and High Speed Rail Authorities for a less hot under the collar reaction..

One piece of information is that in California, when one public authority has the funding for sufficient staff and another doesn't, and it comes to a fight, it is considered fair game for the staffed up authority to toss up spin and red herrings and biased analyses, confident that the other authority does not have the capacity to answer promptly.

Late on in the piece is the technical guy from the Transbay Terminal (TBT) project putting in all sorts arguments against changing the design of the TBT "train box", without concern or regard for whether the arguments would be considered fair or impartial by a disinterested third party. He compared:

  • the station stopping time at a through platform to the station stopping time at a terminal platform;
  • the terminal turn-around for a regional rail service running between San Jose and San Francisco with a long distance rail service for a train that arrived from Anaheim/LA (or even, in Stage 2, San Diego); and
  • the number of services on single routes in Japan and Europe with the number of services for the main northern terminal for multiple routes in California

He wasn't the only one with tricks up his sleeve ... one of the Senators asked after the terminal capacity at Anaheim. The answer, by the way, was six platform tracks for what is supposed to be the secondary Southern Terminus when the full system is complete ... two more than the TBT proposes for HSR, at what is supposed to be the primary Northern Terminus.

How many HSR services should be allowed for?

On the other hand, the California High Speed Rail Authority (CAHSRA) seems to be making claims that are difficult to support. They seem to have asked for an ability to support 12 trains per hour (tph), with 40 minute platform dwell times, claiming that they need 9 to 10 platform tracks.

First, the basic math ... 12tph with 40 minute platform dwells, inclusive of headways, is 8 platforms.

More fundamentally, though, where are 12 trains per hour coming from? That is a technical requirement for through stations, because the HSR line is designed to permit trains to pass at five minute intervals, and along the line, you cannot allow one service to block the next in line.

For the terminal station, the question is the number of services that might start or end at the station. For the California HSR system:
  • LA and the Bay is the backbone transport market for the HSR system ... there could well be demand for one LA/Anaheim Express and, half an hour later, one LA/San Diego Express
  • Running Express Routes drops off very useful trip pairs, so there will be demand for a Semi-Express, and the likelihood is that the hour that supports two Bay/LA Express services will support a Bay/LA/San Diego Semi-Express.
  • The Central Valley will be within three hours by an all-stops HSR to both LA and the Bay ... and within two hours of one, the other, or both. So in addition, one all-stations LA/Anaheim HSR service per hour providing access to and from the CV ... bearing in mind that while this is a smaller transport market, the HSR will grab a larger share of the total market

So this is 4 trains per hour ... 1 train per hour on four distinct services ... without even considering a Fresno special, or a spur at Mojave for Las Vegas.

And when the 400 seat single level, single set trains start filling up, its better for building ridership to increase frequency than to increase capacity. 2 LA/Anaheim Expresses per hour, split the all-stops CV into Express Fresno then all-stops to LA and all-stops to Fresno and Express to LA, and a mix of Express LA/San Diego and Express to LA then all stops to San Diego, and we are already at 6tph.

Indeed, as blogger DoDo on the European Tribune notes, the service schedule that the HSR ridership modeling is based upon (pdf) implies up to 8 trains per hour at the TBT.

Twelve trains per hour may be aiming too high, but six trains per hour clearly risks aiming too low.

Given the massive cost of building more capacity after the original foundation has been laid, the capability for eight (8) HSR trains per hour seems to be a perfectly reasonable expectation for the primary northern terminus for the system.

Following the trail of red herring

Now, when someone deploys deceptive comparisons and unbalanced comparisons, I have a reflex reaction ... a pile of red herring is normally used to cover something up.

And that something seems to be is a design flaw.

The TBT "train box" includes two "tail tracks", allows trains to get off the platform, either for overnight parking or for non-passenger operations like restocking and cleaning, without using up space in the tunnel.

What this means in theory is a train can arrive at an arrival platform, unload passengers (which is a very quick operation, since trains have far fewer passengers per door than airplanes), move to the tail platform to make room for the next train, get trach unloaded, seats needing deep cleaning looked after, food and beverage restocked, and then get move to the departure platform.

And the TBT tunnel access is designed with three tracks, which eliminates all sorts of potential bottlenecks:
  • Both Caltrain and HSR services arrive in the TBT on the central track
  • A Caltrain service departs from the Caltrain island platform using the "inner" tunnel track, which opens up the platform for an arriving Caltrain service
  • After the departing Caltrain service has left, the arriving Caltrain service switches over to the "inner" track to get to the Caltrain island platform
  • HSR services run directly to the central arrival island
  • Passengers depart the HSR services, the train goes to the tail track for restocking and to clear the platform for the next service, from the tail track to the departure platform, then depart using the "outer" tunnel track

This is a system that allows three different islands to be accessed with little interference, because only two islands receive incoming trains, and because each type of service has its own dedicated departure track ... so they can arrive in sync, dwell in station for different lengths of time, and leave on their own schedule.

In particular, it allows the HSR trains to be in the station for over 40 minutes, while only occupying the platforms for 30 minutes each, raising the capacity of four platforms from six trains per hour to eight trains per hour.

This also makes it easier to organize efficient movement of passengers, since passengers are either leaving or arriving at each HSR platform ... there isn't a the problem of departing passengers getting in the way of arriving passengers.

With this approach, 4 platform tracks support 8 trains per hour ... which is to say, adequate to the needs of the planned HSR system with enough spare capacity to allow for some growth.

The Design Flaw

The design flaw ... for supporting 8tph, that is ... may not jump out at you, but its in the picture, taken from the 2003 "locally preferred option" design for the TBT Environmental Impact Report. The right hand side is the tunnel from the present end of the rail line. The left hand is the turn to the tail tracks.

Now, the HSR platforms have to be designed for long trains ... once the capacity is filled with 8 car, 400 seat trains, they can be extended to 16-car, 800 seat trains, and then by moving to bi-level trains, 1400 seat trains. That means a 1,320 foot long platform. That means that the bottom two islands are for the HSR and the top island is for Caltrain. The bottom two platforms need to be stretched a bit, and the middle one straightened somehow ... but the TBT technical person said that that had been fixed up.

So, stepping through the pictured design:
  • Counting access tracks, three tunnel tracks split up to make six platform tracks. So far, so good.
  • For Caltrain to operate as described above, a switch will have to be added so Caltrain services can get from the middle tunnel track to the "inner" tunnel track which leads to the Caltrain platform. And since the outbound train has to leave the Caltrain platform before the inbound train can arrive, that will work just fine.
  • The two Caltrain platform tracks are connected directly to the Caltrain exit track, so that will work just fine.
  • For the HSR trains to operate as described above, the central island is the arriving platform, connected directly to the tunnel track that brings trains in. That will work just fine.
  • That leaves the bottom island as the departing platform. The two platform tracks at this island are connected directly to the HSR exit track, so that will work just fine.
  • And the tail tracks ... are not connected to the bottom platform track. Instead, the bottom platform track comes to a dead end. That is not just fine. Indeed, assuming that the TBT has the staff that they likely know all of this already, that might be what the pile of red herring is supposed to cover up.

What can be done to straighten up the mess? One approach is to swap the Caltrain platform from top to bottom ... and trim a substantial piece from the front (right hand side) of the bottom platform. In fact, trim enough from the front that the switch between the two platform tracks is after the single tunnel track has rounded the corner.

