27 September 2018

Growing Caltrain into an 8-Lane Freeway

Caltrain can and should become an eight-lane freeway. Not like an ugly concrete scar tearing loudly through the landscape, but in terms of throughput capacity in people per hour. Today, Caltrain already carries the equivalent of nearly 3 freeway lanes, and more than doubling the system's capacity is hardly a moonshot. For perspective, BART's Transbay Tube carries up to 27000 people per hour, almost double the entire capacity of the Bay Bridge with its ten freeway lanes.

More than doubling Caltrain's capacity has been proposed before and is now being studied by the agency itself, after a decade of not thinking much past electrification.

Capacity calculations can be controversial and rely on many details and assumptions, so the suggested path to expand Caltrain ridership from 3 to 8 lanes of freeway-equivalent is provided in the form of a spreadsheet, embedded below. You can dig into all the numbers and assumptions for each capacity increase and see the underlying formulas for yourself, down to the detailed number of seats in each train car, to understand how it all adds up.

This is a living document, and feedback is appreciated!

45 comments:

  1. 1) Assuming an average of 4 seats per vehicle, each freeway lane has a capacity of 8,000 SEATS/hour.
    2) Assuming a 50-seat coach every 30 seconds, a dedicated transit lane has a capacity of 6,000 seats/hour.
    3) 12 tphpd requires full Peninsula grade separation.
    4) Are you finally admitting that an 8-car STADLER EMU has nearly 200 seats less than a 7-car Bombardier BEFORE removing another 100 seats for another set of doors?

    ReplyDelete
    Replies
    1. 1) The freeway lane equivalent is based on people per hour, not seats per hour. Unoccupied seats don’t count!

      2) The average vehicle occupancy of highway 101 in San Mateo County is 1.3, including busses, vans and coaches of all kinds.

      3) By 2035, we will be much closer to full grade separation. 12 tphpd is not a crazy goal, but you will note that the plan I described steps up to EMU-12 consists before going beyond 6 tphpd, in deference to the difficulty of grade separation.

      4) I haven’t made a claim about seating capacities much different than what you see here, adjusted for the latest available data, so I’m not sure there is anything to admit. You will not find a more precise capacity calculation than the one you’re looking at right here. As you can see in the vehicle worksheet, the Bombardier Bilevel is a marvel of efficient vehicle packaging, a feat that it pulls off by not being powered. Shove 8000 horsepower of traction in there and the game changes. THAT is why the Stadlers have middling seating density, not the extra doors that detractors are fixated on.

      Delete
    2. 1) Ever heard of HOV 3+ lanes?

      2) The average vehicle occupancy of highway 101 in San Mateo County, including buses, vans and coaches of all kinds is about to get fixed, courtesy of 1) above.

      3) EMU-12 will never happen (have you considered EMU-16?). With regards to being "much closer to grade separation", have you considered that SF <-> Bayshore and parts of Santa Clara county could operate mixed traffic (in ADDITION to Caltrain and HSR) without having to go through San Mateo County?

      4) Bombardier (in conjunction with SNCF) designed a marvel of efficient EMU packaging that was pulled off by alternating toiletless, doorless and powerless bi-level railcars with powered single-level railcars complete with doors, ADA bathrooms and NO traction power equipment interfering with seating space in passenger areas.

      Delete
    3. If you think HOV3+ is going to increase the average vehicle occupancy of 101, then please contact me, I have a bridge for sale.

      Delete
    4. HOV anything-at-all will help, since most of the stretch of 101 in San Mateo County does not have HOV lanes at all. Plans are afoot to change that.

      Now, back to the Bombardier Omneo EMU that Anonymous(?) brings up... Yes it is a nice package. In its Paris Transilien configuration (type Z57000) it achieves a seating density of 5.3 seats/meter, an astonishing 45% greater than Caltrain's Stadler EMU-7 seating density of 3.65 seats/meter, and well north of what any Caltrain diesel consist can do after you factor the locomotive into the length of the train. The current Bombardier six-car trains achieve 4.5 seats/meter.

