The Post this weekend ran an article on the use of automated train operations on the L line, and Janet Roth’s story focuses on “aggravating delays for riders and multimillion-dollar repairs for the MTA.” Basically, according to transit workers, ATO has created more problems than it has solved along the L line. It doesn’t always work reliably well, and it has led to some monumental delays. Transit watchdogs too have confirmed that. “It’s trying to fit new technology into an old system,” William Henderson, head of the Permanent Citizens Advisory Committee to the MTA, said. “There is some skepticism about why we are doing this.”
And what do the numbers say? It cost over $340 million to implement, and the L line has suffered from a high number of delays. Not all of those, however, are related to ATO. There have been, however, 100 glitches a year on the L, and the authority had to hire special contractors to fix the problems. Still, the automated control system, says a Transit spokesman, works on more than 20 trains a day, but it has suffered a “higher than expected failure rate.”
As with every MTA technological project, this one hasn’t been the success the authority had hoped. As the authority looks to bring ATO-based train control to the 7 line, it will use a different contractor and a different set of rolling stock. MTA officials are hopeful that they can learn from the mistakes of the L and bring a better service to the IRT Flushing route.
30 comments
Maybe the MTA should’ve just hired the same people who are doing conversion work on the Métro Ligne 1 of Paris. Pay more, worry about it less later. But that goes beyond the senses of most people running the MTA, and could never be accomplished with union workforce, something that would be especially true if outside work was used.
Right, because in France they don’t have unions.
To explain nycpat’s comment to the snark-impaired, practically every job in France is unionized, and the unions are very powerful.
On the other hand, they don’t have must-accept-lowest-bid laws. They also have enough expertise and staff to do *in-house* work.
The MTA is legally obligated to award the contract to the lowest qualified bidder.
I don’t know who’s doing line 1, but wasn’t line 14 done by Siemens?
Siemens has a history of getting things more or less right in Europe and of screwing the MTA over.
Pretty much this. Siemens America are total nitwits.
So it seems. Any idea who got Flushing? It’s not Siemens, fortunately.
It’s Thales.
The original contract went to Matra, and this was supposed to be based on the Meteor line in Paris.
The original idea was to have the L line contractor produce specs that 2 follower companies would use to make compatible systems so future procurements would have multiple bidders with proven compatible systems. Good idea, but it isn’t clear how many competitors are left for future work. If it’s only 2 that’s an invitation to bid rigging.
Press release from 1999:
http://www2.prnewswire.com/cgi.....38;EDATE=#
2004 update from the person who ran the project:
http://www.tsd.org/cbtc/projec.....Status.htm
2000 powerpoint presentation explaining the project plan:
http://www.tsd.org/papers/ieee-asme042000.ppt
in Paris :
Line 14 : driverless system has been installed by Matra/ Siemens
Line 1 : is also refurbished by Matra Siemens with a driverless CBTC ;
Today project is on going with reasonably short delay
The MTA uses almost all non union labor for there construction projects. That is part of the reason the jobs are done so poorly!
No, that’s not true. Contracts require prevailing wage, and on the projects I’ve seen the workers are pretty much all union. If the foreman and general foreman are good they do very good work. If not, the work isn’t so good.
The TA contracts the work out to companies that have union workforces that do good quality work. But then those companies sub-contract the work to non-union companies that do poor work.
I have worked with ALOT of these companies. They are almost all illegal immigrants who are taken advantage of.
All I can tell you is that the subcontractors are legally required to follow the same prevailing wage rules. They file a huge mound of paperwork to prove it if they want to get paid. (Honestly, the amount of paper is amazing, and someone has to look it all over.)
I’ve never seen the situation you describe in the projects I’ve been involved with. If you run into it in the future you should contact the IG.
Wait ’til you see the delays we’re gonna get on the SAS – with only one track in each direction, and a longer run than the L, one stuck train will really kill rush hour.
The SAS will have 3 stations, nowhere near the length of the L.
Total SAS stations: 16
And by 2040 they’ll have ATO figured out.
I love how in Europe and Asia they can retrofit these systems almost seamlessly but here it’s next to impossible. Seriously, MTA?
The terminology here is a bit jumbled. There are really two related but distinct systems here: CBTC and ATO.
CBTC, or Communication Based Train Control, is a signal (or train control) system – a system that ensures safe separation of trains. The rest of the subway uses a fixed block system with wayside signals, which is perfectly safe, but which is somewhat wasteful of capacity, as (by design) it often keeps trains further apart than is necessary for safety – since a fixed block system doesn’t know where exactly a train is; it only knows which blocks it’s in. This can be mitigated somewhat by installing more signals (thereby shrinking the blocks), but that quickly escalates the costs. CBTC has other advantages as well.