Note that this is just a rough sketch

Trimming off the front of the bottom island allows the middle island to straighten up. Straightening up the middle island allows the top island to straighten up.

The middle island can be straightened up a bit by extending the tail track directly from the middle platform, with switches connecting the top island, which gives more room before the platform track must bend to form the rail track..

This might not be enough for 1320 ft. of straight platform, but it'll be a lot closer ... and, after all, the CAHSRA is probably overstating how much straight platform they need, since the platform only needs to be straight for the passenger car portion. A little bit of bend for the driver cars at the front and rear of the train can be tolerated. If this can get 1200 ft.. of straight platform with a 60 ft. curved part on either side, that certainly seems like it ought to be OK.

So, if both sides are wrong, who is going to admit it?

The question that puzzles me the most is not the technical one ... as tight a squeeze as it may be ... but the political one. The TBT authority have made public claims that present a picture of basically being ready to go, except for the fantastical demands of the CAHSR authority. The CAHSR authority has made fallen into the trap of making an ambit claim that they will have to strain to support ... but if that costs them the political argument, the fact that the TBT train box is an inadequate design is likely to be lost in the collapse.

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.

To Be Continued ...

Anyway, that's the puzzle. But there's another possibility ... one which might be of more appeal to Caltrain ... so I am going to end this with an ellipses.




BruceMcF said...

... now the question is whether this was already talked to death in the Open Thread.

Clem said...

@Bruce, I do have a question. What is the basis for your assumption that outbound Caltrain would run on the left-most track (seen from the train) ? They will have to cross over to the right before 4th & Townsend.

Also, indications are that the tail tracks may just go away entirely.

jim said...

The real solution is to loop the tunnel back under king to forth and the tbt a is - it would solve all the problems and be worth the cost by freeing up tbt to be a place where trains stop long enough to get people on and off and then get out of there back to 4th where there is room to do whatever, dweel shuffle trains whatever. This is obvious coomon sense to any random person off the street. That's why it wasn't done.

jim said...

"and leave the tbt as is"

Robert Cruickshank said...

Yes, that is indeed the question. I thought about that, Bruce, but I figured that it would be worth having the actual post, with the facts and figures and images. And if folks decide we've already hashed it out, oh well, tomorrow's another day!

Thanks so much for letting me post this here.

jim said...

trains per hour schedules... lookin at london-paris 20, london brussels 10 total 30 per day each way
paris-lyon 24 per day each way barcelona- madrid 24 per day each way. on average thats one train per hour in each direction. I don't see why ca would be any different.

yeson1a said...

What is the proposed height of the platforms? Caltrain now has all low level and I would think that CAHSR would want high-level. That would give Caltrain 1 set of tracks with no abilty to switch and use the high levels.

BruceMcF said...

@ Clem: "What is the basis for your assumption that outbound Caltrain would run on the left-most track (seen from the train)?"

I'm not assuming that they will ... rather, that is the way of using the three tunnel tracks that does eliminate the interference between HSR access and Caltrain egress, and Caltrain access and HSR egress.

Now this is back of the envelope ... I didn't do any CPM optimization of the different access/egress systems I was looking at - hell, I haven't done any critical path optimization for over twenty years ... but when I did the operation tables by hand, it was always a crossing between a departing service of one type and an arriving service of the other type that threw a monkey wrench into making full use of the platforms..

The other ways to eliminate that interference involve dives in the throat and train-box end of the tunnel, and that means a substantially deeper cut for the dive.

On why central access rather than central egress, the total train capacity constraint is on the train-box side of the tunnel, if there is an imbalance between access and egress capacity, the extra capacity should go to the egress.

As to the cross-over problem ... what makes solutions to dive inside the throat or the tunnel expensive is the extra tunneling. Crossing under at the mouth of the tunnel, where all three tracks have to make the transition from grade to tunnel level anyway, would be substantially cheaper.

Further, its something that would be finalized six to either years in the future, and so the rush to getting the train-box funded in time for including it in the original foundation of the TBT does not apply to the tunnel mouth side of the plan.

Rob Dawg said...

How much is budgeted for the CAHSR share of the TBT?

BruceMcF said...

@ jim: "The real solution is to loop the tunnel back under king to forth and the [leave the TBT as is]."

Precisely ... they worked out the benefit of underground loops when they built the Grand Central Terminal in New York, which started construction in 1903 (platform loop backs, but the same basic operational benefits apply).

I would only be guessing, but I would guess because property that they wanted to demolish for the project would be taken out by putting the throat for access to the train box where it is, while a loop-back would be of no use to them in terms of property redevelopment.

Unless there is a genuine reason for preferring to tunnel at 2nd rather than 3rd (underground water hazard or such), that would also by my guess why the tunnel goes up 2nd instead of 3rd, where the fan-out could be straight into the train-box and would not impose curves on platforms.

jim said...

@Bruce, So again what we have is a project paid for by the public, for the benefit of politicians with the public good at the bottom of the list of conerns

jim said...

The question is how do you get the public to be outraged about the bungling, while keeping them on board with the project.

go modern said...

It time to get modern. Look at what the State of Michigan is planning.

The Wave of the Future

View the video

BruceMcF said...

jim said...
"trains per hour schedules... looking at london-paris 20, london-brussels 10 total 30 per day each way

paris-lyon 24 per day each way barcelona- madrid 24 per day each way. on average thats one train per hour in each direction. I don't see why ca would be any different.

Yes, but spare platform capacity at 3am cannot be stored up to be used at 5 in the evening. Barcelona-Madrid is an average of 1 tph, with no AVE in the middle of the night, then 1 tph except 0700, 0730, 0800, 0830; 1530, 1600, 1630; and 2130, 2200 (last HSR service until 0700 the next day, the 2300 is a sleeper).

With the CAHSR system designed with a branch for either Anaheim or San Diego when in the time of day that the ridership on the Express SF-SJ/LA justifies a full length set, its obviously better for ridership to have two single-set services at half hour intervals, one to proceeding to Anaheim and one proceeding to San Diego.

That will be exactly the same time of day that an all-stations through the Central Valley would be needed. I know that the CAHSR has this complex route schedule optimized for the ridership model, but IMIHO the most effective means of building ridership in the non-Express stations is a memory hourly timetable through the daytime that gives a trip to anywhere on the system, either directly or with a transfer to a Semi-Express in the CV or Express at LA-Union.

And then a Semi-Express slots in there naturally, and from there a growth to 6 tph is no hard to see.

Frequency and simplicity builds ridership more rapidly than low frequency services of the same total number of seats.

A mix of one tph and two tph, combined with the fact that the route branches south of LA, combined with the quite limited number of stations required for SF-SJ / LA in under three hours which argue for a three tier Express/Semi-Express/All-Stations service system, and you get to the 4tph to 6tph range quite quickly, even without considering extensions like a Fresno special or a Las Vegas route.

Now, DoDo tells me on the Eurotrib that my sketch of timetabling is more Spanish than French, but the modeling that Caltrain commissioned on CHSRA ridership more along the French model had 8 trains per hour in a peak three hour period. Skipper scheduling fits a "browse the web for next service" use, but is not as suited for "show and go", and most importantly is a harder target to hit when country transit/transport authorities are trying to integrate their schedules with the HSR.