      But are we comparing apples to apples? No! There are four major differences that must first be accounted for to make the seating comparison a fair one.

      1) the seat count includes four "strapontin" folding seats per door entryway, which could easily be added to the Caltrain design. Without them, the Z57000 Omneo seats 520, for a seating density of 4.7 seats/meter.

      2) The Z57000 Omneo has five-abreast seating. That is worth about 80 seats, and brings the seating capacity down to 4 seats/meter. This is only a matter of interior configuration, as the Caltrain Stadler and Omneo have the same car body width of 3.00 and 2.99 m respectively.

      3) The Omneo in Transilien configuration has relatively few bike spaces. I think the quota is 16 per consist, for a bike density of 0.145 bikes/meter. The Caltrain EMU-7 will have 0.394 bikes/meter, nearly three times as much. When you consider that bike storage displaces seats at about 1:1 ratio, over a 110 meter train (length of the Omneo consist) it would cost (0.394-0.145)*110 = 27 seats to achieve the same bike space density as the Caltrain KISS. Now the Omneo is down to 3.75 seats/meter.

      4) The Z57000 Omneo has rather anemic power of 4.8 MW per double consist (that's a 220 m long train). That consist length is longer than a Caltrain EMU-8, which will have 8 MW of installed power. While the Omneo does a nice job of squirreling away the traction equipment on the roof and outside of passenger spaces, there probably isn't the space to nearly double the number of transformers and traction inverters. So, to make the Omneo perform like the KISS (and it would need to, since we're blending with HSR and need to maintain high average speeds) the expanded traction equipment would inevitably cut into the passenger space.

      Apples to apples, I claim the Omneo is not really any denser than the KISS. Notice the first three issues are interior configuration issues that could be easily changed by Caltrain. You could add French-style "strapontins," go five-abreast, and reduce the bike capacity in favor of more seating. If you look at the KISS layout and modify it to conform to the specifications of the Transilien Omneo, all these changes would goose the capacity of a KISS EMU-7 by 112 seats ("strapontins") + 114 seats (5-abreast) + 45 seats (remove 45 bikes). The total seating would rise to 667 + 112 + 114 + 45 = 938 seats, for a seating density of 5.1 seats/meter. That's denser than the Bombardier consists currently in use, and on par with the Transilien Omneo!

      I look forward to your detailed rebuttal. Ain't seat math fun?

      Delete
  2. Is it time to do a dedicated post exploring the platform length across along the corridor? We know Millbrae, Palo Alto and San Jose are in good shape. Transbay is good for 16-car EMU on most platform - or TWO 8-car EMUs if we have some operational creativity to do that. We don't need EMU-16 capacity everywhere, but perhaps EMU-12 capacity is worth exploring across "weekend baby bullet" stations.

    ReplyDelete
    Replies
    1. Clem did explore this in some depth last year, though not in the context of 12-car trains.
      http://caltrain-hsr.blogspot.com/2017/05/calmod-11.html

      Seems that the longest platforms are 740-750ft. 12-car trains would require somewhere between 1050-1100 feet.

      Delete
    2. Thanks for digging that up. Yes, EMU-12 would require lengthening platforms at the top six or eight high-ridership stations. This is far from the trickiest project that Caltrain might confront. Someone else said "EMU-12 will never happen," to which I respond with this photo, taken a decade ago.

      Delete
    3. EMU-12 would be formed of an EMU-8 consist temporarily coupled to an EMU-4 consist. The EMU-4s would do off-peak duty when there is no need to rack up maintenance costs by running too many empty car-miles.

      Delete
    4. This could be totally of topic or not, but most (all?) Caltrain consists have a second cab car as either 2nd or 3rd car from the locomotive. In addition to utilizing all rolling stock and having spare cab cars that take the brunt of car impact, was it ever considered to "cut" the consist and run 2-3 car trains off-peak?

      Obviously, this involves more work than just connecting brake lines to include power and controls, so operationally might require too much work and testing to provide any savings.

