19 October 2014

Level Boarding Plan B

Plan B: four doors per car.
Based on a photo by Yevgeny Gromov
What if the recent talk of platform compatibility was just a bunch of lip service, and the high-speed rail authority remained uncompromising on their requirement for platforms and train floors at a 51" (1295 mm) height above the rails?

Then we need to be prepared for Caltrain Level Boarding Plan B.

By definition, Plan B will never be as good as Plan A.  Plan A is a workable compromise solution that would enable a gradual transition to a fully compatible blended system where Caltrain and HSR can share the same station platform tracks with 100% level boarding.

Plan B is to join HSR in their choice of 51" platforms, however misguided it may be.  The transition to 51" platforms, from today's 8" platform height, seems at first a much more complicated problem.  But is it really?  What if you did this to Caltrain's new EMUs?

The train depicted here has two sets of doors, one pair for 51" platforms located on the mid level at the ends of each car, and another pair (quite similar to the Bombardier cars, with two steps up from an 8" platform) on the lower level.  This is only a minor tweak to Caltrain's plan, taking advantage of the vehicles to facilitate a gradual transition from today's 8" platforms to level boarding at 51".  During the transition to level boarding, only one set of doors opens at each stop, depending on the platform height at that stop.

Because it is a very specific solution, it's easy to shoot full of holes.  Plan B elicits a number of objections:

Extra doors take away seating space.  Additional vestibule areas will consume the space for at least 16 seats per car, or about 12% of train's overall seating capacity.  However, Caltrain already needs better standing areas to handle peak loads, and these vestibules could be a good way to comfortably accommodate standees.  The loss of seating could also be compensated by going five-abreast in extra-wide trains.

Extra doors will cause breakdowns.  While the overall reliability of a train certainly depends on how many doors it has, the failure rate of any given door is better measured in mean cycles between failures, rather than mean time between failures.  There is not a single additional door cycle since only one pair of doors opens at any given stop, so breakdowns and maintenance expenses will certainly not double.

Wheelchairs need the ability to change levels.  During the transition, when some platforms are at 8" and others at 51", wheelchairs may need to board and alight at two different heights.  A large ADA-compliant bathroom may also need to be placed in the roomy lower level.  This implies a requirement for a wheelchair lift inside at least one vehicle in the train.  It would be a packaging challenge, but is certainly not unprecedented.

Bicyclists need to navigate interior steps.  The large contiguous areas required for efficient bicycle storage (sorting bicycles by destination as is currently the practice) would most likely be located on the lower level.  When boarding and alighting at a 51" platform, bicyclists would need to negotiate 3 steps inside the train, possibly while it is moving.  This is certainly a challenge, but must be considered in the context of today's situation, where bicyclists have to maneuver inside a moving train to access a 40-bike storage area through a single 3-foot narrow entrance after climbing up four steep steps and turning the corner around a pole--sometimes in the middle of a Giants game crowd.  Providing stair gutters in the bicycle car steps (which, by the way, would be longitudinal steps that could be built far less steep than those transverse gallery car steps) could make interior navigation far easier than it is today.

Extra doors will make the trains more expensive.  There is no question that this extra complication will lead to extra expense, but the key question is how much?  The extra pair of doors will be responsible for perhaps an extra 5% capital cost.  On a half-billion-dollar fleet purchase, this amounts to $25 million, an amount that sounds enormous to anyone with a mortgage.  But $25 million is a pittance in the context of the thousands of millions (billions!) required to build separate station infrastructure for Caltrain and HSR.  The extra cost is a rounding error, and a good case can be made for HSR picking up the compatibility tab.

Trains are difficult to build with that many doors.  Structurally speaking, each door opening compromises the strength of the car body structure, reducing its ability to withstand the enormous loads during a train wreck.  The crashworthiness of rail cars is highly regulated by the FRA, and achieving compliance for a four-door car could be quite an engineering challenge.  This is a question best left for car builders to answer.

So yes, admittedly, Plan B is sub-optimal for a number of reasons as described above.  It is quite controversial even among Caltrain advocates, many of whom harbor a visceral dislike of 51" platforms.  This dislike goes so far as to lead them to a very strange advocacy position: that Caltrain and HSR should have separate platforms!