BruceMcF said...

@ yeson1a said ... "What is the proposed height of the platforms? Caltrain now has all low level and I would think that CAHSR would want high-level. That would give Caltrain 1 set of tracks with no abilty to switch and use the high levels."

The TBT technical fellow (watching this video one more time is probably against doctor's order for blood pressure, but I am guessing he is the staff engineer) said quite definitively that Caltrain platform and HSR platforms would be different heights, and therefore Caltrain would have one island and HSR two.

So, yes, precisely, that would given Caltrain one island. However, they still have their existing terminus, that can continue to be used as secondary terminus, and so have more flexibility on that score than HSR, which would launch with the TBT as their main northern terminus and sole urban San Francisco station.

If Caltrain are satisfied with 2 platforms and no tail tracks ... that opens up the possibility of a redesign of the train-box that would only push the envelope of the box down by a couple of feet.

BruceMcF said...

"It time to get modern. Look at what the State of Michigan is planning."

Thanks for the comic relief. 200mph Magnetic Trains Powered by Solar Powered Hydrogen Batteries. Woot! Because nobody goes anywhere in Michigan in the middle of winter when they can help it, so having transport capacity expand in August and collapse in January makes perfect sense!

And even better, a train station at every Interstate Highway Interchange, because nothing says "walk to the train station" like an Interstate On-Ramp!

Remember, if living between Great Lakes gets too boring for a Michigan state legislator, there's always always the option for a trip to the Fanta Sea.

Rafael said...

@ BruceMcF -

we have been over this quite a bit. First off, let me remind everyone that tunneling south of market is severely complicated by the historical Mission and Hayes Creeks - the area is basically hardened bay mud, not rock.


My take on this whole tiff is that the TBT is a good idea in the wrong location. All talk about tph is basically smoke and mirrors by both parties to avoid this 800lb gorilla in the room.

SF started the TJPA project with an old bus terminal that has to be torn down because it isn't up to seismic code. So actually, it is starting with some left-over access ramps to the Bay Bridge (originally used by trains) and a few parking lots. On that basis, it came up with a concept for transit-oriented redevelopment of the district.

That's all fine and good but a transit hub needs to be anchored by where all the rail service can easily be connected. Buses and even real estate need to follow the rails, not vice versa.

The ideal location for a new Central Station would have been Market & 7th (h/t to commenter Andrew). A short, straight is all that would have been needed to get both Caltrain and HSR to a terminus with as many platforms as needed, right next to BART, SF Muni Subway and SF Muni streetcars. Unfortunately, recent building construction at that location means it is de facto no longer an option.


But perhaps we should ignore SF's real estate ambitions and zoom out one level for a minute.

CHSRA want to be as close as possible to downtown SF in order to maximize the catchment area for the northern terminus of its starter line. That means proximity to BART, buses to Marin and Alameda from across the Bay and Golden Gate plus SF Muni's local transit.

In that context, the greatest deficiency is arguably the long walk between BART and the TBT site. That can be addressed with new side platforms at Embarcadero station to supplement the existing island platform. Trains stopping there would open doors on both sides, which conveniently also eliminates the existing rush hour bottleneck in the BART system.

These new side platforms would be connected to a concourse level at Market and Fremont/Front via gently sloped moving walkways. Ideally, the design would provide access to similar side platforms for SF Muni subway as well. An already planned but "optional" pedestrian tunnel under Fremont Street - presumably also featuring moving walkways - would complete the connection to the TBT's concourse level.

Long pedestrian connections are far from optimal but perhaps acceptable if equipped with moving sidewalks and sheltered from the elements and traffic.

SF Muni streetcars stop at Market and 1st. It might make sense to run some new tracks down Mission and 5th to provide a new service linking 4th & King, multiple high-value properties along Mission, the plaza in front of the TBT, all of the ferry terminals and Pier 39.

The wildcard here is actually Caltrain. It has been stuck at 4th & King forever, far from the financial district. Electrification is now supposed to bring it closer, making the service a lot more attractive for commuters in both directions. At first glance, giving Caltrain a couple of platforms in the TBT train box is the most logical solution.

Perhaps that assumption needs to be challenged. What Caltrain needs in SF is access to the financial district and better connections to SF Muni local transit. A second intermodal with BART north of Millbrae and, improved connections to buses from across the Bay and Golden Gate would all be nice to have but it not strictly necessary. An intermodal with HSR in downtown SF is not required.

Conclusion: Caltrain doesn't need to be in the TBT. It just needs to reach Market Street and preferably, beyond.

At this point, two things pop into my head:

a) Caltrain's four tunnels between the 4th & King terminus and Bayshore station and,

b) SF Muni's $1.3 billion plan for a Central Subway.

Point A is important because UPRR has limited trackage rights to switch cars for its remaining freight customers in the Port of San Francisco. This switching happens during the day and, Caltrain is contractually obliged to provide UPRR with a number of 30-minute windows (p14, clause 4.3) during which it can run freight trains on the main line at commuter rail speeds. Fortunately, all of UPRR's customers are east of the line, so HSR can avoid problems by staying west of Caltrain's tracks (FFSS configuration), at least north of Millbrae.

Meanwhile, CHSRA's Google Map of the preferred route reveals that it plans to run its tracks in the Caltrain ROW all the way to 4th & King. The Authority's cross-sections for the Caltrain corridor imply an SFFS track configuration south of Bayshore and FSSF north of it (instead of FFSS). The switch is supposed to happen in a trench just south of tunnel 4 but needs to be corrected. That implies the existing tunnels will contain the southbound Caltrain and the northbound HSR track. They've also completely missed tunnel #2 between Army/Cesar Chavez and 22nd.

Upshot: Caltrain can fork both of its tracks off to the east anywhere north of Oakdale (north end tunnel 3) where UPRR veers off - in streetcar mode! - out to pier 96.

Here's where Blackadder finds out that Baldric has a cunning plan. What if Caltrain actually descends directly underneath the HSR track between tunnels 3 and 2? Caltrain's 22nd Street station would go underground, but so what.

The important thing is that no reconfiguration of I-280's supports at tunnel 1 would be necessary. Caltrain would veer off at Mariposa Street, continue across the lot to under 4th Street and cross under Mission Creek.

At this point, it hijacks the Central Subway project and stops at 4th & King (with a direct pedestrian connection to the SF Muni stop for the T line), Moscone Center, Stockton/Union Square and Stockton/Washington, with tail tracks out to Broadway.

Meanwhile, HSR gets the DTX tunnel and the entire train box in the TBT all to itself. Tail tracks would not be needed, because it would get all of 4th & King as well. That facility would be used for overflow parking and high speed cargo transshipment.

Caveat 1: HSR and SF Muni would need to buy Caltrain the Brisbane yard for overflow parking.

Caveat 2: Caltrain would be limited to EMU equipment north of Bayshore. In addition, it would have to use that equipment to provide service comparable to what SF Muni had in mind between Bayshore and Chinatown.


Alternative: Caltrain could veer further east in China Basin, run around the AT&T ballpark and under 2nd and Montgomery instead, right into the heart of the financial district. Some deep cut-and-cover excavation at 2nd could be shared with the DTX tunnel for HSR, which would run a level above. The building at 44 Montgomery might pose a problem for tunnel construction.