One of those advantages is Automatic Train Operation, or ATO. ATO is not necessarily a function of CBTC – a CBTC system doesn’t need to have ATO, and rail systems have had ATO well before there was such a thing as CBTC. But once CBTC is installed, ATO is a fairly trivial add-on. What are the benefits of ATO? Aside from the obvious potential reductions in crew costs, ATO brings improvements in speed and in capacity (above and beyond the basic CBTC improvements).
On the Canarsie line in particular, the old signals were very, very old. They were overdue for replacement. They could be replaced by a new wayside signal system, or they could be replaced by CBTC. The cost of installing CBTC can’t be fairly critiqued without comparing it to the cost of installing a new wayside system (with sufficient capacity for the line in question). Many of the Flushing line signals are also very old. They need to be replaced. Before criticizing the decision to install CBTC on another line, what would the cost of a new wayside signal system be?
There’s no question that CBTC on the Canarsie line has been problematic. I don’t know how much is attributable to the contractor (Siemens) and how much is attributable to NYCT, but the plan all along was that the Canarsie line would be NYCT’s CBTC test bed. Hopefully somebody’s paying attention and making sure that the problems on the Canarsie aren’t repeated on the Flushing.
Also keep in mind this is a proving ground for the future. Full CBTC capability may not be such a big deal on the L and 7 which are isolated. But after that comes replacement of a good chunk of the IND where it would be a real service improvement anytime things don’t work perfectly but allowing trains to move closer and providing centralized overview and control.
The isolated lines do make good proving grounds so it makes sense work this out here.
OOOPS – you already said that in your last paragraph but I didn’t catch it.
I hope they do work out the kinks on the isolated lines first. There are bound to be lots of tricky bits. Fundamentally it’s a complete resignalling; I’m surprised that they’re trying to do entire lines at once (rather than having “signal system switchover locations”).
I’d like to add a little extra technical detail.
Every working implementation of so-called “CBTC” *still* relies on fixed blocks. They have no wayside signals (saving a huge amount of money) so the signal indications are all sent to the train by radio waves (often through the tracks). They also have *very short* blocks, which allows for close train spacing and high speeds. They also refine signal indications into more than “stop”, “go”, and “go very slowly”, by telling the train exactly how fast it’s allowed to go.
ATO was implemented with perfect success on the Docklands Light Railway.
The real problem with implementing CBTC on an old line is the switchover between old and new signals. This has to be planned *very carefully*, and is very finicky to do while keeping the trains running. The Central Line in London is the best example I can think of of doing this successfully; there are lots of examples of screwing it up.
The *only* good way to do it is to have boundaries between “old” and “new” signalling systems and to cut over one section of the line at a time (while making sure that the trains can handle both signalling systems, and that they have a switch to switch modes). Oddly this doesn’t seem to have been done with the L.
Getting back to the Post article…let’s keep in mind that “train operators and union leaders” are not unbiased. They see this system as an attack on their jobs (with reason). But they really missed the boat when it was in planning and are now fighting a rear-guard action.
At the end of the article it says that the TA doesn’t have maintenance personnel trained for this system. How are you going to get people smart enough, and keep them, if you cut salaries? Walder & Prendergast need to think very hard about that. Your basically keeping very high availability computer systems running, and Wall St firms (among others) will pay a lot for those skills.
The problem at the mta on the salary front is that low skill workers such as cleaners get paid far too much and skilled position gets paid so so. If you add in benefits they get paid VERY WELL with fairly good job securty. We can not pay cleaners $25 per hour plus benefits especially if their job role is very limites(won’t clean metrocard swipes) Overall all mta titles need to have their job roles expanded to match real world needs not narrow union job padding. Current signal maintainers skills may not match the CBTC skill set. A Fortran computer program writer is clueless to write in perl or other languages
CBTC increases capacity on a given line for a fraction of the cost of building new lines or extending all the platforms on an existing line. In a city like NY which will have continued and growing reliance on its subways, CBTC (and ATO) are an obvious investment to make, even if they are way over budget and way behind schedule. The whole point of doing it on the L line first was so they could iron out the kinks and do it on the shortest, most isolated line. Hopefully, they have learned from the experience and will be better prepared when they install the system on other lines.
well said. they were experimenting with different set ups on the L. That is part of what took so long
[…] this month, The Post highlighted the mixed results for the MTA’s attempts at bringing automatic train operation to the L line. Various glitches […]
[…] Mixed results for ATO on the L :: Second Ave. Sagas […]