      Delete
    5. I know, it is a rhetoric question, but I still wonder why Caltrain is not considering a consistent fleet of EMU-4, as bls does, or parts of SBB for Zürich S-Bahn … (the latter also have EMU-6 trains which they run either single or double)

      Well, it may have non-logical reasons why Caltrain aims for EMU-8…

      Delete
    6. I can think of two reasons: each cab costs a good portion of $1 million in PTC avionics, and culturally they’re unprepared for the Schaku.

      Delete
    7. "...EMU-12 would require lengthening platforms at the top six or eight high-ridership stations."

      What stations would those be?

      I assume Transbay does not need platform lengthening so it doesn't count towards the 6~8.

      4th/Townsend is an obvious one, but DTX plans for this station call for 700 foot platforms. Presumably even post-DTX, this will be one of the highest ridership stations, so this seems to be the most pressing call to action regarding enabling future EMU-12 service, give that whenever it's built, it's going to be set in concrete underground and difficult to change. Or maybe changing this would be too expensive so EMU-12's should just skip it? What do you think?

      The other stations are on the surface with fewer constraints so lengthening platforms would be "LoE Medium" (and in some cases could be undertaken simultaneously with raising the platforms to save a buck or two.) Are the existing Baby Bullet stations essentially the candidates? Besides 4th/King, how many stations have some degree of Baby Bullet service now - 10? Which ones would you pick?

      Delete
    8. Also, Diridon already has 7 tracks with platforms long enough (>1050ft) and only two that aren't. I'm not really sure how the platforms there are (or will be) used, but surely that should be enough?

      Delete
    9. "This photo, taken a decade ago" depicts an ALP-46 locomotive pushing/pulling 12 bilevel railcars.
      Any relevance to an 8+4 EMU configuration is not immediately obvious.

      Delete
    10. Perhaps it would be incumbent on the person who wrote “EMU-12 will never happen” to explain why, when 12-car bilevel regional trains are in common use elsewhere in the U.S. let alone other countries where rail service is more developed. How the EMU/locomotive-hauled distinction is remotely relevant would also be interesting to find out.

      Delete
    11. @Max Caltrain will have a consistent EMU-8 fleet which will run double during peak and single off-peak.

      @Clem The person who wrote “EMU-12 will never happen” also asked you if you had considered EMU-16 (AKA EMU-8 x 2). If and when you do, you may discover that the platform length discussion is incredibly simple and that the EMU/locomotive-hauled distinction is somewhat relevant (clue: it has something to do with coupling/decoupling).

      Delete
    12. Re "Caltrain will have a consistent EMU-8 fleet which will run double during peak and single off-peak." & "platform length discussion is incredibly simple"

      AKA, damn the costs!!!

      Delete
    13. Well, one benefit of SFFS setup is that extending platforms from EMU-8 to EMU-12 or EMU-16 doesn't require moving track or catenary. This means that Caltrain isn't necessarily locking itself to a certain service pattern and can choose platform lengths at a future time.

      Delete
    14. @Peter: Take your pick: either spend $200M now and double Caltrain capacity or find $10B and wait 50 years for full grade separation.

      @Martin: Correct as long as America's Finest Transportation Planning Professionals don't go around peppering future platform extension space with paralleling stations.
      Example in point: Sunnyvale.

      Delete
    15. First, grab your copy of the latest DTX track plans (30 March 2018).

      EMU-16 is an interesting idea but it won't fit at Transbay, even after the train box is extended. The two tracks notionally reserved for Caltrain (T-25 and T-26) have a usable platform edge of 404 m and 257 m respectively, after allowing for 9 m of clearance between the head of the train and the wall (a length that includes the bumper post). An EMU-16 is 417 m, and an EMU-12 is 315 m... you do the math.

      The latest DTX plans show 4th & Townsend platform is sized just 700 ft or 213 m, just barely enough to dock an EMU-8. (What an utter lack of foresight and failure of requirement specification!)