This alternative would leave SF Muni's plans for the Central Subway unchanged.

Devil's Advocate said...

@Jim: In EU a train every half hour is more typical than 1 tph, especially at peak time. There may be only 24 trains a day, but they are virtually all between 0700 and 2200. Don't remember ever seeing train sets over 12 cars in EU. Actually 12 I think seems the magic number on most EU lines. Likely to be the same in SF. I wouldn't count on a lot of volume of passengers for trips to Central Valley cities. The low cost of auto travel in US (cheap gas, no highway tolls) compared to EU, is a big competitive advantage that the car has over HSR on shorter trips under 200 miles.

BruceMcF said...

Interesting proposal ... but San Francisco 1999 Prop H requires Caltrain service to the TBT. And with limited time to apply for the HSR money, the San Francisco third of the Caltrain Board have a club to use to argue against a speculative plan with no EIS.

There is the "final engineering" proviso, but that does not seem like it would come anywhere close to stretching to fit that.

Now, in reality, if crude oil prices bite another two or three times, satisfying rail transport priorities are likely to become an essential requirement for property development, rather than a secondary or tertiary concern.

But that is five to ten years in the future, and the funding submission clock is ticking. The opportunity for "a very good solution" has likely passed us by, and we are in the "avoiding the worst possible outcome" stage.

Rafael said...

@ go modern, BruceMcF -

and I thought I had a cunning plan! That was hilarious.

Alon Levy said...

Jim, we've been over the ridership projections already. The EU has no city nearly as large as LA, and only two larger than SF. It's likely that ridership will be in the ballpark of the Tokaido Shinkansen, which runs 14 tph, rather than in this of the Eurostar, which runs 2.

BruceMcF said...

Devil's Advocate said... I wouldn't count on a lot of volume of passengers for trips to Central Valley cities. The low cost of auto travel in US (cheap gas, no highway tolls) compared to EU, is a big competitive advantage that the car has over HSR on shorter trips under 200 miles.

Yes, in the sense that the larger part of the CV market will be trips from the Central Valley cities ... the majority of trips to the CV cities will be return trips.

However, it is a mistake to treat decisions made with few realistic alternatives as a revealed preference ... the dominance of cars over planes in CV travel to larger metro areas is normal for those distances, given the inconvenience of air travel ... but some of those trips are to destinations where having a car is an inconvenience rather than a convenience. HSR at this range has competitive advantages against car travel for some trips that air travel simply cannot match.

And, again, you cannot save idle platform capacity at one time of day for a boost in capacity in a different time of day. One could assert, for instance, "oh, there'll only be a need for five 400-seat CV all-stations per day each way" ... but the best time to run many of those services will coincide with the period of strongest demand for the SF-SJ/LA services.

And all of that is independent of the obvious point that when the system opens, the relative cost of a gallon of gas in terms of an hour's wage will be closer to current European effective cost of gasoline than to current American effective cost. So if the argument is that its the cost of gas in Europe that makes HSR so effective in attracting people to 2 hour trips by rail ... that's an argument that the market potential will be rising in the next decade more rapidly than the rate of population growth.

Rafael said...

@ BruceMcF -

what's your take on the extent to which WiFi on board would allow operators to scale back the number of prime-time expresses in favor of more semi-express trains, precisely because the CV has needs of its own?

Devil's Advocate said...

@Alon Levy: your population stats are not correct. Madrid has a population of 3.6M, therefore comparable to LA. Of course LA has a larger regional pop. if you include the neighboring counties, but if you look at the 25 mile radius from the station LA is no match for Madrid, which has a higher density. As far as SF is concerned, the pop. in the city is only 800,000. You need to add a great portion of the Bay Area to match the pop of Barcelona proper, which boasts 1.8M in the city proper (and all nicely packed in 100 SqKm, an area smaller than SF city proper). Barcelona has a metro area of over 3M, but if you want to compare the entire Greater Bay ARea with that, then you probably need to include most of Catalonia with Barca, which has over 7.3M people. In conclusion, Madrid/Barca and LA/SF are very comparable in population (although not in density). If 2 trains/hour work for Madrid/Barcelona, I wouldn't expect much more for SF to LA with similar populations but lower densities. Also none of the above cities match in population and density of Japanese cities, therefore projecting figures similar to the Shinkansen may be a little too optimistic.
@BruceMcF: I hope you're right, but I see the cheap US gas as a big competitor for trips from the CV. Of course Obama might invade Iran and things might change in the price of crude oil, but I don't see a family of four traveling from Fresno to Disneyland on a HST at $50pp one way (I'm quoting EU prices for similar distances by train)+ cost of a rental car, just to save an hour or two. Chances are they'll drive their own car for a fraction of the cost. I see the CV-LA (orCV-SF) market limited primarily to airport goers (LAX or SFO), or to business travelers, for whom time and ability to work online on board is more important than cost. But I'm just playing the 'Devil's Advocate' here.

BruceMcF said...

@ Rafeal: "what's your take on the extent to which WiFi on board would allow operators to scale back the number of prime-time expresses in favor of more semi-express trains, precisely because the CV has needs of its own?"

The driving force here ... and lost in apples and oranges comparisons ... is that getting the SF-SJ / LA trip time inside 3 hours, given the populations of LA and the Bay area, is the core justification for building the HSR line.

Then the incremental cost of adding trips to intermediate centers with Limited and Local services more than justifies the incremental costs.

Arrival within 3 hours of departure and arrival more than 3 hours from departure would be hitting where transport demand has a fairly significant elasticity per minute trip time. Appropriate amenities ... WiFi, seat low amperage DC or AC plugs, Red Box DVD rentals and drop offs, food, beverage ... will boost ridership, but for any given level of ridership, the transport demand SF-SJ/LA will be strongest with Express service, and given the size of the cities, the best financial performance is going to come from tapping that demand effectively.

For small stations, the strongest demand will come from regular transit connections, and the easiest schedule for transit to integrate with is an hourly or alternate-hourly memory timetable.

And demand from the largest sources of origins to LA and the Bay can be leveraged by the faster travel time of Limited services, which might run hourly during the morning, afternoon and evening peaks, and every second hour in the rest of the day.

WiFi services on board will help with marginal time differences, and if Fresno/SF is 1:20 at the best time if it should happen to be in the time of day you can travel in some complicated station skipper timetable ... versus 1:30 on a regular memory scheduled Limited ... onboard amenities will reduce the impact of that +0:10 for some but not all passengers.

More important, though, is that regular memory scheduling and consistent connections between specific stations ... a Local, Limited, or Express connection one time of day is a Local, Limited, or Express connection at any other time of day ... is that it provides an easier target for local transport authorities to aim at.

And very tight resource bottlenecks on the system can easily make it impossible to offer something that local transport can integrate to ... either because of the complexity of the timetabling or because of constant service delays because of efforts to run the system beyond a normal operating capacity.

Alon Levy said...

