Seth Deegan

Designing successful railway systems

This is a page that gets updated like a wiki. It represents my thoughts at the particular time I last edited it. The thoughts expressed here may not be the thoughts I have now.

This page focuses primarily on designing successful passenger heavy railway systems. It may expand to freight systems in the future.

Tokyo and Japan have the best railway system in the world. NYC and the US is plagued by delays. Here are some lessons from these 2 systems as well as others’.

Through ongoing capital investment, the world should continuously work towards achieving the aspects of railway systems described on this page.

Why the entire world? Well rail should be able to expand beyond city and country borders and thus all systems should be compatible with each other as much as possible. As cities, states, and countries continue to become more integrated with each other, ensuring transportation compatibility will help as achieve new heights in progress. Imagine rail links between Korea and Japan; Europe and Africa; North, Central, and South America.

Universal train compatibility

The goal is to allow as many passenger and freight trains be able to run on each other’s tracks. Subway style services should be able to run on regional passenger service’s tracks. Freight trains should be able to run on regional train tracks.

We want to do this to achieve what is possible in Japan. Where Keisei trains can run on the the Asakusa line. Where freight trains can run on the Musashino line.

Doing this allows for rolling stock to be easily repurposed and moved between lines – both to efficiently allocate rolling stock based on current demand, and to repurpose older rolling stock for less-used lines. It allows regional trains to become subway trains near the core. It allows us to maximize track utility by running passenger trains during the day, and freight trains during the night.

The exception to this are high speed trains. These can run on their own track and loading gauge. This is because their goal is sheer speed and capacity and can afford to compromise compatibility for this. Electrical compatibility could be maintained.

  • Standardize
    • Loading gauge
      • Over-ground lines should be mostly be able to support freight, double-stacked containers, where possible.
      • Underground lines should be able to support double-decker trains with low height like what we see with JR’s double-decker midsection trains and Paris’s RER.
    • Platform height to achieve level and gapless boarding
      • Utilize ramps as temporary measures
      • Utilize gap fillers
        • Especially important in the US where freight train loading gauge makes high platforms difficult
    • Track curvature
  • Track compatibility. Choose from:
    • Converting to standard gauge
    • Purchase gauge-changing trains
  • Electrical compatibility
    • Catenary preferred over third rail
      • Transition existing third-rail installations
        • Purchase dual-system trains
        • Replace electrical infrastructure

Planning

https://minitokyo3d.com/
  • Reverse branching with shared trackage like what happens in NYC is not good. Extremely prone to delays, inner city branches are quite close, and TPH (trains per hour) are reduced through the core.
  • Interlining from rural places not likely to see major population growth is fine. Good examples of this are the Elizabeth line (though this is seeing massive growth).
  • Services should always use their own tracks. Tokyo does this really well. Although Yamanote, Keihin/Tohoku lines run parallel to each other, they maintain their own tracks.
  • Subway style services that are single lines with their own tracks and minimal branching like we see in China, Korea are optimal.
  • Build circle lines, establish many city work centers/central business districts
  • Allocate right-of-way:
    • Of existing lines to ensure possible expansion in number of tracks both along the line as well as stations.
    • Of planned lines so that building them is less costly. New rail right of way so many times is often repurposed freight lines, which there’s only so many of. Or they are highway medians which have poor landuse. Or they go for fully underground services requiring expensive tunneling. Using existing arterial road ROWs that are planned for rail and building with cut-and-cover or as elevated should be preferred.
  • Prioritize expanding per-train capacity by removing seating, adding more cars, or adding bi-level cars before pushing the limits of trains per hour (20+), as higher TPH can cause magnify the delays on following trains (domino effect). This is the strategy Japan uses, and it helps keep their trains on time.
  • Transit and housing planning need to be coordinated.
    • Very important: Rail expansion should not occur without housing development coupled with it. Many rail expansions in the US fail to recoup their investment because transit oriented development is not allowed after they are built.
    • Less important: Housing expansion should not occur without transit expansion. Mostly a problem for expanding Asian cities like Tokyo where new housing development is so common that lines can get overcrowded within a decade. City planners must work with railway operators to match railroad capacity with housing trends so that goods can move efficiently and the economy can continue to expand.
  • Rail lines need to have redundancy in case one goes out of service due to an accident or construction. Lines should parallel other rail lines within a reasonable distance so that they can be used as alternatives. Take the BNSF and UP West Metra lines in Chicago. There was an accident on one that stopped all service and I was able to take the other home and have someone pick me up. When people use rail, they use it because it is often the best option by a large margin. Making sure there is an alternative of similar capacity when it goes down is critical to ensuring people get to their destinations on time.
    • You wouldn’t want to rely on transit to get to a flight on time if there’s a chance it’s not reliable and there’s no other transit alternative to get to the airport.
    • An example of a system that does not have redundancy and suffers extreme pain because of it is Bart’s Transbay Tube in San Francisco. There is only one set of tracks across the north side of the bay and the only other option is driving the Bay Bridge. When everyone takes that because there is a disruption in the tube, there are horrible delays. This is a primary reason why a second Transbay Tube needs to be built.
    • Not having redundant transportation is one of the reasons car-oriented cities are so inefficient and prone to delays: people have no alternative to driving / commuting by a highway. When one highway goes down, the others become full and the effects are substantially worse compared to rail redundancy.
    • Bus substitution is not a viable replacement for heavy rail routes. Their capacity difference is incomparable and it shows when they are used as a replacement.