As we have often discussed, separate platforms are an operational disaster waiting to happen at San Francisco Transbay, since every inbound Caltrain movement will conflict with every outbound HSR movement.  This constraint will  limit the capacity and future growth of the blended system.  And it's not just a Transbay issue: separate HSR stations at Millbrae, the mid-peninsula, and San Jose will require billions of additional infrastructure spending that would not otherwise be necessary.  Is that a better outcome than Plan B?

The Best of the Rest

Supposing Plan A fails and HSR insists on 51" platforms, then Plan B is the best of the rest.  It is a simple plan, and a reasonable solution for not precluding common platforms in the future.  Anyone who takes issue with it owes a detailed description of their own specific plan to make Caltrain and HSR more compatible.


  1. As sub-optimal as "Plan B" is, at least it's a lot less crazy than demanding level-boarding for both Caltrain and HSR, at the height of the current fleet of Bombardier cars. Which will be long, long gone, by the time HSR operations start.

  2. A strategy for transitioning to level boarding:
    Caltrain should take advantage of the fact that they can hook cars with different boarding requirements together into one train. Initially build high level platforms that could accommodate two EMUs at the north end of every 8 inch platform from San Jose to the San Francisco Terminal. The first EMUs to arrive should be placed one car at a time at the north end of Caltrain’s rush-hour diesel hauled consists in order to eliminate all system wheel-chair-loading-delays. Current diesel hauled rolling stock should continue to haul and board passengers at 8 inch platform sections during high demand periods. During moderate and low demand periods the same EMUs could be converted to one or two car high-level boarding trains. Two-car trains would initially poses sufficient capacity to accommodate likely passenger loads during mid-days, evenings, and weekends if during these low demand periods train frequency was drastically increased.
    A careful run-time level boarding EMU average speed performance simulation (stipulating that peak traction power to be 20 kw/tonne, peak speed to be 100 mph and average boarding time to be 20 seconds, assuming the lift/drag ratio is one thousand to one, the wind constant is triple that for the TVG-A and F = M*A) a home brew train performance simulation indicates the following run times look possible between the San Jose Diridon and San Francisco’s 22nd Street stations. The run time performance of an EMU train calling on all stations served during the weekend along the above line segment is: 50.8 minutes; skipping half the stations produces a run time simulation equal to: 39.4 minutes. The present scheduled Caltrain run period with at least three operating crew members along the same line segment is 87 minutes. Could the same number of crew member hours consumed operating one slow three-man diesel hourly train today be the same as a two local and four express one-man EMU train hourly schedule after electrification?
    3*87 ≈ 2*50.8 + 4*39.4
    The first morning passenger train could send a signal at every station to raise thirty-three inch wide platform extensions 43 inches with the similar hydraulic lift design now supplied to the automobile repair industry.
    Permanently linking 4 EMUs with one common transformer will save 1.2 % in total rolling stock weight. The Philadelphia suburban Lindenwold High Speed EMU Line has the operator sitting among the passengers. If an operator is not needed in any given car the operator’s position is normally occupied by a seated passenger. Therefore little efficiency can be gained by permanently linking cars. But adjusting permanently coupled EMU consists that come anywhere near matching Caltrain’s highly variable demand curve with capacity changes while retaining a constant headway schedule becomes impossible.

  3. Another transition strategy: Convert from diesel/low floor to electric/high floor, section by section.

    Phase 1: One week shutdown 22nd, Bayshore, SSF and San Bruno for raise the platform.
    Express: SF - nonstop -Milbrae - SJ (Diesel)
    Local: SF - Milbrae (EMU) -> transfer Milbrae - SJ (Diesel)

    Phase 2: One week shutdown Burlingame~Menlo Park to raise platform. Palo Alto Station have enough space for temporary platform in between NB and SB tracks.
    Express: SF - Milbrae - (nonstop)- Palo Alto - SJ (Diesel)
    Local: SF - Palo Alto (EMU) -> transfer Palo Alto - SJ (Diesel)

    Phase 3: One week shutdown California Avenue~Santa Clara to raise platform. Tamien have long enough platform can be partially raise the height without shutdown.

    One week shut down and transfer should be acceptable.

    1. Like Keisei converted the track gauge from 4' 6" to the standard without service disruption?

      What about Millbrae?

    2. Caltrain Milbrae have one extra track which never utilized: Platform 4s. In addition, one of BART track can be converted to Caltrain. There will be no ridership growth in BART SFO extension, especially after open the Central Subawa and Transbay Terminal.
      So, this is good opportunity for BART to sell their ROW to Caltrain.