Devil's Advocate, I'm using regional populations. European metro areas, I believe, correspond to American MSAs. By that definition, we need to add the SF and SJ MSAs, which total 6 million people. Japanese metro areas are more like American CSAs, but if we apply the Japanese definition to the Bay Area, then it ends up including Stockton, Modesto, and Merced as suburbs and has 9 million people; LA by that definition has close to 18 million. It doesn't really matter which definition you choose, since CAHSR will hit most satellite metro areas, such as the Inland Empire, as well as primary ones.

Densities in Japan are lower than you think. The Osaka region isn't that dense, and its Shinkansen stop is away from downtown. In the central cities car ownership is low, but in the suburbs most people own cars, which already have a commute modal share of about 60-70%. It's lower than in US suburbs, but far higher than what you'd expect by the standards of New York or Washington.

BruceMcF said...

Devil's Advocate said...
"If 2 trains/hour work for Madrid/Barcelona, I wouldn't expect much more for SF to LA with similar populations but lower densities.

But the Express trip is 2:43, the Limited trip is 2:57, and the All-Stations is 3:24, so running Limiteds through the day does not leave a lot of untapped passenger demand that would be picked up by an Express service.

And Barcelona is the end of the line at the moment ... when the extensions into France are completed, it seems likely that there will be more than 2 tph on that routes, since the benefit of the Express to Barcelona is leveraged by through services or transfers to Lyon and Paris.

So the current, recently opened Madrid / Barcelona is much more comparable to the early years of CA-HSR Stage 1 than to a built out system ... either a built out AVE system or a built out CA-HSR.

Aaron said...

@Alon - I think we've addressed the issue of the Tōkaidō Shinkansen issue before - CAHSR will be popular, but it's not the Tōkaidō Shinksansen. Tokyo is significantly larger than LA (in fact, significantly larger than anywhere else, being that it's by far the largest city in the world). The Keihanshin (Kyoto/Osaka/Kobe) has 20m people (still larger than the NYC metro area); the Bay Area has 7m people. In addition to that, Japan was built on its rail system. Shinkansen and derivative JR services are simply in a league of their own. I wouldn't even attempt to compare it to TGV, let alone CAHSR, and the fact that the Tōkaidō service tops out at 14tph shows that CAHSR is probably awhile away from 12tph.

Bear in mind that Japan is what happens when you cram half of the population of the United States into California, lop off 10% of the state, elongate California, and then render most of the state uninhabitable due to volcanic mountain ranges. Yeah, the comparison starts to sound less apt.

As I said, this is not the Tōkaidō Shinkansen.

Alon Levy said...

First, the Keihanshin has 18.6 million people, not 20. Second, Japanese metro area definitions are laxer than American ones; applying them to the US gives about 22-23 million people for New York (more, if you include Hartford and Albany, with which it shares suburbs) and 17-18 for LA.

Second, these numbers are correct as of 2008. California's cities are growing faster than Japan's; by 2030, 25 million people in LA + SD and 11 million in the Bay Area + its CV exurbs are not out of the realm of possibility.

Alon Levy said...

Another question I'd like to ask is whether three tracks are enough for the DTX tunnel. The New York City subway needs two access tracks to each two-track line's terminal, and even that creates some capacity issues. The capacity New York wants is 30 tph rather than 12, but it is also more flexible about schedules and run trains closer than HSR can. Shouldn't there be four tracks, just to keep Caltrain and HSR separate?

Anonymous said...

More confirmation today that Caltrain's latest scheme is to run ONLY locals, not level boarding. HSRA would more or less own the central tracks.

BruceMcF said...

@ Alon Levy ... enough for what?

For 8tph HSR and 8tph Caltrain ... which is more than the six platform tracks can probably support ... a single access track with 3 minute headways would have 4 spare slots, which is to say 0:45 average slack per slot.

At the slow speeds that they will be going, because of the very tight curves, there's no effective difference between a modern urban EMU and an HSR.

The more you try to squeeze in, the harder it is to see how it can support them ... a two level train box with 12 platform tracks total, eight dedicated to HSR and four to Caltrain, would be enough platform capacity that its hard to see three tunnel tracks being enough.

But, on the other hand, the curves are quite tight ... I have seen claimed too tight ... as they are, so adding another track to the footprint of the tunnel means that the tunnel might need to start curving earlier to ease the curve of the inner-most track, and since its cut and cover at that point, that may mean more property demolition.

Since the project is proceeding on the basis of a three track tunnel, I've been accepting that as an institutional constraint at this point. But then, 8tph HSR and, say, 6tph Caltrain ... would be an 8 platform track box, 6 for HSR, 2 for Caltrain. And 8 platforms is a lot easier to fit into something vaguely similar to the existing train-box design than 12 platforms.

Brandon in San Diego said...

I don't know where to start. So, I'll write and try to re-organize later...

CHSRA requirement of a station accomodating 12 tph is justifiable. One, assuming the argument that ridership numbers and/or horizon year service levels only have 8 tph is correct... there is still life beyond the sunset. If the TBT agrees to spend BILLIONS on a terminal... it better have useful life beyond that horizon year without additional and substantial more sums needed to create another station elsewhere.

Additionally, Prop 1A creates the criteria that HSR should be designed for... which is a network capable of trains spaced as close 5 minutes apart.

I have doubts that the CHSRA has the administrative ability to deviate from the12 tph design criteria.

About the designs contained in this blog post... it is just a sketch, and I understand that. However, it appears the platforms have been elongated at the expense of providing the necessary turning radii for HSR trains. The CHSRA has specified elsewhere that 500ft radii is the minimum design criteria.

As an aside, only the French TGV appears to require substantially less (about 410 feet).

As I understand it, past engineering efforts have already included countless efforts to try and horse-shoe in the platforms.... and I doubt much can be improved on those efforts.

All that said, the Transpay JPA staff presented a powerpoint report on the design of the center, or DTX, and accomodation of HSR and Caltrain. Among points made in that presentation was platform curves adn the gap created between HSR doors and platforms... and past practices to bridge those gaps.

Here's the report: Transbay Transit Center, DTX Progress

Brandon in San Diego said...

And I almost forgot... the above cited Transbay JPA report cites/admits to design flaws. They include:

- insufficient track geometrics
- operationally defecient

Of course, I re-worded their statements to provide a more accurate understanding of likely conditions.

Eric said...

said quite definitively that Caltrain platform and HSR platforms would be different heights

Oh Jesus Mary and Joseph. Really? With 10 years of planning and construction ahead of them they can't get this squared away?

Now, it's pretty common worldwide that at large terminal stations different services have different dedicated tracks, e.g. Penn Station in NYC or most London terminals, but this is in many cases due to historical differences in scheduling, ticketing practices, and operational patterns between formerly or currently different companies that share stations - you mean given 10 years Caltrain and CAHSR aren't even going to try to get on the same page?

Eric said...

I mean even Amtrak, ConnDOT, NJ Transit, MTA, and whoever else I'm leaving out managed to pull of common-height platforms on the NEC.

Aaron said...

I don't quite know why I've taken an interest in this topic ;p.

This is really largely academic - 1A [the good one] has, as Brandon pointed out, bound CAHSRA to 12tph. For good or ill, we own that.

Having said that, given the doubts placed in the number, I'm perfectly comfortable with that 12tph being the absolute maximum that can be reached. However bad this sounds, if HSR proves to be popular enough to exceed 12tph, it may make more sense to go back in later on to expand.