Train design

  • Subway interiors should be upgradable. AKA being able to switch to longitudinal seating, removing seating altogether.
  • Open-gangway / fully-walkthrough trains
  • Internet connectivity (cellular, WiFi)
  • USB-C PD ports on all trains, add outlets on regional trains

Stations

  • Always build through-running stations, with expansion possibilities like more and longer platforms in-mind.
  • Subway stations should be a 20 minute walk from each other. Allows for 10 minute walk catchment area.
  • Platform-track barriers (platform screen doors). Choose/upgrade from:
    • Simple barriers
    • Descending wires barriers
    • Midsection-height doors
    • Full doors
  • Simple in design / floor plan. Good signage.
  • Good transfer connections to nearby stations
  • Consider cross platform transfers.
  • 100% accessible

Running trains on time

  • Automated train control
  • Attendants managing crowds on platforms
  • Indicators on platforms / apps of passenger load per car
  • Keeping to a schedule, down to the second (Japan)!

Fares

  • Tap in and tap out at all stations for both subway and regional-style services.
    • Proof of fare at 2 points helps reduce fare evasion and crime
    • Helps railways better monitor ridership trends
    • Better than manual collection like on Metra because fare management is easier and services more flexible.
    • Allows for charging by distance
    • Worth the physical infrastructure cost
  • Fares by distance
    • Fares by zone are annoying to worry about for the consumer when it comes to staying within zones, and they create market distortion. To be able to go on Google Maps and see various fares for different travel methods in Japan is extremely useful and fares by distance help well-allocate resources.
  • Random fare checks. To compensate for possible evasion from stations with poor physical security or strategic offenders, employ random fare checkers.
  • Have attendants at stations actively monitoring and enforcing gates (attendants just sitting in booths watching fare evaders go by is extremely frustrating).
  • Fare card standard with speed and offline capability for system resiliency (Japan’s FeliCa has fastest speeds)
  • Fare gates that make sense
    • Upgradable
    • High capacity like Japan’s at big stations
    • Hardened physical gates at stations where harder to enforce
    • Smart gates that efficiently change direction based on passenger demand
  • Peak static or dynamic pricing to manage crowds (as done by Singapore’s MRT and Netherlands’ NS.
  • Fare caps like done by London’s TfL are dumb. Pay per fare and passes as it’s done by Japan better price travel.
  • Railcards / passes like the UK’s, Netherlands’, Germany’s, are smart and work well. Implementing them helps drive demand.

21st century preferred new heavy rail line specs

  • Grade separated
  • Built as cut and cover or elevated to reduce costs
  • Full automation + CBTC
  • At least midsection platform screen doors
  • Catenary-powered
  • Wireless connectivity (cellular, WiFi)
  • Tap in, tap out, distance-based fares with hardened fare gates and reliable enforcement
  • Open-gangway trains
  • Simple stations that minimize cost and maximize utility. We don’t need a bunch of expensive flamboyant architecture as modern stations are already extremely expensive

High speed rail

When building high speed systems, you should aim for

  • Maximum speed given the utility of the line
    • The CR450 is achieving speeds of 450 km/h with an operating speed of 400 km/h. Lines between major city pairs should be built to 400 km/h+ standards if rail speeds continue to increase.
  • Built on their own tracks that are only shared with other high speed trains. The Shinkansen system being built mostly entirely on its own grade-seaparated, viaduct spanning, completely new trackage allows for unparalleled reliability and speed compared to countries like Germany which mix regional and high speed services, leading to massive delays.
  • Total grade separation, or at least extremely protective crossings if there are level crossings like near city centers.
  • 3+2 seating
  • High speed stations at or with good rail connections to airports and city centers. China has taken the approach like with Chongquin’s high speed stations to locate them outside of the city center where they can be built large without tearing down existing infrastructure, but give them very good connections to the rest of the city through multiple subway lines running to the station. Likewise, Japan’s Tokaido Shinkansen doesn’t run directly to Haneda Airport, but it does make a stop a Shinagawa Station where there is a quick connection to the airport via the Keikyu Main Line.