  4. My understanding is that most Caltrain stations are 7 cars long. If it's true, then the phasing could involve raising half the length of each platform, having the low-floor trains only open doors on the low-platform half (that's 3-3.5 cars out of 5 on a train, which involves less walking forward or backward than is practiced every day on Metro-North), and then buying high-floor trains and having them only open doors on the high-platform half until the platforms are raised completely. At what is now the busiest and most important station, there are enough platforms that Caltrain can raise the entire length of some of them, allowing for all-door boarding and disembarking.

    The problem is with the shorter stations, the ones next to the grade crossings. There, only 2-2.5 cars might open doors, which is sucky but still in line with what Metro-North does on some of the less busy outer New Haven Line stations.

    1. This is good opportunity to eliminate station grade crossing completely. Grade crossing remain mostly in popular station like Mountain View, Redwood city and Sunnyvale, where many people cross the track.
      Diesel is OK because of slow accerelation. Think about if BART paltform need to acess by grade crossing. Fast accerelation of EMU will be quite dangerous for pedestrian.

    2. "The problem is with the shorter stations"

      Dunno about Broadway-Atherton-College Park, but looks like every other station has 600 ft or more.

    3. Forgot about 22nd St, which I haven't checked. I'm betting it's 600+ ft, but if you want I can look.

    4. 545 feet northbound. 519 feet southbound.

    5. Broadway 504 feet. Atherton 600 feet. College Park 201 feet

    6. College Park should be closed anyway.

      Of course, this discussion isn't about current platform length per se ... It's about how long the platforms can possibly be without running into obstacles like grade crossings.

    7. Agreed, Bellarmine's de-facto private College Park station should be closed. They can run a few extra shuttle runs for a tiny fraction of the cost of rebuilding the station to be Caltrain/HSR level-boarding compliant.

      Quoting Bellarmine's website:
      "CalTrain stops at Bellarmine's College Park station four times a day, Monday-Friday. There is one train coming from the south and one train coming from the north before school, and there is one train going each direction after school as well. Bellarmine provides a free shuttle service that leaves Bellarmine at 5:20 each afternoon and takes students to Diridon station, so that they can catch a train home after afternoon practice or other obligations. Bellarmine's administration coordinates with athletic coaches to ensure that commuters are done early enough to catch the train."

    8. In fact, some years ago now, Caltrain staff proposed to close College Park. Bellarmine parents, students and administrators mounted a major lobbying campaign opposing the closure. They packed board meetings with parent and student speakers appealing to emotion, history and even Jack London. Letters and emails flooded in too, and Caltrain backed down.

    9. Quoting Wikipedia's College Park Caltrain station page:
      "College Park is a lightly used Caltrain station located in San Jose, California. It is only served by two trains in each direction Monday through Friday, and no train stops there on weekends or holidays. It is located in Caltrain Fare Zone 4.

      College Park is principally used to serve Bellarmine College Preparatory, a boys' secondary school, resulting in the school-related service times.

      The station existed prior to the advent of Caltrain, and is mentioned in Jack London's 1903 novel The Call of the Wild.

    10. When Caltrain converts to electric power north of Tamian it will need separate diesel train to serve all stations south of Tamian. I propose that the diesel Gilroy trains run up to Great America (ACE & Capitols) Station to stay off of the electrified track north of Santa Clara and access the golden triangle jobs. This new train can serve a new stop at a new Agnew/Oracle station. Maybe it could serve an Alviso/Gold Street station too. It can also be extended south to Hollister. Finally this new train could take over service to Collage Park so Caltrain does not need to. I propose we call this new train the "San Beniton" as it would be the only train serving San Benito County.

    11. When Caltrain electrifies north of Tamien, it should abandon service farther south (unless it can wrangle an hourly or better service to Blossom Hill from UP) and let Amtrak's CC extension take care of it. A diesel-only branch with as little service as Caltrain runs to Gilroy has extremely high operating costs. It's a serious problem on the LIRR, with its high-frequency service by US standards in electric territory and few trips per day to Greenport and Montauk. There's so little service in the LIRR's diesel territory (except to some extent Port Jefferson) that equipment and crew utilization is horrendous: one of the crews, going to Greenport if I remember correctly, has one one-way trip per day, deadheading in the other direction.

      In general, in urban areas, a rule of thumb is that if you can't provide hourly service all day, you shouldn't provide any service, unless it's an overlay for a high-frequency service, for example the BART lines that run only on weekdays.