Yes, I am aware of how difficult that is. But as a political realist, my largest fear is that if the TBT turns into a massive municial vs. regional vs. statewide pissing match (pardon my language), the whole thing will fall apart at the seams. Build the TBT in such a fashion that expansion is at least not impossible, and that 12tph is at least mathematically possible, and then move on.

@Eric: I don't mean to sound foolish, but why the fervor for Caltrain level boarding? Have there been systemic problems with the wheelchair lifts? I mean, the MBTA still probably loathes my name for all of the hell I gave them over the Green Line, but from all the information that I have, Caltrain's accessibility features tend to work as a general rule, though it would help if they would make the remaining stations accessible (I'm not familiar enough with Caltrain to know what is rendering them inaccessible). Many of the inaccessible stations are weekend-only (why on earth have weekend-only stations?) and isn't 22nd Street slated for long-term removal? Could be totally wrong on the 22nd Street part of it.

As I said, though, please correct me if I'm in error on any of these points.

Brandon in San Diego said...

Btw, I just read a portion of the 2002 DEIR for the Transbay Terminal. In it, it was stated that the CHSRA requested tracks with at least 650 radii curves.

Presumably, this would be to not preclude the type of trains used on the Tiawan system... which have a design criteria for curves at about 650 feet.

And other thoughts I have... A platform can be on a curve. Curves are not fatal. The tradeoff is that a gap is then created between platform and trains. That gap can be as much as 2 feet depending on door location on a car (per DEIR). Doors on a car are typically at the ends of car and closer to couplers rather than the middle. Unfortunately, the ends of cars are furthest from a platform... when car is on a curve.

Gaps can be mitigated with movable bridging structures; however, at the expense of valuable platform space and timeliness of serving trains. Greater safety concerns also come into play. I can imagine maintenace of those bridges can be an additional concern.

Clem said...

More confirmation today that Caltrain's latest scheme is to run ONLY locals

@Anon, could you please elaborate? Or contact me?

bossyman15 said...

just wanted to let you all know the whitehouse.gov now have Open For Questions feature.

sign up and vote on train and transit related questions!

on thursday morning he will speak and answer most popular questions.

so hurry and vote them all up!

jim said...

@Bruce -With the CAHSR system designed with a branch for either Anaheim or San Diego when in the time of day that the ridership on the Express SF-SJ/LA justifies a full length set, its obviously better for ridership to have two single-set services at half hour intervals" Yes OR, you run both trains as one set, and split them at LA. We do that with several trains already. 5 cars are lax cars and 5 cars are sd cars - there are ways to even split them while moving.

BruceMcF said...

"One, assuming the argument that ridership numbers and/or horizon year service levels only have 8 tph is correct... there is still life beyond the sunset."

Yes, that's why the 6tph is not taken as acceptable, even if there are various strategies (all-stations interchanging with expresses, doubled sets splitting en-route, etc) that could be used to squeeze an additional variety of services out of 6tph, some headroom is required.

"Additionally, Prop 1A creates the criteria that HSR should be designed for... which is a network capable of trains spaced as close 5 minutes apart."

Now, while that language does not, of course, require all terminal stations be built for 12 tph, it does imply that:

(1) use of the tunnel tracks as independent bidirectional tracks is out of the question, since 5 minute for tunnel entrance + headway + tunnel exit + headway is not feasible.

(2) Use of the tunnel tracks as suggested above requires 2.5 min headways, not 3 minutes as assumed in the post, since the operation rests upon HSR and Caltrain services alternating, so 3 minutes headways is effectively 6 minute headway for the HSR network.

"However, it appears the platforms have been elongated at the expense of providing the necessary turning radii for HSR trains."

Don't take whatever curves I can coax out of Microsoft Word as illustrating actual turn radii.

The thing is, first, the central platform curve is eased by, essentially, swinging it so its platform tracks are entering along the sweep that is the #2 and #3 tracks from the outside edge. Similarly for the inner platform.

Of course, that assumes that the original design was workable, which I do not know to be a safe assumption. Therefore, as stated, I don't know if its possible. Easing the curve only buys an extension if the curve was not too tight to begin with.

And, after all, sometimes proofs of concept are disproofs ...

(1) HSR turn-arounds in Europe seem to be around an hour or an hour and a half, with the tightest in Germany being about 34 minutes. So banking on reliably keeping turn-around under 27 minutes is not justifiable.

(2) an empty movement per train breaks the tunnel track capacity, so the only place to take extra turn-around time is the tail track

(3) there is a constraint on the EIS envelope at the back of the box that denies the tail track to the outer platform track

(4) so the only 4 platform tracks equals 8tph layout with any operational basis puts the HSR on the middle and the inside platforms

(5) the only way I can see to buy any more length to the inner platform track is to straighten the middle track ... the only way to straighten the middle track is to steal space originally used in the throat by the outer platform track, which can be done by shortening the platform and taking its direct access to the tail track. Then the central island can be straightened a bit on front and back, and then then inner track can be straightened a bit.

(6) A bit. Enough? Obviously I am not convinced its enough, since I noted that I did not know if the layout was physically possible.

(7) If its not enough, that eliminates the four platform for HSR from consideration. If it is just barely enough but it is rejected as being an ugly kludge that takes too much away from Caltrain ...well, whether its rejected as infeasible or impossibly ugly, either way says the 4 platform for HSR is not enough.

Regarding Curved Platforms, DoDo introduces visual evidence. Such an extreme curve that gap bridging equipment is needed is rather uncommon ... its just that the first and last car in a doubled set would not be wheelchair RORO. But if 80% plus of a long train and all of a short train is wheelchair RORO, that's be enough, so least 1,000 ft. straight and a mild curve on the 60 ft. on the ends ought to be fine.

BruceMcF said...

"you run both trains as one set, and split them at LA. We do that with several trains already. 5 cars are lax cars and 5 cars are sd cars - there are ways to even split them while moving."

But if there is the demand to run 1 train per hour and split it, there are more riders available from running one service on the hour and the other on the half hour.

More ridership = fewer cars on the interstate, fewer short haul flights in the air = more social good.

Not being free to take advantage of commercial opportunities because the Transitbay people and the CAHSR could not work out adequate HSR capacity at its main northern terminus is crazy.

jim said...

First, someone please get france on the phone and get them over here to fix this mess. Americans are not capable of doing anything but making a mess of this whole project. Second, the day hsr is running full trains between sf and la every 5 or 10 minutes, successfully I will eat liver and onions. ( in other Ill be dead before that day ever comes) Third, All of this trying to make everyone happy and consider all the scenarios, and satisfy authorities and egos is giong to get in the way of geting the project done. I just get so tired of this with every damn thing that tries to get done in this state. 37 million whiny-ass crybabies have to have their pacifiers before anything can be done. How bout this. Lay the tracks, by the trainsets and run the trains to the chosen tops and tell everyone take it or leave it. trust me theyll ride. We have got to quit giving so many people so many choices. This is a purely american thing that has, int he past 30 years, gotten way out of control. The post reagan generation doesn't even know what it means to compromise, go without, or that sometimes you don't get your way. the way things should be goes like this: We the state of cali are buidlning this train. This is where it will go , this is how many trains there will be and and when its finished we will let you ride it. the soviets did do some things right you know. i hate this democratic process. It doesn't work for infrastructure. Thats why we had eminent domain to begin with -" sorry mr and mrs smith the freeway will be coming though, heres 100k you ll need to bew out by june. SO much simpler.

jim said...


jim said...

following this project is going to turn my hair gray.