      EMU-8 + EMU-4 is the best fleet mix because you can mix and match to make EMU-4, 8 and 12 to suit demand. An all EMU-8 fleet won't work, since EMU-8 is too long for off-peak and too short for peak, while EMU-16 is too long for peak. All this of course presupposes that Caltrain can actually be bothered to use their new fully automatic couplers.

      Delete
  3. Right now, the train box is 1,543 feet long so the $1/2B Second Street crater is about to get some serious headwind: https://youtu.be/CmzplpXg5hU?t=229

    ReplyDelete
    Replies
    1. Can you explain what your comment is supposed to mean? It went right over my head.

      Delete
    2. https://www.sfcta.org/sites/default/files/content/Executive/Meetings/board/2018/09-Sept-25/revised/Item%207%20-%20DTX%20Penn%20Alignment%20v3.pdf (page 18).

      Delete
    3. TLDR: if the train box were approached from Minna/Natoma instead of the turn from 2nd, then the platforms could be made longer.

      Delete
    4. Correct but that alignment also makes it possible to connect the TTC to the East Bay without condemning > $2B in prime SF real estate, including the entire Rincon Center:

      http://default.sfplanning.org/Citywide/railyard_blvd/RAB_TechReport_052118_DRAFT-AppendixB.pdf (page 4).

      Delete
    5. Breaking News: Ms. Gygi's last day is today...

      Delete
  4. Shocking News: Precast sections of platform were observed at Sunnyvale!!!

    ReplyDelete
  5. BREAKING NEWS:

    http://www.caltrain.com/Assets/__Agendas+and+Minutes/JPB/CAC/Presentations/2018/October+2018+Quarterly+Board+Presentation.pdf

    Caltrain finally does the obvious, and finds that giving the corridor BART like frequencies will yield BART like ridership. This is LOL HUEG. Will Caltrain actually commit to quadrupling service as layed out in the proposal, that is the 25 Billion Dollar (and counting) question.

    ReplyDelete
    Replies
    1. One can't exactly begrudge an organization for displaying a little bit of vision and ambition once in a while.

      Delete
    2. Though I can't but wonder if Caltrain's management was really motivated by Metrolink's SCORE project, and trying to one-up it by saying "BART frequency on Caltrain is a good idea". Needless to say, I still don't think is enough service. I think Caltrain is seriously underestimating the kind of off-peak ridership they could get in SF if they had more local service there (And added infill stations like Oakdale and Paul). That's why I will continue to drum beat for more local service in SF.

      I also think that this proposal is assuming a Caltrain crossing over to the East Bay, because I don't think Transbay can really handle the kind of frequency they're implying. (Like 22tph at peak)

      Delete
    3. There is absolutely no way 22 TPH will happen until:

      1) The entire Peninsula is grade-separated (20-30 years after the second tube).

      2) There are AT LEAST another 1/2 dozen passing stations (in addition to Lawrence and Bayshore) sprinkled at strategic locations between SF and San Jose.

      Having said that, the section between Transbay and Brisbane (former Bayshore station) will offer fascinating opportunities the day 16th Street is grade-separated (and Brisbane gets two more tracks)...

      https://mtc.ca.gov/our-work/plans-projects/horizon/request-transformative-projects



      Delete
    4. I really dislike this concept of passing stations. It's a cheap recipe for excessively long dwells. What's the point of saving precious minutes en route only to lose them as you wait five minutes for the express to overtake? Especially when the train that dwells carries 1200 people and the one overtaking has just 300.

      Lawrence and Bayshore are too short to be useful, and Caltrain shows it by how few overtakes are scheduled there.

      Delete
    5. In order for additional passing stations to work, the passing tracks need to pass through at least two or three stations. If the Bayshore passing tracks included South San Francisco and possibly San Bruno or Lawrence included Santa Clara and Sunnyvale, the slower trains could be serving/stopping at those stations while the express trains pass. Caltrain had a study that showed such a scenario back in the 1990’s Caltrain has mentioned a few “overtake” locations as part of the so-called “blended” plan.

