Monday, July 20, 2009

Cost Effectiveness Strikes Again

Is it worth it to single track a major section of transit investment that isn't towards the end of the line? This seems really insane but it's also what the cost-effectiveness measure has done to transit projects. It's stuffed them in this current box instead of thinking about complete life-cycle of the corridor. Baltimore single tracked its first light rail line only to spend much much more later on as well as give a lot of commuters headaches. It also depressed ridership greatly sucking the wind out of an existing line while the upgrade was made.

I'm not against single tracking in all situations as Denver's West Corridor single tracking at the end of the line seems like a good cost cutting measure that can easily be remedied later. But single tracking a tunnel for a mile in a more central section only seems like asking for train delays if the schedule gets bumped even a little bit. Let's get rid of this cost-effectiveness measure. This would just push costs down the line, instead of truly being effective.


njh said...

I think single track is equally bad at any point on the line because the generally best way to run a service is stopping all stations, run every x minutes. My heirachy for good choices, based on the VTA are (from lowest quality to highest):
1. on street running mixed with traffic
2. divided, but shares traffic lights
3. single track, separate right of way
4. double track, separate right of way

The first case is the F-line in SF down market, terrible choice, yet still popular because people don't like subways. Every bus, taxi and smart arse car driver will cause another jolt to the passengers, and timetables are basically optional.

The second case is like the SJ mountain view line, which keeps fairly good time, but the tram is forced to slow down often for lights which haven't changed, and car drivers who still don't understand that trams are bigger than them.

The third case is seen on the campbell line, which manages to keep very tightly to schedule, until something goes wrong. Then you get bunching where they run multiple trams in one direction to try and restore the order.

The fourth case you see on the santa teressa line which has superlative timing as a result, and can scale up far beyond current demand.

We can extend this with experience from other cites to include:

5. Triple or quad track, which is used on higher capacity lines in places like Tokyo and Seoul. This allows for leapfrogging services, and of course even more redundancy in the case of failure.

6. Strongly connected network. This is probably the ideal for tram networks. Rather than having a few lines with almost no opportunities for avoiding problems, build a network on a grid or similar like the street network. Melbourne and Toronto use this strategy, which allows for some amazingly flexible rerouting (and some amazingly hard scheduling).

Matt Fisher said...

Single tracking a rail tunnel is a crappy idea in a downtown, in the name of "cost effectiveness" (whatever that means).

Bob Davis said...

The San Diego Trolley opened in 1981 with several single-track segments on the Downtown-San Ysidro "starter" line. Over the years it has become all double track. The only remaining single track is the Gillespie-Santee section, which runs with 15 minute headways. 15 minutes seems to be the shortest interval for a line with single track sections. Sacramento still has a few (I haven't been there for a while) and 15 min. is their typical headway. They cope with rush hours (when the State bureaus close) by running impressive four-car trains. The last few miles of the new line to Folsom (which always makes me think of Johnny Cash) are single track and run on 30 min. headways. San Francisco Muni is planning to extend track to Fort Mason, which can be reached by an existing single track tunnel, and plans are already assuming a signalling system to prevent "cornfield meets".

ben said...

This is near the end of the line. I'm not familiar enough with the project to say whether it is a second-best idea (everyone seems to agree with that) or a really terrible idea, but they still plan to double-track the tunnel through downtown

Mad Park said...

Single tracking always and only ends up costing more - mass transit requires 5 or 7.5 or 10 minute headways, and when gas is US$10 per gallon, Baltimore and any other system even contemplating single tracking will kick themselves. These systems are 100 year investments and under no circumstance should single tracking be seen as a way to save money or as a compromise. Build it right from the start.

Matt Fisher said...

Oh, yeah. I forgot this: In Ottawa, an expanded O-Train would likely require a closure if it were to be double tracked, since the entire route is single track, except at Carleton University, where there is a passing siding. "Siding" is the term for such a short passing track that appears to be in established practice.

And I couldn't forget "smart arse" like they'd use in Australia. :)