無名 - wu ming said...

i wonder if some of the ridership projection might be taking into account the effect of peak oil on gas and airline fuel prices, and that as we think about what SF might possibly need, we aren't still assuming that the economic of driving or flying might be forcing a whole lot more people into trains than would be the case is a $3 or $4/gallon world, instead of a $10-15/gallon or higher world.

not that this has much to do with the TBT design or curve radius, just that perhaps our assumed projection of the present indefinitely into the not so distant future might be leading us to lowball the numbers somewhat.

peak oil not being publicly recognized as a imminent problem, even if the CAHSR folks were factoring it in, they might not be upfront about that for fear of people dismissing it.

BruceMcF said...

Jim: "Second, the day hsr is running full trains between sf and la every 5 or 10 minutes, successfully I will eat liver and onions."

But that is talking as if there are 6 Express to 12 Express trains per hour, and as if the peak travel demand period runs around the clock.

When they get ramped up, 18 or more Express services between LA and SF is no stretch, some SF-SJ/LA/SD, some SF-SJ/LA/OC.

Obviously, that is 2tph at its peak, 1tph through the day, nothing in the middle of the night.

But travel between LA and SF-SJ is not the only travel demand along the corridors. There will also be demand for travel to and from LA and SF-SJ.

In catering to that demand, the old, familiar, "you can get by with fewer services" chorus always means running bigger trains less frequently. And running bigger trains less frequently means fewer riders and less profit ... because, of course, the familiar chorus is from habits of thinking where each run is subsidized and finding ways of cutting passenger service is finding ways of cutting subsidy.

Where ridership is covering operating costs and contributing to capital costs, that thinking is backward. If you build the ridership that supports running 400 seat All-Stations train 1tph during peak travel demand and 1 train every second hour through the day ... running half as many 800 seat trains is not a cost cutting measure, its an operating surplus cutting measure.

And a number of intermediate centers ... SFO, the Peninsula, Fresno come to mind ... will have the ridership to justify more frequent service than that.

I see 2tph average either between or from and to SF/LA average as a perfectly reasonable expectation when the San Diego route opens, which means 4tph peak when that stage opens. 3tph average, which means 6tph peak, seems perfectly reasonable to me as ridership builds up.

And that does not include Specials, a possible Las Vegas branch, establishment of complementary Rapid Rail routes that would benefit from terminating with the HSR, the commuter overlay which would benefit from terminating with the HSR and etc. 6tph peak ... 2tph between LA and SF-SJ and 4 trains to and from various points between LA and SF-SJ ... the only way I can see that as liver and onions territory is if averages are confused with peaks.

@ 無名 - wu ming "i wonder if some of the ridership projection might be taking into account the effect of peak oil on gas and airline fuel prices"

The Ridership and Revenue (pdf) is Appendix 5 to the Business Plan. They assume constant cost of automobile and airplane travel in real terms.

That, of course, won't happen ... with limited capacity to expand petroleum production beyond the recent high water mark, the only scenario in which car and air travel costs do not face repeated crude oil price shocks between now and 2030 is one of extended global economic stagnation. So auto and air travel affordability is going to drop on either the cost side, the incomes side, or both.

But that is not built into the ridership forecast.

Devil's Advocate said...

Jim: I'll give you another reason to call the French. Today, and it's already the second instance this month, the French have started to take company executives hostage to obtain what they wanted in severance negotiations. We could call them to do the same with our politicians. Kidnap them, take them hostage and release them only when the trains start running.

BruceMcF said...

@ Brandon: "Here's the report: Transbay Transit Center, DTX Progress"

Thanks for that. On the PowerPoint, they are going on 500ft. curve radii, on the basis of 2004 advice (when given different advice, take the advice that best suits your own purposes), and then trying to extend that to the platforms despite the very wide gaps it implies, arguing for the ability to use measures other than close gaps to provide level boarding ... without admitting that measures other than RORO wheelchair boarding will slow access and egress and contradicts the argument presented to the Senate hearing of "just cut platform dwells to increase train capacity".

@ jim "Third, All of this trying to make everyone happy and consider all the scenarios, and satisfy authorities and egos is giong to get in the way of geting the project done."

That goes in cycles ... adopting "just get'r'dun" has its own downsides ... the "make sure everything is right beforehand" is a reaction to past abuses of "just get'r'dun". There is no perfect balance, so we always go overboard whichever direction we are swinging.

jim said...

@ Devil's Advocate said...
Jim: I'll give you another reason to call the French. Today," Sounds like a plan.

jim said...

ill bet I ould go out to market street right now and ask anyone on the street including layabouts - about the tailtrack design, show thee the map and they woudl say... "why doesn't it go all the way back around."

jim said...


anti hsr from ventrua county, lots of comments of support following the article.

BruceMcF said...

@ jim: "ill bet I ould go out to market street right now and ask anyone on the street including layabouts - about the tailtrack design, show thee the map and they woudl say... "why doesn't it go all the way back around.""

Still think the likely answer is because the train-box throat as designed runs underneath properties they wanted to redevelop, while they don't want to redevelop any properties where the loop back tunnel would go.

jim said...

So many of the european platforms are curved.. why cn they esxtend the length and capacity by curving the platform and using more of the radius.

jim said...

would it be possible, to using deep boring from main to king withoug distrubing anything? just wondering.

jim said...

The swiss got under a mountain range. I think we can get under six blocks of san francisco.

jim said...

I imagine the tail tracks are also ending and poised for a future route under the piers and bay.

Randy said...

I'm thinking that we should consider what the expansion options would be IF the train box as designed reached capacity. In 15 years (or never) what are the options? I figure there is enough room for a Caltrain station under 2nd Street between the curve to market.

This would free up the Caltrain platform for HSR, and give a underground mezzanine connection(with moving walks)to Montgomery Bart and Muni metro. Obviously there is no money today, but if planned for now it is rather logical future solution if (as we hope) ridership keeps growing. (I think that Caltrain with a much more central station is going to grow and exceed platform capacity before HSR anyway.)

I sketched it out here:

Alon Levy said...

Bruce, I'm not convinced this will be enough. At slow speeds, long trains have substantially longer headways than short trains. This is because the distance between trains is governed by both the length of the train and a safe stopping distance, and at slow speeds the train length is the more important factor.

For instance, at 10 m/s (= 22.5 mph), the minimal stopping distance for trains is about 60 meters, far smaller than a train length. The safety margin for stopping distance is inherently larger than for train length, but it's not going to equalize 60 meters and 400 meters.

Clem said...

distance between trains is governed by both the length of the train and a safe stopping distance, and at slow speeds the train length is the more important factor.

In practice that won't happen. The headway is determined before the trains arrive into the DTX tunnel. It will likely be a minimum of about 3 minutes, or 1800 m (over a mile) when crawling and screeching through that twisty DTX tunnel.

Alon Levy said...