      Delete
    6. The question is whether there is a scheduled transfer between the express and the local at that particular station. I am kind of familiar with such a station, where the difference is that the express becomes local after that staition, and the local proceeds to another line (plus there is one more train ending at that station).

      In this situation, the scheduling must be set so that the express follows the local in block distance after its last stop before the transfer station. And, there must be a platform between the tracks these trains are using, allowing transfer on the same platform (that's the easiest to accomplish). Finally, the block distances on the outgoing side of the station must be so short that the local can safely depart as soon as possible after the express left. If the routes split, it must be possible to arrive or depart at the same time on both branches. (for the ones interested, the station I have in mind is Wetzikon).

      The situation is different, of course, for non-connecting passing. In this case, a longer passing track would be needed, but still, the scheduling must be set up that there is minimum interference between the express and the local.

      Delete
    7. >> I really dislike this concept of passing stations. It's a cheap recipe for excessively long dwells. What's the point of saving precious minutes en route only to lose them as you wait five minutes for the express to overtake? Especially when the train that dwells carries 1200 people and the one overtaking has just 300.>>

      Clem, Its depends on signal, train acceleration and dwell time of express train. If Caltrain have good signal system and short dwell time, it will not takes long minutes to wait.
      In Japan, it is normal tht Local train arrive 1 min before express arrival, 30 second express train dwell time and then local train can depart 1 min after express departure. Total of 2.5 minutes.

      Delete
    8. I think what the other anon had in mind was “passing stations” where the express (HSR) passes without stopping while the local (Caltrain) passengers wait. Essentially this is a short siding track with a platform. Hard to imagine dwells much shorter than five minutes, especially since precise adherence to a timetable (as practiced by our Swiss and Japanese friends) is not in the cultural norms of U.S. rail operators.

      Delete
    9. No.
      I means passing station where Caltrain express stop while Caltrain local wait. Passenger from local train can use express train from that station (Express passenger going to local stop can use local train). Location of passing station is depending on train frequency both local and express.
      If Caltrain build 2 location: Milbrae and Redwood City then Caltrain can operate both local and express train every 15 minutes and provide local-express transfer at both station. Passenger from San Francisco to San Antonio can use express train and then transfer to local train with minimum waiting.
      I understand US railroad culture is far from such operation but this method can minimize capital investment of long passing track and platform extension. They can spend more money where revenue generate. At the same time, this will satisfy passenger to/from small station.
      This is not high technology where silicone valley engineer is needed. Just need management's motivation for building better transit system.

      Delete
    10. Would Milbrae and Redwood City stations also be able to serve as a node for local/regional buses?

      (I guess Milbrae could act as connector to SFO, but to what else?)

      Delete
    11. Millbrae is kind of nowhere in particular, other than the BART connection. Redwood City is more of a “somewhere” with a demonstrated willingness to densify. Bus service is fragmented along county boundaries.

      Redwood City makes for an excellent cross-platform transfer point in a census-driven service pattern with an efficient connection to Dumbarton and points East (Altamont). I have previously written about how Redwood City ought to be laid out to function efficiently... a very ambitious plan that I should blog about again, as a new grade separation effort (with possible precluding effects arising from lack of imagination) is now getting underway.

      Delete
    12. Milbrae already have space of two island platform 4 trucks. SB platform have enough space to add one more track on the west side. For NB platform, utilize Platform 5 or convert BART truck #3 to Caltrain.
      How about Palo Alto to have same station configuration? It will require major construction but seems already have space for (=without opposition from local) and looks easier than Redwood City.
      Since Palo Alto is second largest Station of Caltrain, it is good to have 4 trucks and provide local-express transfer.

      Delete
    13. So, the starting point should be Redwood City because it can act as a real node, and having all trains meeting at the symmetry time is desirable.

      So, from there, it is relatively easy to find the next places where a full node has to be built…

      Delete
  6. Going back to the living document, would it be possible to add a column stating the exact length (in feet) of the various consists in the table?

    ReplyDelete
    Replies
    1. That information (and much more) is already in there. You need to go to the vehicle worksheets (tabs along the bottom)

      Delete