09 July 2014

The Path to Level Boarding

The California High-Speed Rail Authority recently abandoned its plan to procure trains by piggybacking on an Amtrak order of new trains for the Northeast Corridor.  This news isn't all that surprising: the two have such different requirements that it never made sense for California to hitch its cart to the wrong end of Amtrak's horse.

The effect is to free California from the prescribed platform interface in use on the Northeast Corridor, where the platform edges are located 48 inches above the top of the rail and offset laterally by 67 inches from the center line of the track.  An earlier recommendation that Caltrain should use high platforms to ensure compatibility with high-speed rail, especially in highly constrained stations like San Francisco Transbay, still stands.  The specific dimensions of the platform interface, however, can be optimized for California at a height less than 48 inches and offset greater than 67 inches.

Let us briefly review the Unique Local Conditions present in California:
  • "blended" commuter / HSR service
  • bi-level commuter rolling stock, both in the north (Caltrain) and south (Metrolink)
  • single-level HSR rolling stock, at the CHSRA consultant's insistence
  • a strict requirement for level boarding, per Americans with Disabilities Act, unlike Europe
  • no pre-existing level boarding standard, unlike Amtrak's NEC
  • a very tall (17 feet, AAR Plate F) loading gauge, unlike Amtrak's NEC
These conditions have direct implications for the upcoming procurement of Caltrain EMU trains.  Caltrain recently sent out a Request For Information to train manufacturers.  This RFI discusses the platform interface issue but considers only a high option (~50") and a low option (~25").  Are those really the only options?

Future trends in high-speed train design

Stadler EC 250, with a 760 mm floor
Europe, where "blended" high-speed rail is the rule rather than the exception, is gradually standardizing around two platform heights: 550 mm (21.7") and 760 mm (29.9").  Accessibility laws are becoming more stringent, forcing train floor heights to match the platforms for seamless level boarding.  The traditional single-deck high-speed train designs now operating in Europe, including recent models designed in the last decade such as the AGV and Velaro-D, still don't allow level boarding. Two or three steps are necessary depending on platform height, to the increasing dismay of advocacy groups for persons with reduced mobility.  The next generation of single-deck high-speed trains, of the sort that California might order later this decade, are designed for level boarding with entry floor heights of 760 mm; for example, the Talgo Avril and the Stadler EC 250 (shown at right).  Accessibility requirements will eventually leave the big three manufacturers (Alstom, Bombardier and Siemens) with no choice but to follow suit and offer high-speed trains capable of level boarding, most likely at 760 mm.

Not too high, not too low, just right

760 mm is a good compromise platform height because it satisfies two conflicting requirements:
  1. it is low enough that a Caltrain or Metrolink double-deck EMU could be designed with entry doors on the lower floor without making the train excessively tall, and
  2. it is high enough that a single-deck high-speed train can be designed with a continuously accessible floor with no interior steps throughout the entire length of the train, per ADA requirements.
Floor heights lower than 760 mm quickly become impractical for providing accessibility throughout all cars of a single-deck high-speed train, as required by Federal Railroad Administration regulations that implement the ADA (specifically 49 CFR 38.175).  The train's floor must rise over wheels and traction gear that can be significantly taller than 760 mm, so lower floors lead to impractically long interior ramps or lifts to enable wheelchairs to move about between train cars.

Floor heights higher than 760 mm are not practical for boarding and alighting from the lower level of a bi-level EMU, and push the entry doors out over the wheels at the ends of train cars, resulting in unevenly-spaced doors that may impede passenger flows on station platforms and increase station dwell times.

Stepping up to 760 mm (30 inches)

The transition to level boarding, regardless of the selected platform height, will not take place overnight.  For logistical and financial reasons, there will be a period of several years during which commuter trains will serve a mish-mash of old and new platforms at differing heights.  To allow uninterrupted service through this transition, the trains will require built-in movable steps to serve both heights.

This isn't a new problem.  Numerous trains worldwide have been designed to address it, as seen in the image and videos below.
YouTube videos of moving steps:

Step deployed for a low platform on a Paris
commuter train. Photo credit: Poudou99

This is however a deceptively difficult engineering problem, because the step mechanism and controls must be incredibly reliable to prevent trains from breaking down and disrupting service.  Consider that a single train opens a dozen doors at every stop; if 100 trains a day make an average of 16 stops each, there will be about 20,000 door cycles per day.  If each door cycle has a failure probability of just 1 in 100,000, we are still left with an 18% chance that one or more doors will fail and disrupt service on any given day.  That's why even a 1 in 100,000 failure probability is not acceptable for a door mechanism, as a number of operators have found out the hard way.

The lesson is clear: these mechanisms must be designed with the utmost simplicity and reliability.  For Caltrain's new EMU fleet, that could be nothing fancier than a single step deploying from each door, as shown in the cross-section diagrams below.  Train floors are drawn in dark red.

The EMU steps, shown in the third diagram from the top, are a bit taller than one would like with an 11-inch rise.  That's 1 inch taller than today's first step, but probably acceptable for a temporary transition period until all platforms are raised and the step mechanisms can be permanently retired.

Note also that as drawn here, the extra width of the new cars would place the new platform edge 70 inches from the track center line, a full six inches outside of the nominal Plate F loading gauge.  This dimension might ease any concerns from Union Pacific that high platforms would interfere with freight service, or from the government about STRACNET clearances.

This is a compromise solution, and as such it isn't ideal under every criterion:
  • the floor height is a bit higher than one would like for a bilevel EMU
  • the floor height is a bit lower than one would like for a single deck high-speed train
  • there is a failure-prone moving step mechanism
  • the step height is awkwardly tall
Nevertheless, this compromise, or one very similar to it, provides the only viable path toward a high-speed rail system that is seamlessly interoperable with Caltrain and Metrolink in Southern California.  The guiding principle of "any train, any track, any platform" will pay off with billions in infrastructure savings state-wide, more efficient utilization of station platforms, and valuable minutes saved on every connecting trip.

Gradual Transition Strategy

Getting from today's 8-inch platforms to 100% level boarding isn't something that requires a mega-project right after (or worse, during) electrification.  It can be done piecemeal, on a station-by-station basis as funding becomes available for each, exactly like the 37 platforms built in the last 15 years.  The timeline can expand or contract to match any budget.  Here's the simplest way to get from here to level boarding:

Thinking about this transition ahead of time, and designing the EMU fleet around the new platform interface, has essentially zero up-front cost.  In the long run, doing it right the first time saves money because we won't have to do it over in order to achieve the numerous benefits of level boarding.

Caltrain may not fully realize this, but their decision about the platform interface for their upcoming EMU procurement could set a generational precedent for the entire California rail system.  Will they give it the consideration it deserves?