The headway is determined by what the tunnel can support, though. And there, 20 tph for both Caltrain and HSR is a severe constraint. It will be like with the two-track North River Tunnels in New York, which are running 24 tph and bursting at the seams. I don't think Caltrain will reach the volume of the entire New Jersey Transit system, but 12 tph, on a par with the Northeast Corridor Line, sounds fairly easy, and even 16 tph won't surprise me.

BruceMcF said...

Alon Levy said...
"The headway is determined by what the tunnel can support, though. And there, 20 tph for both Caltrain and HSR is a severe constraint. It will be like with the two-track North River Tunnels in New York, which are running 24 tph and bursting at the seams. I don't think Caltrain will reach the volume of the entire New Jersey Transit system, but 12 tph, on a par with the Northeast Corridor Line, sounds fairly easy, and even 16 tph won't surprise me."

But it seems like Caltrain is being accommodating. Of course, 1/3 of the Caltrain board represents San Francisco, and the TBT is San Francisco's baby. Caltrain can always say, "ah, but we only need X trains to actually run through to the TBT", and, voila, capacity constraint "solved".

The DTX update to the TPJA claims 30mph "design speed" (p. 9). They also make a big deal of examples of curved platforms and lines shared between regional rail and HSR in Europe. Oddly, they describe Cologne as "12 tph (6 in/out)" ... which seems like another widget from the spin factor ... and do not mention that Cologne has 11 mainline platform tracks.

Alon Levy said...

It doesn't really matter what Caltrain says... Popular rail services can generate a huge amount of demand, especially when they're electrified. In that case, tph count quickly increases to the maximum permitted by capacity. The Metro-North's New Haven Line, which serves a smaller population than the Peninsula or the New Jersey-side Northeast Corridor Line, runs 20 tph. The Northeast Corridor Line makes do with 12 only because of the two-track bottleneck near Penn Station, which New Jersey is spending billions of stimulus money on four-tracking.

We can look at it in terms of commuter volumes. There are 28,326 people who live in Connecticut and work in Manhattan, compared with 14 tph from Connecticut into Grand Central. This compares with 79,648 commuters from San Mateo and Santa Clara Counties into San Francisco. In other words, if you expect Caltrain to have even one half the mode share as the Metro-North, you need 20 tph to satisfy demand.

Clem said...

For comparison, the BART transbay tube runs at 20 tph during the AM and PM rush.

BruceMcF said...

@ Alon Levy: "In other words, if you expect Caltrain to have even one half the mode share as the Metro-North, you need 20 tph to satisfy demand."

However, Caltrain can terminate some of those services at 4th and King.

And indeed, with two platforms, its hard to see how they could turn-around over eight or ten trains per hour at two simple terminal platforms unless their lines get upgraded to a through line to somewhere else and the trains do not have to turn-around at all.

Aaron said...

@BruceMcF: That's what I said at the beginning of this thing - that Caltrain seems to have a whole lot more to scream about than CAHSR does, but the response that I got was effectively that Caltrain lacked the lungs to scream about anything to start with.

Hey Clem, if you're still following this monstrous thread - does Caltrain hope to sell 4th & Townsend!King for development, or do they plan to maintain it as a secondary SF terminal? I know there's been talk of saving it, but I also know that Caltrain is bound to bring the terminal to the TBT by previous City&County proposition?

Clem said...

@Aaron, yes Caltrain is planning to keep 4th & King and terminate 60% of their service there. More on this in an upcoming blog post.

BruceMcF said...

@ Aaron ... Caltrain board structure ... if San Francisco says "don't do anything to undermine getting money for the TBT" and Santa Clara County says, "Don't make demands that will get in the way of HSR", that's a 2/3 majority for, "shut up and take it".

BruceMcF said...

Thanks to all for the contributions, many of which are going to be included in Part 2, on the 8 platform track options that seem likely to fit inside the train-box EIS envelope.

Alon Levy said...

I've always been skeptical of the need for many platforms for high-volume commuter rail. Termini have their own capacity issues, but the New York City Subway is capable of running 20 tph out of the two-track stub-end terminals of the 1 train and 24 on the 7; with CBTC technology, the L is slated to run 25 tph out of two-track terminals. They achieve these by keeping terminal dwells low, in several ways:

1. Forgoing cleaning during rush hour. This cuts dwells from 5-10 minutes during normal operation to 1 minute.

2. Dedicated platforms. Penn Station and Grand Central both have long dwells because of the need to announce platform numbers in advance. The subway has island platforms for its termini; for those that have multiple island platforms, such as the 7's Queens terminus and the 42nd Street Shuttle, the platforms are close and easy to move between. SFTT will have a serious problem keeping dwells low if commuters need to scramble when track numbers are announced. It's best to either keep the two Caltrain tracks served by the same platform, or else ensure both tracks' platforms are easily accessible from the center of the main concourse.

3. Out of the way termini. This is not an available option for commuter rail, most of whose passengers get on or off at one central station, but it does help relieve pressure on the 1 and L. However, the situation of Caltrain won't be too different from that of the 7, which only has three stations in Manhattan, comparable to the two Caltrain will have in downtown San Francisco. Likewise, the two-stop 42nd Street Shuttle keeps rush hour dwells to one minute, even though the entire train empties at each station.

4. Multiple doors per train. The 1, 7, and 42nd Street Shuttle have 3 doors per car, the L 4. Caltrain will be constrained with its two-door trains. Having more doors comes at the expense of seating capacity, but Caltrain can compensate for that by having both aisle seating and multiple doors, on the model of the Washington Metro and the RER. (Both have ensured their busiest stations are through, but this dwell-cutting measure helps for both kind of stations).

Of course, you can make an argument from incompetence, and say Caltrain will never implement those features and thus will be condemned to having LIRR-level dwells, with low terminus capacity.

BruceMcF said...

Also note that 3 doors per carriage does not foreclose bi-level sets ... some German EMU's include both.

2+2 seating across on the top level and metro seating on the bottom level maintains the seated capacity of a single level while substantially increasing the crush capacity of the train.

If there is a single service type brought into the TBT ... AFAIR Clem has said Caltrain is saying all locals ... and a single island, there's no need to announce platforms ... just announce "the next Caltrain departure is on track #X", "the train on track #Y is going out of service".

In trying to hash out operation tables, I have been basically assuming that Caltrain services can turn-around about as fast as they can move them into and out of the station.

Even with different services sharing the platform, video monitors at intervals along the platform showing which service is arriving on which platform track and in how many minutes eliminates any "passengers chasing the train" effect.

Alon Levy said...

Well, if you can combine quick boarding with reasonable seated capacity, there's no excuse to keep TBT dwells high at all. With 3 doors per train, even crush-loaded trains where everyone gets off at one station, such as those on the 42nd Street Shuttle, can stay at a station in only one minute. In that case, 20 tph should be achievable with only two tracks, which means Caltrain will clog the DTX tunnel.

Anonymous said...

Ow ow ow. Shouldn't the tail tracks extend in the direction of a future pair of New Transbay Tubes? Isn't that the long-term desire? How does this plan succeed at doing that?

Actually, given the mess involved here, and in the Peninsula, would it be cheaper to just build the freaking second Transbay Tube now, make the New Transbay Terminal a through station with four tracks, put a nice Caltrain terminal in Oakland, and reverse HSR trains in the current Caltrain yard?

Also, it's vital that Caltrain and HSR share platform heights. Anything else is total, mindless, idiocy.